Plasmids for therapy and vaccination

Plasmids for Therapy and Vaccination

Edited by M. Schleef

Plasmids for Therapy and Vaccination Edited by M.Schleef

@WILEY-VCH Weinheim - New York - Chichester - Brisbane - Singapore - Toronto

Editor. Or. Martin Schleef Plasmid Factory GmbH & Co. KG Meisenstrasse 96 D-33607 Bielefeld Germany

1

This book was careful produced. Nevertheless, authors, editor, and publisher do not warrant the information contained therein to be free of errors. Readers are advised to keep in mind that statements, data, illustrations, procedural details or other items inadvertently be inaccurate. Libraly of Congress Card No: applied for British Library Cataloguing-in-Publication Data: A catalogue record for this book is available from the British Library Die Deutsche Bibliothek -

CIP-Cataloguingin-PublicationData A catalogue record for this publication is available from Die Deutsche Bibliothek

0 WILEY-VCH Verlag GmbH, D-694G9 Weinheim (Federal Republic of Germany), 2001 All rights reserved (including those of translation into other languages). No part of this book may be reproducted in any form nor transmitted or translated into machine language without written permission from the publishers. Registered names, trademarks, etc. used in this book, even when not specifically marked as such, are not to be considered unprotected by law. Printed in the Federal Republic of Germany Printed on acid-free paper. Composition Hagedorn Kommunikation, Viernheim Printing betz-dmck, Darmstadt Bookbinding Wilh. Osswald + Co., Neustadt ISBN

3-527-30269-7

I"

Preface Gene therapy and vaccination with nucleic acids is one of the most striking innovations in medical and veterinary sciences. The treatment of a disease on the level of genes rather than on the phenotype level is a promising option to obtain enhanced therapeutics. The pharmaceutical application of genetic material - modified so as to become the so-called active pharmaceutical ingredient (API) - requires developments in biotechnology and pharmacy to obtain systems for the transfer and expression of the API at the right place and time, resulting in the appropriate effects on the organism or cell. The vector for such gene transfer may be as simple as a short piece of DNA. For years now geneticists have been working with plasmids, and the recent developments in vector design and manufacturing point to potent therapeutics and vaccines. In chapters 1 and 2 the background of plasmids is summarized and their structures are presented, since these well-known molecules still have had an unknown potency decades after their discovery. Their different size and structure turned out to be of importance for their function, and the characterization of an API made from DNA required a complete set of quality assurance in manufacturing and quality control. These aspects are explained in chapters 11 and 12. Detailed examples of clinical applications are presented in chapters 4-6, providing an overview of the wide range of preventive and medical applications using plasmid DNA. In chapters 3 and 5, recent overviews on DNA vaccination are presented which should help to oversee the rapid development and published literature in the application of nucleic acid vaccines. Regulatory and quality assurance aspects of such new drugs are considered in chapter 13. Chapters 7-9 describe modified vector systems based on plasmids, as well as the potency of genomic research and vector design by informatics. The link between genomics and the function of genomic information necessarily requires nucleic acids. One example of veterinary health care is presented in chapter 10. The development of veterinary vaccines still requires some effort and the recent worldwide discussion on BSE in (at least) cattle makes obvious that health care in animals is also health care for humans. However, the treatment of animals grown for food production raise further questions on the aspect of genetically modified food.

VI

I

Preface

Plasmid production on an industrial scale is necessarily linked to the question of the expected market size. Chapter 14 guides through the history of gene drug development and the expectations in todays’ and future pharmaceutical markets. The development and preclinical testing of therapeutic or preventive plasmid pharmaceuticals is right at its beginning. The vision of individualized medical treatment might become reality with this type of drug system. The option to have major improvement in comparison to conventional drugs attracts research, industry and politics and is a challenge for all disciplines involved - from genomics to clinical application. Finally, I wish to thank Karin Dembowsky from Wiley-VCH for her continuous collaboration with this project and all authors who contributed to this book, to make it what I hope it will be: A recent overview on the field and a guide through an area of useful innovation. Bielefeld, January 2001

Martin Schleef

Contents Preface

V

List of Contributors XIV

1

The Biology of Plasmids

1 2 2.1 2.2 2.3 2.4 2.4.1 2.4.2 3 3.1 3.2 3.3 3.4 4 4.1

Introduction: What are plasmids? 1 General properties of plasmids 2 Plasmid replication and its control 3 The molecular basis of incompatibility 6 Plasmid inheritance 7 Mechanisms of plasmid spread 8 Conjugation in gram-negative bacteria 9 Conjugation in gram-positive bacteria 11 Plasmid-encoded phenotypes 11 Bacteriocin production and resistance 12 The Ti plasmids 12 Heavy metal resistance 14 Other phenotypical traits 16 The clinical importance of plasmids 18 The spread of antibiotic resistance and the evolution of multiple antibiotic resistance 18 Transfer of antibiotic resistance genes 19 Mechanisms of antibiotic resistance 21 Bacterial virulence genes 23 Plasmid cloning vectors 24 Perspectives 27 References 28

4.2 4.3 4.4 5 6

2

1 2 3 4

1

DNA 29 Introduction 29 Topological structures of plasmids 30 Supercoiling of DNA 32 DNA intercalating dyes 32 Structures o f Plasmid

Vlll

I

Contents

5 5.1 5.2 5.3 5.4 6

Analysis of plasmid structures 33 Electron microscopy (EM) 33 Agarose gel electrophoresis (AGE) 35 Capillary gel electrophoresis (CGE) 38 Analytical chromatography 42 Conclusion 41 References 42

3

Genetic Vaccination with Plasmid Vectors 45

1 2 2.1 2.2 2.3 3 4

Introduction 45 Vector design 45 Plasmid DNA 45 Construction of simple transcription units 47 Construction of complex transcription units 47 Strategies for DNA delivery 49 Priming humoral and cellular immune responses by DNA vaccines 50 Experimental strategies facilitated by DNA vaccination 53 Unique advantages of DNA vaccination 54 DNA vaccines in preclinical animal models 56 DNA vaccines to control infectious diseases 56 Therapeutic tumor vaccines 57 Autoimmune disease 57 Treatment of allergy by therapeutic DNA vaccination 57 Proposed clinical applications of DNA vaccines 58 Risks of nucleic acid vaccination 59 Future perspectives 59 References 61

5 G 7 7.1 7.2

7.3 7.4 8 9 10

4

A Liposomal iNOS-Gene Therapy Approach to Prevent Neointimal Lesion Formation in Porcine Femoral Arteries 75

1 2 2.1 2.2 2.3

Introduction 75 Results and discussion 77 Therapeutic plasmid 77 The gene therapy product has a clinically acceptable format 78 Efficient gene transfer was established in a minipig femoral artery injury model 79 Transfection efficiency is dose dependent 81 Non-viral iNOS gene transfer efficiently inhibits neointimal lesion formation 82 Summary and perspectives 84 References 85

2.4 2.5

3

Contents

5 1 1.1 1.2 1.3 1.4 1.5 2 2.1 2.1.1 2.1.2 2.1.3 2.1.4 2.1.5 2.1.6 3 3.1 4

6

1 2 3 3.1 3.2 3.3 3.4 3.4.1 3.4.2 3.4.3 3.4.4 4 5

7

1 2 2.1 2.2 2.3 2.4 2.5

87

lmmunotherapy o f Chronic Hepatitis B by pCMV-SZ.S DNA Vaccine Introduction 87 Hepatitis B: the disease 87 Hepatitis B: treatments 87 Hepatitis B: immune response to infection 88 What are DNA vaccines? 88 Which DNA vaccines for hepatitis B? 89 DNA vaccines for the prevention of hepatitis B 90

The mouse model 90 Humoral response 90 Cell-mediated response 92 Mechanisms of DNA-induced immune response to HBsAg 93 The primate model 94 DNA-based vaccination of chimpanzees against HBV 95 Neonatal immunization 96 DNA-based vaccination for chronic HBV infections 96 HBsAg transgenic mice as a model for HBV chronic carriers 96 Clinical trials of DNA vaccines 98 References 99 pSG.MEMRAP - A First Generation Malaria DNA Vaccine Vector Parasite life cycle and impact of malaria 103 Concept of vaccination against malaria 104

First-generation plasmid: pSG.MEPfTFL4P 106 Vector backbone 106 Insert 206 Production and formulation 111 Preclinical testing of pSG.MEPfTRAP 112 Toxicity studies 113 Biodistribution 113 Stability testing 113 Potency testing 114 Regulatory aspects 114 Future perspectives 124 References 1 1 5 Polyvalent Vectors for Coexpression of Multiple Genes 129 Introduction 119 Polycistronic expression vectors 121 Mechanisms of translation initiation 121 Characteristics of IRES elements 124 Application of IRES elements in cells and animals 126

Polycistronic vector systems 128 Expression properties of IRES vectors 130

103

I

IX

X

I

Contents

3 3.1 3.2 3.3 4

Bidirectional promoters 130 Natural bidirectional promoters 130 Artificial bidirectional promoters 131 Combining polycistronic and bidirectional expression 132 Perspectives 133 References 133

8

Form Follows Function: The Design o f Minimalistic lmmunogenically Defined Gene Expression (MIDGE') Constructs 139 The problem 139 The solution 141 MIDGE -the concept 142

1 2 2.1 2.2 2.3 2.4 2.5

9

1 2 3

3.1 3.2 3.3 3.4 3.5 4 4.1 4.2 4.3 4.4

Simple MIDGE 143 Smart MIDGE 143 Applications 145 Practical aspects of vector sequence design 145 References 146 Synthetic Genes for Prevention and Therapy: Implications on Safety and Efficacy o f DNA Vaccines and Lentiviral Vectors Introduction 147

147

Paradoxon: HIV-derived vaccines and gene delivery systems 151 Synthetic genes: Novel tools contributing to the understanding of HIV replication 152 Construction of a synthetic, HIV-1 derived gag gene 152 Codon usage modification in the gag gene abolishes Rev dependency and increases expression yields 153 Codon usage modification in the gag gene increases nuclear RNA stability and promotes constitutive nuclear translocation 154 Codon usage modification in the gag gene alters the nuclear export pathway of otherwise CRMl dependent RNAs 1% Codon usage modification increases RNA stability, modulates nuclear RNA export and increases translational efficiency 157 Synthetic genes: Implications on the development of safe and effective DNA vaccines 158 Safety issues to be considered for DNA vaccine development 158 Codon optimization of a gag-specific candidate vaccines results in increased antibody responses 159 Enhanced in vitro cytokine release of splenocytes from mice immunized with synthetic gag plasmid DNA 160 Induction of CTL responses in mice immunized with the modified Gag expression plasmids 161

Contents

5 5.1 5.2 5.3 5.4 5.5 G

Synthetic genes: Implications on the development of safe lentiviral vectors for gene delivery into quiescent cells 162 Safety issues to be considered for lentiviral vector development 163 Construction and characterization of synthetic gugpol expression plasmids 163 Production of lentiviral vectors using synthetic gugpol genes 164 Transduction o f non-dividing cells 164 Absence of replication-competent recombinants (RCRs) 165 Future perspectives 165 References 166

10

Plasmids in Fish Vaccination

1 2 3 3.1 3.2

Introduction 169 Fish 270 Fish immunology 171 Innate defence mechanisms 171 Adaptive defence mechanisms 171 Vaccination o f fish 173 Nucleic acid vaccination of fish 174 Plasmid constructs used in fish studies 175 Routes of plasmid administration 176 Intramuscular injection of plasmid DNA 2 76 Other routes of plasmid administration 178 Fate of injected plasmid DNA 178 Magnitude, distribution and longevity of expressed antigen Responses offish to injection with plasmid DNA 181 Inflammatory responses 181 Avirulent antigens 181 Virulent antigens 182 Regulatory issues and future directions 187 References 188

4 5 G 7 7.1 7.2 8 9 10 10.1 10.2 10.3 11

169

11

Plasmid Manufacturing - An Overview

1 2 2.1 2.2 3 3.1 3.1.1 3.1.2 3.1.3

Introduction 1% Structure of nucleic acids 194 Brief structural description of DNA and RNA structures 294 DNA supercoiling 196 Plasmid DNA Manufacturing 199 Major impurities and main product specifications 201 Host nucleic acids 202 Proteins 203 Endotoxins 203

193

2 79

I

XI1

I

Contents

3.2 3.2.1 3.2.2 3.2.3 3.3 3.3.1 3.3.2 3.3.3 3.4 4

12

1 2 3 4 5 5.1 5.2 5.2.1 5.2.2 5.2.3 5.2.4 5.2.5 5.3 5.3.1 5.3.2

G

13

1 2 3 4 5

G 7

Factors influencing the production of plasmid DNA: Some considerations on the upstream processing and fermentation stages 204 The plasmid vector 204 The bacterial host strain 205 Plasmid fermentation 20G Downstream processing of plasmid DNA 208 Cell lysis 210 Pre-chromatography processing: clarification and concentration 213 Chromatographic processing: purification of supercoiled plasmid DNA 214 Purification Strategies 228 Concluding remarks 229 References 23 1 Quality Control of pDNA Introduction 237

237

Characterization and quality control of pDNA 237 Validation of test procedures 239 GLP 240 Detailed description of the characterization of pDNA (final product) 242 Sterility 241 Purity 243 Content 243 Homogeneity 244 Host DNA 246 Host cell protein impurities 248 Endotoxins 248 Identity 249 Restriction analysis 249 Determination of the DNA sequence 251 Conclusion 251 References 252 From Research Data to Clinical Trials Introduction 255 Approaching regulators 256

255

Vaccine manufacture 256 Predinical safety testing 257 Clinical trials 258 Approval for clinical trials 259 Clinical trial applications in Germany 259 References 260

Contents

262

14

Market Potential for DNA Therapeutics

1 2 3

Definition of biotechnology 261 History 262 Process of pharmaceutical development 263 Human society and technical revolution 264 From sequence to product: Applications of biotechnology 265 Milestones in biotechnology: The Human Genome Project 267 The future is now: Examples for existing therapeutic approaches using gene products 268 Gene therapy in cardiovascular diseases 268 Legal aspects of gene technology and pDNA derived products 270 The extension of patent law to living creatures and their components 270 Impacts of biomedical patents 272 Resisting corporate ownership of life forms 272 Health care in the light of biotech is different in Europe and US 274 Biotech in the US from an economical point of view 274 Emergence of new companies 274 Biotech in Germany 275 Who is the health care industry and their clients, a paradigm? 276 Health care industry share prices 277 HMO enrolment rose 278 Limitations to the access of health care services 278 US uninsured population rose 279 Economical evaluation of the biotech marked in the future 279 Conclusion 280

4 5 5.1 5.2

5.2.1 6 6.1 6.2 6.3 7 7.1 7.2 7.3 8 8.1 8.2 8.3 8.4 9 10

Index

281

I

XIII

I

List of Contributors Jens Alfken MOLOGEN Berlin Fabeclcstrasse 30 D-14195 Berlin Germany Alfl<[email protected] Ruth Baier QIAGEN GmbH Max-Volmer-Strasse 4 D-40724 Hilden Germany [email protected] Kurt Bieler Institut fur Medizinische Mikrobiologie und Hygiene Universitat Regensburg Franz-josef Straufi Allee 11 D-93053 Regensburg Germany [email protected] Joaquim M. S. Cabral Centro de Engenharia Bioldgica e Qumica Instituto Superior Tcnico Av Rovisco Pais 1049-001 Lisboa Portugal gnmferreira @ist.utl.pt

Klaus Cichutek Abteilung fur Medizinische Biotechnologie Paul-Ehrlich-Institut Paul-Ehrlich-Strasse51-59 D-63225 Langen Germany [email protected] Juergen Dobmeyer IMS HEALTH GmbH & Co. OHG Hahnstrasse 30-32 D-60528 Frankfurt/Main Germany dobmeyer@ grnx.net Rita Dobmeyer Johann-IClotz-Strasse3 D-GO5 28 Franltfurt/Main Germany dobmeyeret-online.de Julia Dorge Cardion AG Max-Planck-Strasse 15a D-40699 Erlcrath Germany

doergee cardion-ag.de

xv

XVI

I

List of Contributors

Andreas Fels NewLab Bioquality AG Max-Planck-Strasse 15A D-40699 Erlcrath Germany fels @newlab.de Guilherme N. M. Ferreira Centro de Engenharia Biologica e Qumica Instituto Superior Tcnico Av Rovisco Pais 1049-001 Lisboa Portugal [email protected] Eiwin Flaschel Universitat Bielefeld Technische Fakultat AG Fermentationstechnik D-33501 Bielefeld Germany [email protected] Karl Friehs Universitat Bielefeld Technische Fakultat AG Fermentationstechnik D-33501 Bielefeld Germany kfr @fermtech.techfak.uni-bielefeld.de Sarah C. Gilbert Wellcome Trust Centre for Human Genetics Roosevelt Drive Headington Oxford OX3 7BN UK [email protected]

Hansjorg Hauser Abteilung fur Genregulation and Differenzierung GBF - Gesellschaft fur Biotechnologische Forschung Mascheroder Weg 1 D-38124 Braunschweig Germany hha @ gbf.de Angela Heischmann Cardion AG Max-Planck-Strasse 15a D-40699 Erkrath Germany [email protected] Andreas Herrmann Cardion AG Max-Planck-Strasse 15a D-40699 Erkrath Germany [email protected] Adrian V. S. Hill Level 7 NDM John Radcliffe Hospital Headington Oxford OX3 9DU

uIZ

adrian.hil1@ Molecular-Medicine. oxford.ac.uk Simon R.M. Jones Department of Fisheries and Oceans Pacific Biological Station 3190 Hammond Bay Road Nanaimo, British Columbia V9R 5K6 Canada [email protected]

List of Contributors

Claas Junghans MOLOGEN Berlin Fabeckstrasse 30 D-14195 Berlin Germany [email protected]

Andreas Muhs Cardioii AG Max-Planck-Strasse 15a D-40699 Erltrath Germany muhs @cardion-ag.de

Sven A. IZoenig-Merediz MOLOGEN Molecular Medicines S. L. CNBF Instituto de la Salud Carlos I11 28220 Majadahonda Madrid Spain I
Peter P. Muller Abteilung fur Genregulation and Differenzierung GBF - Gesellschaft fur Biotechnologische Forschung Mascheroder Weg 1 D-38124 Braunschweig Germany hha @gbf.de

Andrea ICroger Abteilung fur Genregulation und Differenzierung GBF - Gesellschaft fur Biotechnologische Forschung Mascheroder Weg 1 D-38124 Braunschweig Germany [email protected] Marcin IZwissa Universitat Ulm Institut fur Medizinische Mikrobiologie Helmholtzstrasse 8/1 D-89081 Ulm [email protected] Mane-Louise Micliel UREG, INSERM U163 Dpartement des Retrovirus Institut Pasteur 75015 Paris France maloum @pasteur.fr

Andre Oumard Abteilung fur Genregulation and Differenzierung GBF - Gesellschaft fur Biotechnologische Forschung Mascheroder Weg 1 D-38124 Braunschweig Germany hha @gbf.de Duarte M. F. Prazeres Centro de Engenharia Biol6gica e Qumica Instituto Superior Tcnico Av Rovisco Pais 1049-001 Lisboa Portugal [email protected] Joerg Reimann Universitat Ulm Institut fur Medizinische Mikrobiologie Helmholtzstrasse 8/1 D-89081 Ulm Germany

[email protected]

I

Xvll

I Andreas Richter

X V I ~ I List of Contributors

Facility Manager NewLab Bioquality AG Max-Planck-Strasse 15A 40699 Erkrath richter @ newlab.de

Martin Schleef PlasmidFactory GmbH & Co. ICG Meisenstrasse 96 D-33607 Bielefeld Germany [email protected]

James S. Robertson National Institute for Biological Standards and Control Blanche Lane South Mimms Potters Bar Herts EN63QG

]ens Schletter Cardion AG Max-Planck-Strasse 15a D-40699 Erkrath Germany [email protected]

u IZ

Torsten Schmidt PlasmidFactory GmbH & Co. KG Meisenstrasse 96 D-33607 Bielefeld Germany [email protected]

[email protected] Manfred Rudiger Cardion AG Max-Planck-Strasse 15a D-40699 Erkrath Germany [email protected] Florian Sack MOLOGEN Berlin Fabeckstrasse 30 D-14195 Berlin Germany [email protected] Reinhold Schirmbeck Universitat Ulm Institut fur Mediczinische Miltrobiologie Helmholtzstrasse 8/1 D-89081 Ulm Germany reinhold.schirmbeclt@medizin. uni-ulm.de

Joerg Schneider Omon Pharmaccines Ltd The Oxford BioBusiness Centre Littlemore Park Littlemore Oxford OX4 4SS UK joerg @oxxonpharmaccines.com Jurgen Schrader Abteilung fur Physiologie Heinrich-Heine-Universitat D- Dusseldorf Germany [email protected] Matthias Schroff MOLOGEN Berlin Fabeckstrasse 30 D-14195 Berlin Germany [email protected]

List of Co~tributors

Wolfgang Schumann Institut fur Genetik Universitat Bayreuth D-95440 Bayreuth Germany [email protected] Colin Smith MOLOGEN Berlin Fabeclcstrasse 30 D-14195 Berlin Germany [email protected]

Dagmar Wirth Abteilung fur Genregulation and Differenzierung GBF - Gesellschaft fur Biotechnologische Forschung Mascheroder Weg 1 D-38124 Braunschweig Germany [email protected]

Burghardt Wittig Abteilung fur MoleICularbiologie und Bioinformatik Fachbereich Humanmedizin Freie Universitat Berlin Heiko E. von der Leyen Arnimallee 22 Cardion AG D-14195 Berlin Max-Planck-Strasse 15a Germany D-40699 Erkrath and Germany MOLOGEN Berlin [email protected] Fabeclcstrasse 30 D-14195 Berlin Ralf Wagner Institut fur Medizinische Mikrobiologie Germany und Hygiene [email protected] Universitat Regensburg Franz-Josef StrauB Allee 11 D-93053 Regensburg Germany and geneart GmbH BioPark Regensburg Josef-Engert-Strasse9 D-93053 Regensburg Germany Ralf.Wagner @ " - "geneart.de

In the text corresponding authors are marked with an asterisk.

I

P/asmids for Therapy and kchation by M.Schleef Copyright 0 WILEY-VCH Verlag GmbH, 2001

1 The Biology of Plasmids Wolfgang Schumann

1

Introduction: What are plasmids?

Plasmids are autonomously replicating entities which can be found essentially in all bacterial species and which play a significant role in bacterial adaptation and evolution. Furthermore, plasmids are studied for their own sake and serve as important tools in studies of molecular biology. Plasmids are normally circular, although linear forms have also been described and vary widely in size from 1 kb to 200 kb. Larger plasmids of up to 1,000 kb termed megaplasmids have also been identified. The copy number per chromosome also varies among plasmids, and bacterial cells can carry more than one type. Like chromosomes, plasmids code for RNA molecules and proteins, replicate as the cell grows, and equal numbers are normally distributed to the two daughter cells upon cell division. However, plasmids do not code for functions which are essential to bacterial growth in the absence of any stressful situation, e. g., in the absence of an antibiotic. The first plasmids which have been described were named after their presence of known phenotypes. Therefore, the first plasmid discovered in the early 50s of the last century has been designated F factor which stands for fertility since this plasmid is involved in the exchange of genetic information in Escherichia coli and related Enterobacteriaceae. In the late 1950s, so-called R factor plasmids have been discovered in Japan, when E. coli and Shigellu strains resistant to a number of antibiotics were isolated from the fecal flora of patients. The ColEl plasmid harbors a gene the product of which named colicin E l can kill bacteria not carrying this plasmid. The Ti plasmid of Agrobacteriurn tumefaciens contains genes which can be transferred to plant cells where they induce tumors. This system of nomenclature has led to some confusion since several plasmids carry genes coding for phenotypes different from that for which they were originally named. To avoid further confusion, the nomenclature of bacterial plasmids has now been standardized. Plasmid names start with a small “p” for plasmid followed by capital letters that either describe the plasmid or the location where it has been

2

I constructed or give the initials of the person or persons who isolated or constructed Wolfgang Schumann

it. These letters are then followed by numbers to identify the particular construct. To give two examples, the plasmid pBR322 was constructed by Bolivar and Rodriguez, pUC19 stands for University of California. Why are there plasmids? If plasmid genes, such as those for antibiotic resistance and toxin synthesis, were part of the chromosome all cells would benefit from those genes. There is one disadvantage and one advantage to explain the presence of plasmids. The disadvantage is that the bacteria would have the burden to replicate and to maintain a larger chromosome. And we know from experiments that bacterial cells with smaller genomes will outgrow those with larger ones, and bacterial cells without a plasmid will outcompete those with a plasmid, provided that the plasmid function(s) is not needed. The advantage of having plasmids is that genetic information not essential for growth under all conditions can be easily distributed among cells of a population and even to cells of other species depending on the host range of the plasmids. While some have a very narrow host range (e.g., enteric species) others can be transferred to and replicated into all gram-negative species. The presence of plasmids in a bacterial host can change its phenotype in a variety of ways. Four basic groups of plasmid genes have been determined: 1. Determinants involved in plasmid replication and segregation to daughter cells: These determinants control plasmid characteristics such as copy number per chromosome, incompatibility and host range. 2. Conjugational transfer determinants: These control plasmid transmissibility and associated characteristics of the plasmid-bearing strain, such as sensitivity to donor-specific phages. 3. Determinants which regulate interactions with other replicons and extrachromosomal elements: These include plasmid sequences involved in chromosome mobilization and genes the products of which inhibit the fertility of other conjugative plasmids in the same host, inhibit the growth of specific bacteriophages, or prevent the host from acting as a recipient in conjugal crosses (surface exclusion). 4. Determinants which affect the interaction of the plasmid-bearing cell with the environment: These include genes for bacteriocin production and immunity, adhesion and pathogenicity factors, resistance to antibacterial agents (antibiotics, ions and radiation), and metabolism of environmental substrates (sugars, organic acids, aromatic and aliphatic hydrocarbons, detergents and pesticides). 2

General properties o f plasmids

Important properties of bacterial plasmids are strict control of their replication to achieve stable coexistence with its host. Besides control of the copy number, plasmids must be equally segregated into the two daughter cells, especially when they

7 The Biology of Plasmids

are present in a few copies only. Since bacterial cells can take up and inherit more than one plasmid, what will happen if two plasmids belong to the same incompatibility group? All plasmids can be transferred to other cells of the same or of other species. Conjugative plasmids code for the genetic information to catalyze their own transfer from the donor to the recipient cell. Mobilizeable plasmids need the presence of a conjugative plasmid within the same cell to achieve transfer to the recipient cell. 2.1 Plasmid replication and its control

When the bacterial cell grows, it replicates its chromosome(s) while plasmids present within the cell are also replicated. While autonomous replication of plasmids is a fundamental characteristic, they must synchronize their replication with the division of the host. To guarantee stable coexistence with its host, each plasmid must replicate, on the average, once every generation. Each plasmid normally has one origin of replication, or oriV site where replication begins. In addition, the plasmid must encode at least one protein that enables replication to initiate at the oriVsite. All other required proteins such as DNA polymerase, ligase, primase and helicase are taken from the host cell. Each type of plasmid replicates by one of two general mechanisms: theta or rolling circle replication. Plasmids replicating via the theta mechanism start by opening the two strands of DNA at the oriV region, creating a structure that looks like the Greek letter 8 hence the name theta replication. Replication can proceed from the oriV in one or in both directions depending on the plasmid. In the first case, one leading primer is synthesized by the primase which is extended by DNA polymerase 111, and one replication fork moves around the whole plasmid DNA (unidirectional replication). In the second case, two leading primers are synthesized within the oriV region, two replication forks start in opposite directions at the oriV which meet somewhere on the other side of the molecule (bidirectional replication). The theta mechanism is the most common form of plasmid DNA replication. It is used by most plasmids such as ColE1, RP4, F factor and P1 prophage. In the other type of replication, a single-strand break (nick) is introduced in one strand of the double-stranded plasmid at the so-called plus origin. The free 3' OH end serves as a primer, and replication proceeds around the circle, displacing the opposite or minus strand. On the displaced strand, replication initiates at specific sites, minus-strand origins, to synthesize the second strand. With some plasmids minus-strand origins do not function properly resulting in the accumulation of single-stranded plasmids. These aberrant forms of single-stranded plasmids are found in some gram-positiveplasmids such as pUBllO and pC194 from Staphylococcus aureus. In most plasmids, the genes coding for replication proteins are located adjacent to the on% Thus, only a very small region surrounding the plasmid oriv site is required for replication. This finding prompted the construction of relatively

4

I

Woifgang Schurnann

small cloning vectors by using the essential backbone of prototype plasmids (see below). The host range of a plasmid includes all types of bacteria in which the plasmid can replicate, and the host range is usually determined by the oriv region. Some plasmids such as ColEl and the derived cloning vectors pBR322 and pUC have very narrow host ranges. These plasmids will replicate only in E. coli and closely related bacteria belonging to the group of Enterobacteriae. In contrast, plasmids with a broad host range include RP4, RSFlOlO and the rolling circle plasmids of gram-positive bacteria. Plasmids with the host range of RP4 will replicate in most, if not all gram-negative bacteria and RSFlOlO will even replicate in some types of gram-positive bacteria. Broad-host range plasmids must encode all of their own proteins required for initiation of replication, therefore being independent from the host cell for any of these functions. Furthermore, they must be able to express these genes in many types of bacteria. This means that the promoters and ribosome initiation sites for the replication genes will be recognized in a wide variety of bacterial species. The copy number of plasmids is determined mostly by their orivregion. All plasmids must regulate their replication. While some plasmids replicate enough to populate the cell with hundreds of copies, others such as the F plasmid of E. coli or the P1 prophage replicate only once or a few times during the cell cycle. Plasmids that have high copy numbers, such as ColE1, only need to have a mechanism that inhibits the initiation of plasmid replication when the copy number of plasmids in the cell reaches a certain number. These molecules are called relaxed plnsmids. By contrast, low-copy number plasmids must have a tighter mechanism for regulating their replication and are therefore called stringent plasmids. Replication is controlled by the plasmid itself. Three different strategies will be discussed: regulation by RNA, regulation by RNA and protein and regulation by a protein. Replication of ColEl is regulated through the effects of a small plasmid-encoded RNA called RNA I. This 108 nt long RNA inhibits plasmid replication by interfering with the processing of another RNA called RNA I1 which serves as the primer for plasmid DNA replication (Figure 1).This RNA is first synthesized as an about 550 nt preprimer and forms an RNA-DNA hybrid at the replication origin. Then, RNA I1 is cleaved by RNase H specific for DNA-RNA hybrids, releasing a 3’ hydroxyl group that serves as the primer for replication catalyzed by DNA polymerase I. RNA I1 will only function as a primer when it is processed properly. RNA I and RNA I1 are complementary to each other, since they are made from the same region of the DNA, but from opposite strands. Both RNAs can base pair to form a double-stranded RNA helix, and this helix is refractory to processing. In addition, base pairing between RNA I and RNA I1 is enhanced by the homodimeric Rom protein, a trans-acting inhibitor of replication. Rom increases the rate at which RNA I binds to the preprimer transcript by stabilizing the initial complex between the two RNAs. RNA I is rapidly synthesized and is extremely unstable (tl,* = 2 min) and this instability is related to polyadenylation carried out by poly(A) polymerase I. More RNA I will be made when the copy number of ColEl increases maintaining the number of ColEl molecules at around 20 per cell.

7 The Biology of Plasmids 15

A -550

-300

-100

5’

oriV +I

t

RNA II

______3’

RNase H

B -550

-300

-100

oriV +I

RNA II 5‘

RNA I Fig. 1. Regulation of replication initiation of the ColEl plasmid. (A) Replication is initiated by synthesis of a primer by the RNA polymerase where its 5’ end i s located 555 nucleotides upstream o f the oriV and i t s 3’ end is variable; this primer RNA has been designated RNA II. Next, RNase H (an RNase which specifically recognizes DNA-RNA hybrids) processes these RNA II molecules o f heterogenous

length so that their 3’ ends are located within the oriV which are then elongated by DNA polymerase I. (B) Processing of RNA II can be inhibited by interaction with RNA I which is derived from the complementary DNA strand and acts as an antisense RNA. Initial contact between RNA I and RNA II occurs via loops of their secondary structures also called the kissing complex.

Another extensively studied plasmid is R 1 where the copy number is regulated by RNA and protein. Initiation of replication is regulated by the RepA protein. The repA gene is transcribed from two promoters, one of them, PcopB,transcribes copB and repA and the second, Prep*, located within copB makes an mRNA that encodes only the RepA protein. When the R1 plasmid enters a cell, a short burst of synthesis of RepA from PrepAcauses the plasmid to replicate until it attains its copy number. Then, PrepAis repressed by the CopB protein, and the repA gene can be transcribed only from the PcopBpromoter. Once the plasmid has attained its copy number, the regulation of synthesis of RepA is regulated by an RNA called CopA. The copA gene is transcribed from its own promoter located within repA and reading into the opposite direction and the RNA product affects the stability of the mRNA made from PcopB.Since both RNAs are complementary, they can pair and, therefore, prevent binding of the ribosomes to the Shine-Dalgarno sequence of the repA. In summary, the replication of R1 will be regulated by the concentration of CopA RNA which in turn will depend on the concentration of the plasmid. The higher the concentration of the plasmid, the more CopA RNA will be made and the less RepA protein will be synthesized maintaining the plasmid copy number.

6

I

Wolfgang Schumann

The third strategy regulates replication by a protein. This mechanism is used by the F factor, the P 1 prophage and by pSC101, RGIC and RP4. The oriVregion of these plasmids contains several tandem DNA repeats termed the iteron sequences which are typically 17 bp to 22 bp long and exist in about three to seven tandem repeats in the oriV region. Additionally, there usually are copies of these repeated sequences a short distance apart. The plasmid pSClOl contains within oriV region the gene repA coding for the protein required for initiation of replication, and three repeated iteron sequences called R1, R2 and R3 through which RepA regulates the copy number. Iteron replication is regulated by two superimposed mechanisms. First, the RepA protein represses its own synthesis by binding to the inverted repeats IR1 and IR2 to autoregulate its own gene. The higher the concentration of pSClOl within the cell, the more RepA will be synthesized leading to a stronger repression. Therefore, the concentration of RepA protein is maintained within narrow limits, and the initiation of replication is strictly regulated. In the second mechanism, RepA regulates the initiation of replication through binding to the three iterons. While binding of RepA to only one iteron favors initiation of replication, binding to all three represses replication. Furthermore, if all three iterons are occupied two plasmids can interact via the RepA molecules, and this is called the coupling or handcuffing model. 2.2

The molecular basis of incompatibility

Many bacteria isolated from their natural habitat carry more than one type of plasmid which stably coexist in the bacterial cell and remain there even after many generations. However, not all types of plasmids can stably coexist in bacterial cells. Some types will interfere with each other’s replication or partitioning so that if two such plasrnids are introduced into the same cell, one or the other will be lost when the cells divide. This phenomenon is called plasmid incompatibility. Two plasmids A and B are compatible with each other when they stably coexist in the same cell in the absence of selective pressure. If plasmid A excludes plasmid B from the cell, both plasmids are incompatible with each other. This finding has led to the classification of plasrnids into different incompatibility (Inc) groups. There may be hundreds of different Inc groups, and, e.g., plasmids RP4 and RK2 both belong to the IncP (incompatibility group P) while RSPlOlO is an IncQ plasmid. This means that RP4 and RSFlOlO can be stably maintained within one single cell since they belong to different incompatibility groups. The knowledge about the Inc phenotype of plasmids is important in gene technology experiments when more than one plasmid is used per cell. What is known about the molecular basis of incompatibility? Plasmids will be incompatible when they either share the same mechanism of replication control or the same par (partitioning) function. If two plasmids are replicated by the same replication control system, they are randomly selected for replication. Assuming that a cell starts with two plasmids A and B belonging to the same incompatibility group which are both present in ten copies each, during the first re-

I The Biology of Plasmids 17

plication round the cell will select seven copies of A which are each replicated four times summing up to 28 copies in total. Furthermore, three copies of B are selected and replicated four times each totaling to twelve copies of B which are then distributed to the two daughter cells. During the next replication round this imbalance is further increased, and after a few generations, plasmid B gets lost from the population. In general, the smaller plasmid is retained while the larger plasmid will disappear. The reason for this finding is that the smaller plasmid is replicated faster than the larger one and, therefore, has a better chance to be reused for the next replication round. The second molecular basis for incompatibility is based on the par function. Many plasmids, especially those with a low copy number, carry partitioning systems ensuring that at least one copy of the plasmid segregates into each daughter cell during cell division. The genetic determinants have been termed par functions. In the case of the P1 plasmid, the par region consists of the cis-acting incB sequence and the two genes parA and parB. How do the partitioning functions operate? The following model has been suggested: The incB site binds three molecules of the 39 ltDa ParB protein. Two molecules interact with a perfect 13 bp palindrome and the third binds to a region separated from the palindrome by a recognition site for the IHF (integration host factor). The IHF is a host-encoded heterodirneric protein which bends the DNA by 140 O upon binding, thereby bringing all three ParB molecules in close proximity This complex then interacts with the membrane at the division site. ParA belongs to the group of ATPases and binds close to the incB site. Possibly, ATP hydrolysis by the ParA-ParB incB complex provides the energy to separate the products of plasmid replication in a process analogous to mitosis in eukaryotic cells. When two plasmids A and B share the same par function, one or the other will always interact with the membrane site and be distributed into the daughter cell during division while the other will randomly segregate into one of the two daughter cells. However, sometimes one daughter cell will receive plasmid A and the other cell will receive plasmid B, producing cells cured of one or the other plasmid. 2.3

Plasmid inheritance

Cells that have lost their plasmid during cell division are said to be cured of the plasmid. Since cells without plasmids will grow a little faster than those carrying plasmids within a population, several mechanisms prevent curing: site-specific recombinases that destroy multirners and addiction systems. The possibility of losing a plasmid during cell division is increased, if the plasmid forms multimers during replication or via homologous replication. Since the total number of origins will always be the same independent whether they occur on many monomers or a few multimers, multimers will increase the chance that daughter cells arise without the plasmid, since a few multirners might segregate within just one daughter cell. To avoid this problem, many plasmids have aquired site-specificrecombination systems to resolve multimers. These resolution

8

1 systems are composed of Wolfgang Schumann

a site-specific recombinase and a short DNA sequence (about 30 bp) to which the recombinase binds. The principle of monomerization will be illustrated with two examples: P1 prophage and ColE1. Prophage P1 codes for the sequence-specific Cre recombinase which interacts with the loxP sequence. When two P1 monomers form a dimer, the two loxP sequences occur as direct repeats. The Cre recombinase will bind to both of them, align both sequences, carry out recombination and thereby resolve the dimer into two monomers. Another example is ColEl where the resolution site is called cer. This cis-acting 240 bp DNA recombination sequence is recognized by the two site-specificrecombinases XerC and XerD which are encoded by the E. coli chromosome. Here this recombination system also resolves ColEl multimers into monomers. Some plasmids code for a system which kills cured cells to prevent their disappearance from a culture due to faster growth of plasmidless cells. These systems are called addiction or post-segregational killing systems and consist of two components. While one component is always a protein toxic to the cell, the other is either an antisense RNA or a second protein. The antisense RNA prevents translation of the mRNA coding for the toxic protein, and the second protein interacts with the toxic protein thereby preventing its lethal action. As long as the cell carries the plasmid, action of the toxic protein is prevented due to the antidote. When a daughter cell without the plasmid arises, the toxic protein becomes active due to the shorter half-life of either the antisense RNA or the second protein. The F factor codes for the antidote CcdA (coupled cell division), a 8.3 kDa polypeptide and the toxic protein CcdB (11.7 kDa). The CcdB toxin alters the activity of DNA gyrase and causes double-stranded breaks in the bacterial chromosome. The addiction system of plasmid R1 consists of the two genes hok (host killing) and sok (suppression gf killing). They are encoded on opposite strands with a 128 bp overlap at their 5' ends. While the Hok polypeptide (5.5 kDa; destroys the cellular membrane potential thereby causing loss of cellular energy) is responsible for the host cell killing, sok codes for an antisense RNA which prevents translation of the hok mRNA. 2.4

Mechanisms o f plasmid spread

The most important characteristic of plasmids is their ability to transfer themselves to other cells of the same or even different bacterial species. These mechanisms are called conjugation and mobilization and will be described in detail. Besides this mechanism of self-transmission, plasmids released from lysed cells can also be taken up by surrounding cells in a process termed transformation. And the mechanims of plasmid spread involve phages which may package plasmid DNA into their heads and inject them into new hosts. We will consider here only the conjugative mechanisms which need genetic information encoded by the plasmid itself.

7 The Biology of Plasmids 19

2.4.1

Conjugation in gram-negative bacteria

One remarkable feature of some plasmids is the ability to transfer themselves from one cell to another in a process called conjugation. This process was discovered in 1947, when mixing different of E. coli strains resulted in strains that were genetically different from their originals. Later, it was discovered that conjugative plasmids can cause the transfer of non-conjugativeplasmids into other cells. This process was termed mobilization. During conjugation or mobilization, a plasmid is transferred from the donor to the recipient cell, and the recipient bacterium that has received DNA is called trans- or exconjugant. Conjugative plasmids in gram-negative bacteria such as in E. coli need three different genetic elements for successful transfer of their DNA. These are the cis-active oriT (&gin of transfer), the tra (transfer) and the mob (mobilization) genes. The process of conjugation will be illustrated with the example of the F factor, the most extensively studied conjugative plasmid with a size of 100 kb. Plasmid transfer specifically initiates at the oriT site which is about 250 bp long and is recognized by the TraI protein (in general called Mob protein) which introduces a single-strand break in one strand at a specific site. Then, the TraI protein is covalently linked to via a tyrosine residue to the 5’ phosphoryl group and acts as a 5’-3’ helicase to unwind the donor duplex. At the same time, DNA polymerase I11 binds to the free 3’ end to start replication, and the displaced strand is transferred to the recipient strain. Transfer needs close contact between the donor and the recipient cell. This is accomplished by the F or sex $us, an appendage of the cell surface of the donor cell. At least 14 tru genes are involved in construction of the F pilus, a hollow cylinder of 8 nm in diameter with a 2 nm axial hole. The pilus is constructed from pilin subunits which are encoded by traA and processed by the product of traQ. After the initial contact between the tip of the pilus and the recipient cell, the pilus retracts by depolymerization and thereby brings donor and recipient into close contact. Transfer of the single-stranded DNA occurs either through the hallow pilus or through the membranes of the two attached cells. It has to be stressed that the donor cell contains several pili on its outer surface enabling one donor to make contact with several recipient cells leading to the formation of mating aggregates. When the complete single-strand DNA molecule has arrived within the recipient, the bound TraI protein acts as a DNA ligase, seals the two ends and dissociates from the circular molecule which is then converted to a double-stranded molecule. In summary, the TraI protein has three enzymatic activities: sequence-specificendonuclease, helicase and DNA ligase. It should be mentioned here that the F pilus serves as a receptor for filamentous phages such as M13 and fl, which are used in gene technology experiments. These phages package single-stranded DNA and are used as vectors whenever single-stranded DNA is needed, e. g., during sequencespecific mutagenesis. Besides replicating as an autonomous replicon, the F factor can also become integrated into the E. coli chromosome leading to an Hj’?strain (high frequency of yecombination). Two different mechanisms are responsible for recombination of the F factor into the chromosome, and both depend on one of the three IS (jnser-

10

I tion sequence) elements which are part of the F factor Worgang Schumann

DNA, namely IS2 and IS3. Insertion elements are small DNA elements (0.7 kb to 1.5 ltb in length) which encode only enzymes needed to promote their own movement (transposition) from one place in a replicon to another. These two IS elements also occur in numerous copies on the bacterial chromosome. In one case, integration occurs by homologous recombination between two identical insertion elements present on both the F factor and the E. coli chromosome, and this reaction is catalyzed by the RecA (recornbination) protein. In the other case, which is quite rare, integration occurs by illegitimate recombination, again using either IS2 or IS3 residing on the F factor. During this process of insertion element transposition, the element will be replicated in situ, and the whole F factor will be inserted at any place within the bacterial chromosome. Such a covalent connection of two different replicons is called a cointegrate. Hfr strains can transfer part or even the whole chromosome into a recipient cell. This conjugative transfer follows the same mechanism as described for the autonomous F factor, The TraI protein binds to the oriT site, nicks one strand, is covalently bound to the free 5' end and is transferred into the recipient together with the single strand. The 5' end of the transferred strand always consists of F factor DNA which is followed by chromosomal DNA. Upon arrival within the recipient, RecA protein will bind to the single-stranded DNA, screen the resident chromosome for the homologous region and initiate recombination between the incoming strand and the chromosome. By this mechanism, alleles can be exchanged, e. g., an auxotrophic mutation within the recipient can be replaced by the wild-type allele transferred from the donor. Transfer of the whole chromosome takes about 100 min under optimal conditions, and therefore, the E. coli chromosome is divided into 100 min (normally, circular bacterial chromosomes are divided into 360 "). However, transfer of the entire chromosome is rare, because the mating cells are separated during the process of conjugative transfer. Transfer always occurs in a polarized way depending on the orientation of the oriT relative to the chromosomal genes. Therefore, genes which are transferred first are called early genes and those which are transferred at the end after approximately 100 min are designated as late genes. The F factor can also be excised from the chromosome. Normally this occurs by homologous recombination between the identical copies of two insertion elements flanking the F factor DNA. Rarely recombination does not occur between the flanking insertion elements, but between only one of them and another identical element at some distance on the other side of the F factor. By this imprecise mechanism the excised F factor contains a piece of chromosomal DNA, and these composite F factors are called Fpvime (F') factors. They can contain up to a quarter of the E. coli chromosome, but tend to be rather unstable. On the other hand, F' factors with up to 300 kb of additional DNA are stable, and this observation has resulted in the creation of BACs (see below). As outlined above, a conjugative plasmid needs the three genetic elements oriT, mob and tra for self-transfer.Most plasmids do not carry tra genes, but they contain an oriTand at least one mob gene. This genetic information is not sufficient for self-transmission, but in the presence of a conjugative plasmid such as the F factor,

1 The Biobgy of Plasmids

a non-conjugative plasmid such as ColEl can be mobilized into the recipient. Mobilization is not dependent on the transfer of F factor DNA, we only need the

pilus whose synthesis is controlled by the conjugative plasmid to form the mating aggregates.

2.4.2

Conjugation in gram-positive bacteria

Conjugative plasmids have also been found in many types of gram-positivebacteria such as Bacillus, Enterococcus, Staphylococcus, Clostridium, Nocardia and Streptomyces. However, much less is known about the transfer systems of these bacteria. A feature of conjugation so far unique to enterococci is the involvement of pheromones. Pheromones are hydrophobic peptides, 7-8 amino acids long produced by recipients cells. Each recipient may produce multiple different pheromones, at least five. If a pheromone is taken up by a donor, it induces the synthesis of an adhesin which in turn stimulates cell clumping followed by conjugation. Once a recipient has acquired a particular plasmid it stops secreting the corresponding pheromone. Plasmid PAD1 from Enterococcus faecalis is the best studied pheromone-induced plasmid. About half of its genome (30 ltb) is devoted to conjugation and related functions whose expression is induced by the pheromone cAD1. Conjugative plasmids are also widespread in Streptomyces species. The self-transmissible 8.8 kb plasmid pIJ101 has been studied in detail. It can be transferred to plasmid-free bacteria with almost 100 % efficiency and promotes recombination between the chromosomes of mating bacteria, yielding up to 1% recombinant cells. A single transfer gene (tra) is essential for conjugation, transfer requires the cis-acting clt locus, and the two genes spdA and spdB enhance the frequency of transfer. Several conjugative antibiotic resistance plasmids of 38-57 ltb have been described in staphylococci,and the genes required for transfer occupy 12-15 kb. The mechanism of conjugation is not well-characterized.

3

Plasmid-encoded phenotypes

Depending on their size, plasmids can code for a few, a dozen or hundred of different proteins. Plasmids do not code for functions involved in transcription and translation of their own genes or their replication, but carry one or more genes responsible for control of plasmid replication initiation and segregation. Gene products encoded by plasmids include enzymes for the utilization of unusual carbon sources such as toluene, synthesis of toxins and proteins that permit the successful infection of higher organisms and resistance to substances such as heavy metals and antibiotics. Phenotypic traits of medical importance will be described in chapters 4-6. There are plasmids which do not code for any known function called cryptic plasmids. These cryptic plasmids must increase its host fitness in a way which has not yet been recognized.

I

”

12

I

Wolfgang Schumann

3.1 Bacteriocin production and resistance

In 1925 the Belgian microbiologist Andre Gratia observed that some bacteria released agents (bacteriocins) into the surrounding medium that killed other bacteria, but to which individuals of the producing strain were resistant. This killing is observed on agar plates as clear zones on lawns of sensitive bacteria, each zone originating from a single bacteriocinogenic bacterium. Normally, only a small proportion of the cells of a population actively produces the bacteriocin. Plasmids which encode bacteriocins are widespread among bacteria, and the bacteriocins are called “colicins”when they are produced by coliform bacteria. Enterobacter cloacae produces cloacin, Streptococcus pyogenes streptocin and Staphylococcus epidermis staphylocin. The genetic determinants for bacteriocin production and resistance have been studied most extensively for colicin production and resistance. A well-studied example of colicin production is the multicopy plasrnid ColE1. Production of colicin E l occurs only in a few cells of an E. coli population (one in 10-3-10-4) and requires the simultaneous expression of the colicin structural gene (cea) and the kil gene, which is responsible for cell lysis thereby releasing the colicin E l into the environment. All cells carrying ColEl are protected by the lethal action of colicin E l by the immunity protein encoded by the imm gene. The Imm protein directly interacts with the colicin E l thereby preventing its lethal action. Colicins exert their lethal action by different mechanisms. While many colicins such as ColE1, ColE4, ColA and ColB kill the cells by depolarization of the inner membrane, ColE2 is an endonuclease which introduces single-stranded breaks into chromosomal DNA. ColE3 is an RNase which specifically cleaves the 16s rRNA. Colicin M inhibits incorporation of diaminopimelinic acid into the cell wall of growing cells thereby causing lysis. 3.2 The Ti plasmids

Pathogenic strains of Agrobacterium tumefaciens can infect wound tissue in dicotyledons and incite crown gall tumours in a wide range of dicotyledonous plants. Pathogenicity is associated with the presence of large plasmids (about 200 kb) known as Ti (3mor inducing) plasmids. These plasmids contain four groups of important genes: genes involved in conjugation and virulence, genes responsible for tumor genesis and genes allowing the synthesis and degradation of unusual compounds collectively called opines. They are synthesized within the plant cells and used by the bacterial cells as sole carbon and energy source. A prototype of the octopine group of Ti plasmids shown in Figure 2 carries 155 genes. Three groups of genes are indicated. The T region contains two T-DNAs (T stands for transfer), T, and T,, 13 and 7.8 kb in length, respectively, which can be transferred into plant cells, while the genes of the 35 kb Vir region catalyze this process. The two Tra regions code for functions involved in the conjugative transfer of the Ti

7 The Biology of Plasmids 113

Map of a Ti plasmid. Indicated are the T region with the T, and T, DNA, the Vir region with the six vir operons and the two Tra regions, Fig. 2.

virABCDEG

plasmid to recipient bacterial cells. In addition, the Ti plasmid codes for genes required for replication and for the uptake and catabolism of opines. Infection of wounded plant cells starts with a chemotactic movement of A. tumefaciens towards the plant. The wonded cell releases an exudate containing phenolic substances such as acetosyringone which are sensed by the bacterial cell. Sensing occurs via a two-component system consisting of the two proteins VirA and VirG. VirA is an integral protein of the inner membrane, and when its outside domain has bound the signaling molecule, a histidine residue is autophosphorylated. Next, the phosphoryl group is transferred to an aspartate residue on the response regulator protein VirG. The activated phosphorylated form of VirG coordinates expression of the set of six vir operons. Transfer of the two T-DNAs requires cell-cell contact and resembles plasmid conjugation. The T-DNAs are each flanked by cis-acting 25 bp imperfect repeat sequences called border sequences. Adjacent to the right border of TLis another cis-acting sequence termed overdrive which provides a binding site for the VirCl protein, and this complex is required for transfer efficiency. Transfer is initiated by processing of the T-DNAs. Each border is cleaved on the bottom DNA strand at a site exactly 4 nucleotides from its left end. This reaction is catalyzed by the VirD2 protein which remains covalently linked to the 5’ end of each cleaved strand via a tyrosine residue. Next, the bottom strands are displaced during rolling-circle DNA synthesis that initiates at the free 3’ end of each right border. The single-stranded T-DNA molecules with the VirD2 protein covalently linked to their 5’ ends is transferred through the VirB pore into the plant cell. Through the same or another pore, VirE2 molecules are transferred, too, which cover the single-stranded T-DNAs in the plant cells to protect them from host nucleases. The VirB pore is composed of 10 proteins which have been localized to the inner or outer membrane, and most of them appear to be integral membrane proteins. Upon arrival within the plant cell, additional steps are required to transport the T-DNAs to the nucleoplasm and to integrate them into the host DNA. Both VirD2 and VirE2 possess functional nuclear localization signals that are important to guide nuclear targeting by interacting with two plant proteins present in the cytoplasm.

14

I

Woygang Schumann

Upon integration into the plant genome, the 13 genes of the T-DNAs are expressed. The nontranscribed regions of each gene possess many of the features of plant genes, including typical eultaryotic TATA and CAAT boxes, transcriptional enhancers and poly(A) addition sequences. One group of T-DNA genes directs the production of plant growth hormones that are responsible for the tumorigenic growth of the transformed plant cells. While two gene products are responsible for the production of auxin, another protein together with a host protein synthesizes the cytokinin zeatin. A second set of transferred genes is responsible for the production of the opines. Opines are subdivided into different groups called octopines, nopalines, agropines, mannopines, etc. The octopine synthase reductively condenses pyruvate with either arginine, lysine, histidine or ornithine to produce octopine, lysopine, histopine or octopinic acid, respectively. The opines are excreted from the plant cell, taken up by A. tumejaciens and serve as carbon, nitrogen and energy source. Opine uptake genes coding for permeases and opine catabolism genes are encoded by the Ti plasmid. It should be mentioned that there exists another closely related species of A. tumefaciens called A. rhizogenes where the pathogenic strains stimulate root proliferation thereby causing the hairy root disease. Here, pathogenicity is associated with the presence of the Ri (zoot-inducing)plasmid. 3.3

Heavy metal resistance

Microbial resistance to a wide range of toxic metal ions has been reported. This includes resistance to metals that are purely toxic, with no ascribed biological function, such as mercury and cadmium, and to metals that are toxic in excess but are required in small amounts for biological activities, sch as copper and zinc. Metal resistance systems are well known in many bacterial species. The genes governing these resistances are generally (but not always) found on plasmids and encode resistance to toxic metal(1oid)ions including Ag', As02-, AsO:-, Cd", Co2+,CrO;-, Cu2+, Hg2+, Ni2+, Pb2+, Sb3+,TeO:-, T1' and Zn2+. While the largest group of metal resistance systems function by energy-dependent efflux of toxic ions, some involve enzymatic transformations. As examples, the metal resistance mechanisms directed against cadmium, copper and mercury will be described. In gram-positive bacteria such as Staphylococcus aureus, Bacillus and Listeria cadmium resistance is caused by an efflux ATPase and in gram-negativebacteria such as Ralstonia by a three polypeptide chemiosmotic antiporter. The cadA operon represents one of the two known mechanisms of plasmid-mediated cadmium resistance in S. aureus and is associated with plasmid pI258. This plasmid codes within a 3.5 ltb operon for the two genes cadA and cadC. While cadA encodes a 727-amino acid protein that shows sequence similarity to the P class of ATPases (they are the only transport ATPases that have a covalent phospho-protein intermediate) and affords protection by functioning as an energy-dependent cadmium efflux ATPase, the CadC protein is smaller, consisting of 122 amino acids, and serves as a transcription regulator of the cadmium operon. Cadmium enters s. aureus through

I The Biology of Plasmids

an Mn*+-specificactive transport system and accumulates to toxic levels. The ATPase is an integral innner membrane protein with six predicted membrane-spanning segments. Upon binding of Cd2+to the N-terminal functional domain, the CadA protein is phosphorylated at an invariant aspartate residue from ATP leading to the activation of the transport activity. The best understood cadmium resistance mechanism in gram-negative bacteria is mediated by the Czc system of Ralstonia sp. strain CH34 (formerly Alcaligenes eutrophus) and codes for the cation diffusion facilitator protein family. The czc genes (for Cd2+,Zn2+and Co2+)are located on the indigenous megaplasmid pMOL30. The efflux system consists of the three proteins CzcA, CzcB and CzcC which are all integral membrane proteins and form a complex. While CzcA is inserted into the inner membrane and CzcC into the outer membrane, CzcB is anchored in both membranes and thereby bridges the two membranes of gram-negative bacteria. The trans-envelope efflux is mediated by the CzcCB2A protein complex, a proton-cation antiporter and is able to export Cd2+,Zn2+and Co2+and at the same times imports 2H' for each cation. Regulation of intracellular copper activity is crucially important for cell viability. Copper is an essential growth element through its function as cofactors in various redox enzymes such as lysyl oxidase, cytochrome c oxidase and superoxide dismutase. On the other side, copper is very toxic to proltaryotic cells. Genes involved in the control of cytoplasmic copper levels have been identified in different bacteria including Pseudomonas, Xanthomonas and E. coli. Regulation of cytoplasmic copper activity appears to involve influx and efflux pathways in addition to copper modification in the cytoplasm. In copper-resistant strains of Pseudomonas syringae pv. tomato, isolated from copper-treatedtomato culture, four plasmid-borne genes have been implicated in copper metabolism. Two of the gene products, CopA and CopC, were shown to be periplasmic copper binding proteins that accumulate in cells challenged with high copper levels, thus acting as copper scavengers. In E. coli, the conjugative plasmid pRJ1004, isolated from the gut flora of pigs fed on a diet supplemented with copper sulfate as growth promotant, confers copper resistance. The resistance determinant involves the pco gene cluster with the seven genes pcoABCDRSE. The proteins PcoS and PcoR show sequence similarity to two-component systems where PcoS senses Co2+,activates PcoR through phosphorylation, and the phosphorylated PcoR-P binds to an upstream motif in the pcoABCD pronioter region. The four structural proteins are the inner membrane protein PcoD, the outer membrane protein PcoB and the two periplasmic proteins PcoA and PcoC. PcoA and PcoC can bind 11 and 1 Cu2+ions, respectively, and are involved in storage of excess copper in the periplasmic space. How the two membrane proteins PcoB and PcoD are involved in copper efflux across the two membranes remains elusive. The best known metal resistance involving enzymatic transformation is mercurial resistance and involves one or two enzymes: mercuric reductase which converts soluble inorganic Hg2+to Hgo which is volatile and, therefore, rapidly eliminated from aerobic microbial cultures as a gas, and organomercurial lyase, which cleaves the Hg-C bond of more toxic methylmercury, phenylmercury and other organomercurials to less toxic inorganic Hg2+. In addition, all mercurial resistance sys-

16

I tems have genes for HgZftransport to bring extracellular Hg" Wovgang Schumann

into the cell, where mercuric reductase is found. The logic for this counter-intuitive finding of a transport system to bring a toxic compound into the cell is that extracellular Hg2+itself is highly toxic and needs to be chaperoned from the initial binding site outside the cell to the intracellular reductase enzyme that depends on the high-energy intracellular cofactor NADPH. A regulatory protein, MerR, provides tight control of expression so that the gene products are made only at times of need. MerR is a positively acting regulatory protein that binds to the transcriptional mRNA start site, and on addition of Hg2+ MerR twists and bends the DNA to a conformation suitable for opening and initiation of mRNA synthesis.

3.4 Other phenotypical traits

Most of the naturally occurring organic molecules can be easily degraded. Part of these organic molecules are aromatic and heterocyclic compounds and those containing halogenic substitutions. Many of these compounds can be degraded by soil and aquatic bacteria, which is called biodegradation. Analysis of bacteria active in biodegradation which use the organic compounds as carbon and energy source has revealed that they carry large degradative plasmids coding for the enzymes responsible for the catabolism of these compounds. The first of these plasmids to be discovered was the SAL plasmid (salicylateutilization) in 1972. Other plasmids are involved in the biodegradation of campher (CAM), benzoate and toluate (TOL),octane (OCT) and naphthalene (NAH). Most of these plasmids have been found in Pseudomonas putida. Soil bacteria of the genus Azorhizobium, Bradyrhizobium,Mesorhizobium and Sinorhizobium (collectively referred to as rhizobia) are capable of forming symbiotic associations with leguminous plants. Symbiotic associations may be specific and result in the formation of specialized root organs called nodules, in which rhizobia are released, where they form nitrogen-fmingbacteroids involved in reducing atmospheric dinitrogen to ammonia. In most legumes, nodule development begins when rhizobia in the rhizosphere bind to the root hairs of the legume provoking deformation and curling of the root hairs. This is followed by bacterial invasion of the root hair via an infection thread, division of cortical cells of the root, and development of the nodule. Signal exchange between the two symbionts largely controls specificity in these interactions. Legume roots secrete various flavonoids, some of which regulate the expression of nod (nodulation) genes in the target rhizobia. The genetic information for most of the genes involved in symbiosis are located on large plasmids called pSym plasmids the length of which ranges between 180 kb and 1600 kb. These plasmids carry the genes involved in nodule formation (besides nod two other groups called nol and noe) as well as those involved in nitrogen formation (ngandfix). To give one example, the 536 kb pSym plasmid pNGR234a of the broad-host range Rhizobium sp. NGR234 codes for 416 predicted open reading frames where most of them code for proteins involved in nodulation and nitrogen fxation.

7 The Biology of Plasmids

Bacteria have the possibility to recognize foreign DNA taken up by transformation, conjugation or transduction by their methylation pattern. If it does not match with that of their own DNA, the incoming DNA will be destroyed. Recognition and destruction occurs by so-called restriction-modification systems. These systems code for two enzymatic activities, an endonuclease and a modifying enzyme with methyltransferase activity. Both enzymes, forming either a complex or acting independently recognize a specific DNA sequence. If both strands within this sequence are methylated the DNA will not become attacked. If only one strand is methylated (which is the case immediately after replication), the second strand will be methylated. If both strands are unmethylated within the recognition sequence, the DNA molecules will be cut either within the or outside of the recognition sequence depending on the type the endonuclease. Most restriction-modification systems are encoded by the chromosome, but some have been found associated with plasmids such as EcoRI, EcoRII, EcoRV, PvuII and PaeR7. The EcoRI system is coded for by the plasmid pMBl which is identical with ColEl but a 2 ltb region which carries the genes coding for the restriction-modification system. Another group of important plasmids codes for enhanced protection against the lethal effect of UV light and alkylating compounds all of which cause DNA damage. The best studied of these plasmids is pICM101 with a size of 35 kb. This plasmid carries the two genes mucA and mucB forming one operon. The products of both genes exhibit significant similarity with the UmuC and UmuD proteins of E. coli which are involved in error-prone DNA repair. UmuD undergoes self-cleavage under the influence of activated RecA (RecK;) protein resulting in UmuD’. The UmuD’,UmuC complex now called DNA polymerase V together with Rec@< is capable of DNA synthesis along lesions in the template strand, and this process may result in incorporation of non-complementary nucleotides. Therefore, this process is termed error-prone. Based on the observations made with UmuC and UmuD, the mucA and mucB genes should also code for a DNA polymerase of the V type. A comparable new DNA polymerase is encoded by the Pseudomonas aeruginosa plasmid pMG2. The gram-positive, aerobic, spore-forming bacterium Bacillus thuringiensis produces proteinaceous parasporal inclusions during sporulation which contain one or more insecticidal delta-endotoxins also called Cry proteins. These proteins are toxic to insects of several orders including Lepidoptera, Diptera and Coleoptera, many of which are important in agriculture and forestry as well as in animal and human health management. The members of the cry gene family encode proteins which show homology in their primary sequence and probably have similar three-dimensional structures. Nontheless, Cry proteins show a great deal of host specificity, with each protein being toxic for only one or a few insect species. The crystal inclusions dissolve in the larval midgut, where one or more protoxins are released and proteolytically converted into smaller toxic polypeptides. Insect specificity is determined to a large extent, although not entirely, by the interaction of the gut protease-activated toxin with receptors on the target insect gut epithelial cells, Following binding to such a receptor, a toxin can insert into the epithelial cell membranes and form pores, eventually killing the insect.

18

I

Wolfgang Schurnann

4

The clinical importance o f plasmids

One of the most impressive examples of plasmid-driven evolution is the spread of bacterial antibiotic resistance since the middle of the last century. The main reason for the increase in resistant bacteria is the mobility of plasmids which is further enhanced by the presence of antibiotic resistance genes on mobile elements such as transposons, conjugative transposons and integrons. Work carried out over the last 10-20 years further revealed that many virulence factors are encoded by plasmids which again favors their spread. Virulence factors are involved in pathogenicity. 4.1 The spread of antibiotic resistance and the evolution of multiple antibiotic resistance

Early reports of bacterial resistance to sulphanilamide and penicillin occurred 1939 and 1941, respectively. Typically, resistance was due to spontaneous mutations which mapped to the bacterial chromosome and affected the site of action of the antibiotic. One prominent example are mutants of E. coli resistant to high levels of streptomycin where the mutation maps to the rspL gene coding for a component of the 30s ribosomal subunit. Most of the work on plasmid-borne antibiotic resistance was done in Japan studying the epidemiology of bacterial dysentery. Unfortunately, these observations were published in the Japanese language and, therefore, escaped the attention of the Western world. It was the Japanse scientist T. Watanabe who summarized all the data in an American journal in 1963. Plasmid-specified resistance to antimicrobial agents has become the most widespread mechanism of bacterial resistance to antibiotics and is mainly associated with the acquisition of an additional property usually in the form of an enzyme. In general terms, these enzymes modify antibiotics, circumvent the target, or result in reduced accumulation of the antibiotic. In contrast, chromosomally encoded antibiotic resistance generally results only in alteration of target sites. Sulphanilamide derivatives had been used against Shigella since 1945, but these were effective for only about five years before resistant strains appeared. By the early 1950s resistant strains occurred for 80-90% of clinical isolates. At the time, new antibiotics such as streptomycin, tetracycline and chloramphenicol promised a solution to the problem. While multiple resistant Shigella accounted for less than 0.02% of clinical isolates in 1955, 74% of the isolates were resistant by 1967. In addition, description of the transfer of multiple resistance from one bacterial species to another in the intestine of mice started in the mid-1960s.

1 The Biology of Plasmids

4.2

Transfer o f antibiotic resistance genes

m e rapid appearance of multiple resistant strains and evidence for the intra- and inter-speciestransfer of antibiotic resistance genes stimulated interest in the underlying mechanisms. When it became clear that most resistance genes were transmitted independently of the chromosome and to be lost from cultures during storage or upon treatment with certain chemicals such as acriflavin known to intercalate into the DNA, it could be shown that these genes are part of conjugative plasmids, then called R factors. Self-transmissible R factors of gram-negative bacteria are spread by the mechanism described for the F factor, but non-conjugative R factors can be mobilized provided a conjugative plasmid is present within the same cell. Later, it was noticed that antibiotic resistance genes could be either lost from an R factor or transferred to another replicon such as a coresident plasmid or the chromosome. It became apparent that most resistance genes are part of so-called transposons, mobile genetic elements. Transposons are DNA sequences that promote their own movement from one place in a replicon to another in the same or in a different replicon. Two different transposition mechanisms have been described, replicative and conservative. Replicative transposons are doubled by in situ replication, and the copy moves to a new place. Conservative transposons are excised from their original place and integrate at a new location. While the first mechanism leads to a duplication of the transposon, the second just changes the location of the genetic element. All transposons code for an enzyme which is involved in their transposition, the transposase, encoded by the tnpA gene. This enzyme recognizes both ends of the transposon, imperfect inverted repeats of up the 30 bp, introduces one nick at each end involving both strands. Then, the transposase selects a target site and introduces a staggered cut, where the two nicks are at a certain distance that is specific for each element. The transposon is next spliced into the target site, and the flanking regions are filled. Therefore, each transposon is flanked by a direct repeat the length ofwhich is characteristicof the particular element. Besides the transposase gene, the mobile element normally codes for a second gene whose product is involved in regulation of the transposition frequency by either interferring with the expression of the tnpA gene or with its activity. Most transposons encode antibiotic resistance genes, normally one (Tn3, Tn5 and TnZO code for resistance to ampicillin, kanamycin and tetracycline, respectively) and rarely several (Tn22 carries genes which confer resistance to sulphonamides and spectinomycin). In should be mentioned that transposons are divided into two classes according to their DNA structure. Simple transposons are flanked by imperfect inverted repeats (Figure 3A), while composite transposons are flanked by IS elements, which are themselves flanked by inverted repeats (Figure 3B). Composite transposons can arise when two IS elements flank at least one gene and transpose together. TnlO is the best-studied example. Here, both IS elements called IS10 left and IS20 right are able to transpose independently (which occurs rarely), or the whole Tn20 transposes as a unit.

I

79

20

I

woygang Schumafln

A

w

B +

C

int/

tnpA

Fig. 3. Schematic structure o f mobile elements. (A) Simple transposon with imperfect inverted repeats at i t s ends, the tnpA

l

.+ A br

Abr

+

Ab'

tnpA

.+

poson where the antibiotic resistance gene

catalyses the integration o f gene cassettes at receptor sites (attl, hatched regions),

Another important source of mobile antibiotic resistance genes are integrons. An integron is a specific integration and excision system for the capture, mobilization and expression of gene cassettes that usually encode antibiotic resistances. An integron consists of a gene for a site-specificrecombinase, intI, and a receptor site, attl, where the captured genes are inserted (Figure 3C). Capture of genes cassettes is performed by an intl-mediated recombination between the attl site of the integron and a 59 bp element (also known as attC) of gene cassettes. Integrons are unique among integrationlexcision systems, since multiple gene cassettes can be tandemly inserted into a single attl site and the intl gene is located adjacent to the attl receptor site, but not in a gene cassette. Plasmids are not the only DNA elements in gram-positivebacteria that are capable of transferring themselves. Some transposons, designated as conjugative transposons encode Tra functions to promote their own transfer. These transposons are normally integrated into a bacterial genome and excised only during the conjugal transfer process and have been described in three distantly related phylogenetic groups of bacteria: the gram-positive bacteria (Tn916, Tn927, Tn918 and Tn920 from Enterococcus faecalis, Tn1545 from Streptococcus pneumoniae and Tn919 from Streptococcus sanguis), the E. coli-heudomonas group of gram-negative bacteria and the Bacteroides group of anaerobic gram-negative bacteria (DOT, XBU4422). The conjugative transposons range in size from 18 kb to over 150 kb, are excised precisely and transferred as covalently closed double-stranded circles which are incapable of autonomous replication. Transfer of the circular intermediate starts with a single-stranded nick at the oriTand results in the transfer of a single-stranded copy of the conjugative transposon to the recipient cell where it will become double-stranded and then integrates into the chromosome. The best-studied example is Tn916 with a size of 18 kb, codes for tetracycline resistance ( t e t M ) and functions involved in its excision from a replicon, conjugative transfer and reintegration into a replicon in the recipient (chromosome or plasmid). In contrast to the classical transposons, integration of the conjugative trans-

7 The Biology of Plasmids

posons does not result in duplication of DNA sequences at their location. Conjugative plasmids can mobilize co-resident plasmids. While some transposons can mobilize plasmids only i n trans, others act i n cis. Tn916 belongs to the former group of conjugative transposons, and during in trans mobilization, the plasmid takes advantage of the mating apparatus encoded by Tn916. On the contrary, XBU4422, a Bacteroides conjugative transposon, mobilizes plasmids in cis. Here, the transposon first integrates into the plasmid, and then the whole element is mobilized to the recipient. 4.3

Mechanisms of antibiotic resistance

Plasmid-mediated antibiotic resistance is caused by one of four different mechanisms:

. cleavage or modification of the antibiotic (p-lactamantibiotics, aminoglycosides, chloramphenicol), . alteration of the target site for the antibiotic (MLS antibiotics), .. bypass decreased accumulation of the antibiotic (tetracycline), mechanism where the enzyme inactivated by the antibiotic is replaced by a new enzyme not acting as a target (trimethoprim).

/?-Lactamantibiotics such as penicillin, ampicillin and its numerous derivatives disrupt the synthesis of the cell wall peptidoglycan by acylation of the active site serine of the DD-transpeptidase. Resistance results from enzymatic hydrolysis of the /3lactam ring of penicillins and cephalosporins. The most common p-lactamases specified by plasmids are the TEM-1 and TEM-2 enzymes. Detoxification by p-lactamases is achieved in different ways by different bacteria. Gram-positive bacteria like Staphylococcus aureus liberate the p-lactamase into the growth medium and steadily destroy susceptible B-lactam molecules present. The level of resistance will depend on the amount of p-lactamase activity available which in turn depends on the number of bacteria. In gram-negative bacteria like Escherichia coli, most enzymatic destruction of p-lactams occurs in the periplasmic space. Under these circumstances, the level of resistance depends on the rate at which P-lactam molecules penetrate the outer membrane, and the rate of hydrolysis. Aminoglycosides such as ltanamycin constitute a large group of substances which are of considerable importance in the treatment of infections, and include streptomycin, ltanamycin, neomycin and amiltacin. They exert their effects through interaction with ribosomes. Plasmid-mediated resistance to aminoglycoside antibiotics is associated either with enzymatic modification of aminoglycosides or with apparent impermeability to these drugs in the absence of enzymatic modification. A large number of different enzymes exist which are capable of N-acetylation or 0-phosphorylation or 0-nucleotidylation of various aminoglycosides. Enzymes are classified by the nature of the transfer reaction which occurs and by the site of the aminoglycoside molecule on which modification takes place.

I

21

22

1

Wargang Schumann

The most widely described form of plasmid-specifiedchloramphenicol resistance has been associated with the presence of chloramphenicol acetyltransferase. This enzyme has been found in a wide variety of bacteria and in many instances is plasmid-specified.Enzymes have been classified on the basis of plasmid linkage, mode of synthesis (constitutive or inducible), subunit molecular weight, thiol reagent sensitivity and immunological class. Resistance to chloramphenicol has also been described in the absence of chloramphenicol acetyltransferase activity, and may be related with a reduced permeability to chloramphenicol in these types of strains. MLS antibiotics act by binding to the 50s ribosomal subunit, and the three letters stand for macrolide (erythromycin, leultomycin), lincomycin and streptogramin B. MLS resistance is widespread in gram-positive bacteria and is caused by the specific methylation of two adenine residues in the 23s ribosomal RNA. The plasmid pE194-encoded methylase acts on free 23s rRNA and on the 50s subunit, but not on the 70s particle. The tetracyclines are a group of antibiotics which are produced naturally by members of the genus Streptomyces and share a four-ring structure with varying substitutions at some carbon atoms. Tetracycline inhibits the binding of aminoacyl-tRNA to the A site on the 30s ribosome, but is somewhat toxic to humans because it also inhibits the eultaryotic translation apparatus. Three major classes of tetracycline resistance determinants have been recognized. One class which is present in gram-negative bacteria and exemplified by the already mentioned transposon Tn 10 involves the synthesis of the membrane-spanning TetA protein which mediates the efflux of the antibiotic. The second class found in gram-positiveaerobic bacteria including Bacillus and Staphylococcus, employs a different active efflux system. The third class acts by a different mechanism and protects ribosomes from the inhibition by the tetracycline. This system (TetM) is present in a diversity of bacteria including mycoplasmas, the gram-positive streptococci (here, resistance is mediated by the conjugative transposon Tn916) and gram-negative species such as Neisseria. The bypass mechanism is exemplified by trimethoprim and sulfonamides. Trimethoprim affects DNA replication by blocking the synthesis of deoxythymidine monophosphate (dTMP). This occurs by interaction with dihydrofolate reductase (DHR) which in turn produces tetrahydrofolate. This compound is needed in large amounts by the thymidylate synthetase enzyme which transfers a methyl group from tetrahydrofolate to the uracil ring of dUMP to make dTMP. Trimethoprim has a very high affinity for DHRs of many, but not all bacteria. Plasmidspecified resistance to trimethoprim is carried on several transposons and results in the production of a new DHR resistance to trimethoprim. Plasmid-specified enzymes differ distinctly from the chromosomally-specified enzymes in several properties. Sulfonamides probably exert their antibacterial activity by inhibiting the enzyme dihydropteroate synthase. This enzyme catalyzes the condensation between p-aminobenzoic acid and dihydropteridin. Inhibition of the enzyme relies on the structure of sulfonamides which act as structural analoga of p-aminobenzoic acid. Resis-

1 The Biology of Plasmids

tance against sulfonamides is caused by plasmid-encoded enzymes which are active even in the presence of high concentrations of the drug. 4.4 Bacterial virulence genes

Virulence genes allow pathogenic bacteria to adapt to their eultaryotic hosts and cause disease. Most pathogenic bacteria express their virulence genes only in the eultaryotic host, and some virulence genes are plasmid-encoded. Plasmid-encoded virulence determinants have been implicated in adhesion, high-affinity iron transport, serum resistance, toxinogenesis, host-cell penetration and invasiveness. These findings will be illustrated with three examples. The virulence plasmids of Salmonella enterica vary in their size, ranging from 50 kb to 285 kb, but they all invariably carry the spu (Salmonella plasmid virulence) plasmid. The spu genes occupying approximately 8 kb and consisting of six genes organized in two transcription units (SPURand s p v A B C D o f l are involved in infection of extra-intestinal tissues, such as mesenteric lymph nodes, spleen and liver, by increasing the growth rate of Salmonella in the intracellular compartment and for survival within macrophages. In addition, the plasmid possesses a number of other genetic determinants that contribute to virulence. These include the rck locus, which codes for a protein required for serum resistance and cell division, and the pef operon, which encodes a plasmid-encoded fimbrial structure which mediates adhesion to the murine small intestine and is required for fluid accumulation in infant mice. Many enteroinvasive Shigella strains harbor a 200 ltb virulence plasmid which codes for 32 genes within a 30 kb invasion region. These genes are arranged in two divergently transcribed regions. The first, referred to as ipa region, encodes the IpaB, IpaC, IpaD and IpC proteins, which have been shown to be essential for the entry process. The second region contains two blocks each of 11contiguous genes, called mxi (membrane egression of invasion plasmid antigen) and spa (Surface presentation of antigen). These mxi-spa genes code for components of a type I11 secretory system which are involved in the secretion of the Ipa proteins. Mutations have been created in most of the mxi and spa genes, and they all abolish the ability of Shigella to secrete Ipa proteins and to invade epithelial cells. The 70 kb virulence plasmid pYV enables all three Yersinia species (I:pestis, I.: enterocolitica and I: pseudotuberculosis)to survive and multiply in the lymphoid tissues of their host. It encodes the Yop virulon, an integrated system allowing extracellular bacteria to disarm the cells involved in the immune response, to disrupt their communications, or even to induce apoptosis by the injection of bacterial effector proteins. This system consists of the Yop proteins and their dedicated type I11 secretion apparatus, called Ysc. The Ysc apparatus is composed of some 25 proteins including a secretin. Most of the Yops fall into two groups. Some of them are the intracellular effectors (YopE, YopH, YplcA/YopO, YopP/YopJ, YopM, and YopT), while the others (YopB, YopD, and LcrV) form the translocation apparatus that is deployed at the bacterial surface to deliver the effectors into the eukaryotic cells,

I

23

24

I across their plasma membrane. Yop secretion is triggered by contact with eultaryoWofgang Schumann

tic cells and controlled by proteins of the vimlon including YopN, TyeA, and LcrG, which are thought to form a plug complex closing the bacterial secretion channel. The proper operation of the system also requires small individual chaperones, called the Syc proteins, in the bacterial cytosol. Transcription of the genes is controlled both by temperature and by the activity of the secretion apparatus.

5 Plasmid cloning vectors

A plasmid cloning vector is defined as an autonomously replicating DNA molecule into which foreign DNA can be inserted. This foreign DNA will then replicate passively with the cloning vector. Plasmids offer many advantages as cloning vectors since they are relatively small and easy to purify away from the chromosomal DNA. Most naturally occurring plasmids are not convenient cloning vectors, but they can be modified and should fulfill the following requirements: 1. It must be able of autonomous replication carrying an orivand the gene(s) in-

volved in its replication. 2. A cloning vector should be relatively small (around 5 kb) to allow easy isolation and introduction into recipient cells by transformation. 3. It should have a relatively high copy number to allow its easy purification in sufficient quantities. 4. It should carry an antibiotic resistance marker which is used to select for transformants carrying the plasmid. 5. It should have a few unique sites for restriction endonucleases to allow insertion of foreign DNA fragments. Many plasmid vectors contain other special properties that aid in particular experiments. Some carry DNA sequences recognized by phage packing systems (pac and cos, recognized by phages P1 and A, respectively). Mobilizable plasmids have a oriT site and can be transferred in the presence of a conjugative plasmid to other cells of the same or different species. Other plasmids contain strong promoters recognized by phage RNA polymerases such as those encoded by T7, T3 and SPG. These promoters can be used either in vivo or in vitro. Most cloning vectors are derived from pMB1, a ColEl-like plasmid. The prototype vector is pBR322 (Figure 4A). It is composed of the oriV from the ColEl derivative pMB1, the ampicillin resistance gene from the transposon Tn3 and the tetracycline resistance gene from the plasmid pSC101. It is considered a high copy-number plasmid with about 20 copies per chromosome. It carries a number of unique cloning sites within either the ampicillin ( M I , ScaI) or the tetracycline resistance gene (EcoRV, BamHI, N d ) .Insertion of DNA into a restriction site in either drug-resistance gene usually inactivates it and allows colonies bearing plasmids with such insertions to be identified by their inability to grow on medium with that antibiotic. This technique is called insertional inactivation.

1 The Biology of Plasmids

A

I Sca

I

-

RV

\

/

HI D-...

B

polylinker

puc19 pBR322

2.086 bp

4.361 bp

I Sca

D

C

I

Pst

I

Ap pBluescript SK (+/-)

\]

pACYC 177 3.940 bp

E

LotN

dlll

HI

Hin

Barn

T7+

If

--. -.--. -.

%.*

Fig. 4. Important prototype cloning vectors. (A) pBR322; (B) pUCl9; (C) pBluescript; (D) pACYCl77; (E) pBAC108L. The cosN sequence allows linerization by A terminase, rep€ codes for the protein that initiates unidirectional replication at orW, and parA and parB maintain the copy number o f the plasmid at 1-2 per chromosome.

I Not

I f *---*

fSP6

*.--

*.-*.-cI..**

I Sfi

I

25

26

I

Wolfgang Schurnann

The pUC vectors (Figure 4B) define a series of relatively small, versatile plasmid vectors containing the oriV region of pBR322 with the ampicillin resistance gene (derived from Tn3) and a sequence coding for the a-peptide of the lac2 gene. The different pUC vectors are distinguished by the polycloning site which is present at the beginning of the truncated lucZ gene. Insertion of DNA into one of the unique restriction sites leads to an interruption of the a-peptide gene, and no functional protein can be synthesized. The a-peptide gene ( l a c 2 ) plays an important role in a reaction called a-complementation. Active IJ-galactosidaseis a homotetramer, and each monomer consists of 1023 amino acid residues. Cells expressing only the a-peptide with a length of 146 amino acid residues cannot metabolize lactose or one of its chromogenic substrates such as X-gal (5-bromo-4-chloro-3-indolyl-/3-D-galactopyranoside).There is another mutant allele of lac2 called lacZMZ5 which carries an internal in-frame deletion removing the coding region for the amino acid residues 11-41. This p-galactosidase is also devoid of any enzymatic activity, but if E. coli cells express both the lacZMZ5 and the lacZ alleles, active P-galactosidase will be formed. This behavior is called a-complementation and can be visualized on plates containing X-gal. X-gal i s a colorless chromogenic substrate for P-galactosidase which i s converted into a blue indolyl derivative (indigo) through cleavage. E. coli colonies exhibiting a-complementation form blue colonies on Xgal plates, those which fail to express active p-galactosidase show white colonies. Therefore, successful cloning of restriction fragments into a pUC vector can be easily monitored by the color of the colonies. While transformants with empty vectors exhibit blue colonies, those with an insert form white colonies. This visual screening is also called blue-white screening. pUC vectors are present in about 1000 copies per chromosome and are, therefore, very high-copy number plasmids. pBluescript (Figure 4C) belongs to the group of phagemid plasmids. This group of cloning vectors consists of two replicons, a and a p l a s m replicon allowing their replication either as a plasmid or a phage. The phage replicon which is present in most phagemids is the filamentous phage M13 or one of its close relatives, fd or f1. This plasmid carries the oriVof pMB1, and either the fl(+)ori or fl(-) o r i and the lacZ gene with a polylinlter. Furthermore, it contains two different strong promoters recognized by the phage RNA polymerases T3 and T7, respectively. The phagemid with the fl(+)ori of replication allows for the recovery of the sense strand of the lacZ gene as single-stranded DNA, while those with the f l (-)oui facilitates recovery of the complementary strand. Recombinant singlestranded DNAs can be used, e. g., for sequence-specificmutagenesis experiments. The T3 and T7 promoters produce sense and antisense RNA, respectively, by in uitro incubation with the appropriate phage RNA polymerase. Plasmid pACYC177 (Figure 4D) is a low-copy number cloning vector which carries the oriv from plasmid pl5A enabling it to coexist with ColEl-derivative plasrnids. This feature of compatibility makes it useful for cloning experiments that require the presence of more than one recombinant plasmid per cell. Furthermore, pACYC177 is used to clone genes the products of which are toxic in high amounts to E. coli cells, e.g., all genes coding for integral membrane proteins cannot be stably inserted into pBR322. While pACYC177 carries antibiotic resistance genes

1 The Biology of Plasmids

for ampicillin and lianamycin its close relative pACYCl84 is equipped with those for chloramphenicol and tetracycline. Another important class of plasmids are the so-called mini-F vectors. These are small deletion derivatives of the F factor which replicate at 1-2 copies per chromosome. Based on the observation that the F factor stably replicates with several 100 kb of chromosomal DNA (the F’ factors mentioned above) the BACs (bacterial artificial chromosomes) were developed. BACs are min-F vectors they can stably propagate foreign DNA sequences of up to 300 kb. They are easy to handle and have proved to be relatively stable. As an example, the map of pBAC108L is indicated (Figure 4E). A particular group of vectors replicate not only in E. coli, but also in another species, e. g., Bacillus subtilis, Streptomyces, Saccharonzyces cereuisiae or animal cells. This type of vectors are called shuttle vectors. A shuttle vector has two origins of replication, one that functions in E. coli and the other that is required for replication in the second host. Cloning and analysis of the inserted fragments occurs in E. coli, and the desired recombinant plasmid is then shuttled into the second organism.

6 Perspectives

Where do plasmids come from? Are they parasites or symbionts? Plasmids may derive from bacteriophages which lost their ability to package their genome, or phages have evolved from plasmids by the acquisition of their replicon. But these possibilities are not mutually exclusive. It can also be envisaged that plasmids evolved from chromosomal replicons through transformation. Lysis of bacterial cells in their natural habitats occurs quite frequently leading to fragmentation of their chromosome. These fragments can be taken up by competent bacteria, and if such a fragment containing the oriC (the bacterial origin of replication is called o r i C ) is able to replicate in the new species a plasmid-like element will evolve. Are plasmids symbionts conferring increased fitness to its host or just selfish DNA parasites which proliferate at the expense of the host? Plasmids are totally dependent on their host for replication, transcription and translation. On the other hand, plasmids code for functions which allow bacteria to survive stress. In this sense, plasmids have to be considered as symbionts. What about cryptic plasmids? These seem to be entirely parasitic, but this view might be misleading because of two reasons. First, cryptic plasmids may encode functions not yet discovered, and second, these plasmids may be novel genetic elements just escaped from phage or chromosomal replicons and not yet aquired a discernible marker gene. In any case, plasmids are important agents of gene flow within and between bacterial species via conjugation and mobilization. Furthermore, mobile genetic elements such as transposons and conjugative transposons further increase the exchange of genes between the chromosome and the plasmid. Plasmids provide a reservoir of genetic information which compensates for the restricted genome

28

I size of prokaryotes, and, in addition, provide the raw material for astonishingly Worgang Schumann

rapid adaptation of bacterial populations faced with changing environments. Acknowledgements

The authers would like to thank Ingo Schumann for preparing the illustrations and the Deutsche Forschungsgemeinschaft and the Fonds der Chemischen Industrie for continuous financial support.

References CLEWEL, D. B. (1993), Bacterial Conjugation. Plenum Press, New York. CORNELIS, G. R., BOLAND, A., BOYD, A. P., GEUIJEN,C., IRIARTE, M. et al. (1998), The virulence plasmid of Yersinia, an antihost genome, Microbiol. Mol. Bid. Rev. 62, 1315-52. GUINEY,D. G., FANG, F. C., KRAUSE,M., LIBBY, S., BUCHMEIER, N.A., FIERER,J. (1995), Biology and clinical significance of virulence plasmids in Salmonella serovars, Clin. Infect. Dis. 21 (Suppl. 2), S146-51.

SALYERS, A.A., SHOEMAKER, N. B. (1997), Conjugative plasmids, Gen. Eng. 19, 89-100. SCHUMANN, W (1990), Biologie bakterieller Plasmide. Vieweg Verlag, Braunschweig. SILVER,S., PHUNG,LE T. (1996), Bacterial heavy metal resistance: new surprises, Annu. Rev. Microbiol. 50, 753-789. SUMMERS, D. I<. (1996), The B i o l o ~ojPlasmids. , Blackwell Science, Oxford. ZHU, J, OGER,P. M., SCHRAMMEIJER, B., HOOYICAAS, P. J. J., FARRAND, S. I<., WINANS, S. C. (ZOOO), The bases of crown gall tumorigenesis, J . Bacteriol. 182, 3885-3895.

P/asmids for Therapy and kchation by M.Schleef Copyright 0 WILEY-VCH Verlag GmbH, 2001

2 Structures o f Plasmid DNA Torsten Schmidt':" Karl Friehs and Erwin Flaschel

1 Introduction

Plasmids are extrachromosomal, double-stranded DNA molecules with a size from 200 up to more than 100,000 base pairs, which naturally exist in both, eultaryotic and proltaryotic cells. Plasmids replicate independently from the chromosome of the host strain. They encode a variety of phenotypic functions which are not essential for the host organisms, but give them properties for a higher adaptability to distinct environments, e. g., antibiotic resistance or metabolism of special compounds. Plasmids with inherent functions encode for such being responsible for replication and stability. Using modern gene technology, a multitude of artificial plasmids, like pBR322 or the pUC family, were designed in the laboratories from naturally occurring species. In biotechnology these plasmids are used as cloning vectors for recornbinant DNA technology and the production of recombinant proteins in prokaryotic and eultaryotic hosts. Plasmid vectors are smaller than natural plasmids (typical size < 10,000 base pairs) and consist of a replication system, one or more selection marker genes and a multiple cloning site (Figure 1).The replication system includes the origin of replication (ori) which regulates the plasmid copy number. Low-copy plasmids with copy numbers of up to 10 and high-copy plasmids with copy numbers of many hundred plasmids per cell must be distinguished. The selection marker genes are mostly antibiotic resistance genes which allow selection of the recipient cells containing the plasmid. The multiple cloning site with different restriction sites for endonucleases permits the introduction of genes for the desired plasmid-encoded, recombinant product. In gene therapy or genetic vaccination, plasmids are used as non-viral vectors for the transfer of therapeutic genes.

30

I

Torsten Schmidt et al

, NdeI

Eco0109; AatU. I I

.

I

j, ;

.BbeI . :,me1

ApoI ECOFU

BanII EC1136 Sac1 Asp718 KpnI AvaI SmI XInaI BamHl XbaI AccI HincLI SalI BspMI S~8387 es tI SphI Hindm

2

Topological structures o f plasmids

Double-stranded DNA has a helical structure. In addition to this inner structure DNA and especially plasmids of identical nucleotide sequence occur in different shapes and forms which differ in topology and size. The overwhelming majority of plasmids isolated from bacterial cells are covalently closed circles (ccc) which are negatively supercoiled and called “ccc form” or “form I” (Sinden, 1994). As shown in Figure 2, supercoiled plasmids have a compact structure, the DNA double-strand helix of which is wound around itself. Both strands are intact and thus covalently closed. The breakage (nicking) of one DNA strand by nucleases or mechanical stress results in an open circular form (oc form) by loss of the molecular coiling. The oc form, which is also called “form II”, is totally relaxed and thus less compact. However, a relaxed covalently closed circular form without supercoiling would also be possible, but this form cannot be distinguished from the oc form by analytical methods. Linear DNA molecules (form 111) result from breakage of both strands at the same position or by restriction endonuclease cleavage. In nature linear plasmids are preferably found in bacteria such as Borrelia, Streptomyces, Trtiobacillus, Nocardia, Rhodococcus and even Eschevichia (Hinnebusch et al., 1993).

2 Structures of Plasmid DNA

Fig. 2. Topological structures of

plasmid DNA.

ccc Form

linear Form

Catenane

oc Form

Concatemer (oc Dimer)

DNA knot

In addition, the topological plasmid forms 1-111 described here can exist as equivalent oligomeric forms which are called concatemers in this case. Dimeric plasmid molecules with doubling in size are found very often in plasmid preparations. Concatemers can arise during replication, but mainly they are formed after replication by homologous recombination (Summers et al., 1993).The formation of oligomers is plasmid-specific.It is supposed that the oligomer-monomer ratio increases with the size of the integrated DNA fragment (Summers et al., 1993). Oligomers also exist in ccc and oc forms. Treated with restriction enzymes, they are disintegrated into linear monomers. Furthermore, plasmids can be intertwined. If monomeric plasmids consisting of isolated circular double strands are interlocked as chain links, they are called catenanes (Kreuzer et al., 1980) or concatenates (Martin, 1996).Catenanes arise during plasmid replication. DNA h o t s are single molecules, where accidentally the DNA double strand is interwoven in itself. Typically they will not be found in plasmid preparations, but they can be synthesized in vitro using DNA topoisomerases (Dean et al., 1996).

I

31

32

I

Torsten Schmidt et al 3

Supercoiling of DNA

DNA topoisomerases catalyze supercoiling and relaxation of circular DNA. These nuclease-active enzymes are divided in two groups. Type I topoisomerases introduce temporary single-strand breaks in DNA. Thus, supercoiled molecules can be relaxed by removal of supercoils (Luttinger,1995).Type I1 topoisomerases cleave both DNA strands for different functions. The enzymes called DNA gyrases which are found in bacteria only introduce supercoiling in DNA molecules by ATP hydrolysis (Gellert et al., 1976). Other type I1 enzymes can create or also resolve DNA knots or catenanes (Krasnow et a]., 1982). Two distinct topological types of supercoiling are possible - a positive or negative one. The majority of circular plasmids in bacteria is negatively supercoiled, which is the opposite rotation to the right-handed double helix structure (Travers, 1993). Through negative supercoiling the DNA is underwound, because the number of helical turns is smaller than in relaxed or linear DNA molecules. The underwinding in the molecule creates a torsional tension. This torsional tension drives the temporary partial separation of the DNA double strand energetically which is necessary for DNA replication, recombination and transcription (Sinden, 1994). Supercoiling of chromosomal and plasmid DNA in bacteria is driven by the ATPdependent DNA gyrase. The degree of supercoiling is regulated by the opposing activities of DNA gyrases and DNA topoisomerases (Drlica, 1992). The balance of these two activities keeps the circular DNA at a precisely tuned level of supercoiling in the cells. Generally, one negative turn per 200 base pairs is found in supercoiled circular DNA, e. g. the 4362 base pairs sized plasmid pBR322 contains 20 to 25 negative supercoils (Goldstein et al., 1984) which depends on cultivation temperature. In E. coli the degree of negative supercoiling increases with increasing temperature (for a review, see Tse-Dinh et al., 1997). In eukaryotic cells DNA is wound around histones resulting in negative and positive supercoils in the same molecule. The positive supercoils will be relaxed by topoisomerases of types I and 11, so at least the molecule is only negatively snpercoiled (Sinden, 1994).

4

DNA intercalating dyes

Certain planar aromatic molecules are able to intercalate into double-stranded DNA. In this process the intercalating molecule inserts between two neighboring base pairs in the center of the helix by hydrophobic and electrostatic interaction. Some intercalators as ethidium bromide form fluorescing complexes with DNA which enables the visualization of DNA by irradiation with UV light. However, the intercalation of these molecules changes the form and shape of DNA. The DNA helix is stretched and unwinds in the presence of ethidium bromide by 26 per intercalating molecule as an effect of the increasing distance of O

2 Structures of Piasmid DNA

neighboring base pairs (Pulleyblank et al., 1975). If the DNA is negatively supercoiled, unwinding of the DNA helix results in partial relaxation of the molecule. With an increasing number of intercalated ethidium bromide molecules the DNA molecule becomes fully relaxed and at last positively supercoiled. The binding of intercalating molecules is proportional to the negative supercoil density of the DNA. Thereby, it is possible to determine the supercoil density of DNA by dye titration of supercoiled DNA and analysis by agarose gel electrophoresis (DeLeys et al., 1975, Iceller, 1975), by viscometric titration (Smit et al., 1971) or sedimentation in an ethidium bromide-sucrose gradient (Bauer et al., 1968).

5

Analysis of plasmid structures

The routine analysis of plasmid structures became a field of interest since plasmids were used as vectors for therapeutic genes in gene therapy or DNA vaccination. The plasmid DNA for clinical trails is manufactured in Escherichia coli host strains following GMP (good manufacturing practice) rules (see also chapter 11).On an industrial scale plasmids are purified from bacterial cells using anion exchange (Colpan et al., 1999, Bussey et al., 1998), reversed-phase (Green et al., 1997) or size exclusion chromatography (Horn et al., 1995; for a review, see Schleef, 1999). In these process steps bacterial contaminants, like proteins, RNA, genomic DNA and endotoxins (lipopolysaccharides)have to be removed significantly. In the final product the residual amounts of these contaminants as well as homogeneity are important quality criteria for the use of DNA in clinical trials (e.g., Schorr et al., 1995, Marquet et al., 1997). Homogeneity refers to the fact that plasmids can exist in different topological structures as described before. According to the recommendations of regulatory bodies, e. g., the Food and Drug Administration (FDA) in the United States or the Paul Ehrlich Institute in Germany, therapeutics based on plasmid DNA must have an adequate homogeneity actually characterized by a content of at least 90 % ccc form (see also chapters 11 and 12). Therefore, exact methods for determining the form distribution of plasmid samples are of great interest. The following techniques have been used to determine the distribution of plasmid isoforms. 5.1 Electron microscopy (EM)

DNA molecules can be visualized by electron microscopy after treatment with urany1 acetate and contrasting with Pt/C. Figure 3 shows an EM picture of a 5.7 kbp pCMV-S2S plasmid sample (Michelet al., 1995)in a 15,000-foldenlargement. Four different plasmid forms can be identified, which differ in size and conformity: supercoiled ccc monomers and dimers, relaxed oc monomers and dimers. The difference in plasmid size can be demonstrated by irradiation of this sample with UV light. The irradiation results in single-strand breaks so that the

I

33

34

I

Torsten Schmidt et a/.

Fig. 3. Electron micrograph of a pCMV-S2S (5.7 kbp) plasmid DNA sample (15,000 fold enlargement).

Fig. 4. Electron micrograph of the same pCMV-S2S (5.7 kbp) plasmid DNA sample shown in Figure 3 after irradiation with UV light (6,000-fold enlargement).

2 Structures of Plasmid D N A

supercoiled ccc molecules relax to an open circular topology. An EM image of this irradiated sample shows oc monomers and oc dimers with a doubling in size (Figure 4). The dimers are concatemers, because they consist ofone double strand only. Although the plasmid molecules can be visualized directly, the electron microscopy method is not suitable for a routine analysis of plasmid DNA forms, because this analytical method is expensive in time and material. Another aspect is a great uncertainty in the statistical evaluation of EM images: The photos represent a very small part of the sample with a few molecules only. 5.2

Agarose gel electrophoresis (AGE)

The most common method used to determine plasmid form distributions is agarose gel electrophoresis. In agarose gel electrophoresis the different plasmid structures are not visualized directly, but they are separated in the presence of an electric field force owing to their different mobilities in polymer matrices. Nucleic acids are negatively charged, because of their sugar phosphate backbone. In an electric field the nucleic acid molecules migrate to the positively charged anode. Polymer matrices like agarose or polyacrylamide gels form a network hindering the mobility of large molecules, so that differences in size or topological structure of nucleic acids have a great influence on the migration velocity of these molecules. When a plasmid DNA sample is analyzed by this electrophoretic method, the dig ferent bands in the agarose gel may be assigned to different plasmid sizes and topologies, like ccc, oc or linear plasmid molecules. The assignment of bands to plasmid forms, however, is not at all easy, since the electrophoretic mobility of plasmids of different structures changes with the electrophoresis operating conditions. This is comprehensively described by Meyers et al. (1976), Johnson et al. (1977), Senver et al. (1984) and Garner et al. (1992). Plasmids of the most compact ccc monomer form are migrating fastest, but the electrophoretic mobility of linear and oc structures depend on the electrophoretic conditions, like gel concentration, electric field strength, temperature, ionic strength of the electrophoresis buffer and dye concentration. In addition, the migration order changes when the electrophoresis is performed in the presence or absence of an intercalating dye. Thus, a direct assignment of the distinct bands to different plasmid structures is not possible in the case of agarose gel electrophoresis. Therefore, untreated, linearized and UV irradiated plasmid samples have to be analyzed in parallel (Schmidt et al., 1999). An untreated sample of pUC19 DNA (2.7 kbp) typically shows one major band corresponding to the ccc form and a slower migrating band which is commonly thought to be the oc form (lane 2 in Figure Sa). However, this is not true. UV irradiation creates single strand breaks in DNA molecules (nicking). Analysis of such a UV treated sample (lane 4) shows two major bands with different electrophoretic velocities than those in lane 1. In consequence, the slower migrating form in lane 2 cannot correspond to the oc form, but can be identified as ccc dimer. The UV irradiated sample consists of oc monomers and oc dimers, too. The reason is the appearance of linear dimers after partial restriction by endonucleases. Com-

I

35

36

I

Jorsten Schmidt et a/.

B

A 1

2

3

4

1

2

3

4

14140 8548 7242 6369 5686 4822 4324 3675

YOC

Dimer

lin. Dimer -ccc

-

Dimer

-

lin. Monomer

+ oc Monomer 2323 1929

ccc Monomer

1371

Fig. 5. Analysis o f different pUC19 plasmid samples (300 n g each) by agarose gel electrophoresis (1.0%). Electrophoresis was performed in TAE buffer at 5 V c m ’. DNA was stained with 1 m g L-’ ethidium bromide for 30 min after

electrophoresis. (A) 1: 2 BstEll molar-mass standard, 2: untreated pUCl9, 3: pUC19 linearized with EcoRI, 4: pUC19 nicked by UV light; (6) 1: 1 U EcoRI, 2: 0.1 U EcoRI. 3: 0.01 U EcoRI, 4: 2 BstEll molar-mass standard.

pletely digested samples show only one band (lane 3 in Figure 5a and lane 1 in Figure 5b) corresponding to the linear monomer. When lowering the reaction time and/or the amount of restriction activity, a new band occurs in addition of the ccc monomer and the linear monomer (lanes 2 and 3 in Figure 5b). This new band corresponds to the linear dimer, which has been identified owing to a similar mobility as the 5.6 ltbp fragment of a molar mass standard. The order of migration of the different plasmid structures on agarose gels under these analytical conditions is as follows: the monomeric ccc form is the fastest plasmid structure followed by the nicked oc monomer, and the linear monomer form. Subsequently, the dimers follow in the order: ccc, oc and linear form. However, as described above this order of migration is not always the same. In general, the resolution of the different plasmid structures in agarose gel electrophoresis decreases with increasing plasmid size. The resolution between oc monomers and ccc dimers becomes insufficient with increasing plasmid size (see Figure 6). The quantification of the different plasmid structures is based on the fluorescence intensity of the distinct bands in the agarose gel. The DNA is stained with intercalating dyes, like ethidium bromide, and these dye-DNA complexes show fluorescence by irradiation with UV light of a transilluminator (A = 366 nm).

2 Structures of Plasmid D N A

mide at 5 V cm I . Densitometric determination of band volumes was performed using a gel documentation system with a 366 nm transilluminator.

Fig. 6. Analysis of different amounts of pCMVS2S plasmid DNA by agarose gel electrophoresis (0.8%). Electrophoresis was performed in TAE buffer containing 1 mg L - ’ ethidium bra-

Figure 6 shows the electrophoretic separation of different amounts of the pCMV-S2S sample (same as Figure 4) on an agarose gel stained with ethidium bromide. This densitometric quantification of the agarose gel has the major disadvantage to work insufficiently. The linearity of the ratio of the DNA amount to the fluorescence intensity of the bands is limited to a small range of masses from 125 ng up to 2 ,ug (Figure 7). This corresponds to DNA concentrations from 12.5 mg L- to 200 mg L because sample sizes of 10 pL are commonly loaded onto the gel. At higher concentrations the bands appear blurred or oc monomers cannot be separated from ccc dimers - as observed in particular for larger plasmids. At higher DNA concentrations the fluorescence intensity reaches a saturation level, whereas at lower concentrations no band is detected. Generally, the gel staining requires experience and shows low reproducibility.

’,

1000:

a’

F’;

y ;v/’

;; 100:

5

v

v

8,’p/’

U

Fig. 7. Correlation of the amount of DNA and the band volume of each

0

0

g L 7

0

m5

B

g

10:

- v 1

I

I , , , ,

i”’

cccMonomer 0

ccc Dimer, oc Monomer

v

oc Dimer I

I

37

38

I

Jorsten Schmidt et a/.

5.3

Capillary gel electrophoresis (CCE)

A novel powerful routine technology for the quantification of plasmid structures is capillary gel electrophoresis (CGE) (Schmidt et al., 1999). With this automated technique the electrophoretic separation of different plasmid structures takes place in thin capillaries (100 pm inner diameter) filled with a liquid polymer solution. For DNA analysis the capillary is neutrally coated and has a typical length from 30GO cm. It is surrounded with a cooling system, so that high voltages up to 30 kV can be supplied for the electrophoretic separation resulting in short analysis times. Small amounts of a sample are injected into the buffer-filled capillary hydrodynamically by pressure or electrokinetically by an electric field. During electrophoresis both ends of the capillary each dip together with each electrode in the two buffer vials. In the presence of an electric field the plasmid molecules migrate through the polymer solution and pass the detector window at the end of the capillary. Detection of DNA is best performed by laser induced fluorescence (LIF) using an argon ion laser (A = 488 nm). Therefore, the electrophoresis buffer and the DNA sample have to be supplemented with novel intercalating fluorescent dyes like YOYO (Glazer et al., 1992). In DNA analysis the detection of UV absorbence is also possible, because the nucleotide bases have an absorption maximum at A = 254 nm, but the sensitivity is 10,000 times less in comparison to LIF detection (Zhu et al., 1994). The separation of DNA molecules, e. g., plasmid structures in gel-filled capillaries results from their different sizes and shapes. In most cases the replaceable non-crosslinked gel in the capillary is a diluted aqueous polymer solution of cellulose derivates or linear polyacrylamides in electrophoresis buffer (e. g., Kleemiss et a!., 1993, Courtney et al., 1995, Ulfelder et al., 1992, Hammond et al., 1997). Therefore, the sieving effect of the liquid gel is determined by chain length, concentration and the kind of polymer. The resolution of large DNA molecules in dilute polymer solutions increases with increasing polymer chain length and decreasing concentration (Barron et al., 1996). Simple separations of plasmid structures by CGE have first been reported by Courtney and coworkers (1995). Using linear polyacrylamide as a separation matrix this method was used for the determination of ligation efficiency. Thereby, linearized DNA was recircularized by a ligase to open circular plasmids. The separation of these forms could be demonstrated for a 3.7 kbp plasmid. Nackerdien et al. (1996) used the same sieving gel for the analysis of laser-induced DNA photolysis with capillary gel electrophoresis. Using a 3.9 kbp plasmid separation of the linear form from the circular forms could be demonstrated. Hammond et al. (1997) reported the separation of a supercoiled plasmid DNA ladder from 2 kbp to 1G kbp by CGE in a diluted hydroxyethyl cellulose. The separation and determination of all plasmid structures which are found in a typical bacterial plasmid preparation has first been described by the authors (Schmidt et al., 1999). In quality control for therapeutics based on plasmid DNA, this CGE method allows the determination of the structural homogeneity

CGE electropherogram o f a mixture o f untreated, linearized and UV irradiated pUC19 plasmid DNA (2.9 kbp). The plasmid sample was pre-stained in 1.44 pM YOYO solution. CCE was performed in a 3 7 cm capillary filled with 0.1 % hydroxypropylmethylcellulose run buffer containing YOYO a t 100 v cm-’ as described by Schmidt et al. (1999). Fig. 8.

i

2 Structures of Plasmid D N A

39

ccc Monomer

linear Monomer

h

K

Monomer OC

10 0

U

20

Wh

30 Migration time I rnin 25

35

(www.CGEservice.com),while in other CGE methods described the dimeric concatemers are not considered. Figure 8 shows an electropherogram of a mixture of untreated, UV irradiated and linearized pUC19 plasmid samples, as described in Figure 5. All monomeric and dimeric plasmid structures can be separated with baseline resolution using diluted HPMC (hydroxypropylmethylcellulose)solution as the liquid gel matrix in the capillary. The order of migration in CGE is determined by analyzing the differently treated pUC19 samples separately and not as a mixture. The electrophoretic mobilities of the plasmid structures in CGE differ from those described for the AGE method. In the case of AGE, all monomeric structures (ccc, oc, and linear) of pUC19 migrate faster than any dimeric form (Figures 5A and 5B). Thus, the order of migration is mainly determined by the molecular size. In contrast, in CGE the ccc forms of monomer and dimer appear earlier than the linear monomer and the linear dimer, followed by the oc forms of the monomer and dimer. Using the CGE method for analyzing plasmid structures the topology of the plasmids govern the order of migration. This rule for the CGE order of migration in diluted HPMC solution also holds for larger plasmids - as shown in Figure 9 for pCMV-S2S (5.7 kbp). The most compact ccc structures migrate faster than linear and relaxed open circular structures. For larger plasmids the resolution of monomeric and dimeric oc structures decreases until only one signal is obtained. The reduced resolution of oc forms is no 501

Fig. 9. CCE electropherogram o f a mixture of untreated, linearized and UV irradiated pCMV-S2S plasmid DNA (5.7 kbp). The plasmid sample was pre-stained in 1.44 pM YOYO solution. CGE was performed in a 47 cm capillary filled with 0.1 % hydroxypropylmethylcellulose run buffer containing YOYO a t 100 v cm-’ as described by Schmidt et al. (1999).

I

ccc Monomer

401

< 303 LL

.

LT 20-

linear oc Forms Monomer

100 -

1

40

40

I

Torsten Schmidt et a/.

problem for quality control of plasmid DNA for clinical trials. Both oc forms are undesired in clinical-grade plasmid preparations. Therefore, it is not relevant, if these forms are quantified separately or together. In AGE a resolution problem appears with respect to oc monomer and ccc dimer structures, which is a major disadvantage. Thus, in general CGE is superior to AGE for analyzing different plasmid structures. In addition, it offers higher sensitivity and reproducibility in quantification. Linearized plasmid DNA can be quantified with CGE in 10 min (Schmidt et al., 1996). The corrected peak area correlates linearly with the DNA concentration over 3 orders of magnitude. This fast and reliable quantification offers an in-time analysis of the plasmid copy number during bacterial cultivation. Contrary to chromatographic methods, where different molecules pass the detector with always the same velocity, the velocities of the molecules passing a CGE detector are different, due to their mobility in the electric field. Therefore, in quantification of DNA with CGE, the integrated detector signal has to be divided by the migration time resulting in a corrected peak area, which is proportional to the DNA Concentration. To quantify the different structures of plasmid DNA, a linear correlation between the corrected peak area and the concentration of each structure should be demonstrated (Figure 10). By analyzing different amounts of pCMV-S2S plasmid DNA samples containing both ccc and oc forms a wide range of DNA concentration from 0.06-4.0 mg L-' shows an excellent linear correlation with the corrected peak area for each plasmid structure. The supercoiling ratio of this plasmid sample was determined to be 86 f 1.5 % independent of the concentration applied. With CGE a routine analysis and quantification of different structures for quality control of pharmaceutical-gradeplasmids for gene therapy or genetic vaccination is feasible in a fast, automated and highly reliable way. In addition this analytical method allows the analysis of plasmid samples during the cultivation of plasmid-bearing cells, during the isolation and purification of plasmid DNA and during the formulation of plasmid based therapeutics. It allows the establishment of quality assurance standards for plasmid DNA form distribution. In comparison to

1

DNA concentration / mg L-'

10

Fig. 10. Correlation of DNA concentration and corrected peak area o f each plasmid structure of a pCMVS2S sample determined by CGE. CCE operating conditions were as described i n Figure 9.

2 Structures of Plasmid D N A

AGE, CGE offers high resolution, high sensitivity and the amount of DNA required is very low. Independent from plasmid size, always the same order of migration was obtained with CGE, thus simplifying the identification of CGE signals. 5.4 Analytical chromatography

Analytical chromatography methods for analyzing and quantifying different plasmid structures have rarely been described. HPLC methods based on anion exchange are often used for the quantification of purified plasmid DNA in industrial manufacturing processes (Chen et al., 1997; Lahijani et al., 1996). This quantification of total plasmid DNA by HPLC can be performed in the presence of other nucleic acid contamination, like RNA or genomic DNA (Ferreira et al., 1999). The separation and quantitative analysis of different plasmid structures using analytical anion exchange chromatography is shown by Marquet et al. (1995). The chromatogram shows the separation of supercoiled plasmid molecules from relaxed ones. Hydrophobic interaction chromatography (HIC) is able to separate relaxed and supercoiled molecules for quantification. At certain salt concentrations, the plasmid structures bind different to HIC resins (TosoHaas, Stuttgart, Germany). With both methods, a separation of different plasmid topologies is possible. However, the separation of monomers and dimers i s not considered. In consequence, HPLC is not the exclusive analytical tool for the assessment of homogeneity in quality control of pharmaceutical-grade plasmid DNA for gene therapy or DNA vaccination.

6

Conclusion

Plasmid DNA, commonly produced in bacteria, may appear in different structures differing in topology (supercoiled ccc, linear or nicked oc) and size (monomeric, dimeric, etc.). The quantification of these different structures is an important quality criterion for the homogeneity in clinical-gradeplasmid manufacturing (see also www.PlasmidFactory.com).A novel method based on capillary gel electrophoresis (CGE) allows the routine determination of structural homogeneity of plasmid DNA by separation of all plasmid structures. This CGE method allows an automated, easy, fast, and highly reliable quantification. CGE offers high resolution in addition to low amounts of DNA samples. Classical agarose gel electrophoresis is inferior to this CGE method, because the resolution as well as the quantification is poorer, less reproducible and less sensitive. HPLC is no exclusive method for analyzing plasmid structures, but it is an excellent complementary method to CGE. The reliable analysis of plasmid DNA will be performed with CGE technology, especially if this plasmid DNA may be used for pharmaceutical applications. Quality assurance, in-process control and stability studies require this technology immediately.

I

41

42

I

Torsten Schmidt et a/.

References

BARRON,A. E., SUNADA,W. M., BLANCH,H. W. ing numbers vary with growth temperature, (1996), The effects of polymer properties on Proc. Natl. Acad. Sci. U S A 81, 4046-4050. DNA separations by capillary electrophoresis GREEN,A. P.,PRIOR,G . M., HELVESTON, N. M., in uncross-linked polymer solutions, Electro- TAITTINGEK B. E., LIU, X., THOMPSON, J. A. phoresis 17, 744-757. (1997), Preparative purification of supercoiled plasmid DNA for therapeutic applications, BAUER,W, VINOGRAD,J. (1968),The interacBiopharm 10, 52-62. tion of closed circular DNA with intercalating HAMMOND, R. W., OANA,H., SCHWINEFUS, J. J., dyes, /. Mol. Biol. 33, 141-171. BUSSEY,L. B., ADAMSON, R., ATCHLEY, A. (1998), BONADIO,J., LEW, R. J., MORRIS,M. D. Methods for purifying nucleic acids, WO 98,’ (1997), Capillary electrophoresis of super05673. coiled and linear DNA in dilute hydroxyethyl COLPAN,M., SCHORR,J., MORITZ,P. (1999), cellulose solution, Anal. Chem. 69, 1192Process for the separation and purification of 1196. HINNEBUSCH, J., TILLY,I<. (1993), Linear plasnucleic acids from biological sources, US 5990301. mids and chromosomes in bacteria, Mol. COURTNEY, B. C., WILLIAMS,I<. C., BING, Q.A., Microbiol. 10, 917-922. SCHLAGER,I. J. (1995), Capillary gel electro- HORN,N.A., MARQUET,M., MEEK,J. A., phoresis as a method to determine ligation BUDAHAZI,G. (1995),Cancer gene therapy using plasmid DNA: purification of DNA for effency, Anal. Biochem. 228, 281-286. DEAN,F. B., STASIAI~, A,, KOLLER, T., COZZAhuman clinical trials, H u m . Gene Tner. 6, RELLI,N. R. (1985), Duplex DNA knots pro565-573. duced by Escherichia coli Topoisomerase I, J O H N S O N , P. H., GROSSMAN, L. I. (1977), Elec/. Bid. Chem. 260, 4975-4983. trophoresis of DNA in agarose gels. OptiDELEYS,R. J., JACKSON, D.A. (1975), Dye titra- mizing separations of conformational isotions of covalently closed supercoiled DNA mers of double- and single-stranded DNAs, analysed by agarose gel electrophoresis, BioBiochemistry 16, 4217-4225. chem. Biophys. Res. Commun. 69, 446-454. KLEEMISS, M. H., GILGES,M., SCHOMBURG, G. DRLICA,I<. (1992), Control of bacterial super(1993), Capillary electrophoresis of DNA coiling, Mol. Microbiol. 6, 425-433. restriction fragments with solutions of enFERREIRA, G. N. M., CABRAL,J. M. S., PRAZERES, tangled polymers, Electrophoresis 14, 515-522. KRASNOW,M.A., COZZARELLI, N. R. (1982), D. M. F. (1999), Monitoring of process Catenation of DNA-rings by topoisomerases: streams in the large-scale purificaton of plasmid DNA for gene therapy applications, mechanism of control by spermidine, 1.Biol. Chem. 257, 2687-2693. Pharm. Pharmacol. Commun. 5, 57-59. GARNER,M. M., CHRAMBACH, A. (1992), Reso- KREUZER, I<. N., COZZARELLI, N. R. (1980), Forlution of circular, nicked circular and linear mation and resolution of DNA catenanes by DNA, 4 kb in length, by electrophoresis in DNA gyrase, Cell 20, 245-254. G., polyacrylamide solutions, Electrophoresis 13, LAHIJANI,R., MARQUET,M., HULLEY, 176-178. SORIANO,G., HORN,N. A. (1996), High-yield production of pBR322-derived plasmids for GELLERT, M., MIZUUCHI,I<., O’DEA, M. H., NASH,H.A. (1976), DNA gyrase: An enzyme human gene therapy by employing a temthat introduces superhelical turns into DNA, perature-controllable point mutation, H u m . Proc. Natl. Acad. Sci. U S A 73, 3872-3876. Gene Ther. 7, 1971-1980. GLAZER, A. N., RYE,H. S. (1992),Stable dyeLUTTINGER A. (1995), The twisted life of DNA DNA intercalation complexes as reagents for in the cell - bacterial topoisomerases, Mol. high-sensitivity fluorescence detection, Microbiol. 15, 601-606. Nature 359, 859-861. MARQUET,M., H O R N ,N.A., MEEK,J.A. (1995), Process development for the manufacture of GOLDSTEIN,E., DRLICA,I<. (1984), Regulation of bacterial DNA supercoiling: Plasmid link- plasmid DNA vectors for use in gene therapy, Biopharm 8, 26-37.

2 Structures of Plasmid D N A

MARQUET, M., HORN,N.A., MEEK,J.A. (1997), Characterization of plasmid DNA vectors for use in human gene therapy, Parts 1 & 2, Biopharm 5, 42-50. MARTIN, R. (1996), Gel Electrophoresis: Nucleic Acids. Bios Scientific, London. MEYERS, J.A., SANCHEZ,D., ELWELL,L.P., FALKOW, S. (1976), Simple agarose gel electrophoretic method for the identification and characterization of plasmid deoxyribonucleic acid, /. Bacterial. 127, 1529-1537. MICHEL, M.-L., DAVIS, H. L., SCHLEEF, M., MANCINI, M., TIOLLAIS, P., WHALEN, R. G. (l995), DNA-mediated immunization to the hepatitis B surface antigen in mice, Proc. Nat. Acad. Sci. USA 92, 5307-5311. NACRERDIEN,Z., MORRIS, S., CHOQUETTE, S., RAMOS, B., ATHA, D. (1996). Analysis oflaserinduced plasmid DNA photolysis by capillary electrophoresis, /. Chromatogr. B 683, 91-96. PULLEYBLANK, D. E., MORGAN,A. R. (1975),The sense of naturally occurring superhelices and the unwinding angle of intercalated ethidium, /. Mol. B i d . 91, 1-13. SCHLEEF, M. (1999), Issues of large-scale plasmid manufacturing. In: Biotechnology 2nd Edn. Vol. 5a: Recombinant Proteins, Monoclonal Antibodies and Therapeutic Genes (Rehm, H,-J.,Reed, G., Puhler, A,, Stadler, P., Eds.), pp. 443-470. Wiley-VCH, Weinheim. SCHMIDT,T., FRIEHS,I<., FLASCHEL, E. (19961, Rapid determination of plasmid copy number, /. Biotechnol. 49, 219-229. SCHMIDT, T., FRIEHS, I<., SCHLEEF,M., Voss, C., FLASCHEL, E. (1999), Quantitative analysis of plasmid forms by agarose and capillary gel electrophoresis, Anal. Biochem. 274, 235.240.

I

SCHORR,J., MORITZ,P., SEDDON, T., SCHLEEF,

M. (1995), Plasmid-DNA for human gene therapy and DNA vaccines: pi-oduction and quality assurance, Ann. N. Y Acad. Sci. 772, 271-273. SERWER,P., ALLEN, J.A. (19841, Conformation of double-stranded DNA during agarose gel elctrophoresis: fractination of linear and circular molecules with molecular weights between 3.106and 26.106, Biochemistry 23, 922-927. SINDEN, R. R. (1994), D N A Structure and Function. Academic Press, San Diego, CA. SMIT, E., BORST, P. (1971), The superhelix density of bacteriophage PM2 DNA determined by a viscosimetric method, FEBS Lett. 14, 125-129. SUMMERS, D. I<. (1996), The Biology ofPlasmids. Blackwell Science, Oxford. TRAVERS A. A. (1993), DNA-Protein Interaction. Chapman & Hall, London. R (1997), TSE-DINH, Y.-C., QI, H., MENZEL, DNA supercoiling and bacterial adaptation: therinotolerance and thermoresistance, Trends Microbiol. 5, 323-326. ULFELDER, K. J. SCHWARTZ,H. E., HALL,J. M., SUNZERI, F. J. (1992), Restriction fiagment length polymorphism analysis of ERBB2 oncogene by capillary electrophoresis, Anal. Biochem. 200, 260-267. VIEIRA,J., MESSING, J. (1982), The pUC plasmids, an M13mp7-derived system for insertion mutagenesis and sequencing with synthetic universal primers, Gene 19, 259-268. ZHU,H., CLARK, S. M., BENSON,S. C., RYE, A. N., MATHIES, R. A. (1994), H. S., GLAZER, High-sensitivity capillary electrophoresis of double-stranded DNA fragments using monomeric and dimeric fluorescent intercalating dyes, Anal. Chem. 6 6 , 1941-1948.

43

P/asmids for Therapy and kchation by M.Schleef Copyright 0 WILEY-VCH Verlag GmbH, 2001

I

3 Genetic Vaccination with Plasmid Vectors Jorg Reimann”, Marcin Kwissa and Reinhold Schirmbeck

1

Introduction

In nucleic acid immunization, expression plasmid DNA containing antigen-encoding sequences is delivered by different techniques that support the in uiuo expression of the antigen and its immunogenic presentation. Humoral and cellular immune responses to protein antigens can be efficiently primed by nucleic acid or DNA vaccination. A key feature of DNA-based vaccination is the in situ expression of immunogenic proteins with correct posttranslational modifications, three-dimensional conformations or oligomerizations ensuring the integrity of conformational epitopes that can stimulate neutralizing antibody (B cell) responses. Furthermore, DNA (or RNA) immunization is exceptionally potent in stimulating T cell responses that makes it attractive for vaccinations against intracellular pathogens and cancer. Peptides generated in (endogenous or exogenous) processing pathways (without interference by viral proteins) from intracellular or extracellular protein antigens expressed after transient in uiuo transfection efficiently stimulate cellular (T cell) immune responses. Both features are difficult to achieve with recombinant antigens that are produced on a large scale in eukaryotic or prokaryotic expression systems. Excellent reviews on DNA-based vaccination have been published recently (Donnelly et al., 1997b; Koprowski et al., 1998; Gurunathan et al., 2000).

2 Vector design 2.1 Plasrnid DNA

The “vaccine” delivered in nucleic acid immunization is usually plasmid DNA although successful vaccination with antigen-encoding mRNA has been reported (Martinon et al., 1993; Conry et al., 1995c; Boczkowski et al., 1996; Qiu et al., 1996;

45

46

I

Jorg Reimann et a/.

Bacterial backbone

1/

j

Transcription unit

PLASMIDVECTOR

1 of intanat

W i n of replication

;

PolyATail

BsAg (Hepatitis B surface antigen)

n f l origin

j

Fig. 1. Expression plasmid constructs used i n DNA vaccination. (A) Prototype vector: the plasmid contains a bacterial backbone (with sequences of the origin o f replication and the antibiotic selection marker) and the eukaryotic transcription unit (with promoter/enhancer

\\

‘SV4OpolyATail

sequences, sequences encoding the antigen o f interest, and a poly A signal sequence). (B) An example for an expression vector encoding the hepatitis B surface antigen (HBsAg) that was successfully used as a DNA vaccine.

3 Genetic Vaccination with Piasmid Vectors

Hoerr et al., 2000). The plasmid DNA vector is from < 1 kb to > 20 lcb in size and contains two units: the bacterial backbone (with the origin of replication and markers for prokaryotic and/or eukaryotic selection), and the transcription unit (encoding the antigen under appropriate promoter control) (Figure 1).Most plasrnid vectors are closed circular, double-stranded DNA. Supercoiled plasmid DNA (see also chapter 2) seems to ensure the most efficient, transient transfection in vivo into immunologically relevant cells. Plasmid DNA vectors do not replicate in eukaryotic cells and are designed to minimize the chances for integration into the host cell genome. 2.2 Construction of simple transcription units

In the transcription unit, the antigen-encoding sequence is cloned downstream from strong promoter/enhancer sequences that support optimal expression of the immunogenic protein. Different promoters have been used successfully for constructing vectors for DNA vaccination. These include promoter/enhancer sequences from viruses (e. g., cytomegalovirus, papova virus, retrovirus), bacteria (e.g., Bowelia), or mammalian cells (promoter/enhancer sequences controlling, e. g., the expression of elongation factor-la, desmin, or metallothionin) (Tang et al., 1992; Ulmer et al., 1993; Geissler et al., 1994; Johnston et al., 1994; Raz et al., 1994; Davis et al., 1995; Michel et al., 1995; Bohm et al., 1996; Simon et al., 1996; Kwissa et al., 2000a). The mRNA transcripts are stabilized by incorporating a polyadenylation signal and an intron into the expression construct. In Semliki forest virus-derived constructs, mRNA is dramatically amplified in the cytoplasm thereby enhancing the level of antigen expression (Zhou et al., 1995).These simple transcription units are constructed following standard rules for expression of genes in eukaryotic cells. Many commercially available expression vectors have been used successfully in DNA vaccination. 2.3 Construction o f complex transcription units

To produce polyvalent DNA vaccines, or DNA vaccines with inherent adjuvant activity, it is desirable to construct vectors that allow the coexpression of different antigens, or of antigen and cytoltine/chemokine. This can be achieved with relative ease at the DNA level using different cloning strategies. The most simple way is to construct fusion proteins as recombinant “polyepitope vaccines” (Thomson et al., 1996),“multivalent minigene vaccines” (An et al., 1997), antigen/cytoline fusion vaccines (Kim et al., 1997; Maecker et al., 1997), or antigen/costimulator fusion constructs (Boyle et al., 1998) (Figure 2A). Fusion proteins can be designed in a way that antigen is processed immediately after synthesis (Valmori et al., 1999). Such instable, nascent protein can generate multiple, bioactive protein fragments during or after translation that, e. g., are antigenic, facilitate processing, target the antigenic domain to particular subcellular compartments, cells or tissues, or convey costimulator or cytokine effects.

I

47

48

I

Jorg Reimann et a/.

Three simple methods have been used to codeliver different antigens, or an antigen and a cytokine with a DNA vaccine. These include -

-

the injection of a mixture of different expression plasmids (Figure 2B), the construction of large plasmids that contain multiple, independent transcription units (Figures 2C, D), or the coating of different expression plasmids on particles that are used to deliver DNA via the skin with the gene gun (Xiang et al., 1995; Bueler et al., 199G; He et al., 199Gb; He et al., 1996a; Kwak et al., 1996; Mahvi et al., 1996; Chow et al., 1997; Corr et al., 1997; Geissler et al., 1997; Iwasaki et al., 1997; Kim et al., 1997; Kim et al., 1997; Oltada et al., 1997; Tsuji et al., 1997; Tsuji et al., 1997; Geissler et al., 1998; Gurunathan et al., 1998; Kim et al., 1998; Kim et al., 1998; Larsen et al., 1998; Sin et al., 1998; Kimura et al., 1999; Kipps et al., 1999; Lu et al., 1999; Sin et al., 1999).

Many of these approaches have demonstrated adjuvant effects of the codelivered cytokine, chemokine or costimulator molecule that enhance the magnitude of the elicited immune response and modify its polarization. If the coordinated expression of different proteins (different antigens, or antigen and cytokine/chemoFusion construct

I

Plasmid mix

+I

II B Independenttranscription units on one plasmid

C

Polycistronic construct

I

Bidirectional promoter

R

' F I

Fig. 2. Complex expression vectors support expression of multiple antigens, or antigen plus cytokines. Two transcription units can be fused (A), or separate plasmids with independent transcription units can be mixed before injection (B).Alternatively, t w o transcription units

can be cloned as independent units into a plasmid using different strategies (C, D). Coordinated expression o f two gene products is achieved using polycistronic constructs (E) or expression under bidirectional promoter control (F).

3 Genetic Vaccination with Plasmid Vectors

kine) at stochiometrically defined ratios is required, more complex vector systems have to be used. Intact antigen and/or cytokine proteins can be coexpressed in polycistronic constructs (Dirks et al., 1993, Johanning et al., 1995; Huang, 1996; Clarke et al., 1997; Wild et al., 1998) (Figure 2E), or constructs with bidirectional promoters (Kwissa et al., in press) (Figure 2F). Complex expression vectors have been designed that contain multiple polycistronic expression units under bidirectional promoter control (H. Hauser, unpublished data; see also chapter 7). This approach may eventually allow the package of the complete immunogenic information of a virus together with the optimal adjuvant effect into a single expression plasmid DNA vector. This could generate new generations of DNA vaccines that can bypass many non-responder problems and thus protect a large fraction of the population.

3

Strategies for DNA delivery

To be an effective DNA vaccine, vector DNA has to be injected in a way that ensures -

-

the efficient in vivo transfection of somatic cells (transfection efficiency in vivo), and the access of the gene product to antigen-presenting cells (APC) that can prime an immune response (i. e., efficiency of handling of antigen expressed from the injected vector DNA by somatic cells and/or APC).

Three techniques are currently available to deliver DNA vaccines:

.

injection of a high dose (e.g., 50-100 pg DNA per mouse) of non-packaged (“naked) plasmid DNA (Jiao et al., 1992; Wolff et al., 1992; Davis et al., 1993) injection of variable doses (1-50 pg per mouse) of plasmid DNA packaged into liposomes, lipoplexes, polymers or virosomes (Nabel et al., 1992; Martinon et al., 1993; Harrison et al., 1995; Yoltoyama et al., 1996; Ishii et al., 1997; Liu et al., 1997; Toda et al., 1997; Dow et al., 1999; Klavinsltis et al., 1999; Goldman et al., 1997; Chen et al., 1998; Kwoh et al., 1999) coating of plasmid DNA onto 0.5-3 pm gold particles (50-300DNA molecules per particles) that are ‘shot’ with the gene gun into the skin to deliver a low dose of 0.1-1 pg DNA per mouse (Williams et al., 1991; Tang et al., 1992; Eisenbraun et al., 1993; Fynan et al., 1993; Johnston et al., 1994; Vahlsing et al., 1994; Fuller et al., 1995; Jenkins et al., 1995; Pertmer et al., 1995; Sun et al., 1995; Yang et al., 1995; Zarozinski et al., 1995; Fuller et al., 1996; Haynes et a]., 1996; Keller et al., 1996; Mahvi et al., 1996; Qiu et al., 1996; Choi et al., 1997; Feltquate et al., 1997; Leitner et al., 1997; Prayaga et al., 1997; Tanelian et a]., 1997; Torres et al., 1997; Macklin et al., 1998; Porgador et al., 1998).

Some routes are preferentially used to deliver these plasmid DNA formulations as “vaccines”by these techniques. Non-packaged (“naked)plasmid DNA is often in-

I

49

I jected intramuscularly (into the

50 Jorg Reimann et a/

rn. tibialis anterior or rn. quadriceps of the mouse), but can also efficiently prime murine T and B cell responses when injected intradermally or subcutaneously (Bohm et al., 1998). Packaged plasmid DNA can prime immune responses when injected intramuscularly, subcutaneously, intradermally, intravenously, intraperitoneally or into tumors, or when the formulation is applied onto mucosal surfaces (Plautz et al., 1994; Goldman et al., 1997; Ishii et al., 1997; Liu et al., 1997; Chen et al., 1998; Dow et al., 1999; Klavinsltis et al., 1999; Icwoh et al., 1999). For the delivery of particle-coated plasmid DNA with the gene gun, the skin is the preferred site for inoculation but exposed muscle surfaces and mucosal surfaces have been successfully used for particle bombardment (Tang et al., 1992; Eisenbraun et al., 1993; Fynan et al., 1993; Vahlsing et al., 1994; Jenkins et al., 1995; Pertmer et al., 1995; Zarozinski et al., 1995; Keller et al., 1996; Qiu et al., 1996; Choi et al., 1997; Feltquate et al., 1997; Leitner et al., 1997; Prayaga et al., 1997; Torres et al., 1997; Macklin et al., 1998). Hence, a variety of delivery techniques and of different delivery routes have been developed to vaccinate animals and man with plasmid DNA. Unfortunately, all delivery techniques currently available are clearly suboptimal. It i s uncertain which vaccination techniques will eventually find wide spread use in medical practice.

4

Priming humoral and cellular immune responses by DNA vaccines

DNA vaccination efficiently primes humoral (antibody) and cellular (T cell) immune responses against a wide spectrum of viral, bacterial, parasitic and tumor antigens in different animal species (e. g., mouse, rat, woodchuck, pig, rabbit, sheep, cattle, dog, monkey, chimp, fish, chicken) and in man. Often, a single inoculation of the DNA vaccine using an appropriate delivery technique and a suitable route primes readily detectable and long-lasting serum antibody responses and T cell responses. The immunological data generated with different DNA vaccines have been reviewed during the last years (Yang, 1992;Waine et al., 1993; Donnelly et al., 1994; Johnston et al., 1994; Montgomery et al., 1994; Vahlsing et al., 1994; Dixon, 1995; Fynan et al., 1995; Krishnan et al., 1995; Liu, 1995; Marwick, 1995; Pardoll et al., 1995; Vogel et al., 1995; Weiner, 1995; Wilkinson et al., 1995; Conry et a]., 1996b; Cornelissen et al., 1996; Ertl et al., 1996b; Ertl et al., 1996a; Hassett et al., 1996; Haynes et al., 1996; Kumar et al., 1996; McCarthy, 1996; McDonnell et al., 1996; Robinson et al., 1996; Siegrist et al., 1996; Ulmer et al., 1996; Ulmer et al., 1996b; Ulmer et al., 1996; Barry et al., 1997; Butler et al., 1997; Chattergoon et al., 1997; Donnelly et al., 1997b; Donnelly et al., 1997a; Levitsky, 1997; Liu et al., 1997; Mor et al., 1997; Pisetsky, 1997; Ramsay et al., 1997; Raz, 1997; Siegrist et al., 1997; Icoprowski et al., 1998; Lowrie, 1998; Tighe et al., 1998; Kanellos et al., 1999; Icipps et al., 1999; IUinman et al., 1999; Gurunathan et al., 2000). In addition, data on the prevention or treatment of particular diseases by DNA vaccination have been reviewed in detail, e. g., DNA vaccination against hepatitis B virus (HBV) (Michel, 1995; Wands et al., 1997; Davis, 1998)

3 Genetic Vaccination with Plasrnid Vectors

and hepatitis C virus (HCV) (Inchauspe, 1997), HIV/SIV (Liu et al., 1996; Lu et al., 1996; Shiver et al., 1996; Ishii et al., 1997; Kim et al., 1997; Shiver et al., 1997; Letvin, 1998),tuberculosis (Lowrie et al., 1997; Lowrie et al., 1999),or tumors (Conry et al., 1995a; Felgner et al., 1995; Spooner et al., 1995; Stevenson et al., 1995; Yang et al., 1995; Conry et al., 1996a; Zhu et al., 1997; Wang et al., 1998; Ying et al., 1999). Because DNA vaccination allows protein antigens to be expressed in situ with all posttranslational modifications (e. g., glycosylation, proteolytic processing, lipid conjugations) that determine or stabilize its native conformation, antibody responses against native epitopes predominate the humoral immune responses elicited by these vaccines. This allows the generation of high protective (neutralizing) antibody titers that is usually only achieved by using vaccines based on attenuated live vectors. DNA vaccination is an exceptionallypotent strategy to stimulate T cell responses. This makes DNA-based immunization an interesting strategy for vaccination against intracellular pathogens and cancer. Different mechanisms seem to be involved in priming T cells by genetic vaccination. Evidence has been presented that intramuscular vaccination with a high dose of plasmid DNA depends on “cross-priming” (Doe et al., 1996; Ulmer et al., 1996a; Corr et al., 1996). Antigenic material produced by transiently transfected myocytes or lteratinocytes becomes immunogenic by gaining access to professional, bone marrow-derived APC. In contrast, the intradermal delivery of low amounts of plasmid DNA with the gene gun results in the transfection of resident dendritic cells (DC) that rapidly migrate to regional lymph nodes where they initiate the immune response (Porgador et al., 1998). Induction of CD8’ CTL responses by cross priming requires cognate CD4’ T cell help. CTL priming to antigen by DNA vaccination is partly CD4+ T cell-dependent although immunostimulatory bacterial CpG motifs in bacterial plasmid DNA facilitate CD4’ T cell-independent priming of na’ive CD8+ CTL precursors (Wild et al., 1999). The CD4’ T cell-dependence of immune responses stimulated by DNAbased vaccination points to a regulatory control that determines the polarization of the response, i. e., the spectrum of immune effector specificities that can be specifically elicited. Anti-viral immune responses with different polarization profiles can be primed using alternative DNA vaccination approaches. A single intramuscular or subcutaneous injection of 50-100 pg non-packaged plasmid DNA encoding antigen, or a single intradermal injection of 1 pg particle-coated plasmid DNA often prime serum antibody responses of similar magnitude and longevity but strikingly different isotype profiles. Priming mice by an intramuscular injection of 100 pg plasmid DNA stimulates long-lasting Thl immune responses that include specific activation of potent MHC-I-restricted CTL responses. In contrast, priming mice with an intradermal injection of a low dose of plasmid DNA (0.1-1 pg per mouse) stimulates long-lasting Th2 immune responses but no CTL responses. This polarization pattern is confirmed by analyses on the cytoline expression profile of T cells primed in vivo by the two alternative DNA delivery strategies: while interferon-y (IFN-y)-producingT cells are preferentially primed by the

I

51

I injection of high doses of non-packagedplasmid DNA,

52 Jorg Reimann et a/.

IL-4-, IL-5- and IL-10-producing T cells predominate after intradermal delivery of low doses of particle-bound plasmid DNA with the gene gun (Feltquateet al., 1997). But the skin area to which plasmid DNA is delivered can be conditioned to support priming of Thl immune reactivity. Subcutaneous injections of cytoltines (interferons, IL-12, IL-18) or oligonucleotides (CpG-containing oligonucleotides, poly I/C) at the site of intradermal DNA delivery with the gene gun shifts the polarization of the elicited immune response towards a Thl phenotype. To change the polarization of the immune response, these reagents have to be delivered locally to the site of particle injection from 48 h prior to until 48 h after the gene gun vaccination. The technique of conditioning the skin to support priming of T1 responses by the gene gun is of practical interest for the design of CTL-stimulating DNA vaccines against persistent virus infections or cancer. Protective anti-viral T and B cell responses have been primed in neonatal mice by a single injection of plasmid DNA, but not recombinant antigen-based vaccine formulations, within 24 h after birth (Forsthuber et al., 1996; Sarzotti et al., 1996; Bot et al., 1997; Hassett et al., 1997; Manickan et al., 1997; Martinez et al., 1997; Sarzotti et al., 1997; Siegrist, 1997; Adltins et al., 1998; Ichino et al., 1999; Pertmer et al., 1999). In different viral antigen systems studied, newborn mice develop balanced Thl/Th2 primary effector responses in vivo. The technique of DNA vaccination is thus an attractive option to prime protective immune responses in neonatal mice, to bias the induction of specific immunity towards effective anti-viralThl immune effector functions, and to override the bias of the neonatal immune system towards specific tolerance induction. Intrinsic properties of the antigen (that are not yet identified biochemically) can have a decisive influence on the type of specific immunity they establish after priming by DNA vaccination. Some antigens prime preferentially Thl immune effector functions when immunity to them is stimulated by delivering intradermally low doses of particle-bound DNA. This points to the possibility of constructing by recombinant DNA technology chimeric antigens that enhance and modulate their immunogenicity. An interesting example has been reported recently in which mice were vaccinated with alternative DNA constructs encoding chimeric antigens in which the same immunogenic domain was fused to different immunomodulating domains (Boyle et al., 1998). The immunomodulating domains were ligands that bind receptors either on high endothelial cells (to target the antigen to lymph nodes), or on APC (to provide costimulatory signals). This antigen-targeting strategy not only enhanced the immune responses to the chimeric vaccines but also altered strikingly the polarization patterns of immune responses elicited against the same antigenic determinant. Hence, antigens can be constructed that preferentially prime Thl or Th2 immune responses. Codelivery of cytokines, chemolines or immune receptor-binding domains with antigens can enhance and/or modulate their immunogenicity. DNA vaccination protocols have been reported in which these enhancing or modulating factors are either delivered as expression constructs mixed with antigen-encoding DNA, or coexpressed with antigen as fusion constructs or in polycistronic vector systems,

3 Genetic Vaccination with Plasmid Vectors

or coated onto particles for gene gun delivery together with the antigen-encoding DNA. The magnitude and type of immunity that can be primed by DNA vaccination is thus influenced by many factors, the most important ones are

. the vaccination technique und the route of D N A delivery used, the age (maturity of the immune system) of the immunized animal,

. the condition of the local tissue environment to which DNA is delivered, . the codelivery of immune-modulating factors (cytokines, chemoltines, immune receptor-binding ligands),

. the type of antigen encoded by the plasmid DNA vaccine. 5 Experimental strategies facilitated

by DNA vaccination

An exciting new technique of DNA vaccination is the expression library immunization (ELI) (Barry et al., 1995; Lai et al., 1995; Piedrafita et al., 1999). With this technique, immunogenic proteins from complex viral, bacterial or parasitic pathogens have been rapidly defined. Because of the relative simplicity with which recombinant DNA can be modified, DNA vaccination has been successfully used to identify antibody-defined domains as well as T cell-defined epitopes of large, complex antigens. This can be achieved through site-directed mutagenesis, in frame deletions, truncations or transfer of domains to heterologous proteins. This can lead to the construction of polyvalent vaccines in which immunodominant domains from different antigens and/or different pathogens are fused to a chimeric, polyvalent fusion protein. DNA vaccination furthermore offers an attractive alternative to establish specific probes (monoclonal antibodies or restricted T cell lines) for well-defined antigenic determinants (Martinon et al., 1993; Watanabe et al., 1993; Barry et al., 1994; Hawltins et al., 1994; Shiver et al., 1995; Krasemann et al., 1996). By fusing antigenic domains with immune receptor-binding domains from costimulator or homing molecules, the antigen can be targeted to secondary lymphoid organs or specific types of APC. Within cells transfected by a plasmid DNA vaccine, the expressed antigen can be targeted to different subcellular compartments. Antigen can reside in the cytosol at the site of synthesis; it can be transported to endolysosomes; it can be translocated into the endoplasmic reticulum; it can be displayed on the cell surface as an integral membrane protein; or it can be secreted as a soluble protein. These different traffic patterns of a protein antigen are expected to influence its immunogenicity for B cells and T cells. A key factor that determines immunogenicity of an antigen for T cells operates through its susceptibility and type of processing (partial proteolytic degradation) that generates peptides binding to major histocompatibility (MHC) molecules; it is only in this form that antigens becomes visible for T cells (Reimann et al., 1999). Efficiency (enhancement of proteolyhc degradation) and type (proteolytic cleavage pattern) of processing are the major factors that determine the efficiency and specificity of

I

53

54

I

Jorg Reimann et a/.

T cell priming. These key factors of a DNA vaccine that determine its T cell-stimulating potency can be readily manipulated in experimental DNA vaccines for testing in preclinical models. As described above, DNA vaccines can be readily designed to assay the adjuvants effect of codelivered cytokines, chemokines, or costimulator molecules on the magnitude and type of the humoral and cellular immunity primed. The increasing number of molecularly cloned cytokines, growth factors, chemoltines, adhesion molecules, homing molecules or costimulator molecules that have potential adjuvants activity make this approach particularly attractive as a screening tool. DNA vaccines allow us to fuse a stable label to the expressed antigen (e.g., GFP fusion constructs) to trace its in vivo the biodistribution, uptake by cells, or processing by APC. Areas in which DNA vaccination can help to define generic principles to optimize novel vaccine and adjuvant approaches thus comprise the following areas:

the identification of antigens, their immunogenic domains or epitopes, the generation of immunological probes (monoclonal antibodies, T cell clones) to define immunogenic domains or epitopes of an antigen, the construction of polyvalent vaccines from different immunogenic domains or epitopes, the design of chimeric antigens with optimized immunogenicity facilitating B cell priming and processing for the MHC-restricted presentation to T cells, the optimization of adjuvant codelivery by identifying costimulating ligands of the specific or innate immune system and finding ways to codeliver them with the antigen, the study of the biodistribution of antigen expressed from a DNA vaccine to optimize its targeting to immunologically relevant secondary lymphoid tissues and APC. 6

Unique advantages of DNA vaccination

Experience with DNA vaccination within the last years has revealed a number of advantages of this novel approach to vaccination that make it a very attractive alternative for therapeutic and prophylactic vaccination approaches. In terms of their potent immunogenicity for T cells, in particular MHC-I-restrictedcytotoxic T lymphocytes (CTL),and their tendency to prime Thl-biased immune responses, DNA vaccines mimic the immunogenicity of live attenuated vaccines although they do not share many of the disadvantages of the latter (Gurunathan et al., 2000). DNA vaccines can override non-responder status in different preclinical animal models being more effective than recombinant vaccinia virus in revealing CTLdefined epitopes (Schirmbeck et al., 1995). This offers the chance to reduce the fraction of non-responders to the vaccine in a population, the complete protection of which is the goal of the intervention.

3 Genetic Vaccination with Plasrnid Vectors

As described above, priming protective T cell responses in neonatal animals has been shown to be exceptionally encouraging with DNA vaccines (reviewed in (Siegrist et al., 1997; Siegrist, 1997; Kovarik et al., 1998). For early protection in life, this vaccine seems to have advantages that are difficult to obtain with more conventional approaches. Because DNA vaccination is a potent way to prime CTL to internal (non-variant) viral antigens (such as nucleocapsid, matrix or polymerase proteins), it is easier to obtain cross-strain protection against pathogenic viruses with DNA vaccines than with conventional vaccines that rely on the induction of neutralizing antibody responses against variant envelope proteins of a virus. A case that strikingly illustrates this point is the vaccination against influenza virus. DNA vaccination may lead to better protection against epidemic virus infections. The partly helper-independent priming of CTL achieved by DNA vaccination is of value in immunocompromized patients that can not (because of genetic or acquired defects) generate adequate CD4’ T cell helper activity for priming protective CTL responses. DNA vaccines may offer a way to protect such individual against infectious diseases. Conventional vaccination protocols usually use the same vaccine for priming and boosting an immune response. Recent data have revealed the surprising observation that more effective T cell responses can be generated following a vaccination protocol in which a primed response is boosted by a different vaccine formulation containing the same (or a similar) antigen (Hanlte et al., 1999; Plebanski et al., 1999; Robinson et al., 1999; Hanke et al., 1998; Richmond et al., 1998; Schneider et al., 1998; Sedegah et al., 1998). In these protocols, DNA vaccines are usually combined with vaccines based on recombinant proteins or viruses. We might see in “difficult”infections (such as ,e. g., malaria) the emergence of primelboost protocols in which DNA vaccines play a role. The large scale production of DNA vaccines is cost-effective,and storage conditions for the vaccine will be less demanding than for many alternative vaccine formulations. There thus seem to emerge a number of advantages of DNA vaccines that will facilitate their acceptance in medical practice. These are summarized as follows:

9

potent inducer of protective anti-viral CTL responses, override non- or low responder status in anti-viral T cell reactivity, overcome low responsiveness or tolerance induction in neonatal immune systems, convey cross-strain protection against viruses, support priming of anti-viral CTL despite defbcient CD4’ T cell helper activity, are components of prirne/boost vaccination protocols that strikingly enhance antiinfectious T cell reactivities, are cost eflective and easy to store.

I

55

56

I

Jorg Reimam et a1

7

DNA vaccines in preclinical animal models

The experimental work on DNA vaccines in preclinical animal models explores its potential for prophylaxis and therapy in four areas that are of interest in human and veterinary medicine. These comprise the following: control of infectious disease, immune therapy of cancer, specific treatment of autoimmune disease, immunmodulation of allergic disease. 7.1

DNA vaccines to control infectious diseases

DNA vaccines have been developed to protect different animal species (mice, rats, woodchucks, dogs, cats, cattle, sheep, pigs, chicken, fish, monkeys, chimps) against a wide variety of viral, bacterial or parasitic pathogens (Cox et al., 1993; Fynan et a]., 1993; Montgomery et al., 1993; Robinson et al., 1993; Ulmer et al., 1993; Wang et al., 1993; Rouse et al., 1994; Sedegah et al., 1994; Ulmer et al., 1994; Xiang et a]., 1994; Xu et al., 1994; Barry et a]., 1995; Donnelly et al., 1995; Ghiasi et al., 1995; Jenkins et al., 1995; Lai et al., 1995; Manickan et al., 1995; Martins et al., 1995; Xu et al., 1995; Yokoyama et al., 1995; An et al., 1996; Anderson et al., 1996; Bourne et al., 1996; Bourne et al., 1996; Boyer et a]., 1996; Donnelly et al., 1996; Doolan et al., 1996; Gonzalez Armas et al., 1996; Herrmann et al., 1996; Hildeman et al., 1996; Huygen et al., 1996; Lu et al., 1996; McClements et al., 1996; Saltaguchi et al., 1996; Simon et al., 1996; Tascon et al., 1996; Zhong et al., 1996; An et al., 1997; Choi et al., 1997; Donnelly et al., 1997; Gerdts et al., 1997; Gurunathan et al., 1997; Hassett et al., 1997; Kodihalli et al., 1997; Lai et al., 1997; Letvin et al., 1997; Lozes et al., 1997; McClements et al., 1997; McDaniel et al., 1997; Sundaram et al., 1997; Ugen et al., 1997; Ward et al., 1997; Beclcer et al., 1998; Bender et al., 1998; Bonato et al., 1998; Chen et al., 1998; Larsen et al., 1998; Li et al., 1998; Lin et al., 1998; Macklin et al., 1998; Schneider et al., 1998; Sedegah et al., 1998; Sixt et al., 1998; Triyatni et al., 1998; Xu et al., 1998; Youssef et al., 1998; Gu et al., 1999; Icamath et al., 1999b; Kamath et al., 1999a; Tanghe et al., 1999) . In some of these models, immune parameters that correlate with specific protective immunity have been characterized in detail. In many cases, it was demonstrated that it is the specific MHC-I-restricted CTL reactivity primed by the DNA vaccine that was of critical importance in its prophylactic efficacy. In addition, therapeutic DNA vaccines have been tested in animal models of persistent, pathogenic virus infections. In some studies, limited therapeutic efficacy could be demonstrated. Though promising, this area will need a intensive efforts and informative, well defined animal models to elucidate the therapeutic potential and the immunological mechanism of action of therapeutic DNA vaccines.

3 Genetic Vaccination with Plasmid Vectors

7.2

Therapeutic tumor vaccines

As DNA vaccines are very effective in priming CTL responses, and CTL are considered a major defense mechanism against tumors, many groups have developed therapeutic DNA vaccines against cancers in immunologically well defined animal models using transplantable tumors with defined tumor-associated antigens (Plautz et al., 1993; Watanabe et al., 1993; Conry et al., 1994; Hawltins et al., 1994; Conry et al., 199513; Conry et al., 1995a; Felgner et al., 1995; Nicolet et al., 1995; Spooner et al., 1995; Stevenson et al., 1995; Sun et a]., 1995; Wang et al., 1995; Yang et al., 1995; Bueler et al., 1996; Ciernik et al., 1996; Kwak et al., 1996; Mahvi et al., 1996; Rakhmilevich et al., 1996; Schirmbeck et al., 1996; Syrengelas et al., 1996; Tan et al., 1996; Bohm et al., 1997; Corr et al., 1997; Geissler et al., 1997; Rosato et al., 1997; Spellerberg et al., 1997; Tuting et al., 1997; Agadjanyan et al., 1998; Gurunathan et al., 1998; Iwasalti et al., 1998; Neglia et a]., 1999; Wei et al., 1999; Ying et al., 1999). Of particular interest are carcinoembryonicantigen-expressing tumors, idiotypic determinants of B cell lymphomas, and an increasing array of oncogenes as antigens in these DNA vaccines. Though often successful in mouse models, it remains to be shown that a therapeutic DNA vaccine can control growth of a “spontaneous” tumor. 7.3 Autoimmune disease

DNA vaccines have been used for the therapy of established autoimmune disease using two different approaches. Based on the idea of idiotype/anti-idiotyperegulation of specific immune responsiveness, therapeutic DNA vaccines have been developed for the clonotype-specific suppression of autoaggressive antibodies or T cells (Watanabe et al., 1993; Kuhrober et al., 1994; Waisman et al., 1996). Suppressive immunization with DNA vaccines encoding a self-peptidethrough modulation of T cell costimulation have been demonstrated to prevent autoimmune disease (Ruiz et al., 1999). Hence, modulation of the spectrum of immune effector functions expressed by an autoaggressive immune response by a therapeutic DNA vaccine might offer a way for the specific immunotherapy of autoimmune disease. 7.4

Treatment o f allergy by therapeutic DNA vaccination

Immune reactions underlying allergy are characterized by a strong Th2 phenotype, and DNA vaccination often introduces a strong Thl-bias into the immune responses it primes or boosts. In mouse models, allergy-inducing Th2 responses have been converted into non-pathogenic Thl responses against the allergen by DNA vaccination (Hsu et al., 1996; Roman et al., 1997; Spiegelberg et al., 1997). The inhibition of IgE antibody formation by plasmid DNA immunization has been shown to be mediated by both CD4’ and CD8+ T cells (Lee et al., 1997).

I

57

58

l More data are needed to critically evaluate in preclinical models the therapeutic poJorg Reimanri et a/.

tential of DNA vaccines in this clinical condition.

8 Proposed clinical applications o f DNA vaccines

Potential clinical applications have emerged from studies in preclinical animal models in the areas of prevention or treatment of infectious diseases, tumors, autoimmune diseases and allergies. An overview and representative examples are listed in Table 1.

Table 1.

Clinical applications o f DNA vaccines

Area

Indication

prophylactic

therapeutic

Example

I

human papilloma virus (HPV) human immunodeficiency virus (HIV) influenza virus hepatitis b virus (HBV) hepatitis c virus (HCV) human papilloma virus (HPV) human immunodeficiency virus (HIV) malaria m ycobacteria

Tumor Immunotherapy

specific immunotherapy

Induction of CTL responses against tumor-associated antigens (CEA, MAGE, BAGE, immunoglobulin idiotypes)

Autoimmune disease

specific suppression of autoaggressive immune reactions

Induction of Th2 T cell responses to autoantigens against which pathogenic Thl responses are ongoing

Allergy

immunomodulation to suppress allergy

Induction of Thl T cell response to allergens against which pathogenic Th2 responses are ongoing

I

3 Genetic Vaccination with flusmid Vectors

9

Risks of nucleic acid vaccination

DNA vaccination includes risks that are at this early stage of development of the novel vaccination technique difficult to assess critically. The limited experience with safety considerations obtained in animal models suggests the following potential risks: insertional mutagenesis (integration of plasmid DNA into coding or regulatory sequences of the eukaryotic genome), (low-zone)tolerance (induced by exposure of the immune system to suboptimal doses of antigen). This has not been experimentally verified; the available experimental data indicate that suboptimal immunization can prime responses that are revealed by boost injections, induction of autoimmune responses (e.g., autoantibodies against cytokines). This occurs but is also observed in the course of, e. g., chronic infections, extensive immune-mediateddestruction of tissue transfected in vivo. This may actually be of advantage to delete genetically altered cells. 10

Future perspectives

DNA immunization is an exciting new option for designing vaccines that efficiently stimulate T cell responses. Only a few years have passed since its first description in 1992 but many experimental DNA vaccines have moved from preclinical animal models into clinical trials. The main factors that contribute to the success of a DNA vaccine are schematically shown in Figure 3 . These are -

the design of the DNA vaccine, its delivery, and conditions of the treated subject.

Most principles of the design of plasmid vectors, cloning of antigens of interest and codelivery of immuno-modulators have been defined. The choice of the site for inoculation, the conditioning of the injected tissue, the age of the vaccinated subject, and the type of preexisting immunity are factors that have to be taken into account but are readily controlled in vaccination protocols. The major challenge is to find better ways to deliver the DNA vaccine, including the techniques used for inoculation, the packaging of the DNA vaccine, and the use of low doses of DNA in such vaccines. These problems are largely unresolved and will decide on the eventual success of this vaccination strategy.

I

59

60

I

Jorg Reimann et a/.

1. Expression vector 2. Plasmidencodedantigen

3. Co-delivery of immunomodulators

I DNA delivery

I

I 4. Delivery technique

B

5. Packaging of DNA 6. Dose of plasmid DNA inoculated

I I

C 7. Site of injection \\

/,'

8. Condition of tissue chosen for DNA immunization 9. Age of individual (maturity of immune system)

"

I

; Fig. 3. Critical factors that have t o be considered i n designing DNA vaccination strategies include (A) the design o f the vector (expression construct, choice of antigen, codelivery o f different antigens, or antigen plus cytokine), (B) DNA delivery (delivery technique, injection

$0. Type of preexisting immunity

of 'naked' or packaged DNA, choice o f t h e dose injected), and (C) the selection o f the site of inoculation, and tissue condition used for injection, or consideration of preexisting immunity.

3 Genetic Vaccination with Plasmid Vectors

References ADI~INS, B., Du, R. Q. (1998), Newborn mice develop balanced Thl/Th2 primary effector responses in uivo but are biased to Th2 secondary responses, /. Immunol. 160, 42174224. AGADJANYAN, M. G., WANG,B., NYLAND,S. B., WEINER,D. B., UGEN,I<. E. (1998), DNA plasmid based vaccination against the oncogenic human T cell leultemia virus type 1, CUT. Top. Microbiol. Immunol. 226, 175-192. AN, L.L., PAMER,E.G., WHITTON, J.L. (1996), A recombinant minigene vaccine containing a nonameric cytotoxic T-lymphocyte epitope confers limited protection against Listeria wlonocytogenes infection, Infect. Immun. 64, 1685-1693. AN, L.L., WHIITON, J. L. (19971, A multivalent minigene vaccine, containing B-cell, cytotoxic T-lymphocyte,and Th epitopes from several microbes, induces appropriate responses in uivo and confers protection against more than one pathogen, J . Virol. 71, 2292-2302. ANDERSON, E. D., MOURICH,D.V., FAHRENICRUG,S. C., LAPATRA,S., SHEPHERD,J., LEONG,J. A. (1996),Genetic immunization of rainbow trout (Oncorhynchus mykiss) against infectious hernatopoietic necrosis virus, Mol. Mar. Biol. Biotechnol. 5, 114-122. BARRY,M. A,, BARRY,M. E., JOHNSTON, S. A. (1994), Production of monoclonal antibodies by genetic immunization, Biotechniques 16, 616-8,620. BARRY,M.A., JOHNSTON,S. A. (1997), Biological features of genetic immunization, Vaccine 15, 788-791. BARRY,M.A., LAI, W. C., JOHNSTON,S.A. (1995), Protection against mycoplasma infection using expression-libraryimmunization, Nature 377, 632-635. BECICER,S. I., WANG,R., HEDSTROM,R. C., AGUIAR,J.C., JONES,T. R. et al. (1998), Protection of mice against Plasmodium yoelii sporozoite challenge with P. yoelii merozoite surface protein 1 DNA vaccines, Infect. Immun. 66, 3457-3461. BENDER,B. S., ULMER,J. B., DEWIIT, C. M., COTTEY,R., TAYLOR, S. F. et al. (1998), Immunogenicity and efficacy of DNA vac-

cines encoding influenza A proteins in aged mice, Vaccine 16, 1748-1755. BOCZIZOWSICI, D., NAIR S. K., SNYDER, D., G I L B o A , E. (1996),Dendritic cells pulsed with RNA are potent antigen-presenting cells in vitro and in uivo,1.Exp. Med. 184, 465-472. BONATO,VL.. LIMA,V.M., TASCON, R.E., LOWRIE,D. B., SILVA,C. L. (1998), Identification and characterization of protective T cells in hsp65 DNA-vaccinated and Mycobacterium tuberculosis-infected mice, Infict. Irnmun. 66, 169.175. BOT, A,, ANTOHI, S., BOT, S., GARCIA,S.A., BONA, C. (1997), Induction of humoral and cellular immunity against influenza virns by immunization of newborn mice with a plasmid bearing a hemagglutinin gene, Jut. Immunol. 9, 1641-1650. BOURNE,N., MILLIGAN,G. N., SCHLEISS,M. R., BERNSTEIN,D.I., STANBERRY, L.R. (1996), DNA immunization confers protective immunity on mice challenged intravaginally with herpes simplex virus type 2 Vaccine 14, 1230-1234. BOURNE,N., STANBERRY, L. R., BERNSTEIN, D. I., LEW, D. (1996), DNA immunization against experimental genital herpes simplex vims infection, /. Infct. Dis. 173, 800-807. BOYER,J. D., WANG,B., UGEN, I<.E., AGADJANYAN, M. G., JAVADIAN, A. et a]. (1996), In vivo protective anti-HIV immune responses in nonhuman primates through DNA immunization, 1.Med. Primatol. 25, 242-250. BOYLE,J.S., BRADY,J.L., LEW, A.M. (199S), Enhanced responses to a DNA vaccine encoding a fusion antigen that is directed to sites of immune induction, Nature 392, 408-411. BOHM, W., ICUHROBER,A., PAIER,T., MERTENS, T., REIMANN,J., SCHIRMBECK, R. (1996), DNA vector constructs that prime hepatitis B surface antigen-specific cytotoxic T lymphocytes and antibody responses in mice after intramuscular injection, /. ImuMunol. Methods 193, 29-40. R., BOHM, W., MERTENS,T., SCHIRMBECK, REIMANN,J. (1998), Routes of plasmid DNA vaccination that prime murine humoral and cellular immune responses, Vaccine 16, 949954.

I

I

62 Jorg Reimann et a/.

CONRY, R. M., LOBUGLIO,A. F., LOECHEL, F., W., SCHIRMBECR, R., REIMANN,J. MOORE, S. E., SUMEREL, L.A. et al. (1995b),A (1997),Targeting an anti-viral CD8' T cell carcinoembryonic antigen polynucleotide response to a growing tumor facilitates its vaccine has in viuo antitumor activity, Gene rejection, Cancer Immunol. lmmunother. 44, Ther. 2, 59-65. 230-238. R. M., LOBUGLIO, A. F., LOECHEL,F., BUELER,H., MULLIGAN,R. C. (1996), Induction CONRY, MOORE, S. E., SUMEREL,L. A. et a. (1995a), of antigen-specific tumor immunity by genetic and cellular vaccines against MAGE: A carcinoembryonic antigen polynucleotide vaccine for human clinical use, Cancer Gene enhanced tumor protection by coexpression Tner. 2, 33-38. of granulocyte-macrophage colony-stimulatCONRY, R. M., LOBUGLIO,A. F., WRIGHT,M., ing factor and B7-1, Mol. Med. 2, 545-555. SUMEREL, L., PIKE,M. J. et al. (1995c), BUTLER,D., MAURICE,J., O'BRIEN,C. (1997), Characterization of a messenger RNA polyVaccines: a roller-coaster of hopes, Nature nucleotide vaccine vector, Cancer Res. 55, 386, 537-538. 1397-1400. D. B. CHATTERGOON, M., BOYER,7.. WEINER, G., LOBUGLIO,A. F., R. M., WIDERA, (1997), Genetic immunization: a new era in CONRY, J.T., MOORE, S. E. et al. (1996b), vaccines and immune therapeutics, FASEB J. FULLER, Selected strategies to augment polynucleotide 11, 753-763. vaccination, Gene Ther. 3, 67-74. CHEN,S. C., J O N E S , D. H., FYNAN, E. F., CORNELISSEN, A.W, SCHE-ITERS,T. P. (1996), FARRAR, G. H., CLEGG, J. C. et al. (1998), Protective immunity induced by oral immu- Vaccines against protozoal diseases of veterinary importance, FEMS Immunol. Med. nization with a rotavirus DNA vaccine Microbiol. 15, 61-72. encapsulated in microparticles, J. Virol. 72, D.A., TIGHE,H. CORR,M., LEE, D. J., CARSON, 5757-5761. (1996), Gene vaccination with naked plasmid CHOI,A. H., KNOWLTON, D. R., MCNEAL, M. M., WARD,R. L. (1997), Particle bombard- DNA: mechanism of CTL priming, J. Exp. Med. 184, 1555-1560. ment-mediated DNA vaccination with rotaJ., ~ TIGHE,H., LEE, D., DUDLER, virus VP6 induces high levels of serum rota- C O RM., TRIEU,M. et al. (1997), Costimulation virus IgG but fails to protect mice against provided by DNA immunization enhances challenge, Virology 232, 129-138. antitumor immunity, J. Immunol. 159, 4999CHOW,Y. H., HUANG, w. L., CHI, w. I<., CHU, 5004. Y. D., TAO,M. H. (1997), Improvement of T. J., BABIUIZ, L.A. (1993), hepatitis B virus DNA vaccines by plasmids Cox, G. J., ZAMB, coexpressing hepatitis B surface antigen and Bovine herpesvirus 1: immune responses in mice and cattle injected with plasmid DNA, interleukin-2, J. Virol., 71, 169-178. I. F., BERZOFSKY,J.A., CARBONE, D. P. J. Virol. 67, 5664-5667. CIERNIK, (1996), Induction of cytotoxic T lymphocytes DAVIS,H. L. (1998), DNA-based immunization and antitumor immunity with DNA vaccines against hepatitis B: experience with animal expressing single T cell epitopes, J. Immunol. models, Cum. Top. Microbiol. Immunol. 226, 57-68. 156, 2369-2375. R., REIMANN, J., CLARICE, N. J., HISSEY,P., BUCHAN,I<., HARRIS,DAVIS,H. L., SCHIRMBECR, WHALEN, R. G. (l995), DNA-mediated imS. (1997), pPV: a novel IRES-containing vector to facilitate plasmid immunization and munization in mice induces a potent MHC antibody response characterization, Imrnuno- class I-restricted cytotoxic T lymphocyte response to Hepatitis B virus surface antigen, technology 3, 145-153. Hum. Gene Ther. 6 , 1447-1456. CONRY, R. M., LOBUGLIO, A. F., CURIEL, D.T. DAVIS,H. L., WHALEN, R. G., DEMENEIX, B.A. (1996a), Polynucleotide-mediated immuni(1993), Direct gene transfer into skeletal zation therapy of cancer, Semin. Oncol. 23, muscle in uivo: factors affecting efficiency of 135-147. transfer and stability of expression, Hum. CONRY, R. M., LOBUGLIO, A. F., KANTOR,J., Gene Ther. 4, 151-159. SCHLOM, J., LOECHEL,F. et al. (1994), ImH. (1993), mune response to a carcinoembryonic anti- DIRKS,W, WIRTH,M., HAUSER, Dicistronic transcription units for gene gen polynucleotide vaccine, Cancer Res. 54, 1164-1168. BOHM,

3 Genetic Vaccination with Plasmid Vectors

expression in mammalian cells, Gene 128, ERTL, H. C., XIANG,Z . (1996a), Novel vaccine 247.249. approaches, /. Imrnunol. 156, 3579-3582. DIXON,B. (1995),The third vaccine revolution, ERTL, H. C., XIANG, Z . Q. (1996b), Genetic Biotechnology 13, 420-426. immunization, Viral. Immunol. 9, 1-9. P. L., ZAUGG, R. H., NORMAN,J.A. DOE,B., SELBY, M., BARNETT,S., BAENZIGER,J., FELGNER, WALKER, C. M. (1996), Induction of cytotoxic (1995), Synthetic recombinant DNA delivery T lymphocytes by intramuscular immunizafor cancer therapeutics, Cancer Gene Tner. 2, tion with plasmid DNA is facilitated by bone 61-65. FELTQUATE, D. M., HEANEY, S., WEBSTER, R. G., marrow-derived cells, Proc. Natl. Acad. Sci. USA 93, 8578-8583. ROBINSON, H. L. (1997), Different T helper cell types and antibody isotypes generated by DONNELLY, J .J., FRIEDMAN, A,, MARTINEZ,D., MONTGOMERY, D. L., SHIVER,J. W. et al. saline and gene gun DNA immunization, (1995), Preclinical efficacy of a prototype J. Im~zunol.158, 2278-2284. DNA vaccine: enhanced protection against FORSTHUBER, T., YIP, H. C., LEHMANN,P.V. antigenic drift in influenza vims, Nature Med. (1996),Induction of TH l and TH2 immunity 1, 583-587. in neonatal mice, Science 271, 1728-1730. DONNELLY, J. J., FRIEDMAN, A., ULMEX, J. B., FULLER,D. H., MURPHEY,C. M., CLEMENTS, J., LIU, M.A. (1997). Further protection against BARNETT,S., HAYNES, J, R. (1996), Induction antigenic drift of influenza virus in a ferret of immunodeficiency virus-specificimmune model by DNA vaccination, Vaccine 15, responses in rhesus monkeys following gene 865-868. gun-mediated DNA vaccination, 1. Med. DONNELLY, J. J., MARTINEZ,D., JANSEN, I<. U., Primatol. 25, 236-241. ELLIS, R. W., MONTGOMERY, D. L., LIU, M.A. FULLER J.T., FULLER, D. H., MCCABE,D., (1996), Protection against papillomavirus HAYNES, J. R., WIDERA, G. (1995), Immune with a polynucleotide vaccine, 1.Infect. Dis. responses to hepatitis B virus surface and 713, 314-320. core antigens in mice, monkeys, and pigs DONNELLY, J. J., ULMER, J. B., LIU, M.A. (1994), after Accell particle-mediated DNA immuniImmunization with DNA, J. Immunol. zation, Ann. N Y Acad. Sci. 772, 282-284. Methods 176, 145-152. E. F., WEBSTER, R. G., F U L L ED. ~ H., FYNAN, J. J., ULMER, J. B., LIU, M.A. DONNELLY, HAYNES, J.R., SANTORO, J.C., ROBINSON, (1997a), DNA Vaccines, L@ Sci. 60, 163-172. H. L. (1993), DNA Vaccines: protective DONNELLY, J. J., ULMEQJ. B., SHIVER,J.W, immunizations by parenteral, mucosal, and LIU, M. A. (1997b),DNA vaccines, Annu. Rev. gene-gun inoculations, Proc. Natl. h a d . Sci. Immunol. 15, 617-648. U S A 90, 11478-11482. DOOLAN, D.L., SEDEGAH, M., HEDSTROM, FYNAN,E. F., WEBSTER, R. G., FULLER D. H., R. C., HOBART, I?., CHAROENVIT, Y., HOFFMAN, HAYNES, J. R., SANTORO, J.C., ROBINSON, S. L. (1996),Circumventing genetic restriction H. L. (1995),DNA vaccines: a novel approach of protection against malaria with multigene to immunization, Int. J. Immunopharrrcacol. DNA immunization: CD8' cell-, interferon 17, 79-83. gamma-, and nitric oxide-dependent GEISSLER, E.K., WANG,J., FECHNER,J. H., J R , immunity, J. Exp. Med. 183, 1739-1746. BURLINGHAM,W J., KNECHTLE,S. J. (1994), Dow, S. W., FRADKIN,L. G., LIGGITT, D. H., Immunity to MHC class 1 antigen after direct T. A. WILLSON, A. P., HEATH,T. D., POTTER, DNA transfer into skeletal muscle, /. Immu(1999), Lipid-DNA complexes induce potent nol. 152, 413-421. activation of innate immune responses and G E I S S L EM., ~ GESIEN, A., TOIWSHIGE, K., antitumor activity when administered intraJ. R. (1997), Enhancement of cellular WANDS, venously, J. Irnmunol. 163, 1552-1561. and humoral immune responses to hepatitis EISENBRAUN, M. D., FULLER,D. H., HAYNES, C virus core protein using DNA-based vacJ. R. (1993), Examination of parameters cines augmented with cytokine-expressing affecting the elicitation of humoral immune plasmids, J. hmunol. 158, 1231-1237. M., SCHIRMBECK,R., REIMANN,J., responses by particle bombardment-mediated GEISSLER, genetic immunization, D N A Cell Bid. 12, BLUM,H. E., WANDS, J. R. (1998), Cytokine 791-797. and hepatitis B virus DNA co-immunizations enhance cellular and humoral immune

I

63

64

I

]erg Reimann et al.

responses to the middle but not to the large hepatitis B viius surface antigen in mice, Hepatology 28, 202-210. GEISSLER, M., WANDS,G., GESIEN, A., DE-LA, M. S., BELLET, D., WANDS,J. R. (1997), Genetic immunization with the free human chorionic gonadotropin beta subunit elicits cytotoxic T lymphocyte responses and protects against tumor formation in mice, Lab. Invest. 76, 859-871. V, JONS, A,, MAKOSCHEY, B., VISSER, GERDTS, N., METTENLEITER, T. C. (1997), Protection of pigs against Aujeszky’s disease by DNA vaccination, J. Gen. Virol. 78, 2139-2146. GHIASI,H., CAI, S., SLANINA, S., NESBURN, A. B., WECHSLER, S. L. (1995), Vaccination of mice with herpes simplex virus type 1 glycoprotein D DNA produces low levels of protection against lethal HSV-1 challenge, Antiviral Res. 28, 147-157. C. IZ., SOROCEANU, L., SMITH,N., GOLDMAN, GILLESPIE, G.Y., SHAW,W. et al. (1997), In uitro and in viuo gene delivery mediated by a synthetic polycationic amino polymer, Nature Biotechnol 15, 462-466. GONZALEZ ARMAS,J. C., MORELLO, C. S., CRANMER, L.D., SPECTOR, D.H. (1996), DNA immunization confers protection against murine cytomegalovirus infection, J. Virol. 70, 7921-7928. Gu, M. L., LEPPIA, S. H., KLINMAN,D. M. (1999), Protection against anthrax toxin by vaccination with a DNA plasmid encoding anthrax protective antigen, Vaccine 17, 340-344. GURUNATHAN, S., IRVINE,I<. R., Wu, C.Y., E., PRUSSIN,C., COHEN,J. I., THOMAS, RESTIFO, N. P., SEDER,R.A. (1998), CD40 ligand/trimer DNA enhances both humoral and cellular immune responses and induces protective immunity to infectious and tumor challenge., J. Irnmunol., 161, 4563-4571. S., IZLINMAN, D. M., SEDER, GURUNATHAN, R. A. (ZOOO), DNA vaccines: immunology, application and optimization, Annu. Rev. Irnmunol. 18, 927-974. GURUNATHAN, S., SACKS,D. L., BROWN,D. R., H. et al. (1997), REINER,S. L., CHAREST, Vaccination with DNA encoding the immunodominant LACK parasite antigen confers protective immunity to mice infected with Leishmania major, J. Exp. Med. 186, 1137-1147.

HANIZE, T., BLANCHARD,T. J., SCHNEIDER, J., HANNAN, C. M., BECICER, M. et al. (1998), Enhancement of MHC class I-restricted peptide-specific T cell induction by a DNA prime/MVA boost vaccination regime, Vaccine 16, 439-445. T. J., HANICE, T., NEUMANN,V.C., BLANCHARD, SWEENEY, P., HILL,A.V., SMITH,G. L., MCMICHAEL, A. (1999), Effective induction of HIV-specific CTL by multi-epitope using gene gun in a combined vaccination regime, Vaccine 17, 589-596. HARRISON, G. S., WANG,Y., TOMCZAIZ, J., E. J. et al. (1995), OptiHOGAN,C., SHPALL, mization of gene transfer using cationic lipids in cell lines and primary human CD4 and CD34’ hematopoietic cells, Biotechniques 19, 816-823. HASSETT, D. E., WHITTON,J. L. (1996), DNA immunization, Trends Microbiol. 4, 307-312. HASSETT,D.E., ZHANG,J., WHITTON,J. L. (1997), Neonatal DNA immunization with a plasmid encoding an internal viral protein is effective in the presence of maternal antibodies and protects against subsequent viral challenge, J. Virol. 71, 7881-7888. HAWKINS, R. E., ZHU,D., OVECIU,M., WINTER, G., HAMBLIN, T. J. et al. (1994), Idiotypic vaccination against human B-cell lymphoma. Rescue of variable region gene sequences from biopsy material for assembly as singlechain Fv personal vaccines, Blood 83, 32793288. HAYNES, J.R., MCCABE,D.E., SWAIN,W.F., WIDERA,G., FULLER, J.T. (1996), Particlemediated nucleic acid immunization, J. Biotechnol. 44, 37-42. HE, X. S., CHEN,H. S., CHU, I<., RIVKINA, M., ROBINSON, W. S. (1996b), Costimulatory protein B7-1 enhances the cytotoxic T cell response and antibody response to hepatitis B surface antigen, Proc. Natl. Acad. Sci. U S A 93, 7274-7278. HE, X. S., RIVICINA, M., ROBINSON, W. S. (1996a), Construction of adenoviral and retroviral vectors coexpressing the genes encoding the hepatitis B surface antigen and B7-1 protein, Gene 175, 121-125. HERRMANN, J.E., CHEN,S.C., FYNAN, E. F., J. C., GREENBERG, H. B. et al. SANTORO, (1996), Protection against rotavirus infections by DNA vaccination, J. Infect. Dis. 174 (Suppl. l), s93-7.

’

3 Genetic Vaccination with Plasmid Vectors

HILDEMAN, D., SALVATO,M., WHITTON,J.L., J E N K I N S , M., KERR, D., FAYER, R., WALL,R. MULLER, D. (1996), Vaccination protects b2 (1995),Serum and colostrum antibody microglobulin deficient mice from immune responses induced by jet-injection of sheep mediated mortality but not from persisting with DNA encoding a Cryptosporidium viral infection, Vaccine 14, 1223-1229. pawum antigen, Vaccine 13, 1658-1664. HOERR, I., OBST,R., RAMMENSEE, H. G., J U N G , JIAO, S.,WILLIAMS, I?., BERG, R. I<., HODGEG. (2000), i n vivo application of RNA leads to MAN, B.A., LIU, L. M. et al. (1992), Direct induction of specific cytotoxic T lymphocytes gene transfer into nonhuman primate myofibers in vivo, Hum. Gene n e r . 3, 21-33. and antibodies, Eur. /. Immunol. 30, 1-7. Hsu, C. H., CHUA,K.Y., TAO,M. H., LAI, Y.L., J O H A N N I N G , F. W., CONRY, R. M., LOBUGLIO, L.A. et al. Wu, H. D. et al. (1996),Immunoprophylaxis A. F., WRIGHT,M., SUMEREL, (1995),A Sindbis virus mRNA polynucleotide of allergen-induced immunoglobulin E synthesis and airway hyperresponsiveness in vector achieves prolonged and high level vivo by genetic immunization, Nature Med. 2, heterologous gene expression in vivo, Nucleic 540-544. Acids Res. 23, 1495-1501. HUANG, H.V. (1996), Sindbis virus vectors for JOHNSTON, S.A., TANG,D. C. (1994), Gene gun expression in animal cells, Cur. Opin. transfection of animal cells and genetic Biotechnol. 7 , 531-535. immunization, Methods Cell Biol. 43 Pt A, HUYGEN, K., CONTENT, J., DENIS,O., 353-365. MONTGOMERY, I ~ M A T A.T., H , FENG,C. G., MACDONALD, D. L., YAWMAN,A.M. et al. M., BRISCOE, H., BRITTON,W. J. (1999b), (1996), Imrnunogenicity and protective effiDifferential protective efficacy of DNA vaccacy of tuberculosis DNA vaccine, Nature cines expressing secreted proteins of MycoMed. 2, 893-898. ICHINO, M., MOR,G., CONOVER, J., WEISS, bacterium tuberculosis, Infect. Immun. 67, M. et al. (1999), Factors asso- 1702-1707. W. R., TAKENO, ciated with the development of neonatal KAMATH, A.T., HANKE, T, BRISCOE, H., tolerance after the administration of a plasBRIITON, W. J. (1999a),Co-immunization mid DNA vaccine, /. Imrnunol. 162, 3814with DNA vaccines expressing granulocyte3818. macrophage colony-stimulating factor and INCHAUSPE, G. (1997), Gene vaccination for mycobacterial secreted proteins enhances hepatitis C, Springer Semin. Immunopathol. T-cell immunity, but not protective efficacy against Mycobacterium tuberculosis, Immu19, 211-221. ISHII,N., FUKUSHIMA, J., ICANEKO,T., OICADA, nology 96, 511-516. T., SYLVESTER, I. D., AMBALI, A. G., E., TANI,I<. et al. (1997), Cationic liposomes KANELLOS, are a strong adjuvant for a DNA vaccine of HOWARD, C. R., RUSSELL, P. H. (1999), The human immunodeficiency virus type 1,AIDS safety and longevity of DNA vaccines for fish, Res. Hum. Retroviruses 13, 1421-1428. Immunology 96, 307-313. ISHII, N., SUGITA, Y., NAICAJIMA,H., BUIWWA,KELLER, E.T., BURKHOLDER,J. K., SHI, F., PUGH,T. D., MCCABE, D. et al. (1996), In vivo H., ASAKURA, Y., OKUDA, I<. (1997), Genetic particle-mediated cytokine gene transfer into control of immune responses to HIV-1 env DNA vaccine, Microbiol. Immunol. 41, canine oral mucosa and epidermis, Cancer 421-425. Gene Ther. 3, 186-191. IWASAKI,A,, BARBER,B. H. (1998), Induction KIM, J. J., AYYAVOO, V, BAGARAZZI,M. L., by DNA immunization of a protective antiCHATTERGOON, M.A., DANG,K. et al. (1997), tumor cytotoxic T lymphocyte response In vivo engineering of a cellular immune against a minimal-epitope-expressing tumor, response by coadministration of IL-12 Cancer Immunol. Immunother. 45, 273-279. expression vector with a DNA immunogen, IWASAKI,A., STIERNHOLM,B. J., CHAN,A. K., J. Irnmunol. 158, 816-826. BERINSTEIN,N. L., BARBER,B. H. (1997), KIM, J. J., BAGAEAZZI,M. L.,TRIVEDI, N., Hu, Enhanced CTL responses mediated by Y., KAZAHAYA,K. et al. (1997), Engineering of plasmid DNA immunogens encoding in vivo immune responses to DNA immunicostimulatory molecules and cytolunes, zation via codelivery of costimulatory moleJ . rmmunoi. 158, 4591-4601. cule genes, Nature Biotechnol. 15, 641-646.

I

65

I

66 Jorg Reirnann et al.

bodies against human prion proteins (PrP) KIM, J. J., N O ~ I N G H A ML., K., WILSON,D.M., by DNA-mediated immunization of PrPO/O BAGARAZZI,M. L., TSAI,A. et al. (1998), Engineering DNA vaccines via co-delivery of mice, 1.Imrnuid. Methods 199, 109-118. KRISHNAN, S., HAENSLER, J., MEULIEN,P. co-stimulatory molecule genes, Vaccine 16, (1995), Paving the way towards DNA vac1828-1835. cines, Nature Med. 1, 521-522. KIM, J. J., TRIVEDI, N. N., N O ~ I N G H A ML., K., R., REIMANN,J. MORRISON,L., TSAI,A. et al. (1998), Modu- KUHROBER,A,, SCHIRMBECK, (1994), A selfreactive class I-restricted T cell lation of amplitude and direction of in vivo response of H-2b mice to determinants of the immune responses by co-administration of cytokine gene expression cassettes with DNA V8.2 domain of the T cell receptor for antigen, Irnrnunology 83, 532-539. immunogens, Eur. J. Imrnunol. 28, 1089KUMAR,V., SERCARZ, E. (1996), Genetic 1103. vaccination: the advantages of going naked, KIM, J. J., WEINER, D.B. (1997), DNA gene Nature Med. 2, 857-859. vaccination for HIV, Springer Semin. KWAK, L.W, YOUNG, H.A., PENNINGTON, Immunopathol. 19, 175-194. R. W, WEEKS,S. D. (1996), Vaccination with KIM, T. S., DEKRUYFF,R. H., RUPPER, R., syngeneic, lymphoma-derived immunogloMAECKER,H.T., LEVY, S., UMETSU, D.T. (1997),An ovalbumin-IL-12fusion protein is bulin idiotype combined with granulocyte/ macrophage colony-stimulating factor primes more effective than ovalbumin plus free recombinant IL-12 in inducing a T helper cell mice for a protective T-cell response, Proc. Natl. Acad. Sci. USA 93, 10972-10977. type 1-dominated immune response and KWISSA, M., VON KAMPEN, J., ZURBRIGGEN, R., inhibiting antigen-specific IgE production, GLUECK, R., REIMANN,J. et al. (2000a), 1.Immunol. 158, 4137-4144. Efficient vaccination by intradermal or intraKIMURA,K., NISHIMURA,H., HIROSE,K., muscular inoculation of plasmid DNA MATSUGUCHI, T., NIMURA,Y., YOSHIKAI, Y expressing hepatitis B surface antigen under (1999), Immunogene therapy of murine desmin promoter/enhancer control, Vaccine, fibrosarcoma using IL-15 gene with high 18, 2337-2344. translation efficiency, Eur. J . Immunol. 29, KWOH,D.Y., COFFIN,C. C., LOLLO, C. P., JOVE1532-1542. NAL, J., BANASZCZYK,M. G. et al. (1999), KIPPS, T., MENDOZA, R. (1999), Extending Stabilization of poly-L-lysine/DNApolyplexes genetic vaccines with chemolunes, Nature for in vivo gene delivery to the liver, Biochim. Biotechnol. 17, 226-227. Biophys. Acta 1444, 171-190. KLAVINSKIS, L. S., BARNFIELD,C., GAO,L., M., UNSINGER, J., SCHIRMBECK, R., PARKER, S. (1999), Intranasal immunization KWISSA, HAUSER, H., REIMANN, J. (in press), Polywith plasmid DNA-lipid complexes elicits valent DNA vaccines with bidirectional mucosal immunity in the female promoters, 1.Mol. Med. genital and rectal tracts, J. Irnmunol. 162, LAI, W C., BENNETT,M., JOHNSTON, S.A., 254-262. S. P. (1995), Protection M. A,, PAICES, KLINMAN,D. M., VERTHELYI,D., TAKESHITA, BARRY, F., ISHII,I<. J. (1999), Immune recognition of against Mycoplama pulmorzis infection by genetic vaccination, DNA Cell Biol. 14, foreign DNA: a cure for bioterrorism? 643- 651. Immunity 11, 123-129. S. P., REN, K., Lu, Y. S., KODIHALLI, S., HAYNES, J. R., ROBINSON,H. L., LAI,W. C., PAKES, BENNETT,M. (1997), Therapeutic effect of WEBSTER, R. G. (1997), Cross-protection DNA immunization of genetically susceptible among lethal H5N2 influenza viruses induced by DNA vaccine to the hemagglutinin, mice infected with virulent Mycoplama pulmonis, J. Immunol. 158, 2513-2516. /. Virol. 71, 3391-3396. LARSEN, D. L., DYBDAHL, S. N., MCGREGOR, KOPROWSKI,H., WEINER,D. B. (1998), DNA M. W., DRAPE,R., NEUMANN,V et al. (1998), vaccination/genetic vaccination, Cum. Top. Coadministration of DNA encoding interMicrobiol. Immunol. 226, V-XIII. leukin-6 and hemagglutinin confers protecKOVARIK,J., SIEGRIST, C. A. (1998), Immunity tion from influenza virus challenge in mice, in early life, Immunol. Today 19, 150-152. KRASEMANN,S., GROSCHUP, M., HUNSMANN, J . Virol. 72, 1704-1708. G., BODEMER,W. (1996), Induction of anti-

3 Genetic Vaccination with Plasmid Vectors

LEE, D. J., TIGHE,H., CORR,M., ROMAN,M., CARSON,D. A. et al. (1997), Inhibition of IgE antibody formation by plasmid DNA immunization is mediated by both CD4' and CD8' T cells, Int. Arch. Allergy Immunol. 113, 227.230. LEITNER,W. W., SEGUIN,M. C., BALLOU, W. R., SEITZ,J, P., SCHULTZ,A.M. et al. (1997), Immune responses induced by intramuscular or gene gun injection of protective deoxyribonucleic acid vaccines that express the circumsporozoite protein from Plasmodium berghei malaria parasites, /. Immunol. 159, 6112-6119. LETVIN,N. L. (1998), Progress in the development of an HIV-1 vaccine, Science 280, 1875-1880. LETVIN,N. L., MONTEFIORI, D.C., YASUTOMI, Y., PERRY,H.C., DAVIES,M. E. et al. (1997), Potent, protective anti-HIV immune responses generated by bimodal HIV envelope DNA plus protein vaccination, Proc. Natl. Acad. Sci. U S A 94, 9378-9383. LEVITSKY, H . I. (1997), Accessories for naked DNA vaccines, Nature Biotechnol 15,619-620. LI, X., SAMBHARA,S., LI, C.X., EWASYSHYN, M., PARRINGTON, M. et al. (1998), Protection against respiratory syncytial virus infection by DNA immunization., J. Exp. Med. 188, 681-688. LIN, Y. L., CHEN,L. K., LIAO, C. L., YEH, C.T., MA, S. H . et al. (1998), DNA immunization with Japanese encephalitis virus nonsttuctural protein NS1 elicits protective immunity in mice, /. Virol. 72, 191-200. LIU, M.A. (1995), Overview of DNA vaccines, Ann. N Y A c a d . Sci. 772, 15-20. LIU, M. A,, McClements, W, Ulmer, J. B., Shiver, J., Donnelly, J. J. (1997), Irnmunization of non-human primates with DNA vaccines, Vaccine 15, 909-912. LIU, M.A., Yasutomi, Y., Davies, M. E., Perry, H. C., Freed, D. C. et al. (1996),Vaccination of mice and nonhuman primates using HIVgene-containing DNA, Antibiot. Chemother. 48, 100-104. LIU, Y., MOUNKES,L. C., LIGGITT,H.D., BROWN,C. S., SOLODIN,I. et al. (1997), Factors influencing the efficiency of cationic liposome-mediated intravenous gene delivery, Nature Biotechnob 15, 167-173. LOWRIE,D. B. (1998), DNA vaccination exploits normal biology, Nature Med. 4, 147-148.

LOWRIE,D. B., SILVA,C. L., TASCON, R. E. (1997), DNA vaccines against tuberculosis, Immunol. Cell Biol. 75, 591-594. LOWRIE,D. B., TASCON,R. E., BONATO,V L., LIMA,V M . , FACCIOLI,L. H. et al. (1999), Therapy of tuberculosis in mice by DNA vaccination, Nature 400, 269-271. LOZES,E., HUYGEN, I<., CONTENT,J., DENIS,O., MONTGOMERY, D. L. et al. (1997), Immunogenicity and efficacy of a tuberculosis DNA vaccine encoding the components of the secreted antigen 85 complex, Vaccine 15, 830-833. Lu, S., ARTHOS,J., MONTEFIORI, D.C., YASUTOMI, Y., MANSON,K. et al. (1996), Simian immunodeficiency virus DNA vaccine trial in macaques, J. Virol. 70, 3978-3991. Lu, Y., XIN, K. Q., HAMAJIMA, K., TSUJI,T., AOKI, I. et al. (1999), Macrophage inflammatory protein-1 alpha (MIP-1 alpha) expression plasmid enhances DNA vaccineinduced immune response against HIV-1, Clin. Exp. Immunol. 115, 335-341. MACKLIN,M. D., MCCABE,D., MCGREGOR, V, MEYER,T. et al. (1998), M. W., NEUMANN, Immunization of pigs with a particlemediated DNA vaccine to influenza A virus protects against challenge with homologous virus, J . Virol. 72, 1491-1496. MAECI~ER, H. T., UMETSU,D.T., DEKRUYFF, R. H., LEVY,S. (1997), DNA vaccination with cytoltine fusion constructs biases the immune response to ovalbumin, Vaccine 15, 1687-1696. MAHVI, D.M., BURKHOLDER, J. K., TURNER, I., CULP,J., MALTER,J. S. et al. (1996), Particlemediated gene transfer of granulocytemacrophage colony- stimulating factor cDNA to tumor cells: implications for a clinically relevant tumor vaccine, Hum. Gene Ther. 7 , 1535-1543. MANICKAN,E., ROUSE,R.J., Yu, Z., WIRE, W. S., ROUSE,B.T. (l995), Genetic immunization against herpes simplex virus. Protection is mediated by CD4' T lymphocytes, J. Immunol. 155, 259-265. MANICICAN,E., Yu, Z., ROUSE,B.T. (1997), DNA immunization of neonates induces immunity despite the presence of maternal antibody, J. Clin. Invest. 100, 2371-2375. MARTINEZ,X., BRANDT,C., SADDALLAH, F., TOUGNE,C., BARRIOS,C. et al. (1997), DNA immunization circumvents deficient induction of T helper type 1 and cytotoxic T

I

67

l

68 Jdrg Reimann et a\.

lymphocyte responses in neonates and during early life, Proc. Natl. Acad. Sci. U S A 94, 8726-8731. MARTINON, F., KRISHNAN,S., LENZEN,G., MAGNE,R., GOMARD, E. et al. (1993), Induction of virus-specific cytotoxic T lymphocytes in vivo by liposome-entrapped mRNA, Eur. J. Immunol. 23, 1719-1722. MARTINS,L. P., LAW,L. L., ASANO,M. S., AHMED,R. (1995), DNA vaccination against persistent viral infection, J. Virol. G9, 2574-2582. MARWICK, C. (1995), Exciting potential of DNA vaccines explored, JAMA 273, 1403-1404. MCCARTHY, M. (1996), DNA vaccination: a direct line to the immune system, Lancet 348, 1232-1232. MCCLEMENTS, W. L., ARMSTRONG, M. E., KEYS, R. D., LIU, M.A. (1996), Immunization with DNA vaccines encoding glycoprotein D or glycoprotein B, alone or in combination, induces protective immunity in animal models of herpes simplex virus-2 disease, Proc. Natl. Acad. Sci. U S A 93, 11414-11420. M. E., KEYS, MCCLEMENTS, W. L., ARMSTRONG, R. D., LIU, M. A. (1997),The prophylactic effect of immunization with DNA encoding herpes simplex virus glycoproteins on HSVinduced disease in guinea pigs, Vaccine 15, 857-860. MCDANIEL, L. S., LOECHEL, F., BENEDICT,C., GREENWAY, T., BRILES,D. E. et al. (1997), Immunization with a plasmid expressing pneumococcal surface protein A (PspA) can elicit protection against fatal infection with Streptococcus pneumoniae, Gene Ther. 4, 375-377. MCDONNELL, W.M., ASICARI, F. K. (1996), DNA vaccines, N. Engl. J. Med. 334, 42-45. MICHEL,M.-L. (1995), DNA-mediated immunization: prospects for hepatitis B vaccination, Res. Virol. 146, 261-265. M., MICHEL,M.-L., DAVIS,H. L., SCHLEEF, MANCINI,M., TIOLLAIS, P., WHALEN, R. G. (1995), DNA-mediated immunization to the hepatitis B surface antigen in mice: aspects of the humoral response mimic hepatitis B viral infection in humans, Proc. Natl. Acad. Sci. U S A 92, 5307-5311. J. J., SHIVER, MONTGOMERY, D. L., DONNELLY, J.W., LIU, M.A., ULMER, J. B. (1994), Protein expression i n vivo by injection of polynucleotides, Curr. Opin. Biotechnol. 5 , 505-510.

MONTGOMERY, D.L., SHIVER,J. W., LEANDER, I<. R., PERRY, H. C., FRIEDMAN, A. et al. (1993), Heterologous and homologous protection against influenza A by DNA vaccination: optimization of DNA vectors, DNA Cell Biol. 12, 777-783. MOR, G., SINGLA,M., STEINBERG,A.D., HOFFMAN, S. L., OIUJDA,K., I
3 Genetic Vaccination with Plasrnid Vectors with an expression genomic library of Leishmania major, J. Immunol. 163, 1467-1472. PISETSKY, D. S. (1997), DNA and the immune system, Ann. Intern. Med. 126, 169-171. PLAUTZ,G.E., NABEL,E.G., Fox, B., YANG, Z.Y., JAFFE, M., GORDON, D., CHANG,A., NABEL,G . J , (1994), Direct gene transfer for the understanding and treatment of human disease, Ann.NY Acad.Sci., 716, 144-153. PLAUTZ,G. E., YANG, Z.Y., Wu, €3. Y., GAO,X., HUANG,L., NABEL,G. J. (1993), Immunotherapy of malignancy by i n vim gene transfer into tumors, Proc. Natl. Acad. Sci. U S A 90, 4645-4649. PLEBANSKI, M., LEE, E.A., HANNAN, C. M., I<. L., GILBERT,S. C. et al. (1999), FLANAGAN, Altered peptide ligands narrow the repertoire of cellular immune responses by interfering with T-cell priming, Nature Med. 5, 565-571. PORGADOR, A,, IRVINE,K. R., IWASAIU, A., BARBER, B. H., RESTIFO,N. P., GERMAIN, R. N. (1998), Predominant role for directly transfected dendritic cells in antigen presentation to CD8' T cells after gene gun immunization, J. Exp. Med. 188, 1075-1082. PRAYAGA, S. I<., FORD,M. J., HAYNES, J. R. (1997),Manipulation of HIV-1 gp120-specific immune responses elicited via gene gunbased DNA immunization, Vaccine 15, 1349-1352. QIU, P., ZIEGELHOFFER, P., SUN,J., YANG,N. S. (1996), Gene gun delivery of mRNA in situ results in efficient transgene expression and genetic immunization, Gene Thev. 3, 262-268. RAKHMILEVICH, A. L., TURNER, J., FORD,M. J., MCCABE,D., SUN,W H. et al. (1996), Gene gun-mediated skin transfedion with interleukin 12 gene results in regression of established primary and metastatic murine tumors, Proc. Natl. Acad. Sci. U S A 93, 62916296. RAMSAY, A. J., RAMSHAW,LA., ADA,G. L. (1997), DNA immunization, Immunol. Cell Biol. 75, 360-363. RAZ, E. (1997), Introduction: gene vaccination, current concepts and future directions, Springer Semin. Irnmunopathol. 19, 131-137. RAz,E., CARSON, D. A,, PARKER, S. E., PARQ T. B., ABAI,A.M. et al. 1994), Intradermal gene immunization: the possible role of DNA uptake in the induction of cellular immunity to viruses, Proc. Natl. Acad. Sci. U S A 91, 9519-9523.

REIMANN, J., SCHIRMBECK, R. (l999), Alternative processing pathways for exogenous and endogenous antigens that can generate peptides for MHC class I-restricted presentation, Immunol. Rev. 171, 131-152. RICHMOND, J. F., Lu, S., SANTORO, J. C., WENG, J., Hu, S. L. et al. (1998), Studies of the neutralizing activity and avidity of anti-human immunodeficiency virus type 1 Env antibody elicited by DNA priming and protein boosting, J. Virol. 72, 9092-9100. ROBINSON, H. L., HUNT,L.A., WEBSTER,R. G. (1993), Protection against a lethal influenza virus challenge by immunization with a haemagglutinin-expressing plasmid DNA, Vaccine 11, 957-960. ROBINSON, H.L., Lu,S., FELTQUATE, D.M., TORRES,C.T., RICHMOND, J. et al. (1996), DNA vaccines, AIDS Res. Hum. Retroviruses 12,455-457. ROBINSON, H. L., MONTEFIORI, D. C., J O H N S O N , R.P., MANSON, I<.H., I ~ L I S H , M. L. et al. (1999),Neutralizing antibodyindependent containment of immunodeficiency virus challenges by DNA priming and recombinant pox virus booster immunizations, Nature Med. 5, 526-534. ROMAN,M., SPIEGELBERG, H . L., BROIDE,D., RAz, E. (1997), Gene immunization for allergic disorders, Springer Semin. Irnmunopathol. 19, 223-232. ROSATO, A,, ZAMBON, A., MILAN,G., CIMINALE, V., D'AGOSTINO,D. M. et al. (1997), CTL response and protection against P815 tumor challenge in mice immunized with DNA expressing the tumor-specific antigen P815A, H u m . Gene Trter. 8, 1451-1458. ROUSE,R. J., NAIQ S. I<., LYDY,S. L., BOWEN, J. C., ROUSE,B.T. (1994),Induction in vitro of primary cytotoxic T-lymphocyte responses with DNA encoding herpes simplex virus proteins, J. Virol. 68, 5685-5689. RUIZ, P. J., GARREN, H., RUIZ, I. U., HIRSCHBERG, D. L., NGUYEN,L.V. et al. (1999), Suppressive immunization with DNA encoding a self-peptide prevents autoimmune disease: modulation of T cell costimulation, J. Immunol. 162, 3336-3341. SAIZAGUCHI, M., NAIZAMURA, H., SONODA, K., HAMADA, F., HIRAI,I<. (1996), Protection of chickens from Newcastle disease by vaccination with a linear plasmid DNA expressing

I

69

70

I

Jorg Reirnann et a\

the F protein of Newcastle disease virns, Vaccine 14, 747-752. SARZOITI, M., DEAN,T.A., REMINGTON,M.P., LY, C. D., FURTH,P.A., ROBBINS,D. S. (1997), Induction of cytotoxic T cell responses in newborn mice by DNA immunization, Vaccine 15, 795-797. SARZOITI, M., ROBBINS,D. S., HOFFMAN, P.M. (1996),Induction of protective CTL responses in newborn mice by a murine retrovirus, Science 271, 1726-1728. SCHIRMBECK, R., BOHM, W., ANDO,I<.-I., CHISARI, F.V., REIMANN,J. (1995), Nucleic acid vaccination primes hepatitis B surface antigen-specific cytotoxic T lymphocytes in nonresponder mice, /. Virol. G9, 5929-5934. SCHIRMBECK, R., BOHM, W., REIMANN,J. (1996), DNA vaccination primes MHC class I-restricted, simian virus 40 large tumor antigen-specific cytotoxic T lymphocytes in H-2d mice that reject syiigeneic tumors, /. Immunol. 157, 3550-3558. SCHNEIDER, J., GILBERT, S. C., BLANCHARD, T. J., HANKE, T., ROBSON,I<. J. et al. (l998), Enhanced immunogenicity for CD8' T cell induction and complete protective efficacy of malaria DNA vaccination by boosting with modified vaccinia virus Ankara, Nature Med. 4, 397-402. SEDEGAH, M., HEDSTROM, R., HOBART, P., HOFFMAN, S. L. (1994), Protection against malaria by immunization with plasmid DNA encoding circumsporozoite protein, Proc. Natl. Acad. Sci. U S A 91, 9866-9870. SEDEGAH, M., JONES, T. R., KAUS M., HEDSTROM, R., HOBART, P. et al. (1998), Boosting with recombinant vaccinia increases immunogenicity and protective efficacy of malaria DNA vaccine, Proc. Natl. Acad. Sci. U S A 95, 7648-7653. SHIVER,J. W., DAVIES, M.E., PERRY, H.C., FREED,D. C., LIU, M. A. (1996), Humoral and cellular immunities elicited by HIV-1 vaccination, J. Pham. Sci. 85, 1317-1324. SHIVER, J. W., DAVIES, M. E., YASUTOMI,Y., PERRY, H. C., FREED,D. C. et al. (1997), AntiHIV env immunities elicited by nucleic acid vaccines, Vaccine 15, 884-887. SHIVER,J.W., PERRY, H.C., DAVIES, M.E., FREED,D. C., LIU, M. A. (1995), Cytotoxic T lymphocyte and helper T cell responses following HIV polynucleotide vaccination, Ann. NYAcad. Sci. 772, 198-208.

SIEGRIST, C.-A. (1997), Potential advantages and risks of nucleic acid vaccines for infant immunization, Vaccine 15, 798-800. SIEGRIST,C.-A,, LAMBERT,P. H. (1996), DNA vaccines: what can we expect? In@. Agents Dis. 5, 55-59. SIEGRIST,C. A. (1997), Vaccination strategies for children with specific medical conditions: a paediatrician's viewpoint, Eur. J. Pediatr. 156,899-904. SIEGRIST,C.A., LAMBERT, P. H. (1997), Immunization with DNA vaccines in early life: advantages and limitations as compared to conventional vaccines, Springer Semin. Immunopatlzol. 19, 233-243. SIMON, M. M., GERN,L., HAUSER, P., ZHONG, W., NIELSEN,P. J. et al. (1996), Protective immunization with plasmid DNA containing the outer surface lipoprotein A gene of Borrelia burgdorferi is independent of an eularyotic promoter, Eur. /. Im,munol. 26, 2831-2840. SIN,J.I., KIM, J. J.. ARNOLD,R.L., SHROFF, K. E., MCCALLUS, D. et al. (1999), IL-12 gene as a DNA vaccine adjuvant in a herpes mouse model: IL-12 enhances Thl-type CD4+ Tcellmediated protective immunity against herpes simplex virus-2 challenge, J. Immunol. 162, 2912-2921. SIN,J.I., KIM, J. J., UGEN,I(. E., CICCARELLI, R. B., HIGGINS, T. J., WEINER, D. B. (1998), Enhancement of protective humoral (Th2) and cell-mediated (Thl) immune responses against herpes simplex virus-2 through codelivery of granulocyte-macrophage colonystimulating factor expression cassettes, Eur. J. Immunol. 28, 3530-3540. SIXT, N., CARDOSO, A,, VALLIER, A,, FAYOLLE, J., BUCI~UND, R., WILD,T. F. (1998), Canine distemper virus DNA vaccination induces humoral and cellular immunity and protects against a lethal intracerebral challenge, J. Virol. 72, 8472-8476. SPELLERBERG, M.B., ZHU, D., THOMPSETT, A., KING, C.A., HAMBLIN, T. J., STEVENSON, F. K. (1997), DNA vaccines against lymphoma: promotion of anti-idiotypic antibody responses induced by single chain Fv genes by fusion to tetanus toxin fragment C, J . Immunol. 159, 1885-1892. SPIEGELBERG, H. L., OROZCO, E. M., ROMAN, M., RAZ, E. (1997), DNA immunization: a novel approach to allergen-specific immunotherapy, Allergy 52, 964-970.

3 Genetic Vaccination with Plasrnid Vectors

SPOONER,R.A., DEONARAIN, M.P., EPENETOS, TODA,S., ISHII, N., OIWDA, E., I
I

71

72

I

Jorg Reirnann et a\.

ULMER, J. B., DONNELLY, J. J., LIU, M.A. (1996), WARD,G., RIEDER,E., MASON,P.W. (1997), Toward the development of DNA vaccines, Plasmid DNA encoding replicating foot-andCuw.Opin. Biotechnol. 7, 653-658. mouth disease virus genomes induces antiviral immune responses in swine, ]. Virol. 71, ULMER,J. B., DONNELLY, J. J., PARKER, S.E., 7442 -7447. RHODES, G. H., FELGNER,P. L. et al. 1993), Heterologous protection against influenza by WATANABE, A,, RAZ,E., KOHSAKA,H., TIGHE, H., BAIRD,S. M. et al. (1993), Induction of injection of DNA encoding a viral protein, antibodies to a kappa V region by gene Science 259, 1745-1749. ULMER,J.B., SADOFF, J. C., LIU, M.A. (1996), immunization, J. Immunol. 151, 2871-2876. DNA vaccines, Cuw.Opin. Immunol. 8, 531- WEI,WZ., SHI, W. P., GALY,A,, LICHLYTER, D., HERNANDEZ, S. et al. (1999), Protection 536. VAHLSING, H. L., YANKAUCKAS,M.A., SAWDEY, against mammary tumor growth by vaccinaS. H., MANTHORPE, M. M., GROMKOWSRI, tion with full-length, modified human ErbB2 DNA, Int. J. Cancer 81, 748-754. (1994), Immunization with plasmid DNA using a pneumatic gun, J . Immunol. Methods WEINER,D. B. (19953, New vaccine strategies, Mol. Med. Today 1, 108-109. 175, 11-22. VALMORI, D., GILEADI, U., SERVIS, C., DUNBAR,WILD,J., GXUSBY, M. J., SCHIRMBECK, R., REIMANN, J. (1999), Priming MHC-II?. R., CEROTTINI, J.C. et al. (1999), Modularestricted, cytotoxic T lymphocyte responses tion of proteasomal activity required for the to exogenous hepatitis B surface antigen is generation of a cytotoxic T lymphocyteCD4' T cell-dependent, /. Immunol. 163, defined peptide derived from the tumor antigen MACE-3, J. E q . Med. 189, 895-906. 1880-1887. K., KUHROBER, VOGEL,F. R., SARVER, N. (1995), Nucleic acid WILD,J., GRUNER,B., METZGER, vaccines, Clin. Microbiol. Rev. 8, 406-410. A,, PUDOLLEK, H.-P. et al. (1998), Polyvalent WAINE,G. J., MCMANUS,D. P. (1993), DNA vaccination against hepatitis B surface and core antigen using dicistronic expression vaccination, Parasitol. Today 11, 113-116. WAISMAN, A,, RUIZ,P. J., HIRSCHBERG, D. L., plasmids, Vaccine 16, 353-360. GELMAN, A,, OKSENBERG, J. R. et al. (1996), WILKINSON, G. W., BORYSIEWICZ, L. I<. (1995), Suppressive vaccination with DNA encoding Gene therapy and viral vaccination: the ina variable region gene of the T-cell receptor terface, Br. Med. Bull. 51, 205-216. R. S., JOHNSTON, S.A., RIEDY,M., prevents autoimmune encephalomyelitis and WILLIAMS, DEVIT,M. J., MCELLIGOTI,S. G., SANFORD, activates Th2 immunity, Nature Med. 2, J. C. (1991), Introduction of foreign genes 899-905. WANDS,J. R., GEISSLER, M., PUTLITZ,J. Z., into tissues of living mice by DNA-coated microprojectiles, Proc. Natl. Acad. Sci. U S A BLUM,H., WEIZSACKER, F.V. et al. (1997), Nucleic acid-based antiviral and gene therapy 88, 2726-2730. of chronic hepatitis B infection, WOLFF,J.A., LUDTKE, J. J., ACSADI,G., /. Gastroenterol. Hepatol. 12, S354-S369. WILLIAMS, P., JANI, A. (19921, Long-term WANG,B., BOYER,J.D.. SRIICANTAN, V, CONEY, persistence of plasmid DNA and foreign gene L., CARRANO, R. et al. 1993), DNA inoculation expression in mouse muscle, Hum. Mol. Genet. 1, 363-369. induces neutralizing immune responses against human immunodeficiency virus type XIANG,2.Q., ERTL, H. C. J. (1995), Manipula1in mice and nonhuman primates, D N A Cell tion of the immune response to a plasmidBid. 12, 799-805. encoded viral antigen by coinoculation with plasmids expressing cytokines, Immunity 2, WANG,B., GODILLOT, A. P., MADAIO,M.P., WEINER,D. B., WILLIAMS,W. V. (1998), 129-135. S., TUN, M., Vaccination against pathogenic cells by DNA XIANG,2.Q., SPITALNIR, WUNNERW.H., CHENG,J., EXTL,H.C. J. inoculation, Cum Top. Microbiol. Immunol. (1994),Vaccination with a plasmid vector 226, 21-35. carrying the rabies virus glycoprotein gene WANG,B., MERVA,M., DANG,I<., UGEN,I<. E., induces protective immunity against rabies W.V., WEINER,D. B. (1995), WILLIAMS, virus, Virology 199, 132-140. Immunization by direct DNA inoculation Xu, D., LIEW, F.Y. (1994), Genetic vaccination induces rejection of tumor cell challenge, against leishmaniasis, Vaccine 12, 1534-1536. Hum. Gene Theu. 6, 407-418.

3 Genetic Vaccination with Plasmid Vectors

Xu, D., LIEW, F.Y. (1995), Protection against leishmaniasis by injection of DNA encoding a major surface glycoprotein, gp63, of L. major, Immunology 84, 173-176. X U , L., SANCHEZ, A., YANG, Z., ZAKl, S. R., NABEL, E. G. et al. (1998), Immunization for Ebola virus infection, Nature Med. 4, 37-42. YANG, N. S. (1992), Gene transfer into mammalian somatic cells i n viuo, Crit. Rev. Biotechnol. 12, 335-356. YANG, N. S., SUN,W. H. (1995), Gene gun and other non-viral approaches for cancer gene therapy, Nature Med. 1, 481-483. YING, H., ZAIG, T. Z., WANG,R. F., IRVINE, I<. R., KAMMULA, U. S. et al. (1999), Cancer therapy using a self-replicating RNA vaccine, Nature Med. 5, 823-827. YOKOYAMA,M., ZHANG,J., WHIITON,J. L. (1995), DNA immunization confers protection against lethal lymphocytic choriomeningitis virus infection, I . Viral. 69, 2684-2688. YOICOYAMA, M., ZHANG,J., W n I n o N , J. L. (1996),DNA immunization: effects of vehicle and route of administration on the induction of protective antiviral immunity, FEMS rmmunol. Med. Microbiol. 14, 221-230. YOUSSEF,S., WILDBAUM, G., MAOR,G., LANIR, N., GOUR,L.A. et al. (1998), Long-lasting protective immunity to experimental auto.

immune encephalomyelitis following vaccination with naked DNA encoding C-C chemokines., J. Immunol. 161, 3870.3879, ZAROZINSIU, C. C., FYNAN,E. F., SELIN,L. I<., ROBINSON, H. L., WELSH,R. M. (1995), Protective CTL-dependentimmunity and enhanced immunopathology in mice immunized by particle bombardment with DNA encoding an internal virion protein, J. Immunol. 154, 4010-4017. ZHONG,W;, WIESMULLER, IZ.H., KRAMER, M. D., WALLICH, R., SIMON,M. M. (1996), Plasmid DNA and protein vaccination of mice to the outer surface protein A of Borrelia burgdorfeti leads to induction of T helper cells with specificity for a major epitope and augmentation of protective IgG antibodies i n vivo, Eur. J . Immunol. 26, 2749-2757. ZHoU, X., BERGLUND,P., ZHAO,H., LILJESTROM,P., JONDAL, M. (1995), Generation of cytotoxic and humoral immune responses by nonreplicative recombinant Semliki Forest virus, Proc. Natl. Acad. Sci. U S A 92, 3009-3013. ZHU, D., SPELLERBERG, M. B., THOMPSEIT, A., KING,C.A., HAMBLIN, T. J., STEVENSON, F. I<. (1997), DNA vaccination as cancer immunotherapy, Biochem. SOC.Trans. 25, 743-747.

I

73

P/asmids for Therapy and kchation by M.Schleef Copyright 0 WILEY-VCH Verlag GmbH, 2001

I 4 A Liposomal iNOS-Gene Therapy Approach to Prevent Neointimal Lesion Formation in Porcine Femoral Arteries Manfred Rudiger ”, Andreas Muhs, Jens Schletter, Angela Heischmann, Andreas Herrmann, Julia Dorge, Jurgen Schrader and Heiko E. von der Leyen

1

Introduction

Restenosis, the recurrent re-occlusion of blood vessels after balloon dilatation (angioplasty) is one of the most frequent complications in today’s interventional cardiology. Although stenting technology has significantly reduced the risk of restenosis, an average of 20 % of all angioplasty patients will suffer from this complication. In certain subpopulations, like diabetics, this rate may even be considerably higher. Restenosis is most frequently caused by a proliferative (hyperplastic)reaction of the smooth muscle cell layer of the vascular wall. Angioplasty causes an overdilation and thus an injury stimulus to the vessel wall which triggers the onset of restenosis. In a healthy vessel, the innermost cell monolayer of the vessel (endothelium) partly would compensate for this overdilation by producing nitric oxide (NO). In a diseased vessel, however, the endothelium is either destroyed or non-functional. NO (also described as EDRF - endothelium-derived relaxing factor - before its identity was elucidated) generally functions as a signaling molecule in endothelial and nerve cells as well as a killer molecule for activated immune cells. NO has an important regulatory function in maintaining vascular homeostasis (Vane et al., 1990) and injury to the endothelium plays an essential role in the pathogenesis of vascular disease (Ross, 1993).The general role of NO has recently been reviewed by Coolte and Dzau (1997).Among the most prominent roles of NO are the following: 1. NO mediates vasorelaxation/vasodilation. 2. NO inhibits smooth muscle cell migration and proliferation. 3 . NO attenuates platelet activation and adhesion. 4. NO reduces vascular inflammation.

According to these many facets, it has been hypothesized that the impairment of NO production and/or bioactivity may be the underlying mechanism for the var-

75

76

I ious risk factors such as hypercholesterolemia, hypertension, diabetes, and cigaManfied Rudiger et al.

rette srnolting in the generation of vascular disease (Harrison, 1997). Experimental studies have shown that vascular injury induces the local expression of mitogens and chemotactic factors that stimulate vascular smooth muscle and leukocyte migration and proliferation (Ross, 1993). Thus, a deficiency in NO production or bioactivity and/or excess of growth promoting factors favors the development of vascular lesions. Apparently, these deficiencies might be therapeutically addressed, by transferring a gene that results in enhanced production of nitric oxide, and vascular diseases such as restenosis, atherosclerosis, bypass graft failure, and transplant vasculopathy characterized by absence or diminished activity of NO production constitute unique opportunities for nitric oxide synthase (NOS) gene transfer as the basis for therapy. Three isoforms of NOS have been described: constitutive-type isoforms (neuronal NOS (NOS I) and endothelial-cellNOS (ecNOS, NOS 111) and an inducible type of enzyme (iNOS, NOS 11). The constitutive isoforms are calcium-dependent and regulated, e. g., by shear stress; the inducible isoform can be rapidly induced by cytoltines to produce high amounts of NO. The three enzyme isoforms share a similar overall catalytic mechanism, in which the homodimeric enzyme catalyzes the oxidation of L-arginine to NO plus L-citrulline. Overexpression after virosome gene transfer of the constitutive isoform of nitric oxide synthase (ecNOS) into balloon-injured rat carotid arteries not only restored NO production within the vessel wall, but also significantly improved vascular reactivity (von der Leyen et al., 1995). Furthermore, ecNOS transgene expression resulted in a 70% inhibition of neointimal lesion formation (von der Leyen et al., 1995). Thus, in uivo ecNOS-gene transfer in this rat model constitutes a “proofof-concept”that a gene therapy to treat vascular proliferative disorders, like restenosis, is a therapeutically feasible approach. Given that the activation of mitogenic factors mediating the cellular processes essential for lesion formation occurs within the first few days after injury, the “short term” expression of ecNOS transgene during the early period after injury may be critical and sufficient to prevent the subsequent development of neointimal hyperplasia (Mann et al., 1995; Morishita et al., 1994; Simons et al., 1992). Since the initial reports, several groups have confirmed the effects of NOS gene transfer on various experimental models of vascular disease. However, all of these attempts used adenoviral vectors, to achieve gene transfer into the vessel wall. Animal models used by these groups comprise transfer of ecNOS in a rat balloon injury-model (Fang et al., 1999; Janssens et al., 1998),in rabbit carotid arteries (ICullo et al., 1997),in large canine cerebral arteries (Chen et al., 1997), and in coronary arteries using an intewentional delivery device (Varenne et al., 1998; Varenne et al., 2000). Principle of a non-viral iNOS-based gene therapy

Here we describe a non-viral gene therapy in a femoral artery injury model in minipigs. A plasmid containing the gene for human inducible nitric oxide synthase (iNOS) defines the therapeutic DNA, which is complexed with monocationic lipo-

4 A Liposornal ;NOS-Gene Therapy

Fig. 1. Schematic representation of the mod-

tic DNA (iNOS), a delivery vehicle (cationic !iular design o f a gene therapy product t o pre- posomal formulation), and a local drug delivery vent restenosis. Modules comprise a therapeu- catheter (Infiltrator@).

sornes forming multilamellar vesicles (Lipoplexes).These lipoplexes are depoted intramurally via the INFILTRATORa catheter (Interventional Technologies, San Diego, CA). Localized gene transfer with this local drug delivery device is achieved by physically microinjecting the lipoplex solution into the vessel wall. This concomitantly ensures a high local concentration of the therapeutic DNA and guarantees a minimal escape into the perivascular space or into the systemic circulation. This modular therapy design is schematically depicted in Figure 1. A gene therapy to prevent restenosis seems to be feasible with currently available technology and is not hampered by the difficulties typically encountered with many other gene therapies. Specifically, three advantages render a gene therapy against restenosis a uniquely promising approach: Gene expression is only needed transiently (Mann et al., 1995; Morishita et al., 1994; Simons et al., 1992),because an early block of smooth muscle cell migration and proliferation will inhibit restenosis. Gene transfer is only needed locally at the site of injury (local “gain of function”), which is accessible by an interventional procedure and can be accomplished in the course of a routine percutaneous transluminal angioplasty. Only part of the cells at the lesion site need to be transfected, because NO is a diffusible gas and thus also acts on neighboring cells that have not taken up the nitric oxide synthase gene (bystander effect).

Results and discussion 2.1 Therapeutic plasmid

An iNOS expression plasmid vector (pAH9)was constructed using standard molecular biology techniques. pAH9 (Figure 2) carries the cDNA encoding the human iNOS (HsiNOS)under the control of the CMV-immediateearly promoter/enhancer (CMV p/e). The iNOS coding sequence was derived by reverse transcriptase-poly-

I

77

78

I merase chain reaction (RT-PCR)from cytokine-induced human primary hepatoManfred Rudiger et a/.

cytes. The bovine growth hormone-polyadenylation signal, BGH p/(a), and the partially deleted polyadenylation signal of SV 40 are located at the 3’-end of the iNOS coding region. The SV 40 origin of replication (Ori) and promoter and the f l Ori were removed from the plasmid backbone. Furthermore, the vector contains the E. coli origin of replication (Col E l ) as well as a cassette encoding the kanamycin resistance gene (IZanR) instead of an AmpR, to ensure regulatory compliance. 2.2 The gene therapy product has a clinically acceptable format

Our gene therapy product contains pAH9 complexed with a cationic liposome (DC30 or DAC-30). DC-30 is a proprietary mixture of DC-Chol (3P-N-(N’,N’-Dimethylaminoethane)-carbamoyl-Cholesterol)(Farhood et al., 1994) and the neutral colipid DOPE (dioleoyl phosphatidylethanolamine). DAC-30 has a very similar composition, where DC-Chol is replaced by the isomeric DAC-Chol (3/3-N-(N’,N-Dimethylaminoethane)-carbamoyl-Cholesterol)(G. 0.T., Berlin). We have developed a formulation that allows to lyophilize the product and reconstitute the formulation by the addition of sterile water before application. The lyophilized formulation is stable for several months at 4’ and the reconstituted product is stable and thus ready-for-use for more than 4 h at room temperature (data not shown). This formulation thus provides a gene therapy product in a clinically acceptable format.

pAH 9 6915 bps

Fig. 2. Therapeutic plasmid pAH9 was cloned into a commercially available vector. The AmpR was replaced by KanR. The plasmid is routinely purified from E. coli according t o standard operating procedures (Qiagen). The iNOS-

cDNA is driven by the constitutive CMVpromoter. Plasmid identity was verified by sequencing. The control plasmid used in many experiments i s identical to pAH9 but it lacks the iNOS-gene cassette.

4 A Liposomal ;NOS-Gene Therapy

2.3

Efficient gene transfer was established in a minipig femoral artery injury model

We tested our iNOS gene therapy product in the femoral artery model of Gottingen Mini-pigs (20-25 kg; Ellegard, Denmark). The protocols for the surgical procedures in experimental animals were approved by the institutional Administrative Panel on Laboratory Animal Care and performed in accordance with state and federal animal protection laws. Access to the arterial vascular system was achieved by carotid arteriotomy after standard anesthesia. After diagnostic angiography of peripheral arteries, local gene transfection was performed with the Infiltrator device. The Infiltrator catheter i s a triple lumen catheter (for details on the device refer to http:// www.iut.com/iutmed/dvugdeli.html). It comprises a low pressure positioning balloon with three rows of injection ports mounted on its surface. The ports are connected to a fluid channel independent of the inflationldeflation system. The injection ports are recessed during maneuvering in the artery. Upon inflation the ports radially extend and penetrate the lesion and/or vessel wall. The port rows ensure injection of drug over a 15 inm long vessel segment into the medial vessel layer. We typically delivered 400 pL of lipoplex solution directly into the tunica media of the arterial wall. Injection of the drug can be accomplished in less than 10 s, without substantial intimal damage, with efficient deposition along the length of the delivery ports (15 mm), and near zero luminal washout (see also Morishige et al., 2000; Teiger et al., 1999; Varenne et a]., 1998). Local gene transfer was analyzed by using iNOS-specificimmunohistochemistry. Animals were sacrificed at day 3 post transfection, and vessels were removed and stained for iNOS protein. The medial vessel wall was reproducible stained positive for iNOS-protein in transfected vessel segments. A typical histochemical analysis i s shown in Figure 3A. Untransfected vessel areas showed background activity (Figure 3B). Occasionally, transfected vessel segments appeared vasodilated (Figure 3A), whereas untransfected vessel segments were usually vasoconstricted upon preparation (Figure 3B). It was not systematically investigated, however, whether this effect can be attributed to NO-production in iNOS-transfectedvessels. Efficient depoting of the therapeutic plasmid pAH9 and expression of the iNOSgene was also corroborated by PCR analysis and RT-PCR, respectively. To demonstrate depoting of the plasmid, six cryostat sections from iNOS-transfectedand control-transfectedvessel segments were processed for DNA extraction and analyzed by (nested) PCR. As shown in Figure 4A, pAH9 was reproducibly detected in iNOStransfected vessels, while only the control fragment was amplified in other vessels. Expression of the iNOS-gene after transfection into the vessel wall was analyzed by nested RT-PCR using RNA isolated from a small vessel segment. Again, a signal specific for iNOS-mRNA could only be detected in the pAH9-transfected vessels. Thus, both, immunohistochemistry as well as (RT-)PCR analysis demonstrate that the therapeutic DNA i s efficiently deposited in the medial vessel layer and that the vascular smooth muscle cells are efficiently transfected by our non-viral gene transfer product.

I

79

80

I

Manfred Rudiger et a/

A

-

100 00

"llCrOnletPr

Fig. 3. lmmunohistochemical analysis o f iNOS expression in porcine femoral arteries after local gene delivery. (A) Representative picture o f a femoral artery three days after transfection with 2 pg o f lipoplexed iNOS-plasmid. ( 6 ) Representative picture o f a control femoral artery three days after transfection with 2 pg o f lipoplexed control plasmid lacking the iNOS coding sequence. For immunohistochemical localization o f iNOS, arteries were removed, overlaid with OTC and quick frozen in liquid nitrogen. Ten-micrometer cryostat serial sections (0,5mm) were mounted on slides, dried over night, blocked with normal

goat serum for 45 min, incubated over night with anti-iNOS primary monoclonal mouse IgC antibody (Transduction, 1:loo) followed by a 1-hour incubation with biotinylated secondary F(ab),-fragment goat anti-mouse IgC (Dianova, 1:1,000). Sections were then incubated for 30 m i n with horseradish peroxidase-streptavidine complex (Vectastain" Elite ABC-Kit, Vectorm Laboratories) followed by development with DAB peroxidase substrate. Note the homogeneous and circumferential transfection (brown staining of iNOS) o f the vessel in (A).

4

PCR

A

Liposomal iNOS-Gene Therapy

RT-PCR

vector

Deposition o f iNOS lipoplex and expression o f the iNOS-cDNA was analyzed by PCR (A) and RT-PCR (panel B), respectively. (A) Vessels were processed for immunohistochemical staining at day 3 post transfection with 2 p g o f therapeutic or control DNA using the Infiltrator catheter. DNA was extracted according t o standard procedures (Dneasy Tissue Kit, Qiagen) from six adjacent 10 p m cryosections and then further processed for a nested PCR analysis. In the first PCR reaction, primers hybridized t o vector sequences in the BCHpolyA and in the CMV-promoter, respectively, and DNA was amplified for 30 cycles (no visible signal). In the second nested PCR one primer was localized again in the vector while the other was localized in the iNOS-codFig. 4.

ing sequence. A fragment o f the predicted size o f 1,400 bp was amplified from iNOS-transfected specimens, whereas a 1,000 bp predicted fragment was amplified from the control-cryosections. (B) RT-PCR (nested) was performed on RNA extracted from 0.5 cm o f iNOS- or controltransfected vessel segments. In both, the first (40 cycles) and second (18 cycles) amplification reaction one primer hybridized t o the BCHsequence, while the second hybridized t o an iNOS-specific sequence. RNA isolated from iNOS-transfected vessels yielded a fragment of the predicted size o f 400 bp, while no fragment was amplified from control-transfected vessels, as expected.

2.4

Transfection efficiency is dose dependent

To analyze the effect of increasing amounts of plasmid DNA on the transfection efficiency in vivo, groups of 4-8 minipigs each received 0.7-4.0 pg of iNOS lipoplexes or a control vector with no iNOS-gene. All vessels were also stented with the Wiktor stent as described below. Transfection efficiency was analyzed at day three post transfection and quantitated by planimetric computerized evaluation of areas positively stained for iNOS on cryostat serial sections. The results from these studies are summarized in Figure 5. Maximum transfection efficiency was achieved with 2 pg of iNOS-lipoplex resulting in 44% iNOS-positive vessel area. Higher amounts of iNOS-lipoplex do not result in increased transfection efficiencies but may even lead to decreased apparent transfection efficiency, which is attributed to cytotoxic and pro-apoptotic effects of high lipoplex and high NO concentrations. Positive staining in the control is explained by crossreactivity of the antibody with endogenous pig iNOS. These data show, that a dose-response anal-

I

81

82

I

Manfied Riidiger et al.

pg of iNOS-plasmid Fig. 5. Minipig femoral arteries were transfected with 2 pg o f control plasmid or with amounts of therapeutic DNA, as indicated. After infiltrator-mediated transfection, a Wiktor stent was placed. 4 - 8 minipigs were analyzed per dose. Animals were sacrificed at day 3 post transfection and the vessels were processed

for iNOS-specific immunohistochemistry as in Figure 3. iNOS-positive areas i n the tunica media were quantitated using computer assisted morphometry. Note that transfection efficiency increases up t o 2 pg. Larger amounts of plasmid do not result in increased transfection efficiency.

ysis for gene therapy approaches might be very useful, in terms of both, transfection efficiency as well as therapeutic efficacy. In the case of an iNOS-based approach, very high expression levels as might be achievable using viral delivery system, are not necessarily advantageous, because exceedingly high local NO concentrations might kill the transfected (and other) cells. Very low transfection efficiency, however, might not result in therapeutic levels of the active compound (NO). f.2

Non-viral iNOS gene transfer efficiently inhibits neointimal lesion formation

To analyze the effect of the iNOS-gene therapy on in-stent lesion formation, a Wiktorm stent was deployed directly after the local gene transfection to induce neointimal lesion formation. This in-stent restenosis model was established to most closely mimic the ultimate clinical scenario (i. e., PTCA with subsequent stent implantation). Four weeks after injury, animals were sacrificed and vessels processed for histological and morphometric analysis. Typical data of such an experiment are shown in Figure 6. When increasing doses were analyzed in terms of efficiency to inhibit neointimal lesion formation, minipigs treated with 1 pg of therapeutic DNA displayed the most significant inhibition (p < 0.05) of lesion formation (Table 1).

4 A Liposomai ;NOS-Gene Therapy

Fig. 6. Inhibition of neointimal lesion forma- fusion-fixed vessels were sectioned and stained tion. Vascular injury was induced by implanta- with modified elastic van Cieson stains. The tion of a stent (Wiktor) after transfer o f 2 pg holes at the neointimalmedia boundary are of a control plasmid lacking the iNOS-cDNA caused by stent wires. Note that neointima is (panel A) or o f the therapeutic iNOS-plasmid significantly inhibited i n vessels treated with the (pAH9) using the Infiltrator catheter. Animals therapeutic plasmid. were sacrificed 4 weeks post transfection. Per-

I

83

84

I

Manfred Rudiger et al. Table 1.

Peripheral artery study: Morphometric analysis ( n = 7-12)

Dose CgI

Number of

Minipigs

Neointima/Media Ratio (Mean F SEMI

Control

41

2.0 t 0.14

iNOS 0.25

12

1.6

iNOS 0.5

12

1.5 i 0.16

0.08

iNOS 1.0

7

1.3 i 0.27

0.04

iNOS 2.0

10

1.8 k 0.28

0.5

* 0.15

p vs. Control

I

I 0.1

3

Summary and perspectives

Besides its vasodilatory property, NO is a potent inhibitor of platelet activation (Radomski et al., 1990), thrombosis (Kaul et al., 2000), vascular smooth muscle cell proliferation (Garg and Hassid, 1989; Mooradian et al., 1995; von der Leyen et al., 1995) and migration (Dubey et al., 1995; Sarltar et al., 1995).These modes of action of NO constitute the rationale to use NO as a therapeutic agent in ameliorating the pathological development of restenosis following vascular injury. However, approaches based on the systemic delivery of NO suffer from limitations. NO is rapidly inactivated by hemoglobin in the circulating blood which results in limited bioavailability. Unwanted systemic hemodynamic effects (e.g., vasodilation and drop in systolic blood pressure) do not allow the systemic administration of locally effective doses of NO donors (Kaul et al., 2000). The prospect of local vascular gene transfer therefore provides a novel therapeutic approach which allows to locally supply NO at therapeutic doses without the risk of unwanted systemic side effects. The beneficial effect of NO on inhibiting neointimal lesion formation has been shown in several in vivo models of vascular injury (Guo et al., 1994; Lee et al., 1996; von der Leyen et al., 1995).A key advantage of using NOS as a the therapeutic gene is that the therapeutically active substance, the gaseous NO easily diffuses to neighboring cells. Thus, not only cells that have taken up the gene are affected, but also cells in the vicinity. The major advantage of the approach described in this report is seen in the substantial efficiency of the cationic liposomal gene transfer used here. A therapeutically efficient non-viral approach has obvious advantages. For instance, a non-viral approach has a superior safety profile. Many lipids have been administered to patients in clinical trials and demonstrated to be safe. Second, GMP-production of liposomal DNA formulations is easier than the GMP production of viral vectors. Basically all compounds involved are either synthetic or can be purified easily under GMP according to standard procedures (see also chapter 11).Third, a liposoma1 formulation like the one used in this study, can be lyophilized and is easy to use clinically.

4 A Liposomal [NOS-Gene Therapy

Clinical feasibility of this non-viral approach is underlined by the efficient local drug delivery device that was used. Furthermore, the INFILTRATOR has already been used in the clinics by others in the coronary system (Pavlides et al., 1997). In essence, the data described here suggest that this novel gene-therapy approach employing iNOS enables therapeutically sufficient transfection rates using a liposoma1 (non-viral) formulation in conjunction with a commercially available local drug delivery device. The safety profile of this approach and the efficacy in a peripheral artery injury model (descibed here) and in a coronary injury model (to be published elsewhere) thus emphasize the potential clinical utility of this therapeutic approach.

References CHEN,A. F., O’BRIEN, T., TSUTSUI, M., KINOSHITA, H., POMPILI, V. J. et al. (1997), Expression and function of reconibinant endothelial nitric oxide synthase gene in canine basilar artery, Circ. Res. 80, 327-235. COOICE, J. P., DZAU,V. J. (1997), Nitric oxide synthase: role in the genesis of vascular disease. Artnu. Rev. Med. 48, 489-509. DUBEY, R. K., JACKSON, E. K., LUSCHER,T. F. (1995), Nitric oxide inhibits angiotensin 11-induced migration of rat aortic smooth muscle cell. Role of cyclic-nucleotidesand angiotensinl receptors, J. Clin. Invest. 96, 141-149. FANG, S., SHARMA, R.V., BHALLA,R. C. (1999), Enhanced recovery of injury-caused downregulation of paxillin protein by eNOS gene expression in rat carotid artery. Mechanism of NO inhibition of intimal hyperplasia? Arterioscler. Tnromb. Vasc. Bid. 19, 147.152. FARHOOD,H., GAO,X., SON,K., YANG, YY, LAZO, J. S. et al. (1994), Cationic liposomes for direct gene transfer in therapy of cancer and other diseases, Ann. NY Acad. Sci. 716, 23-34; discussion 34-35. A. (1989), Nitric oxideGARC,U. C., HASSID, generating vasodilators and 8-bromo-cyclic guanosine monophosphate inhibit mitogenesis and proliferation of cultured rat vascular smooth muscle cells, /. Clin. Invest. 83, 1774-1777. Guo, J. P., MILHOAN,ICA., TUAN,R. S., A.M. (l994), Beneficial effect of SPMLEFER, 5185, a cysteine-containing nitric oxide

donor, in rat carotid artery intimal injury, Circ. Res. 75, 77-84. D. G. (1997), Cellular and moleHARRISON, cular mechanisms of endothelial cell dysfunction, J . Clin. Invest. 100, 2153-2157. JANSSENS, S., FLAHERTY, D., NONG,Z., VARENNE, O., VAN PELT,N. et al. (1998), Human endothelial nitric oxide synthase gene transfer inhibits vascular smooth muscle cell proliferation and neointima formation after balloon injury in rats, Circulation 97, 1274- 1281. KAUL, S., CERCEIC, B., RENGSTROM, J., Xu, X. P., MOLLOY, M. D. et al. (2000),Polymeric-based perivascular delivery of a nitric oxide donor inhibits intimal thickening after balloon denudation arterial injury: role of nuclear factor-kappaB,/. Am. Coll. Cardiol. 35, 493-501. KULLO,I. J., MOZES,G., SCHWARTZ, R. S., P., CROTIY, T. B. et al. (1997), GLOVICZKI, Adventitial gene transfer of recombinant endothelial nitric oxide synthase to rabbit carotid arteries alters vascular reactivity, Circulation 96, 2254-2261. LEE, J. S., ADRIE,C., JACOB, H. J., ROBERTS, J.D., J R , ZAPOL,W.M., BLOCH,K.D. (1996), Chronic inhalation of nitric oxide inhibits neointimal formation after balloon-induced arterial injury, Circ. Res. 78, 337-342. G. H., KERNOFF, R. S., MANN,M. J., GIBBONS, DIET,F. P., TSAO,P. S. ET al. (1995), Genetic engineering of vein grafts resistant to atherosclerosis, Proc. Natl. Acad. Sci. U S A 92, 4502-4506.

I

85

86

I

Manfied Rudiger et af. MOORADIAN, D. L., HUTSELL, T. C., KEEFER,L. I<. (1995), Nitric oxide (NO) donor molecules: effect of NO release rate on vascular smooth muscle cell proliferation in vitro, J. Cardiovasc. Pharmacol. 25, 674-678. MORISHIGE, IZ., SHIMOKAWA, H., YAMAWAKI, T, MIYATA,K., ETO, Y et al. (ZOOO), Local adenovirus-mediated transfer of C-type natriuretic peptide suppresses vascular remodeling in porcine coronary arteries in vivo, J. Am. Coll. Cardiol. 35, 1040-1047. MORISHITA,R., GIBBONS, G. H., ELLISON, I<. E., NAKAJIMA, M., VON DER LEYEN,H. et al. (1994), Intimal hyperplasia after vascular injury is inhibited by antisense cdlc 2 lcinase oligonucleotides, /. Clin. Invest. 93, 1458.1464. PAVLIDES, G. S., BARATH,P., MAGINAS,A., VASILIKOS, V., COKKINOS, D. V., O’NEILL, W. W. (1997), Intramural drug delivery by direct injection within the arterial wall: first clinical experience with a novel intracoronary delivery-infiltrator system, Cathet. Cardiovasc. Diagn. 41,287-292. RADOMSICI, M. W., PALMERR. M., MONCADA,S. (199O),An L-arginine/nitric oxide pathway present in human platelets regulates aggregation, Proc. Natl. Acad. Sci. USA 87, 5 193-5197. Ross, R. (1993),The pathogenesis of atherosclerosis: a perspective for the 1990% Nature 362, 801-809. SARIZAR, R., WEBB,R.C., STANLEY, J. C. (1995), Nitric oxide inhibition of endothelial cell mitogenesis and proliferation, Surgery 118, 274-279.

SIMONS,M., EDELMAN, E. R., DEKEYSER, J. L., LANGER,R., ROSENBERG, R. D. (1992), Antisense c-myb oligonucleotides inhibit intimal arterial smooth muscle cell accumulation in vivo, Nature 359, 67-70. TEIGER,E., DEPREZ,I., DUPOUY,P., SITEON,M., ADNOT,S., DUBOIS-RANDE, J. L. (1999), Local gene delivery within the media of rabbit iliac arteries by using the infiltrator intramural delivery device, J. Cardiovasc. Phanna601. 33, 726-732. VANE,J. R., ANGGARD, E. E., BO~TING, R. M. (1990), Regulatory functions of the vascular endothelium, N. Engl. J. Med. 323, 27-36. VARENNE, O., PISLARU,S., GILLIJNS, H., VAN PELT,N., GERARD,R. D. et al. (1998), Local adenovirus-mediated transfer of human endothelial nitric oxide synthase reduces luminal narrowing after coronary angioplasty in pigs, Circulation 98, 919-926. VARENNE, O., SINNAEVE, P., GILLIJNS,H., IUNG,B., LAURYSENS, V. et al. (ZOOO), Percutaneous gene therapy using recombinant adenovimses encoding human herpes simplex vims thymidine lcinase, human PAI-1, and human NOS3 in balloon-injured porcine coronary arteries, Hum. Gene Ther. 11, 1329-1339. VON D E R LEYEN,H.E., GIBBONS, G.H., MORISHITA,R., LEWIS,N. P., ZHANG,L. et al. (1995), Gene therapy inhibiting neointimal vascular lesion: in vivo transfer of endothelial cell nitric oxide synthase gene, Proc. Natl. Acad. Sci. U S A 92, 1137-1141.

P/asmids for Therapy and kchation by M.Schleef Copyright 0 WILEY-VCH Verlag GmbH, 2001

tmmunotherapy of Chronic Hepatitis by pCMV-S2.S DNA Vaccine

B

Marie-Louise Michel

1 Introduction 1.1 Hepatitis 6:the disease

Infection with hepatitis B virus (HBV) is one of the most common infectious diseases with an estimated 350 million chronic HBV carriers worldwide (Lee, 1997). The high risk of patients with chronic hepatitis B is to develop liver cirrhosis which is associated with a high rate of mortality due to the development of hepatocellular carcinoma or non-carcinomatous complications of cirrhosis (portal hypertension and liver failure) (Hoofnagle and di Bisceglie, 1997; Hoofnagle and Lau, 1997). Most of adults suffering from acute HBV infection recover spontaneously and completely. Only a small fraction ( 3 - 5 %) of HBV-infected adults become chronic carriers of hepatitis B surface antigen (HBsAg) which is a marker of chronic infection. By contrast, up to 90% of children of HBV-infected mothers or children infected in the first years of life, become chronically infected (Peters et al., 1991). One third of chronically HBV-infected subjects are asymptomatic carriers while two thirds develop chronic liver disease. Although the pathogenesis of chronic liver disease is not well understood, there is a consensus that liver damage is immune mediated. 1.2 Hepatitis 6:treatments

Currently, the only therapy for chronic hepatitis which gives a durable beneficial effect is systemic treatment with interferon-alpha, but a sustained response is achieved in only one third of patients with chronic hepatitis B (Lee, 1997). Nudeoside analogs, such as lamivudine, provide a rapid decrease in serum HBV DNA levels and a histopathological improvement of the liver disease when used as an al-

88

I

Marie-Louise Michel

ternative therapy. However, short-term treatment leads to a rapid relapse of the disease and long-term treatment often results in the selection of resistant viral variants (Lai et al., 1998). These outcomes emphasize the need for novel therapeutic approaches. 1.3 Hepatitis B: immune response to infection

HBV i s a non-cytophatic virus and liver injury is mainly mediated by the host immune response against virus-infected liver cells and by the production of inflammatory cytolines. The cellular immune response also contributes to viral clearance and these two opposite functions have been attributed to the same cells. A vigorous, polyclonal and multispecific cytotoxic (CTL) and helper T (Th) cell response to HBV is readily detectable in the peripheral blood of patients with acute self limited hepatitis B, but is weak, antigenically restricted or undetectable in patients with chronic infection (Chisari, 1995).This Tcell response is believed to be responsible for both the elimination and the control of the hepatitis B virus (Rehermann, 2000). CD8' T cells have the capacity to kill infected hepatocytes via perforine, fas ligand or TNF-a (tumor necrosis factor a) death pathways or to cure them via non cytopathic, cytoltine-mediated inhibition of HBV replication (Guidotti and Chisari, 1999).Moreover, in contrast to what was observed during acute hepatitis B, production of Thl cytokines (interferon-y)following antigen stimulation of T cells in the periphery was weak or absent in chronic hepatitis B (Bertolettiet al., 1997; Jung et al., 1999; Penna et al., 1997).Therefore, it is tempting to speculate that skewing the T cell response into a predominant Thl pathway represents an effective strategy to facilitate eradication of chronic HBV infection. Specific immunotherapeutic strategies have been proposed as possible alternatives to interferon or antiviral drugs to enhance or to broaden the defective T cell responses in chronically infected patients. Specific vaccine therapies, using either currently available recombinant anti-hepatitis B vaccines (Couillin et al., 1999), a lipopeptide-based T cell vaccine designed to induce a nucleocapside-specific CTL response (Vitiello et al., 1995) or newly developed genetic vaccines (Mancini et al., 199Gb; Rollier et al., 1999) have been recently studied in animal models or during clinical trials. 1.4 What are DNA vaccines?

In contrast to the traditional antigen-based vaccines, DNA vaccines involve introduction of plasmid DNA encoding the antigen. A plasmid is a double-stranded, closed circular form of DNA that can easily be produced in bacteria (see also Chapter 11). In addition to a cassette allowing antigen-encoding sequences to be expressed in eultaryotic cells (e. g., promoter, enhancer sequences for gene expression and polyadenylation signal for mRNA), DNA vaccines also contain genetic elements that permit growth of the plasmid in bacteria. Upon intradermal or intra-

5 lmmunotherupy of Chronic Hepatitis B by pCMV-82.5 DNA Vaccine

muscular injection, DNA enters different cell types, uses the cell’s machinery and directs the synthesis of the encoded antigen. By producing the antigen itself, the body becomes its own vaccine factory (Donnelly et al., 1997). 1.5

Which DNA vaccines for hepatitis B?

HBV is the prototype virus for a family of small enveloped DNA viruses called Hepadnaviridae. Besides HBV, this family comprises woodchuck hepatitis virus (WHV), ground squirrel hepatitis virus (GSHV), duck hepatitis B virus (DHBV), and less well characterized viruses that infect woolly monkeys, herons and other hosts (Ganem, 1996; Wei and Tiollais, 1999). The hepadnaviruses of animals are currently not of particular interest to veterinary medicine, but they serve as important models for human HBV. These viruses share common characteristics such as a narrow host range, predominant hepatotropism and similar virion structure. The complete infectious HBV virions or Dane particles, are 42 nm double-shelled particles containing a lipoprotein envelope surrounding an inner core particle or nucleocapsid. Within the nucleocapsid, a circular DNA molecule is attached to the viral polymerase. In addition to Dane particles, the serum of infected patients also contains spherical and filamentous non-infectious particles composed exclusively of lipoprotein envelopes (Ganem, 1991; Wei and Tiollais, 1999). The immune response specific to envelope proteins is very broad and plays an important role at various stages of the infection. Envelope-specificantibodies mediate elimination of virions at an early stage of infection and prevent the spread of the virus. CD4’ T cells specific for envelope stimulate the humoral response, but also provide help to cytotoxic T lymphocytes which are effectors of the clearance of infected hepatocytes. Although both humoral and cell-mediated immunity result from natural HBV infection, the presence of antibodies alone is sufficient to confer protection from infection. There is a clear role for anti-envelope antibodies (antiHBs) in conferring protective immunity and all vaccines used in human to date have been designed to induce anti-HBs antibodies. We have therefore developed DNA vaccines for HBV infection based on injection of plasmid DNA encoding envelope proteins. These proteins are the translational product of a large open reading frame (ORF) of the HBV genome which is divided into three domains preS1, preS2 and S (from 5’ to 3’) by three in frame ATG start codons. Three polypeptides are thus encoded by this ORF: the small or major (S), the middle (preS2 S) and the large (preS1 + preS2 + S) proteins, each of them existing in both glycosylated and unglycosylated forms (see Figure 1).These proteins have the property to self-assemble with cellular lipids into subviral particles carrying the hepatitis B surface antigen (HBsAg). Several B and T cell epitopes have been identified on both the S and preS domains. The S domain encodes the major protective B cell epitope or HBsAg defined by a group-specificdeterminant called a and subtype determinants named d , y, w and y. Combination of determinants results in three major serotypes adw, adr and uyw (Ganem, 1991).

+

I

89

90

I

Marie-Louise Michel

pCMV42.S S

NHZ

.

“ BPYd Y3A

coon

small protein

COOH

middle protein

HBsAg particles

Fig. 1. Structure of the HBV genome and pCMV-SZ.S plasmid expression vector used for DNA-based immunization. This vector uses the cytomegalovirus early gene promoter (CMV) t o drive the expression of the S gene and preS2 region. It contains the viral sequences located

3 ’ t o the envelope coding sequences, the HBV termination and polyadenylation sequences. The encoded HBV envelope proteins are the small and the middle proteins, that selfassemble into HBsAg sub-viral particles.

2

DNA vaccines for the prevention of hepatitis B

We have constructed several vectors expressing one, two or three forms of the HBV envelope. Three of these vectors pCMV-S, pCMV-S2.S and pCMV-Sl.SZ.S utilize the immediate early gene promoter of cytomegalovirus (CMV) and encode one, two or three forms of envelope protein, respectively (Michel et al., 1995). Other vectors use the endogenous HBV promoter present within the preSl domain (Michel et al., 1995), the SV40 early gene promoter (Mancini et a]., 199Ga) or a muscle-specific promoter, the human desmin gene promoter (Loirat et al., 1999). All of these vectors have successfully been used to drive the expression of HBsAg in vitro or in vivo (Davis et al., 1993) and to induce cellular and humoral immune responses in different animal models. 2.1 The mouse model

2.1.1

Humoral response

Systemic immunization of mice by a single intramuscular injection of plasmid DNA expressing HBV envelope proteins and HBsAg induces rapid, strong and sustained humoral immune responses. Antibodies, which are initially of the IgM then

5 lrnrnunotherapy of Chronic Hepatitis B by pCMV-SZ.S D N A Vaccine

IgG isotype (predominantly IgG2a), recognize several of the B cell epitopes present on the S, preS2 or preSl domains of the envelope proteins (see Figure 2). Antibodies specific to both group (a) and subtype (y) determinants of HBsAg were detected. The induction of group-specificantibodies is of particular importance since they could provide cross protection from heterologous HBV strains in humans. Following injection of preS2-encoding DNA, anti-preS2 antibodies were induced very rapidly, even before anti-HBs. Anti-preS2 antibodies can provide protection by themselves (Itoh et al., 1986; Neurath et al., 1986) and are a marker of viral clearance in patients. However, these anti-preS2 antibodies are absent in individuals who progress to a chronic carrier state (Alberti et al., 1988). The early, strong and persistent humoral immune response as elicited by pCMV22.S may be critical for vaccination of infants born to chronically infected mothers or for individuals at high risk of infection. Anti-preS1 antibodies were much more difficult to induce and a delay in the appearance of such antibodies was observed. This was probably due to the fact that when the large protein is expressed from a strong promoter, it remains trapped into the cell. This problem has been solved by the use of a vector based on dual expression of the large or the middle and the small proteins from two promoters in the same plasmid, respectively (Mancini et al., 1996a). Thus, the nature of the encoded sequences and the promoter used for gene expression can modulate the induced immune responses. The highest titers of anti-HBs are reached by 4-8 weeks and persist at a near maximal level for at least 17 months after a single DNA injection. These titers can be boosted ten-fold by a second injection of DNA, or somewhat less by injection of a recombinant HBsAg protein (Davis et al., 1996a). This is a remarkable result since the transfected muscle fibers which produce the antigen are probably destroyed by the cytotoxic T cell response, and suggest that the antigen is still present, presumably as antigen-antibody complexes associated to follicular dendritic cells in the lymph nodes (Davis et al., 1997). In immunocompetent individuals, incidence of non-response or hyporesponse to the current HBV vaccines is less than 5 % for young adults, but increases to about 30 %with advancing age. In humans, the hyporesponse is defined by anti-HBs titers less than the 10 milli International Units (mUI)per mL required to provide protection. This problem of non-response is, in some cases, HLA-linked (ICruskall et al., 1992). To address this problem, we took advantage of the existence of well defined strains of mice which are known to be poor responders to HBsAg at the B cell level, unless preS1 or preS2 domains of the envelope protein are present in the vaccine (Milich et al., 1985). This non-response in mice is H-2 linked. It was shown that DNA immunization with a plasmid containing the S gene only overcomes haplotype-restrictednon-responsiveness to HBsAg in congenic strain of mice, resulting in an earlier and stronger antibody response even after a single injection of DNA. Addition of preS-encodingdomain in the DNA vaccines resulted in antibody responses comparable to those obtained with protein vaccines, but persistence of anti-HBs antibodies after DNA-based immunization did not required booster injection (Davis et al., 1995). These results may have clinical significance for individuals who are poor responders or non-responders to the current HBV vaccines.

I

91

92

I

Marie-Louise Michel 2.1.2 Cell-mediated response A strong cellular immune response is induced by the DNA-based immunization with high level of CTL and CTL precursors being detected by one week and being maintained for several months. After specific in vitro restimulation for five to seven days with HBsAg-expressing cells or with an epitopic peptide (H-2d S28-39), splenocytes from DNA-immunized BALB/c mice were able to efficiently lyse autologous target cells infected with recombinant vaccinia virus expressing HBV envelope proteins or targets loaded with the H-2d-restricted peptide (see Figure 2). This response is detectable by one week after intramuscular injection of as little as 10 pg of pCMV32.S DNA and increases after injection of higher amounts of DNA. The recruitment of antigen presenting cells at the DNA injection site by pretreatment of the muscle with a necrotizing agent significantly increases the precocity and the intensity of the responses (Loirat et al., 1999).The C57BL/6 mice belong to a strain which normally does not mount CTL responses to HBV envelope proteins even after injection of exogenous recombinant HBsAg or infection with a live recombinant vaccinia virus encoding HBsAg. Nevertheless, a cytotoxic response was induced in these mice following a single immunization with plasmids expressing HBsAg (Schirmbeck et al., 1995). Injection of pCMV-S (Hui et al., 1999) or pCMV32.S (Mancini et al., 1998) into C57BL/6 mice resulted in the induction of functional CTL that could be detected either after in vitro stimulation with autologous cells expressing HBsAg and using standard chrome release assay as a readout or directly by using y-interferon ELISPOT assay. The frequency of CD8' T cells detected in C57BL/6 mice four weeks after a single injection of DNA is around 100 per lo6 spleen cells. Before applying genetic immunization approaches to humans, it was important to verify that injection of a plasmid encoding HBsAg allows the generation of cytotoxic T cells of specificities comparable to those observed in infected patients. To address this question, we studied the specificity of cytotoxic responses induced in HLA-A0201 transgenic mice (Pascolo et al., 1997) by DNA immunization as compared to that observed during acute infection in humans. Immunization of mice that, are both transgenic (Tg) for the human HLAA2.1 molecule and knock-out for murine MHC class I molecules, with pCMV-SZ.S DNA shows that epitopes presented after in uivo processing in HLA-A2.1 Tg mice are very similar to those generated in humans during HBV infection (D. Loirat et al., 2000). This suggests that responses induced by DNA immunization could have the same immune potential as those developing during natural HBV infection. DNA vaccines also act as their own adjuvant owing to the presence of bacterial immunostimulatory CpG motifs. Such sequences present in the plasmid backbone promote a Thl like pattern of cytokine production dominated by interleukin-12 (IL-12) and interferon-y (IFN-y) with little secretion of Th2 cytoltines (Roman et al., 1997; Sato et al., 1996). This was illustrated by the strong proliferative responses observed for splenocytes derived from pCMV-S2.S immunized C57BL/6 mice and stimulated with HBsAg or preS2-specific synthetic peptides (Figure 2) (Mancini et a]., 1998). CD4' T cells primed by pCMVS2.S immunization have a

5 lmmunotherapy of Chronic Hepatitis B by pCMV-52.5 D N A Vaccine 100000

.6 0

L I

IgG anti-SZ.S 10000

IgGantl-S

P

-

-

'5?

l o80o 1

-

-e

-

I

2

&

loo0

IgM anti-S2.S I

100

IgM anti%

10

l5Ooo

10000

1

p8 W S ""S "V

s2.s

d VVWT

T

Weeks

-00

E - T ratio

.

m

3000

non-depleted splenocytes CDCdepleted splenocytes 2500

0

CD8-depleted splenocytes 2000 5000

-

,E m a

0

111

1500

rnedmn

HBsAg

~ p 1 2 4 - 1 ~ 1000

ConA

500

0

11.4

Fig. 2. Immune responses in mice following pCMV-SZ.S immunization. (A) Humoral response: IgC and I g M specific for S and/or preS2 antigenic determinants o f HBsAg. (B) Cytotoxic response in BALB/c mice after pCMV-SZ.S immunization. Splenocytes from immunized mice lyse target cells infected with recombinant vaccinia virus expressing S gene (vv S) or preSZ+S gene (vv 5254, p815 transfected cells expressing S gene (p815/S) but not target cells infected with wild type vaccinia virus (vv WT) or untransfected p815 cells. E - T effector t o target ratio. (C) T helper re-

IFN-g

TNF-a

sponse in C57BL/6 mice after pCMV-SZ.S i m munization. Splenocytes from immunized mice proliferate following stimulation with HBsAg or peptides derived from the preS2 region. Splenocytes are either undepleted, CD4.l T cell depleted or CD8' T cell depleted. Results are expressed in count per mn (cpm) of incorporated thymidine. (D) Cytokine production of activated splenocytes as described i n C. IL-4 (interleultine 4), IFN-g (interferon gamma), TNF-a (tumor necrosis factor-alpha). Results are expressed in picog/ml. ConA: concanavaline A stimulated splenocytes. Medium: culture medium.

Thl phenotype since they secrete IFN-p and tumor necrosis factor-a (TNF-a) following specific antigenic stimulation (Figure 2).

2.1.3

I

Mechanisms o f DNA-induced immune response to HBsAg

Intramuscular injection of HBsAg-expressing plasmid results in transfection of approximately 1-2 % of muscle fibers which can be detected by immunofluorescent labeling (Davis et al., 1997). Detection of HBsAg using antibodies recognizing a conformational epitope suggests that the expressed antigen adopts a conformation similar to that resulting from natural virus infection. Low levels of circulating antigen are detected in the sera of mice by one to two weeks after injection indicating that the antigen is secreted or otherwise released from the transfected muscle fibers (Davis et al., 1993). Following recognition by B cells, this antigen allows production of specific anti-HBs antibodies (see Figure 3 ) . However, resident antigen presenting cells (APC),or APC attracted into muscle tissue following pretreatment

93

94

I

Marie-Louise Michel

Injection of muscle fibers or APCwith PCMV-SS.S iberated protein ibet Iysis

&cumulation and secretion of nBsAg

Fig. 3.

Model for t h e

proposed mechanisms of

DNA-induced i m m u n e response to H B ~ A ~ .

Uptake and processing of HBsAg via class I and class II pathways

Restimulation

-cle

V

Fiber or APC>

Amplification of the immune response via transfected cells or APC

with necrotizing agents, may also be directly transfected with DNA and express the antigen (see Figure 3). These cells are the only cells able to present the antigen that induces cell-mediated immunity. Processing of the antigen and presentation as peptides by MHC class I molecules results in the priming of CD8' T cells and cytotoxic responses after migration of APC into lymph nodes. Processing of the secreted HBsAg and presentation via the MHC class I1 pathway results in the priming of the CD4' T cell response and provides efficient T help for both CTL and B lymphocytes (see Figure 3). In addition, exogenous HBsAg in the form of particles can be taken up and processed for class I presentation (Schirmbecli et al., 1994), thus accounting for additional induction of cytotoxic T cell precursors from CD8' lymphocytes. Furthermore, any cell type which expresses class I molecules can potentially restimulate CTL precursors in a secondary phase, thus amplifying the response, This is probably the case for transfected muscle cells, which may serve as a reservoir of antigen production and peptide presentation, but in turn have been shown to be destroyed by the immune response by ten to twenty days following transfection (Davis et a., 1997).Thus, DNA vaccines by the in v i m synthesis of antigen leads to appropriate presentation of antigenic peptides in the context of class I and class I1 MHC molecules, which results in strong humoral and cellular immune responses. The prolonged synthesis of antigen may act as a boost resulting in long-term memory.

2.1.4

The primate model

We have tested the efficiency of genetic immunization with the hepatitis B virus envelope based vectors in naive primates (rhesus macaques). In these experiments, DNA was administered without adjuvant intramuscularly (500 pg in the deltoid muscle) and intradermally (500 pg in the back) at months 0, 2, 8 and 12. The immunizations were well tolerated by all animals. No local reactions were noticed at the sites of injections and also proximal lymph nodes were not reactive.

5 lmmunotherapy of Chronic Hepatitis B by pCMV-82.8 DNA Vaccine

Injections of pCMVS2.S DNA encoding two HBV envelope proteins first induced an HBV-specific cytotoxic response followed by the appearance of potentially protective anti-HBs antibodies (55 mUI per mL after two injections and up to 1,500 mUI per mL after four injections). The frequency of CTL precursor cells as measured by limiting dilution assay was in the same range of magnitude as those found in humans following infection. Moreover, injections of a DNA expression vector (pCMVV3.S) encoding an epitope of the human immunodeficiency virus envelope fused to HBsAg induced strong humoral and cytotoxic responses to antigenic determinants of both viruses in mice and non-human primates alike. In addition, in protein-primed rhesus monkeys B cell memory was successfully boosted by DNA injection and animals subsequently developed a multispecific cellular response. This suggests that DNA-based immunization could be used to boost eficiently and broaden the immune response in individuals immunized with conventional vaccines, regardless of their genetic variability (Le Borgne et al., 1998).

2.1.5

DNA-based vaccination of chimpanzees against HBV

The chimpanzee is the best animal model to evaluate DNA vaccines for their potential use in humans. It is similar to man with respect to susceptibility to HBV infection and antibody titer required for protection. Intramuscular immunization of chimpanzees using pCMVS2.S DNA resulted in an HBsAg-specific humoral response. High anti-HBs levels (> 100 mIU per mL) were induced in a female chimpanzee with 2 mg DNA. However, at least two boost injections of DNA were required to prevent antibody levels from diminishing over time. Nevertheless, extremely high antibody titers were ultimately attained (> 14,000 mUI per mL), and high levels were sustained for at least one year. In a second chimpanzee (male) injected with a lower dose of DNA (400 ,ug) anti-HBs were only detected after the second injection of DNA. Although antibody titers of 60 mUI per mL were attained, these were transient even after three DNA boosts. Despite loss of detectable anti-HBs, this chimpanzee showed a strong anamnestic response to injection of recombinant HBsAg given one year after the initial DNA injection. As observed in mice, DNA-based vaccination of chimpanzees induces HBsAg-specific antibodies. Initially of IgM isotype, these subsequently shift to IgG antibodies, which are predominantly IgG1. The antibodies induced are predominantly against the preS2 domain, but are also specific to the group and subtype determinants of the S protein (Davis et al., 199Gb). Levels of antiHBs and protection afforded correlate well for human and chimpanzee. Based on titers higher than 10 mUI per mL observed in one chimpanzee, this animal would have been protected against HBV infection. Regarding the second animal, the strong and rapid anamnestic response to recombinant HBsAg suggests that this animal would probably resist hepatitis. Regarding the safety issues, no abnormalities were found in the weekly serum assays (routine hematology and liver enzymes), nor were any other untoward reactions noted.

I

95

96

I

M a w L o u i s e Mrchel

2.1.6

Neonatal immunization

Two chimpanzees born to HBsAg-naive mothers were immunized on the day of birth and at 6 and 24 weeks with 500 pg pCMV-Sl.SZ.S. Although only transient and low levels of anti-HBs were detected both animals were protected from subsequent live viral challenge (100 CID). The immunized animals developed anamnestic anti-HBs antibody responses; but neither developed detectable HBsAg or antibody to the core protein. In both animals small amounts of HBV DNA were produced concurrent with the anamnestic response indicating that sterilizing immunity was not achieved (Prince et al., 1997).

3

DNA-based vaccination for chronic HBV infections 3.1

HBsAg transgenic mice as a model for HBV chronic carriers

Owing to the endogenous synthesis of antigen, the processing into relevant epitopes, and the induction of CD8' CTLs, DNA vaccines may be useful for the treatment of individuals chronically infected with HBV As a model to study the possibility of inducing an immune response in HBV chronic carriers that could control the infection, we used transgenic mice that constitutively express the HBsAg in their liver only. The transgene which consists of one copy of the HBV genome, but with the core gene deleted, is expressed before birth under the control of an endogenous HBV promoter (Babinet et al., 1985; Farza et al., 1987). This results in the secretion of large amounts of HBsAg into the serum without antibody production and without apparent liver pathology. These mice may represent a model for HBV chronic carrier infected at birth without HBV replication, generally defined as asymptomatic carriers. Using the pCMV62.S plasmid encoding the small and the middle HBV envelope protein, rapid production of anti-HBs antibodies was induced in these transgenic mice which in turn induce the clearance of circulating HBsAg. Anti-preS2 and anti-HBs antibodies induced after a single immunization in HBsAg-Tg mice reach levels as high as those induced in normal C57BL/G mice after DNA-based immunization. The isotypes of the induced antibodies is comparable in Tg and in non-Tg mice and include IgM and then IgG; predominantly IgG2b and IgGl with some IgG3. Elimination of HBsAg concomitantly with anti-HBs induction occurred in most of the mice in 4-8 weeks and persisted for at least 20 weeks without any further injection of DNA. This clearance is not only due to the neutralization of antigen by antibodies, but is also correlated with a decrease or a disappearance of the HBV messenger RNA from the liver (Mancini et al., 1996b).Down-regulation of transgene expression is likely mediated by HBsAg-specificT cells as shown by adoptive transfer experiments. These experiments, involving fractionated HBsAg primed spleen cells obtained from DNA-immunized mice into Tg mice, showed that both CD4' and CD8' T cells were able to control transgene expression even in the absence of antibody production. Interest-

5 lrnrnumtherapy of Chronic Hepatitis B by pCMV-S2.S D N A Vaccine

ingly, this regulation of transgene expression occurred without increase of liver enzymes and without histological evidence of liver necrosis (Mancini et al., 1998). This suggests that the T cell effect is probably mediated through a non cythopathic mechanism. Results from HBsAg-Tg mice that are knock-out for the y-interferon receptor gene indicate that the regulation of the HBV envelope messenger RNA was mediated by type 1 cytoltines produced by the activated T lymphocytes. Indeed, in vitro activation of spleen cells from DNA-immunized mice with HBV antigens resulted in predominant production of IFN-y, TNF-a and low levels of IL-2, but not IL-4 or IL-10. It was shown in similar Tg mouse models that HBV-specific CTL can abolish viral replication or transgene expression in the hepatocytes by noncytopathic mechanisms that are mediated by IFN-y and TNF-a (Guidotti et al., 1996). The antiviral activity of the cytokines operates probably through the degradation of HBV mRNA in the liver following the induction of endoribonuclease activity cleaving the viral RNA immediately adjacent to a stem loop structure (Heise et al., 1999a, b). The mechanisms by which DNA-based immunization can break B and T cell tolerance, or anergy, in transgenic mice is still unresolved. Systemic injection of recombinant HBsAg alone is not sufficient to induce B or T cell responses able to control transgene expression, unless it contains heterologous B and T cell epitopes or is derived from an heterologous subtype of HBV (Mancini et al., 1993). In contrast, intramuscular injection of a plasmid encoding HBsAg and secreting a very limited amount of antigen in viuo is sufficient to induce a response in these Tg mice. This could be related to the presentation of different antigenic peptides to the immune system following either DNA-based or systemic immunization, or to the direct activation of antigen presenting cells via the CpG motifs presents in the plasmid backbone (Klinman et al., 1997). The non-cytolytic antiviral process described in the Tg mice, if confirmed in humans, could be of importance for the control of the viral infection without liver disease, but could also result in an incomplete elimination of the virus from the liver, thereby promoting viral persistence. These results have been further confirmed in duck, a model for DHBV infection. Three day old infected ducklings develop chronic infection with aspects mimicking what occurs in the HBV chronic carrier state. In these experiments chronically infected ducks were immunized four times with DNA plasmids encoding the large envelope protein. A significant decrease or even elimination of viral replication was observed in immunized ducks. This effect was persistent since no rebound in viral replication was observed after cessation of the treatment. More importantly, analysis of intrahepatic viral DNA showed the clearance of the cccDNA pool in two animals which is tightly associated with rebounds of viral replication after anti-viral treatments (Rollier et al., 1999). These results point to the possibility of designing more effective ways for the prevention and treatment of HBV infections. However, many basic questions of this vaccine approach still need to be answered, namely safety issues associated with the injection of DNA vaccines and potential adverse effects of a too strong cytotoxic response against the infected liver cells.

I

97

98

I

M a r k L o u i s e Miche/

4

Clinical trials o f DNA vaccines

Research efforts to develop DNA vaccine products have progressed to the clinical evaluation of several vaccine candidates. These clinical trials assess safety and immunogenicity and whether different routes and schedules of administration can improve immune responses. Vaccine candidates currently under study include those for hepatitis B, herpes simplex-2, HIV-1, influenza and malaria. Unique safety concerns include the potential for genomic integration, biodistribution, tolerance and autoimmunity. One major concern was that after injection the DNA would integrate into the recipient host’s chromosome leading to mutagenesis and potentially insertional carcinogenesis. Animal studies involving plasmid DNA injections have shown that any mutations from a potential integration event would be infrequent and result in a calculated mutation rate much lower than the spontaneous mutation rate for the mammalian genome (Martin et al., 1999). A second concern was that the immunization with plasmid DNA would induce anti-DNA antibodies and this would accelerate the development of autoimmune diseases. Animal studies have shown non-significant increase in anti-DNA antibodies and no evidence of autoimmune disease induction or acceleration after administration of plasmid DNA injections (Mor et al., 1997). In a Phase I safety and tolerability clinical trial of a malaria DNA vaccine, 28 healthy, malaria-nalveadult volunteers were immunized with plasmid DNA encoding a malaria antigen. It was found that intramuscular administration of three doses of up to 2,500 pg of plasmid DNA was safe, well tolerated and not inducing any obvious hematological or biochemical abnormalities or anti-DNA antibodies. This was the first demonstration of the induction of CD8’ CTLs by DNA vaccines in healthy nalve humans (Wang et al., 1998). However, DNA vaccination via intramuscular injection failed to induce detectable malaria-specificantibodies in any of the volunteers (Le et al., 2000). Two clinical trials of DNA vaccines for the treatment of human immunodeficiency virus-1 (HIV-1)infections were also reported. A DNA-based vaccine containing HIV-1 envelope and rev genes was tested for safety and host immune response in 15 asymptomatic HIV-infectedpatients. Vaccine administration induced no local or systemic reactions and no laboratory abnormalities were detected (no anti-DNA antibodies or muscle enzyme elevations). Injection of plasmid DNA resulted in antibody against the HIV envelope and some increases were observed in envCTL and lymphocytes proliferative activity (MacGregor et al., 1998; Ugen et a]., 1998). Another trial was performed in 9 asymptomatic HIV-infected patients using DNA encoding regulatory genes of HIV-1 (neJ;tat, rev). DNA immunization induced antigen-specific T cell proliferation which persisted up to nine months after the last injection and CD8-mediated cytolytic activities, but did not by itself reduce viral load (Calarota et al., 1998; Calarota et al., 1999). A phase I clinical trial of DNA vaccine for hepatitis B was conducted in seven adult healthy volunteers using a plasmid encoding HBsAg and the “gene gun” as delivery system. Extremely low doses of DNA (0,25 pg) coated onto gold

5 lmmunotherapy of Chronic Hepatitis B by pCMV-S2.S D N A Vaccine

beads were administered two times into the sltin. Results were very disappointing, since the only one of the six volunteers who developed high titers of HBs antibodies may have had previous exposure to hepatitis B (Tadtet et al., 1999). Nevertheless, CTL induction specific for HBV envelope has been reported for some HLA-A2 patients receiving HBsAg-encoding DNA via gene gun immunization (Swain et al., Development and clinical progress of DNA vaccines, Langen, Germany, 1999, personal communication). Altogether, the results from the first clinical trials show that DNA vaccines appear to be well tolerated. The aim of the studies described above is to develop a therapeutic vaccine against persisting HBV infection. Patients suffering from chronic-associated liver disease with the risk of developing hepatocellular carcinoma will greatly benefit from the availability of a therapeutic vaccine against HBV. Vaccination may be the therapeutic procedure with the lowest cost and the greatest potential benefit. As chronic carriers are the major source of spreading the disease, control of infection in this population will be a major advance in controlling overall HBV infection. Acknowledgements

I thank all of my colleagues who have collaborated with me on the studies reported here: H. L. Davis, D. Loirat, S. Le Borgne, M. Mancini, E. Malanchi.re, M. Schleef, R. G. Whalen. I am particularly grateful to P. Tiollais for his constant support. I greatly appreciate the help of Dr. H. Farza and Dr. P. Townsend for the critical

reading of this manuscript.

References DNA vaccination in HIV-1-infected patients, ALBERTI,A,, CAVALLETIO, D., PONTISSO,P., Lancet 351 (9112), 1320-1325. CHEMELLO, L., TAGARIELLO, G., BELUSSI,F. CALAROTA, S. A., LEANDERSSON, A. C., BRAIT, (19881, Antibody response to pre-S2 and G., HINIWLA,J., KLINMAN, D.M. et al. hepatitis B virus induced liver damage, (1999), Immune responses in asymptomatic Lancet 331, 1421-1424. HIV-1-infected patients after HIV-DNA BABINET,C., FARZA,H., MORELLO, D., HADimmunization followed by highly active CHOUEL, M., POURCEL, C. (1985), Specific antiretroviral treatment, I. bnrnunol. 163 (4), expression of hepatitis B surface antigen 2330-2338. (HBsAg) in transgenic mice, Science 230, CHISARI,F. V. (1995), Hepatitis B vims 1160-1163. immunopathogenesis, Annu. Rev. Irnrnunol. BERTOLETII, A., D’ELIOS,M. M., BONI, C., 13, 29-60. DE CARLI,M., ZIGNEGO,A. L. et al. (19971. Different cytolune profiles of intraphepatic T COUILLIN,I., POL, S., MANCINI,M., DRISS,F., BRECHOT, C. et al. (1999), Specific vaccine cells in chronic hepatitis B and hepatitis C therapy in chronic hepatitis E: induction of vims infections, Gastroenterology 112 (l), T cell proliferative responses specific for 193-199. envelope antigens [published erratum CALAROTA, S., BRAT^, G., NORDLUND, S., appears in J Infect Dis 1999 Nov;180(5):1756], HINKULA, J., LEANDERSSON,A. C. el al. (1998),Cellular cytotoxic response inducedby /. Infect. Dis. 180 (l),15-26. I

,

I

99

100

I

MarkLouise Michel

DAVIS,H. L., MANCINI,M., MICHEL, M.-L., R. G. (1996a),DNA-mediated WHALEN, immunization to hepatitis B surface antigen: Longevity of primary response and effect of boost, Vaccine 14 (9), 910-915. DAVIS,H. L., MCCLUSICIE, M. J., GERIN,J. L., PURCELL, R. H. (1996b), DNA vaccine for hepatitis B: evidence for immunogenicity in chimpanzees and comparison with other vaccines, Proc. Natl. Acad. Sci. U S A 93, 72137218. DAVIS,H. L., MICHEL, M.-L., WHALEN, R.G. (1993), DNA based immunization for hepatitis B induces continuous secretion of antigen and high levels of circulating antibody, Hum. Mol. Genet 2, 1847-1851. DAVIS,H.L., MICHEL, M.L., MANCINI, M., SCHLEEF,M., WHALEN, R. G. (1995), DNAbased immunization overcomes H-2 haplotype-restricted non-responsiveness to hepatitis B surface antigen in mice, in: Vaccines 95, Molecular Approches to the Control of Infectious Diseases (Ginsberg, H. S., Brown, F., Chanocck, R. M., Eds.), pp. 111-116. Cold Spring Harbor Laboratory Press, New York. S. C. DAVIS,H. L., MILLAN, C. L., WATKINS, (1997), Immune-mediated destruction of transfected muscle fibers after direct gene transfer with antigen-expressingplasmid DNA, Gene Ther. 4 (3), 181-188. DONNELLY, J. J., ULMER, J. B., SHIVER, J. W, LIU, M.A. (1997), DNA vaccines, Annu. Rev. Immunol. 15, 617-648. M., FARZA, H., SALMON,A,-M., HADCHOUEL, MOREAU, J.-L., BABINET,C. et al. (1987), Hepatitis B surface antigen gene expression is regulated by sex-steroids and glucocorticoids in transgenic mice, Proc. Natl. Acad. Sci. U S A 84, 1187-1191. D. (1991), Assembly of hepadnaviral GANEM, virions and subviral particles, in: Current Topics in Microbiology and Immunology: Hepadnaviruses, Molecular Biology and Pathogenesis Vol. 168 (Mason, W. S., Seeger, C., Eds.), pp. 61-83. Springer-Verlag, Berlin, Heidelberg. D. (1996), Hepadnaviridae: The Viruses GANEM, and their Replication, Third Edition ed. Virology 2. 2 Vols. (Fields, B. N., Ed.). Lippincott-Raven, Philadelphia. GUIDOTTI,L. G., CHISARI,F.V (1999), Cytoltine-induced viral purging - role in viral pathogenesis, Curr. Opin. Microbiol. 2 (4), 388-391.

GUIDOTTI, L.G., ISHIIWWA,T., HOBBS, M.V., B., SCHREIBER,R., CHISARI, F.V. MATZICE, (1996), Intracellular inactivation of the hepatitis B virus By cytotoxic T lymphocytes, Immunity 4 (l),25-36. HEISE,T., G U I D O ~L.G., I , CAVANAUGH, V J., F.V. (1999a), Hepatitis B virus CHISARI, RNA-binding proteins associated with cytoltine-induced clearance of viral RNA from the liver of transgenic mice, /. Virol. 73 (l), 474-481. HEISE,T., GUIDOITI,L. G., CHISARI, F. V. (1999b),La autoantigen specifically recognizes a predicted stem-loop in hepatitis B virus RNA, J . Virol. 73 (7), 5767-5776. J. H., DI BISCEGLIE,A.M. (1997), HOOFNAGLE, The treatment of chronic viral hepatitis, N . Engl. /. Med. 336 (5), 347-356. J. H., LAU, D. (1997),New theraHOOFNAGLE, pies for chronic hepatitis B, J. Viral Hepat. 4 (Suppl. l),41-50. Hut, J., MANCINI, M., LI, G., WANG,Y., P., MICHEL, M. L. (1999), ImmuniTIOLLAIS, zation with a plasmid encoding a modified hepatitis B surface antigen carrying the receptor binding site for hepatocytes, Vaccine 17 (13-14), 1711-1718. ITOH,Y., TAICAI, E., OHNUMA, H., KITAJIMA, K., TSUDA, F. et al. (1986), A synthetic peptide vaccine involving the product of the pre-S(2) region of the hepatitis B virus DNA: protective efficacy in chimpanzees, Proc. Natl. Acad. Sci. U S A 83, 9174-9178. JUNG, M. C., HARTMANN, B., GERLACH, J.T., DIEPOLDER, H., GRUBER, R. et al. (1999), Vims-specific lympholcine production differs quantitatively but not qualitatively in acute and chronic hepatitis B infection, Virology 261 (2), 165-172. KLINMAN,D. M., YAMSHCHIKOV,G., ISHIGATSUBO, Y. (1997), Contribution of CpG motifs to the immunogenicity of DNA vaccines, /. Immunol. 158 (8), 3635-3639. KRUSKALL,M. S., ALPER,C.A., AWDEH, Z., YUNIS, E. J., MARCUS-BAGLEY, D. (1992),The immune response to hepatitis B vaccine in humans: Inheritance patterns in families, /. Exp. Med. 175, 495-502. LAI, C. L., CHIEN,R. N., LEUNG,N. W., CHANG, T. T., GUAN,R. et al. (1998).A one-year trial of lamivudine for chronic hepatitis B. Asia Hepatitis Lamivudine Study Group, N. Engl. /. Med. 339 (2), 61-68.

5 lmmunotherapy of Chronic Hepatitis B by pCMV-52.5 D N A Vaccine

1101

LE BORGNE,S., MANCINI, M., LE GRAND, R., “nonresponse” to HBsAg, J. Med. Virol. 39, SCHLEEF, M., DORMONT,D. et al. (1998). In 67-74. vivo induction of specific cytotoxic T MARTIN,T., PARKER, S. E., HEDSTROM, R., lymphocytes in mice and rhesus macaques LE, T., HOFFMAN, S. L. et al. (1999), Plasmid immunized with DNA vector encoding an DNA malaria vaccine: the potential for HIV epitope fused with hepatitis B surface genomic integration after intramuscular antigen, Virology 240, 304-315. injection, Hum. Gene %her. 10 (S), 759-768. LE, T. P., COONAN, I<. M., HEDSTROM, R. C., M., MICHEL,M.-L., DAVIS,H. L., SCHLEEF, CHAROENVIT, Y.,SEDEGAH,M. et al. (ZOOO), MANCINI, M., TIOLLAIS, P.,WHALEN, R. G. Safety, tolerability and humoral imrnune (1995), DNA-mediated immunization to the responses after intramuscular administration hepatitis B surface antigen in mice: Aspects of a malaria DNA vaccine to healthy adult of the humoral response mimic hepatitis B volunteers, Vaccine 18 (18), 1893-1901. viral infection in humans, Proc. Natl. Acad. LEE, W. M. (1997), Hepatitis B virus infection, Sci, U S A 92, 5307-5311. N. Engl. J. Med. 337 (24), 1733-1745. MILICH,D. R., Mc NAMARA,M. I<., MCLACHAN, LOIRAT, D., LI, Z., MANCINI, M., TIOLLAIS, P., F.V. (1985), A., THORNTON, G., CHISARI, PAULIN, D., MICHEL, M. L. (1999), MuscleDistinct H-2 linked regulation of T-cell specific expression of hepatitis B surface responses to the pre-S and S regions of the antigen: no effect on DNA-raised immune same hepatitis S surface antigen polypeptide responses, Virology 260 (l), 74-83. allows circumvention of nonresponsiveness LOIRAT,D., LE MONNIER F. A,, MICHEL, M-L. to the S region, Proc. N d .Acad. Sci. USA 82, 8168-8172. (2000), Multiepitopic HLA-Pc0201-restricted A.D., immune response against hepatitis B surface MOR, G., SINGLA,M., STEINBERG, antigen after DNA-based immunization, S. L., OI~UDA, K., KLINMAN, D. M. HOFFMAN, J. lmmunol. 165, 4748-4755. (1997), Do DNA vaccines induce autoR. R., BOYER,J. D., UGEN, I<. E., MACGREGOR, immune disease? Hum. Gene Tner. 8 (3), 293-300. LACY, I<. E., GLUCKMAN, S. J. et al. (1998), First human trial of a DNA-based vaccine for NEURATH,A. R., KENT, S. B. H., PARKER, I<., treatment of human immunodeficiency vims PRINCE, A.M., STRICK, N. et al. (1986), type 1 infection: Safety and host response, Antibodies to a synthetic peptide from the /. Infect. Dis. 178 (l),92-100. preS 120-145 region of the hepatitis B virus P., MANCINI,M., DAVIS,H.L., TIOLLAIS, envelope are virus-neutralizing, Vuccixe 4, MICHEL,M:L. (1996a), DNA-based immuni- 35-37. zation against the envelope proteins of the S., BERVAS,N., URE,J.M., SMITH, PASCOLO, hepatitis B virus, J. Biotechnol. 44, 47-57. A. G., LEMONNIER, F.A., PERARNAU, B. MANCINI, M., HADCHOUEL, M., DAVIS,H. L., (1997), HLA-A2.1-restrictededucation and R. G., TIOLLAIS, P., MICHEL, M. L. WHALEN, cytolytic activity of CD8(+) T lymphocytes (1996b), DNA-mediated immunization in a from beta2 microglobulin (beta2m) HLAtransgenic mouse model of the hepatitis B A2.1 monochain transgenic H-2Db beta2m surface antigen chronic carrier state, Proc. double knockout mice, J. Ezp. Med. 185 (12), Natl. Acad. Sci. USA 93 (22), 12496-12501. 2043 -2051. MANCINI, M., HADCHOUEL, M., TIOLJAIS,P., PENNA, A., DELPRETE,G., CAVALLI, A,, MICHEL,M. L. (1998), Regulation of A., D’ELIOS,M.M. et al. (1997), BERTOLETTI, hepatitis B virus mRNA expression in a Predominant T-helper 1 cytoline profile of hepatitis B surface antigen transgenic mouse hepatitis B virus nucleocapsid-specific T cells model by IFN-gamma-secretingT cells after in acute self-limited hepatitis B, Hepatology DNA-based immunization, /. Immunol. 161, 25 (4),1022-1027. 5564-5570. M.E., PETERS, M., VIERLING,J.. GERSHWIN, MANCINI, M., HADCHOUEL, M., TIOLLAIS, P., 1. H. MILICH,D., CHISARI, F. V., HOOFNAGLE, C., MICHEL, M.-L. (l993), InducPOURCEL, (199l),Immunology and the liver, Hepatology tion of anti-hepatitis B surface antigen 13 (51, 977-994. (HBsAg) antibodies in HBsAg transgenic PRINCE, A.M., WHALEN, R., BROTMAN,B. mice: a possible way of circumventing (19971, Successful nucleic acid based immu-

102

I

Marie-Louise Michel protein elicits murine H-2 class I-restricted nization of newborn chimpanzees against CD8’ cytotoxic T lymphocyte responses in hepatitis B virus, Vaccine 15 ( 8 ) ,916-919. , Immunol. 152, 1110-1119. REHERMANN, B. (ZOOO), Intrahepatic T cells in U ~ U O J. TACKET,C. O., ROY,M. J., WIDERA,G., SWAIN, hepatitis B: viral control versus liver cell R.(19991, Phase W. F., BROOME,S., EDELMAN, injury, J. Exp. Med. 191 ( 8 ) , 1263-1268. 1 safety and immune response studies of a ROLLIER,C., SUNYACH, C., BARRAUD, L., MADANI, N., JAMARD, C. et al. (1999), Protective DNA vaccine encoding hepatitis B surface antigen delivered by a gene delivery device, and therapeutic effect of DNA-based immunization against hepadnavims large envelope Vaccine 17 (22), 2826-2829. UGEN,I<. E., NYLAND,S. B., BOYERJ. D., VIDAL, protein, Gastroenterology 116 (3), 658.665. ROMAN,M., MARTIN-OROZCO, E., GOODMAN, C., LEU, L. et al. (1998), DNA vaccination with HIV-1 expressing constructs elicits J. S., NGUYEN,M. D., SATO,Y. et al. (1997), immune responses in humans, Vaccine 16 Immunostimulatory DNA sequences (19), 1818-1821. function as T helper-1-promoting adjuvants, VITIELLO,A., ISHIOI~A, G., GREY,H. M., ROSE, Nature Med. 3 (8),849-854. SATO,Y., ROMAN,M., TIGHE, H., LEE, D., CORR, R., FARNESS,P. et al. (1995), Development of a lipopeptide-based therapeutic vaccine to M. et al. (1996), Immunostimulatory DNA sequences necessary for effective intrademal treat chronic HBV infection. I. Induction of a primary cytotoxic T lymphocyte response in gene immunization, Science 273 (5273), humans, J. Clin. Invest. 95 (l), 341-349. 352 - 354. WANG,R., DOOLAN,D. L., LE, T. P., HEDSTROM, SCHIRMBECK, R., BOHM, W, ANDO, I<., R. C., COONAN,I<. M. et al. (1998), Induction CHISARI,F.V., REIMANN,J . (1995), Nucleic of antigen-specific cytotoxicT lymphocytes in acid vaccination primes hepatitis B virus humans By a malaria DNA vaccine, Science surface antigen-specific cytotoxic T lympho282, 476-480. cytes in nonresponder mice, J. Virol. 69, WEI, Y., TIOLLAIS,P. (Eds.) (19991, Molecular 5929-5934. Biology of Hepatitis B Virus Vol. 3:2. Clinics SCHIRMBECR, R., MELBER,I<., KUHROBER,A., JANOWICZ,Z.A., REIMANN,J. (1994), Immu- in Liver Disease (Lee, W M., Ed.). W. B. nization with soluble hepatitis B virus surface Saunders Company, Philadelphia.

P/asmids for Therapy and kchation by M.Schleef Copyright 0 WILEY-VCH Verlag GmbH, 2001

6 pSC.MEPfrRAP A First Generation Malaria DNA Vaccine Vector Jorg Schneider”, Sarah C. Gilbert and Adrian Hill

1

Parasite life cycle and impact o f malaria

Malaria, caused by the parasite Plasmodiumfalciparum, annually claims 2-3 million lives worldwide. More than 500 million people (40 % of the world’s population) are infected. The extent of the mortality and morbidity inhibits socio-economical development of the respective societies. Malaria particularly affects sub-Saharan Africa. The life cycle of the parasite is complex and involves anopheline mosquitoes as the transmission vector (Figure 1).The infectious stages of the malaria parasite, the sporozoites, are inoculated during the blood meal of an infected mosquito. Within minutes following inoculation, sporozoites home to the liver where they infect hepatocytes. This liver stage of P. fakiparum lasts for up to a week and has no pathological effects. Subsequently, the parasite invades erythrocytes and a different set of genes is expressed. This blood stage of malaria infection results in a massive parasite burden and leads to the typical symptoms of malaria (fever, shivers and apathy). In the case of P. fakiparum malaria, anemia and particularly cerebral malaria are the most common causes of death in infected individuals. During the blood stage sexual stages of the parasite eventually develop. In order to complete the live cycle these sexual stages have to be ingested during the blood meal of another female mosquito. As outlined, the parasite not only lives in two different hosts, but within each host it has several developmental stages. In the 1960s and 1970s international health organizations announced that malaria would be eradicated within 10-20 years (Jackson, 1998). This optimism was based on the initial success of vector control programs using insecticides and other malaria control measurements. However, with the decay of health care systems in malaria endemic regions due to political and social unrest, the implementation and maintenance of vector control programs became difficult. In addition, drug-resistant P. falciparum and insecticide-resistant Anopheles strains developed under the selection pressure. A vaccine against malaria is one of the few remaining underdeveloped tools for control. Vaccination is considered to be one of the most cost-effective health care measurements.

104

I

Jorg Schneider et al.

I I

Fig. 1. Life cycle of the malaria parasite.

The observation that DNA vaccines induce humoral and cellular immune responses raised new hopes for the development of subunit vaccines against intracellular pathogens and tumors. Vaccine research currently experiences a renaissance and the idea of controlling infectious diseases for which there is no traditional vaccine available seems a realistic goal. The need to combat diseases such as malaria and HIV on a global scale released significant public and private funding. The continuing economic boom particularly in the US led to the availability of substantial amounts of private money through philanthropic organizations. For both malaria and HIV two new private funding organizations with dedicated vaccine approaches have been established. DNA vaccines play an important role in the vaccine development programs of the Malaria Vaccine Initiative (MVI, h t t p : / / w w . MalariaVaccine.org) and the International Aids Vaccine Initiative (IAVI; http://w.iavi.org).

2

Concept of vaccination against malaria

The complexity of the parasite’s life cycle makes the development of a malaria vaccine much more difficult than vaccine development against diseases caused by viruses. Plasmodiumfalclparum, the most virulent of four parasite species infecting humans, not only uses two different hosts (humans and anopheline mosquitoes),

G pSG.MEPflRAP - A First Generation Malaria D N A Vaccine Vector I 1 0 5

but it also expresses different proteins at different stages of development (Nussenzweig et al., 1994). These proteins are potential target antigens for vaccines. To complicate matters further different geographical isolates of P. falciparum show considerable sequence variations (Robson et al., 1990). Observations from studies in animal models and humans suggest that different stages of the parasite are targeted by different effector arms of the immune response. Historically,the development of malaria vaccines is divided into targeting the different stages of malaria. Vaccines targeting the blood stage of the parasite need to induce high levels of antibodies. Studies in semi-immune individuals living in malaria endemic areas suggest that antibodies may contribute to decreased incidence and prevalence and levels of blood stage parasitemia. The blood stage vaccine candidate SPf66 which contains antigens and epitopes derived from blood stage antigens failed to achieve protection in the field (Bojang et al., 1998).Antigens expressed by the sexual stages of the parasite offer yet another target for developing transmission blocking vaccines (Kaslow, 1997).Antibodies would prevent the uptake of gametocytes by mosquiotes and, therefore, interrupt the parasite's life cycle at a population level. The most promising approach for the development of a malaria vaccine is targeting the liver stage of the parasite. Repeated immunization with irradiation-attenuated sporozoites induced protection against a subsequent challenge with infectious sporozoites (Nussenzweig et al., 1994). In order to induce protective immunity, sporozoites have to infect hepatocytes and develop into liver stages. However, propagation of sporozoites in vitro for a vaccine remains impossible. Therefore, the development of a subunit malaria vaccine targeting the liver stage is a major goal. Using the advances of modern molecular biology, vaccinologists try to mimick the protective efficacy observed in model immunizations with irradiation-attenuated sporozoites. Direct studies in mice (Romero et al., 1989) and indirect studies in humans (Hill et al., 1992) suggest that liver stage specific CD8' T cells are capable of mediating protection against a subsequent sporozoite challenge. CD8' T cells recognize antigen as a trimeric complex of 8-10 amino acid peptides (eptiopes) bound to the groove of MHC class I molecules stabilized by P2-microglobulin. In malaria a number of CD8' T cell epitopes derived from liver stage specific antigens have been identified (Aidoo et al., 1995; Doolan et al., 1997). These epitopes can be fused to a string of epitopes. Inclusion of many epitopes presented by common MHC class I haplotyes should ensure that the majority of the vaccinated population mounts appropriate CD8' Tcell responses. Several liver stage specific antigens such as the circumsporozoite protein (CSP) (Spitalny et al., 1973) and thrombospondin related adhesive protein (TRAP) (Robson et al., 1988) have been cloned. Efforts to sequence the complete genome of P. falciparum will result in the identification of many more potential candidate antigens for vaccination. The challenge for a malaria vaccine targeting the liver stage is the ability to induce antigen-specificcellular immune responses, particularly CD8' and CD4' T cells. Subunit vaccines based on recombinant proteins alone do not induce cellular immune responses. One approach trying to overcome this limitation is the use of adjuvants together with recombinant proteins. The most developed malaria vaccine candidate is RTS,S a fusion protein of the hepatitis B virus surface antigen and the C-terminal

106

I

Jorg Schneider et a/.

portion of CSP in an adjuvant. Recent clinical studies demonstrate immunogenicity for cellular immune responses (Lalvani et al., 1999) and short-lived protection from challenge (Stoute et al., 1997). DNA vaccines with their ability to induce both cellular and humoral immune responses are promising antigen delivery systems. This chapter describes the generation and development of plasmid DNA vaccines to be tested in humans.

3

First-generation plasmid: pSC.MEPfTRAP 3.1

Vector backbone

Plasmid vector pSG.MEPfTRAP is based on pSG, a standard 6773 bp mammalian expression vector containing regulatory sequences for expression in mammalian cells. pSG derived from vector pTH (Hanke et al., 1998) by exchanging the ampicillin resistance gene with a ltanamycin resistance gene. The human cytomegalovirus (CMV) immediate-early enhancerlpromoter containing intron A drives the expression of the insert. For the expression of malaria antigens we found that this promoter is stronger than the short version of the CMV promoter without the enhancer. DNA vaccines containing the enhancer in the promoter were more immunogenic than constructs containing the short CMV promoter. The bovine growth hormone-derived polyadenylation signal ensures polyadenylation of the mRNA. Additional sequences for propagation in E. coli, inducing high copy numbers and selection markers are described in greater detail in Chapters 1, 7 and 11 of this book. According to FDA guidelines on DNA vaccine vectors pSG carries a kanamycin resistance gene. 3.L

insert

The polypeptide encoded by the pSG.MEPfTRAP vaccine consists of a series of known CD8' T cell epitopes from Plasmodium falciparum liver stage specific antigens (ME: multiple epitopes). This CD8' Tcell epitope string is N-terminally fused to a complete liver stage specific antigen, Thrombospondin Related Adhesion Protein (TRAP) (Robson et al., 1988). The 20 peptide epitopes in the multiple epitope string (Gilbert et al., 1997) were selected as follows. The first criterion in epitope selection was the antigen of origin of the epitope. We attempted to select at least one epitope from all pre-erythrocytic P. fulcipauurn antigens known to be recognized by CD8' T cells. To date, seven antigens have been identified as targets of CD8' T cells responses and included in the epitope string are epitopes from six of these. We included several epitopes from the two leading vaccine candidate antigens, CSP and TRAP. Homologs of these two antigens and a third antigen, exported protein-1, have induced CD8' T cell-mediated

G pSG.MEPfTRAP

-

A First Generation Malaria D N A VacciMe Vector I107

protection against sporozoite challenge in rodent malaria models. Rodent homologs have not been defined for the other three malaria antigens: liver stage antigens-1 and -3, sporozoite threonine and asparagine-rich protein. However, there is evidence in support of a potential protective role for each of these from experimental and field studies. The second criterion was the definition of a peptide as a CD8' T cell epitope for a particular HLA class I molecule. The 14 CD8+ T cell epitopes selected constitute the vast majority of the known CD8+ T cell epitopes for the malaria parasite P. fdciparum (see also Table 1).Thus there was acbally little choice to be made, but the considerations listed here influenced the final selection as well as the types of epitopes that were originally searched for. These CD8' T cell epitopes were defined during studies of immune responses of Africans in endemic areas (Hill et al., 1992; Aidoo et al., 1995). During this work we attempted to identify epitopes for a variety of common HLA types. This represented a third criterion for selection, the rationale being that the more HLA types for which an epitope could be included, the greater the proportion of the vaccinated population that would be able to recognize at least one peptide epitope in the string. Thus epitopes are included for the following HLA types: HLAA2.1, -A2.2, -B8, -B35, -B53, -B58. These are particularly prevalent HLA class I types in both Europeans and West Africans. For example, HLA-A2 is found in about 42 % of Europeans and about 25 % of West Africans; HLA-B35 is the most prevalent HLA class I type in West Africans and is found in 30 % of the population with H U B 5 3 found in about 25 % of the population. It can be calculated that over 70 % of both European and West African populations should have one of these selected HLA types. It is possible that a larger percentage of the population might respond to the epitope string as some peptides are likely to function as epitopes for some other (related) HLA types. Additionally, individuals without one of these HLA types would be expected to respond to an epitope in the P f f RAP antigen to which the epitope string is fused. A fourth criterion was to try to include multiple epitopes for selected common HLA types to assess whether CDS' T cell responses could be induced to multiple epitopes for a single HLA type in the same individual. Thus there are three HLAB35 epitopes and five HLA-A2 epitopes in the string (these latter encompasses two of the common subtypes of HLA-A2: A2.1 and A2.2). In principle, the induction of responses to multiple epitopes should reduce the probability of variant parasites being selected by vaccination. A further criterion was the assessment of the capacity of these epitopes to be successfully processed and presented from the epitope string. In practice this turned out not to be discriminatory in that all epitopes tested were processed effectively (Gilbert et al., 1997). A final consideration for CD8' T cell epitope selection was the degree of conservation of the epitope between strains. Fortunately, most identified epitopes are conserved (Aidoo et al., 1995) and with the exception of Cp2G those included in the epitope string are completely or substantially conserved. Six further peptide epitopes were included for the following reasons: first, the P. berghei epitope, pb9, was included to allow potency and efficacy testing of the

E

D

C

B

KPICDELDY

ICPIVQYDNF

Cp26

Ls6

CD8' T cell CD8' T cell

liver stage antigen 3 circumsporozoite protein P. berghei

IS KYEDEI

SYIPSAEKI

Ls50

Pb9

NANPNANPNA NPNANP

DEWSPCSVTCG ICGTRSRICRE

TRAP AM

LLMDCSGSI

Tr29

NANP

liver stage antigen 3

KSLYDEHI

Ls53

thrombospondin related adhesive protein

circumsporozoite protein

thrombospondin related adhesive protein

thronibospondin related adhesive protein

H LGNVKYLV

Tr26

B58 A2.2

CD8' T cell CD8' T cell

heparin binding motif

B cell

A2.1

CD8' T cell

mouse H2-Kd

B17

A2.2

CD8' T cell

sporozoite threonine- and asparagine-rich protein

MINAYLDKL

B7

St8

circumsporozoite protein

CD8' T cell

MNPNDPNRNV

A2.1

CD8+ T cell

CPG

thrombospondin related adhesive protein

GIAGGLALL

Tr39

B8

CD8' T cell

BS3

CD8' T cell

ASICNKEKALII

B35

B35

HLA Restriction

CD8' T cell

CD8+ T cell

Type

Tr42/43

thrombospondin related adhesive protein

liver stage antigen 1

circumsporozoite protein

liver stage antigen 1

ICPNDICSLY

Ls8

A

Antigen Derived from

Amino Acid Sequence

Epitope

CD8+ T cell epitopes of the malaria multiepitope (ME) string

Cassette

Table 1.

%

v+.

m

s.%

: : z

d

h

H

QFIKANSRFIGI TE

TT

T helper

T helper

M . tuberculosis 3XKd antigen

QVHFQPLPPAV VICL

BCG tetanus toxoid

T helper

circumsporozoite protein

DPNANPNVDPN ANPNV

CSP

universal CD4 epitopes

B58

CDX+ T cell

exported protein 1

ATSVLAGL

ex23

88

CD8+ T cell

HLA Restriction

sporozoite threonine- and asparagine-rich protein

MEKLICELEK

La72

TVpe

A2.1

circumsporozoite protein

Antigen Derived from

CD8' T cell

YLNKIQNSL

cp39

F

Amino Acid Sequence

Epitope

(continued)

Cassette

Table 1.

I

110

I DNA vaccine in mice. HLA transgenic mice were considered for potency assays, Jorg Schneider et a/.

but the different T cell receptors of mice and humans and differences in the interaction of mouse /1’2-microglobulin and HLA heavy chain make this model of vaccine immunogenicity difficult. Two B cell (antibody) epitopes were also included in the epitope string. One is the well-studied major B cell epitope (of sequence NANPNANPNANP) from the circumsporozoite protein of P. falciparurn. Antibodies to this epitope may opsonize sporozoites before entry into hepatocytes and thus induced antibodies to this epitope might be protective. Similarly, antibodies against the TRAP AM antibody epitope present on the surface of sporozoites might impair the entry of these into hepatocytes. Finally, three CD4 T cell or helper T cell epitopes were included in the epitope string. It has been shown that CD4 T cells can augment the levels of CD8’ T cells induced by vaccination (Bennett et al., 1997). Therefore, for induction of CD4 T cell mediated help several CD4 T cell epitopes are included in this string. Thus, we included three such epitopes. One (BCG) is from the 38 Kd mycobacterial antigen present in Mycobacteviuwl tuberculosis and M. bovis. This is an epitope for several HLA-DR types (Vordermeier et al., 1992). The second (TI)is the well-studied epitope from tetanus toxoid (amino acids 830-843) with an additional substitution of phenylalanine for tyrosine at the second position to broaden further the range of HLA class I1 types to which it may bind (Panina-Bordignon et al., 1989). The third epitope (CSP) is a helper Tcell epitope from a conserved segment of the circumsporozoite protein of P.falciparurn (Nardin et al., 1989). Most individuals will have memory T cell responses to one or more of these CD4 T cell epitopes through infection or vaccination. TRAP is an abundant liver stage specific antigen. The plasmid map of pSG.MEPflRAP is shown in Figure 2. The ME epitope string was constructed by fusing different cassettes, each containing three epitopes. Each cassette consists of the epitopes shown in Table 1, in the order shown, with no additional sequence between epitopes within a cassette. Synthetic oligonucleotides were used to generate the individual cassettes. The annealed oligonucleotides were ligated into the pUC-based cloning vector pIC2OR between the BglII and BamHl sites. After seqencing, cassettes were ligated as required. The junction formed by ligating the BamHl site at the 3’ end of one cassette to the BglII site at the 5’ end of the next destroys both restriction sites; facilitating the addition of further cassettes to either end. Between cassettes in an epitope string the BamHI/Bglll junction encodes for glycine and serine. This design of individual cassettes allowed sequential assembly in different orders. All epitopes are from P.fulciparurn antigens except for pb9 (P. berghei), BCG ( M . tuberculosis) and TT (tetanus) (Gilbert et al., 1997). The full-length sequence of PfTRAP was cloned from the P.falciparum strain T9/96. In a final step, the epitope string was fused to the Pff RAP sequence. Human volunteers immunized with irradiated sporozoites and protected against malaria develop T cell responses against TRAP making it a strong candidate for inclusion in a malaria vaccine.

6 pSC.MEPfTRAP

-A

First Generation Malaria D N A Vaccine Vector

n

P.falciparum TRAP Fig. 2.

Schematic diagram of pSC.MEPfTRAP.

3.3 Production and formulation

The plasmid construct pSG.MEPffRAP was made under controlled conditions in the laboratories of Oxford University. Following initial characterization (restriction enzyme digest, sequencing of the insert and potency testing in mice) the material was sent to a contract manufacturer (Qiagen, Germany) for the production of clinical material under cGMP (Schleef et al., 1999) which is a general requirement for such material (Schleef et al., 2000; Chapter 11 of this book). In a first step pSG.MEPflRAP was transduced into a number of different E. coli strains to determine the optimal strain for production of the clinical batch. Following strain selection a research cell bank and a master cell bank were established. The clinical batch was purified using ion exchange chromatography. The specification for the plasmid pSG.MEPffRAP DNA is given in Table 2. The specification is based on tests conducted by Qiagen prior to despatch of the bulk manufactured pDNA and the identity and potency assays. One aim of the Phase I study will be the comparison of intramuscular vs. intradermal gene gun delivery of pSG.MEPfTRAP. For intramuscular administration this batch was formulated in PBS and for intradermal delivery using a “gene gun” into a buffer suitable for further formulation onto gold beads.

I ”’

112

I

Jorg Schneider et a/

Specification for clinical grade pSC.MEPtTRAP plasmid DNA (see also Chapters 11 a n d 12)

Table 2.

Test

Method

Specification

Appearance

visual inspection

clear, colorless solution

Identity: Restriction map

restriction enzyme digestion/ agarose gel electrophoresis

conforms to reference pattern

DNA homogeneity

agarose gel electrophoresis of undigested DNA

2 90%

OD2G0

0.9-1.1 mg mL-'

Host (E. coli) DNA

southern blot hybridization analysis HPLC

< 5%

RNA

BCA colorimetric assay

none detected

0D260/0D280

4 5%

Purity: DNA concentration

Protein

I

ccc

scan 220-320 n m Other impurities (salts/organics)

Limulus amoebocyte lysate (LAL) kinetic-chromogenic assay

1.80-1.95 Peak at 260 nm

Endotoxin

Ph Eur

< IOOEU per mg Sterile

Sterility Identity DNA sequence of insert

mdeotide sequence determination

conforms to consensus sequence ~

Potency: pb9-specific CTL response in mouse

4 x BALB/c mice immunized at days 0 and 14 with 2 x 25 pg pDNA; immunogenicity of pb9 epitope evaluated at day 28 using ELISPOT assay

~~~

~~~~~

mean number of spot-forming cells (SFC)/lOGsplenocytes in ELISPOT assay. 15; 1 non-responding mouse of 4 tested.

3.4

Preclinical testing of pSG.MEPfTRAP

Prior to clinical testing of pSG.MEPflRAP in healthy volunteers the plasmid has to be tested for toxicity, biodistribution, stability and potency. This section describes the studies for the intramuscular administration route. Similar studies will be performed for the intradermal route of administration (gene gun).

6 pSG.MEPfTRAP

3.4.1

-

A First Generation Malaria D N A Vaccine Vector

Toxicity studies

The objective of this study was to test the toxicity of DNA vaccine plasmid pSG.MEPffRAP in BALB/c mice following two intramuscular (i. m.) DNA injections given two weeks apart. DNA injected animals (males and females) received a total of 100 pg pSG.MEPflRAP. They were compared to groups receiving PBS only. The following investigations were carried out: monitoring of weight during the in-life phase of the study and any gross adverse effects were monitored. Two weeks after the last immunization standard hematology on EDTA blood and liver enzymes (alanine transaminase and alkaline phosphatase in serum samples) were determined. The hematological parameters tested were: red blood cell counts, hemoglobin, hematocrit, platelets, mean cell volume, mean cell hematocrit, mean cell hemoglobin concentration and white blood cell counts. The following organs were analyzed using standard histopathology: injection site, liver, spleen, kidneys, heart, gonads and draining lymph nodes. No treatment-related signs were observed in gross pathology, hematology and liver enzymes, Histological analysis of the injection site revealed expected treatment-related changes such as minimal or mild infiltration with lymphocytes. The greater incidence in plasmid DNA sites than PBS at the injection site was expected due to the induction of cellular immune responses following expression of the malaria antigen.

3.4.2

Biodistribution

The objective of this study was to monitor persistence and distribution of the DNA vaccine pSG.MEPfTRAP in vivo after a single i.m. injection. Distribution of the DNA vaccine on day 4 after administration was monitored by PCR analysis of genomic DNA extracted from various organs: liver, spleen, draining lymph nodes, gonads and injection site (m. tibialis). The study was repeated on day 30 to monitor persistence of the DNA. The presence of pSG.MEPfTRAP in these DNA samples was detected using PCR primers binding specifically to the TRAP part of the plasmid. The detection limit of the method was 5 femtograms of pS G .M EPflRAP. Four days after i. m. injection pSG.MEPflRAP could only be detected in the injection site and in draining lymph nodes of DNA-injected animals. No DNA vaccine plasmid was detected in the other organs and in samples from PBS-injected animals. The results of the 32 day study show persistence of pSG2.MEPfTRAP at the site of injection and in draining lymph nodes in some animals. Using similar methods equivalent results were found in studies testing gene gun delivery of pSG.MEPfTRAP.

3.4.3 Stability testing

Research-grade quality plasmid pSG.MEPffRAP was tested for stability using the following assays: agarose gel analysis of restricted and unrestricted plasmid, determining concentration using UV spectrophotometry and potency €or antigen specific CD8' T cell induction in BALB/c mice. In our stability studies pSG.MEPffRAP

I

113

114

I

Jorg Schneider et af.

stored at -20" no apparent degradation (as detected by gel electrophoresis) or loss of potency was observed for more than 20 months. Pilot and cGMP-grade clinical material of pSG.MEPflRAP was tested in parallel using the same assays and no loss of potency was observed over a period of 15 months.

3.4.4

Potency testing

The objective of this study was to monitor the ability of pSG.MEPffRAP to induce antigen specific CD8' T cell responses. To determine potency of plasmid preparations for CD8+ T cell induction a murine reporter epitope was introduced in the ME polyepitope string. BALB/c mice were immunized using one i.m. injection or three intradermal (gene gun) administrations. Fourteen days after the last immunization the number of peptide specific interferon-y secreting CD8' T cells in spleens was determined using ELISPOT assays (Schneider et al., 1998).

4

Regulatory aspects

In order to carry out Phase I studies in the UK, approval will have to be obtained by various regulatory bodies. The studies will address many questions for the first time in healthy human volunteers in the UIC. This will be the first DNA vaccination trial in healthy volunteers in the UIC. It will also be the first study of DNA vaccines using a synthetic polyepitope string. Manufacturing procedures, preclinical testing and the clinical protocol of the study will be reviewed by the Medicines Control Agency. Two routes of review will be used: intramuscular injection of pSG.MEPfTRAP will be submitted as an investigator-initiatedstudy under a Doctors and Dentists Clinical Trial Exemption Certificate (DDX) and the gene gun protocol will be submitted under a general Clinical Trial Exemption Certificate (CTX). The Local Ethics Committee at Oxford University will review the ethical aspects of administering DNA vaccines to healthy individuals. Discussions and/or applications were held not just with the MCA and local ethics committees, but also with the UIC Health and Safety Executive, the Gene Therapy Advisory Committee (GTAC) secretariat and the Medical Devices Agency.

5

Future perspectives

The primary endpoint of the outlined studies with pSG.MEPffFL4P will be to demonstrate safety. A secondary endpoint will be to determine the irnmunogenicity of DNA vaccines for T cell induction in humans. Results from earlier studies showed that DNA vaccines in primates including humans are not as immunogenic as demonstrated in rodent models (Fuller et al., 1997; Wang et al., 1998). This is

G pSG. M E T R A P - A First Generation Malaria D N A Vaccine Vector

now widely recognized and various strategies are employed to improve the immunogenicity of DNA vaccines in humans. Two current strategies are heterologous prime-boost immunization strategies and co-administration of antigen-expressing plasmid DNA with cytokine-expressing plasmids. In the heterologous primeboost strategy an immune response induced by a plasmid DNA vaccine i s boosted with a different antigen delivery system such as proteins, peptides or viral vector expressing the same antigen (Letvin et al., 1997; Schneider et al., 1999).We are currently preparing to test heterologous prime-boost strategies using poxviruses in Phase I studies in malaria (Schneider et al., 1998).Studies in humans in HIV, melanoma and hepatitis B infection will follow. Co-administration of DNA vaccines with cytokine-expressingplasmid DNA shows some encouraging results in animal models (Kim et al., 1999; Kim et al., 2000). However, this strategy has to face additional safety concerns of introducing highly bioactive cytokines with homologous sequences into humans. A further aspect of DNA vaccination particularly aimed at diseases endemic in developing countries is cost per dose. The amount of plasmid DNA currently used in intramuscular administrations is in the milligram range (Wang et al., 1998; Le et al., 2000). Technologies delivering smaller amounts of DNA such as the gene gun require additional formulation processes contributing to increased costs. The ability to upscale these processes will determine the price of DNA vaccines in the future.

References and HLA-B supertype alleles, Immunity 7 (l), AIDOO,M., LALVANI,A., ALLSOPP, C. E. M., 97-112. PLEBANSKI, M., MEISNER,S. J. et al. (1995), D. H., CORB,M.M., BARNEIT, S., Identification of conserved antigenic compo- FULLER, STEIMER, K., HAYNES, J. R. (1997), Enhancenents for a cytotoxic T lymphocyte-inducing ment of immunodeficiency vims-specific vaccine against malaria, Lancet 345 (8956), immune responses in DNA-immunized 1003-1007. S. R., CARBONE, F. R., I ~ R A M A LF., I S , rhesus macaques, Vaccine 1 5 (8), 924-926. BENNETT, S. C., PLEBANSICI, M., HARRIS, S. J., MILLER, J. F., HEATH,W R. (1997), Induction GILBERT, ALLSOPP, C. E., THOMAS, R. et al. (1997), A of a CD8' cytotoxic T lymphocyte response protein particle vaccine containing multiple by cross-priming requires cognate CD4' T malaria epitopes, Nature BiotechnoL 15 (12), cell help, J. Exp. Med. 186 ( l ) ,65-70. BOJANG, K.A., OBARO, S. I<., D'ALESSANDRO, 1280-1284. T., SCHNEIDER, J., GILBERT, S. C., HILL, U., BENNETT, S., LANGEROCIC,P. et al. (1998), HANICE, A. (1998), DNA multiAn efficacy trial of the malaria vaccine SPf66 A. V., MCMICHAEL, in Gambian infants - second year of follow- CTL epitope vaccines for HIV and Plasmodium fakiparum: immunogenicity in mice, up, Vaccine 16 (l),62-G7. S. L., SOUTHWOOD, Vaccine 1G (4), 426-35. DOOLAN, D.L., HOFFMAN, S., WENTWORTH, P.A., SIDNEY,J. et al. (1997), HILL,A.V, ELVIN, J., WILLIS,A. C., AIDOO,M., ALLSOPP, C. E. et al. (1992), Molecular Degenerate cytotoxic T cell epitopes from analysis of the association of HLA-B53 and P. fakiparum restricted by multiple HLA-A

116

I

Jorg Schneider et a/. resistance to severe malaria, Nature 360 (6403), 434-439. JACIGON, J. (1998), Cognition and the global Malaria Eradication Programme, Parasitologia 40 (1-2), 193-216. IGSLOW, D. C. (1997), Transmission-blocking vaccines: uses and current status of development, Int. /. Parasitol. 27 (2), 183-189. KIM, J. J., NOITINGHAM,L. I<., TSAI,A., LEE, D. J., MAGUIRE, H. C. et al. (1999), Antigenspecific humoral and cellular immune responses can be modulated in rhesus macaques through the use of IFN-gamma, IL-12, or IL-18 gene adjuvants, J . Med. Primatol. 28 (4-5),214-223. KIM, J. J., YANG,J. S., VANCOIT,T. C., LEE,D. J., MANSON, K. H. et al. (ZOOO), Modulation of antigen-specific humoral responses in rhesus macaques by using cytoltine cDNAs as DNA vaccine adjuvants, /. Virol. 74 (7), 3427-3429. LALVANI,A,, MORIS,P., Voss, G., PATHAN, A.A., KESTER I<.E. et al. (1999), Potent induction of focused Thl-type cellular and humoral immune responses by RTS,S/SBAS2, a recombinant Plasmodium falciparum malaria vaccine, J. In&. Dis. 180 (5), 1656-1664. LE, T. P., COONAN, I<.M., HEDSTROM, R. C., Y., SEDEGAH, M. et al. (2000), CHAROENVIT, Safety, tolerability and humoral immune responses after intramuscular administration of a malaria DNA vaccine to healthy adult volunteers, Vaccine 18 (18), 1893-1901. LETVIN, N. L., MONTEFIORI,D. c., YASUTOMI, Y., PERRY, H. C., DAVIES, M. E. et al. (1997), Potent, protective anti-HIV immune responses generated by bimodal HIV envelope DNA plus protein vaccination, Proc. Natl. Acad. sci. U S A 94 (171, 9378.9383. N A R D I N , E. H., HERRINGTON, D.A., DAVIS, J., LEVINE,M., STUBER, D. et al. (1989), Conserved repetitive epitope recognized by CD4' clones from a malaria-immunized volunteer, Science 24G (4937), 1603-1606. NUSSENZWEIG, R. S., LONG, C.A. (1994), Malaria vaccines: multiple targets, Science 265 (5177): 1381-1383. PANINA-BORDIGNON, P., TAN.A,, TERMIJTELEN. A., DEMOTZ. S., CORRADIN. G. LANZAVECCHIA, A. (1989). Universally immunogenic T cell epitopes: promiscuous binding to human MHC class I1 and promiscuous recognition by T cells, Eur. /. Immunol. 19 (12), 2237-2242.

ROBSON,I<. J., HALL,J. R.,J E N N I N G S , M. W., HARRIS, T. J., MARSH,I<. et al. (1988), A highly conserved amino-acid sequence in thrombospondin, properdin and in proteins from sporozoites and blood stages of a human malaria parasite, Nature 335 (6185), 79-82. ROBSON,I<. J. H., HALL, J. R. S., DAVIES, L. C., T. E. CRISANTI, A,, HILL,A. V S., WELLEMS, (1990), Polymorphism of the TRAP gene of Plasmodium falciparum. Proc. Royal SOC. London, Series B, Biological Sciences 242 (1305), 205 -216. ROMERO,P., MARYANSIZI, J. L., CORRADIN, G., NUSSENZWEIG, R. S., NUSSENZWEIG, V., ZAVALA, F. (1989), Cloned cytotoxic T cells recognize an epitope in the circumsporozoite protein and protect against malaria, Nature 341 (6240): 323-326. SCHLEEF, M. (1999), Issues of large-scale plasmid DNA manufacturing, in: Biotechnology 2nd Edn. Vol. 5b: Recombinant Proteins, Monoclonal Antibodies and Therapeutic Genes (Rehm, H.-J., Reed, G., Piihler, A., Stadler, P., Eds.) pp. 443-469. Wiley-VCH, Weinheim. SCHLEEF, M., SCHMIDT, T., FLASCHEL, E. (ZOOO), Plasmid D N A for Pharmaceutical Applications. Development and Clinical Progress of D N A Vaccines (Brown, F., Cichutek, I<., Robertson, J., Eds.), pp. 25-31. Karger. Basel. SCHNEIDER,J., GILBERT, S. C., BLANCHARD, T., ROBSON,I<. J. et al. (1998), T. J., HANKE, Enhanced immunogenicity for CD8' T cell induction and complete protective efficacy of malaria DNA vaccination by boosting with modified vaccinia virus Ankara, Nature Med. 4 (4), 397-402. SCHNEIDER, J., GILBERT, S. C., HANNAN, C. M., DEGANO, P., PRIEUR, E. et al. (1999), Induction of CD8' T cells using heterologous prime-boost immunisation strategies, I m m u nol. Rev. 170, 29-38. SPITALNY, G. L., NUSSENZWEIG, R. s. (1973), Plasmodium berghei: relationship between protective immunity and anti-sporozoite (CSP) antibody in mice, Exp. Parasitol. 33 (l), 168-178. M., HEPPNER, D.G., STOUTE, J.A., SLAOUI, M O M I N , P., KESTER, K. E. et al. (1997), A preliminary evaluation of a recombinant circumsporozoite protein vaccine against Plasmodium f a k i p a r u m malaria. RTS,S Malaria Vaccine Evaluation Group, N. Engl. J. Med. 336 (2), 86-91.

G pSG.MEPfTRAP - A First Generation Malaria D N A Vaccine Vector I 1 1 7 VORDERMETER, H. M., HARRIS, D.P., MEHRO- WAWG,R., DOOLAN,D.L., LE, T. P., HEDSTROM, TRA, P. I<., ROMAN, E., ELSAGHIER, A. et al. R.C., COONAN, I<. M.et al. (1998),Induction (1992),M. tuberculosis.- complex specific of antigen-specific cytotoxic T lymphocytes in T-cell stimulation and DTH reactions humans by a malaria DNA vaccine, Science induced with a peptide from the 38-kDa 282 (5388), 476-480. protein, Scan$. 1.Immunol. 35 ( 6 ) , 711-718.

P/asmids for Therapy and kchation by M.Schleef Copyright 0 WILEY-VCH Verlag GmbH, 2001

7 Polyvalent Vectors for Coexpression of Multiple Genes Peter P. Muller, Andri. Oumard, Dagmar Wirth, Andrea Kroger and Hansjorg Hauser"

1

Introduction

In all living organisms gene expression is highly coordinated. There are many examples showing co-regulation of two or more genes in the same cells. In the initial phase of genetic manipulation of cells and animals coordinated expression was not of importance. However, current applications in mammalian cell culture, but also in transgenic animal biology and biotechnology increasingly require the defined co-expression of different genes. Co-expression offers new perspectives in DNA vaccination and gene therapy. Four typical applications are described as follows: 1. Co-expression of a selectable marker together with the protein of interest This is routinely used to establish stable cell lines. For industrial applications overexpression of a recombinant protein in Chinese hamster ovary cells (CHO) is achieved by gene amplification. To obtain such cell lines an amplification marker (mostly the dihydrofolate reductase gene) is co-transfected with the gene of interest. High-level expressing cell clones are successively selected for increasing resistance to the amplification drug (Methotrexate).This is often due to the amplification of a chromosomal domain containing the amplification marker and the gene of interest. The expression level of both heterologous genes increases with the domain copy number present in the genome. 2. Defined, but unequal expression Some applications require that the protein of interest is expressed at a much higher level than the selectable marker. If a defined relationship of co-expression of both genes can be adjusted, cells resistant to the respective selective drug will produce the protein of interest at levels which cannot be below a critical threshold required for survival in the presence of the respective drug concentrations. Also, for DNA immunization adjusting the ratio of co-expression of an antigen and a co-stimulatoryprotein would be of advantage to maximize the immunological effects.

120

I

Peter P. Miiller et al.

3 . Equivalent expression, i. e. expression of genes at a one to one ratio

Subunits of heteromultimeric protein complexes need to be synthesized in equal amounts. Examples are cytolcine receptors, antibodies or other di- or trimeric proteins. An additional application refers to the co-expression at similar amounts of several enzymes forming a metabolic pathway. 4. Quantitation of a protein for which no satisfactory assay method is available If strict coupling of expression of a reporter gene and the gene of interest is achieved, a calibration for reporter product level can serve to indirectly determine the quantity of the protein of interest. Prokaryotes co-express different genes by synthesizing multicistronic mRNAs. However, eulcaryotic genes are usually monocistronic. How then is coupled expression naturally achieved in the mammalian kingdom? It appears that evolution has evolved multiple levels of regulatory mechanisms to ensure stoichiometric expression of genes in mammalian cells. Often these genes are even located on different chromosomes. Various levels of expression control including transcription, posttranscriptional processing, mRNA transport, stability and translational efficiency adjust the correct levels of the synthesized proteins. Simulation of these events for genetic manipulation of cells would imply coordinated engineering efforts of all crucial steps. With the current techniques, this type of co-expression is nearly impossible to achieve. Thus, alternative methods were developed. The currently applied methods for co-expression are summarized in Figure 1. A straightforward way to obtain co-expression of two proteins is to transfect cells with two independent constructs (Figure 1A) or by introducing a single vector harboring two discrete expression cassettes (Figure 1B). The first approach is often limited by the inefficiency and unpredictability of co-transfection. The second approach requires the construction of complex vectors and suffers from unpredictable mutual influences of promoters. An elegant method relies on bidirectional promoters (Figure 1C). Although this method is currently not in common use it opens a new level of co-expression technology. These three methods suffer from the fact that even if the same promoter strength is given, the two transcripts might significantly differ due to variations in processing, half life time and translational effciency of the mRNA and, therefore, the amount of protein expressed is not predictable. Di- and polycistronic mRNAs can be constructed by using internal ribosomal entry sites (IRES) allowing the co-expression from a single mRNA (Figure 1D). In this way variable expression ratios due to unpredictable transcription efficiencies of separate expression cassettes are circumvented. Finally, for special applications such as tagging with fluorescent proteins or antigenic sites, fusion proteins provide the strictest coupling of two protein functions (Figure 1E). However, for various reasons this often is not possible. This chapter focuses on the application of IRES elements and bidirectional promoters for the construction and use of polyvalent protein expression. Since in our opinion they represent the most promising approach to achieve co-expression, background information is provided to permit the application of these principles for solving individual tasks. Then, properties of naturally occurring IRES elements

7 Polyvalent Vectors for Coexpressron of Multiple Genes

Fig. 1. Strategies for construction of polyvalent vectors. Cassettes for the expression o f two different genes (open and striped large boxes) under the control o f unidirectional or bidirectional promoters (small boxes) are shown. Arrows symbolize transcription starts, filled circles depict polyadenylation sites. The filled oval symbolizes an IRES element. Vector sequences are drawn as black lines.

A

-,g--

+

-r,e--~

t

B

c D E

and bidirectional promoters are compared, followed by practical applications. Finally, restrictions, drawbacks and hints to overcome potential problems are outlined.

2

Polycistronic expression vectors 2.1

Mechanisms of translation initiation

One of the most complex reactions in eukaryotic cells is the initiation of protein synthesis. It is the process leading to the establishment of a translation competent complex at the initiation codon on the mRNA. Generally, the rate-limiting step in translation is the initiation process. For the rational design of polyvalent expression vectors it is important to understand some basic factors and mRNA features that influence translation efficiency and how various modes of translation initiation can be employed for optimization of specific applications. Initiation complex assembly can be divided into distinct steps. In the resting state, ribosomes accumulate as 80s particles that are not associated with mRNA (Figure 2). The inactive 80s ribosomes spontaneously dissociate into GOS and 40s subunits. Ribosomal subunit re-associationis then prevented by binding of eukaryotic initiation factors (eIFs) to the individual ribosomal subunits. In addition to translation initiation factors, charged initiator tRNA binds to the small ribosomal subunit as a ternary complex with eIF2 and GTP. This ternary complex is essential for the recognition of the initiation codon on the mRNA. There are two distinct ways by which initiation complexes associate with mRNA: a cap-dependent initiation pathway and internal binding. All cellular cytoplasmic mRNAs are modified at the 5' end by the addition of a 7-methylguanosine group termed cap structure (Shatkin et al., 1976). The vast majority of these

I

lZ1

122

I

Peter P. Miiller et a/.

otnet

Fig. 2. Model for cap-dependent translation. After dissociation o f 805 ribosomes into the 405 and 605 subunits (lower part ofthe figure), initiation factors (3) and initiator elF2-GTP -tRNA (met) ternary complex join the small ribosomal subunit. Binding ofthe mRNA (middle part of the figure) a free mRNA 5' end (left) devoid of strong secondary structures and is facilitated by the cap structure (black dot). The ribosome then migrates towards the 3' end of

the mRNA until the first AUG codon is encountered that is recognized as initiation codon of the encoded reading frame (box) by virtue of its complementarity to the anticodon o f the initiator tRNA. Subsequently, elF-2 bound GTP is hydrolyzed, initiation factors are released and the large ribosomal subunit joins to form a translating 805 ribosome (center). At the termination codon the ribosome dissociates from the mRNA (right).

mRNAs are translated by a cap-dependent mechanism. The cap-binding complex consists of three proteins, eIF4G (Keiper et al., 1999), the RNA helicase eIF4A and the cap-binding protein eIF4E (McKendrick et al., 1999). This complex is responsible for the preferred translation of capped mRNA (Henis-Korenblit et al., 2000; Sonenberg and Pelletier, 1989; Svitkin et al., 1999). A model for cap-dependent initiation proposes that initiation complexes bind first to the mRNA 5' end and then migrate on the mRNA in the 3' direction until the first AUG codon is encountered and recognized as an initiation codon (Kozak, 1987). The movement of the ribosome on the mRNA in search of an initiation codon has been termed scanning. The ribosomal scanning is hindered by strong mRNA secondary structures. Therefore, 5' mRNA non-translated leader regions in expression constructs should be devoid of secondary structure and free of AUG codons. Some natural mRNAs are efficiently expressed despite AUG codons upstream of the initiation codon of the main translated reading frame. Therefore, the basic scanning model was modified to accommodate these exceptional mRNAs (Kozak, 1989, 1999).The modifications include sequence context-dependent recog-

7 Polyvalent Vectors for Coexpression of Multiple Genes

nition of initiation codons (“leaky scanning”). The optimal context for efficient initiation was determined to be (GCC)GCCA/GCCAUGG (Kozalt, 1987). Such efficiently used translation initiation sites are of special interest for high-level expression vectors for recombinant protein production. In addition, ribosomes were found capable of translating two separate consecutive reading frames on a single mRNA, either artificial or even on some naturally occurring viral mRNAs (Horvath et al., 1990). However, downstream reading frames are expressed inefficiently, unless the upstream reading frames were short and terminated 5’ of the main reading frame. Translation of the downstream reading frame is thought to occur by a fraction of re-initiating ribosomes that remain mRNA bound after translating the upstream reading frame. The low level of re-initiation of eukaryotic ribosomes has been used to express selection marker genes at low ratios relative to recombinant upstream reading frames to permit the stringent selection of efficient producer clones (Davies and Kaufman 1992; Kaufman et al., 1987) and isolation of competent retroviral vector packaging cells (Cosset et al., 1995). In the above examples ribosomes bind first to the mRNA 5‘ end. The modified scanning model does not explain the efficient translation of uncapped viral mRNAs with inhibitory secondary structures at the 5’ end that reduce or prevent ribosome binding and migration. Sequences that mediate internal translation initiation (internal ribosomal binding sites, IRES) of such mRNAs independent of their 5‘ end were discovered in picornavirus mRNA (Pelletier and Sonenberg, 1988; Jang et al., 1988). A few years later the first of an ever increasing number of IRES elements in cellular mRNAs was detected in the rnRNA encoding immunoglobulin heavy-chain binding protein BiP (Macejak and Sarnow, 1991). Cellular IRES elements have been suggested to play a role in maintaining translation under certain stress conditions (Altiri et al., 1998; Stein et al., 1998; Svitkin et al., 1999). A number of RNA binding proteins has been demonstrated to support IRES-dependent translation when added to translation extracts in vitro or in vivo when such proteins are overexpressed. Even though these proteins stimulate IRES-mediated translation in some experimental setups, it is not yet clear, whether these proteins are really indispensable for internal initiation in viva Presently, factors supporting IRES-dependent translation initiation are still poorly defined in a physiological context. Whereas re-initiation of ribosomes is generally inefficient, IRES-mediated internal initiation can be as efficient or in rare cases even more efficient than capdependent initiation. Therefore, IRES elements are the method of choice to coexpress multiple reading frames (Martinez-Salas,1999). However, IRES-mediated translation efficiency is highly context-dependent, requiring experimental confirmation of the expression characteristics of each individual vector construct (Attal et al., 1999).

I

123

Peter P. Miiller et a/ L.L

Characteristics o f IRES elements

Internal ribosomal entry sites (IRES) have been identified in viral and cellular eulcaryotic mRNAs. All currently well defined IRES elements are present in the 5‘ untranslated regions (5’ UTR) of mRNAs and range from about 200 to 1,300 nucleotides. They contain multiple non-initiating AUG codons as well as distinct secondary structures. A comparison of the 5’ UTR primary sequences of unrelated origin with an IRES function revealed that between the different IRES elements no conserved sequence homology exists. Therefore, besides the IRES hallmarks of being present in unusually long mRNA 5’ regions containing multiple upstream AUGs, IRES elements cannot be identified by sequence homology. However, a conserved organization of structural elements with loops and Y structures at similar positions in picornaviral and several cellular IRES elements (e. g., BiP and EGF-2) could be identified by computer analysis (Figure 3 ) . Potential IRES elements are generally verified by inserting them between two cistrons in an expression vector and determination whether they stimulate the expression of the second cistron. For applications in expression vectors the type of IRES element used and its efficiency are of major importance. The currently used classification of IRES elements is based on the picomaviral IRES classification (Jackson and Kaminski, 1995; Sachs et al., 1997) that relies on the position of the initiation codon relative to the IRES element and additional downstream sequence requirements (Table 1). Type I IRES elements include rhinoviruses (HRV) and enteroviruses, such as the polio virus IRES element (Pelletier and Sonenberg, 1988) which is regarded as the prototype of this group. In addition, all cellular IRES elements characterized so far are type I. Type I IRES elements can be located at a variable distance from the downstream reading frame and can be as far as 50-100 nucleotides upstream of the initiation codon. The process of translation initiation can therefore be regarded as a two-step process (Figure 4). The first step is the binding of the 40s ribosomal Fig. 3. Conserved secondary structures i n cellular and Picornaviral IRES elements. Simplified drawing o f common computer predicted secondary structural elements of cellular (A) and Picornaviral (6) mRNAs (adapted from Le et al. 1997). The double stranded structures a, b, c and d correspond t o e.g., h and i, respectively. In addition, a potential pseudo loop structure h and a stem loop j can be formed by sequences preceding the initiation codon in Picornaviral mRNA.

7 Polyvalent Vectors for Coexpression of Multiple Genes Table 1.

Classes o f IRES elements Type I

Type I I

IRES elements in their Y-UTRs

rhinoviridae, enteroviridae, cellular IRES elements

cardioviridae

Prototype

polio virus

EMCV

HCV

Requirements for transfation initiation

scanning to the next downstream AUG

initiation at an AUG codon precisely positioned at the 3’ e n d of the IRES element

binding site includes coding region of the downstream reading frame

codon

Type 111

Hepatitis C virus

I

subunit to the IRES, whereby AUG sequences present within the IRES sequence do not serve as initiation codons. The second step i s the migration of the ribosome until the next AUG codon is encountered that serves as initiation codon. Type I1 IRES elements are characterized by the strict position requirements for the initiation codon at the 3’ boundary of the IRES (Pestova et al., 1996). Type I1 IRES elements naturally occur in the 5’ region of the aphthovirus (FMDV) and cardiovirus mRNA. Encephalomyocarditis virus (EMCV; Jang et al., 1988) i s regarded as a prototype for class I1 elements. Translation initiation i s thought to occur by direct binding of the 40s ribosomal subunit precisely at the initiation codon (Figure 4). In type I11 IRES elements the initiation codon position requirements are similar to type I1 IRES. However, additional sequences important for IRES function are located in the downstream coding sequence. The prototype of type I11 IRES is represented by the hepatitis C virus (HCV) IRES. The sequences that contribute to the HCV IRES function extend 28 nucleotides into the HCV polyprotein coding sequence (Reynolds et al., 1995). Fig. 4. Model for translation of dicistronic vectors with type I and type II IRES elements. The first cistron is translated in a cap-dependent manner and the ribosomes are released at the end o f the reading frame (left) In type I IRES dependent initiation ( I ) , the initiation complex scans the mRNA downstream o f the IRES element (dashed arrow) for an AUC initiation codon Type II IRES elements (2) do not permit scanning and require an initiation codon precisely positioned at the end o f the IRES element

I

.,..

I

125

126

I

Peter

P. Mu//er et

a/.

2.3

Application o f IRES elements in cells and animals

Due to the uncritical distance between IRES sequence and AUG, type I IRES elements like the polio virus element are frequently used for the construction of polycistronic expression vectors. These IRES elements permit the use of polylinlters to facilitate the insertion of cloned fragments, but their efficiency varies greatly and their function depends on the cell line (Borman et al., 1995).Type I IRES elements have successfully been used in murine vaccination models (Wild et al., 1998). Intramuscular injection of dicistronic plasmid DNA into mice elicited polyvalent humoral and cytotoxic T lymphocyte responses to HBsAg and HBcAg. However, such applications are still in the initial phase and presently there is little systematic and quantitative expression data available. Because of the frequently poor function of type I IRES elements in several cell lines the EMCV IRES element as prototype for type I1 IRES elements has been the most successful in higher translation efficiency in a variety of cultured cell types as well as in whole animals (Borman et al., 1997). However, due to the strict sequence requirement for the initiation codon, the reading frame must be adapted. Generally, a suitable restriction site with an in-frame initiation codon is inserted making use of the polymerase chain reaction (PCR) methods. This type of IRES element functions in a wide spectrum of cell types. On the other hand, type 111 IRES elements are rarely used for heterologous expression because of their inefficiency of initiation and their need to integrate IRES sequences into the downstream reading frame. The spectrum of cell types in which the so far known type I11 IRES elements function is extremely small. Cellular IRES elements do not require exact positioning of the initiation codon and can lead to translation rates as efficient as type I1 IRES elements. In one case, a highly efficient cellular type I IRES element has been identified that shows a higher relative activity in various cell lines than picornaviral IRES elements (Oumard et al., 2000). Cellular elements were discovered much later than the viral IRES elements (Table 2), and presumably for this reason they currently are not used in such a broad range of applications as the viral elements. The use of different IRES elements facilitates the construction of multicistronic vectors with more than two reading frames by avoiding the potentially genetically unstable repeated insertion of identical IRES sequences. In therapeutic applications safety considerations of cellular IRES elements may be to avoid viral sequences that could lead to recombination events with naturally occurring viruses. An interesting aspect of certain cellular IRES elements is their regulatory potential. The IRES element of the vascular endothelial growth factor (VEGF) mRNA is induced by hypoxia and hypoglycemia (Akiri et al., 1998). However, so far no application for IRES-mediated regulation of gene expression has been published. In addition to viral and cellular sequences artificial IRES elements have been developed (Havenga et al. 1998).A 9 nt segment of a cellular mRNA functions as an IRES and when present in linked multiple copies efficiently directed internal initiation (Chappell et al., 2000).

7 Polyvalent Vectorsfor Coexpression of Multiple Genes Table 2.

Cellular IRES elements

Celfufar IRES Nements

Species

Refireme

BiP

human

Macejak and Samow, 1991

FGF-2

human

Vagner et al., 1995

IGF-I1

human

Teerink et al., 1995

eIF4G

human

Gan and Rhoads, 1996

Antennapedia

Drosophila

Ye et al., 1997

Ultrabithorax

Drosophila

Ye et a]., 1997

PDGF-2/c-sis

human

Bernstein et al., 1997

c-myc

human

Nanbru et al., 1997

Kvl.4

mouse

Negulescu et al., 1998

VEGF

human

Akiri et al., 1998

X-linked Kallmann

human

De Zoysa et al., 1998

MYT2

human

Kim et al., 1998

XIAP

mouse

Holcik et al., 1999

Connexin 43

rat

Schiavi et al., 1999

AML-l/RUNXl

human

Pozner et al., 2000

GtX

mouse

Chappel et al., 2000

NRF

human

Oumard et al., 2000

DAP-5

human

Henis-Korenblit et al., 2000

The use of IRES elements for co-expression in in vitro assays is also of high interest. In general, the type I viral IRES elements function poorly in a standard cellfree system (rabbit reticulocyte lysate). Their translational efficiency can be stimulated by cell extracts or by the addition of certain RNA-binding proteins such as the La protein (Meerovitchet al., 1993; Belsham and Sonenberg, 1996). In addition, the reports so far available do not indicate for strong activity of cellular type I IRES elements in this system. Further, the type 111 hepatitis A virus IRES element is inefficient in directing internal translation in vitro. In contrast, type I1 IRES elements function efficiently in rabbit reticulocyte lysate systems.

I

127

128

I

Peter P. Mii//er et a/.

2.4 Polycistronic vector systems

Most available dicistronic expression cassettes contain a genetic marker, usually a selectable marlier or a reporter gene 3' of an IRES element and a multiple cloning site (MCS) for the insertion of the gene of interest. These cassettes are transcriptionally initiated by a promoterlenhancer for cap-dependent expression of the first cistron and also provide a polyadenylation site (Kaufmann et al., 1991; Gaines and Wojchowski, 1999, Rees et al., 199G). The insertion of genes into such vectors is straightforward and most helpful for many applications. However, it is difficult to replace the second cistron by other genes, if necessary. Dirks et al. (1993) developed a vector system for the construction of dicistronic expression vectors based on two monocistronic vectors, pSBC-1 and pSBC-2. This system has the advantage of free combination of diverse cistrons. In addition, the use of two monocistronic vectors gives the possibility to test the expression of the individual constructs in a cap-dependent manner. In this vector system the first gene is inserted into the MCS of pSBC-1, which is 5' from a polio virus IRES element. The second gene is inserted in the MCS of pSBC-2. The two vectors are fused to a dicistronic expression unit. This is achieved by using the same unique 8 bp restriction site downstream of the IRES element of pSBC-1 and upstream of the MCS of pSBC-2 and another unique restriction site in the ampicillin gene of the vector backbone. Translation of mRNA encoding reading frames of both genes occurs upon capdependent initiation of the first cistron and translation re-initiation of the second cistron mediated by a polio virus IRES element. For this vector system a family of different promoters and vector backbones for multiple use in mammalian cells in vim and in vitro experiments has been created (Dirks et al., 1994). Malting use of IRES elements tri- and polycistronic expression cassettes can be constructed. Some examples for tricistronic expression have been published (Zitvogelet al., 1994; Fussenegger et al., 1998a; Miellte et al., in press). Fussenegger et al. (199813) developed a system for cloning of tricistronic expression plasmids (pTRIDENT; Figure 5A). Further, systematic cloning of polycistronic expression vectors is achieved by the pCI-system (Schirmbeck et al., 1999) (Figure 5B). The pTRIDENT-based tricistronic expression plasmid (Figure 5A) (Fussenegger et al., 199%) is a one-vector system and allows the consecutive cloning of multiple genes. Seven different sites for restriction endonucleases recognizing 8 bp and additional sites for other endonucleases in the three MCSs are available for insertion of the genes of interest. Different promoters and IRES elements from polio virus and EMCV can be combined in this system. Certainly, endonuclease recognition sites in the reading frames of genes inserted in the vector in previous steps that are also contained in the MCS can limit the addition of further genes. The pCI-system (Figure 5B) (Schirmbeck et al., 1999) for polycistronic expression is based on the subsequent transfer of cassettes containing a gene of interest and a downstream IRES element into an expression vector downstream of the first gene of interest. The details are outlined in Figure 5B. The expression of the genes

7 Polyvalent Vectorsfor Coexpression of Multiple Genes

A

pTRlDENT

-,

-

%
MCSl

Not1

Fig. 5. Tri- and polycistronic expression systems. Schematic presentation o f the pTRlDENT and pCI-vector system, promoters (striped boxes), arrows indicate transcription starts, filled circles depict polyadenylation sites, filled oval symbolizes IRES elements and vector sequences are drawn as black lines. (A) pTRIDENT: The tricistronic plasmid was constructed by inserting the Gene of interests in the different multicloning sites MCSI, MCSll and MCSIII.

(6)The pCI-1 plasmid contains a unique Notl-site downstream o f an IRES element. Into this site, the Eagl fragment from pCI-2, containing the gene o f interest followed by another IRES element, was inserted. This eliminates the upstream Not1 site, but leaves the downstream Not1 site intact which can be used for further cloning steps. This strategy can therefore used t o construct tri- or polycistronic plasmids.

from monocistronic cassettes prior to the construction of polycistronic vectors has some advantages: Each expression cassette can be cloned independently of restriction sites existing in other genes of the polycistronic mRNA. The expression cassettes can be tested as a monocistronic expression unit before cloning into the polycistronic expression construct. The expression cassettes can be simply cloned in various combinations. The construction of polycistronic expression vectors is not limited to two or three cistrons.

I

129

130

I

Peter P. M u k r et a/.

2.5 Expression properties o f IRES vectors

Compared to monocistronic mRNAs, artificial dicistronic mRNAs are considerably longer and more complex. To achieve optimal expression, care has to be taken to avoid the unintentional inclusion of special regulatory sequences such as polyadenylation signals or mRNA destabilizing elements (reviewedby Ross, 1995; Day and Tuite, 1998). The expression level depends not only on the particular IRES element used, but also on both the sequence and on the order of the reading frames. A systematic study of dicistronic vectors expressing two unrelated luciferase genes from Renilla and firefly, respectively, showed that the presence of the firefly luciferase in the first cistron has a surprisingly drastic inhibitory effect on internal initiation. This effect is independent of the promoter and IRES element used and is observed in several cell lines (Henneclte et al., unpublished results). Highly efficient IRES elements such as the NRF IRES could advantageously be utilized not only for efficient expression in multicistronic vectors, but also as translational enhancers to increase expression levels from monocistronic mRNAs (Oumard et al., 2000). Overall, the efficiency of IRES elements can vary considerably depending on the particular construct, on the experimental setup or on the host cell used, and in some cases depending even on the physiological status of the cells. Despite unquestionable advantages of multicistronic expression vectors, each individual construct must be tested, and if necessary expression levels must be optimized to achieve satisfactory results.

3 Bidirectional promoters

3.1 Natural bidirectional promoters

An alternative strategy to achieve coordinate expression of two transcription units can be obtained by using bidirectional promoters. While for viral bidirectional promoters (e.g., SV40 and adenovirus) expression of the divergent transcription units is temporarily controlled in the course of infection (Gidoni et al., 1985; Natarajan et al., 1984),mammalian bidirectional promoters allow simultaneous transcription in both directions. Most of the known mammalian bidirectional promoters are TATA box-deficient and mediate low level transcription of housekeeping genes (Table 3). Generally, they are asymmetric and one direction is transcribed preferentially. Despite of the striking bidirectional activity of some promoters in an artificial context the form and functional significance of the second message in vivo has not been provided in all cases (Table 3). While the sharing of promoter elements suggests coordinate regulation of bidirectionally transcribed genes, this cannot be generally as-

7 Polyvalent Vectorsfor Coexpression of Muftipk Genes Table 3.

Examples for bidirectional promoters found in natural genes

Reference

First Direction

Second Direction

Doyen et al., 1989

IgH IIIB

not known

Schilling and Farnham, 1989

DHFR

Rep-1

Johnson and Friedman, 1990

PGK

not known

Johnson and Friedman, 1990

HPRT

not lcnown

Weichselbaum et al., 1990

thyrnidine ltinase

not known

Lennard and Fried, 1991

Surf1

Surf2

Takami et al., 1992

H3-111

H3-I1

Schmidt et al., 1993

Col4A1

Co14A2

Liao et al., 1994

thymidylate synthetase

not known

Wright et al., 1995

TAP1

LMP2

Orii et al., 1999

HADHA

HADHB

I

sumed: Examples for bidirectional promoters resulting in coordinate expression (Wright et al., 1995) and discordant expression (Taltami et Nakayama, 1992; Schilling and Farnham, 1995; Schmidt et al., 1993) are reported. However, none of the natural bidirectional promoters have been developed for practical applications in recombinant gene expression. This might be due to the fact that the expression levels obtained with these promoters are usually lower than those obtained by standard promoters such as the viral CMV or SV40 promoters. 3.2 Artificial bidirectional promoters

Generally, an in vitro design of bidirectional promoters with high and coordinate expression of two mRNA seems feasible. The general potential of bidirectional promoters for simultaneous expression of two genes has recently been shown for the artificial promoter (Baron et al., 1995).This promoter is derived from a unidirectional tetracycline (Tc) promoter (Gossen and Bujard, 1992). The promoter is regulated by a tetracycline-repressibletransactivator (tTA).The tTA dependent promoter system is commercially available (Tet-OnTM/Tet-OffTM; Clontech) and is widely applied in both research and gene therapeutic approaches. Practical application of the tTA system requires some effort, This is due to the fact that the characteristics of recombinant gene expression from regulated promoters to a large degree depends on the chromosomal integration locus, frequently

I

131

132

I resulting in an undesired basal expression level. In addition, the toxicity of the tTA Peter P. Miiller et a\.

transactivator imposes an upper expression limit (Sadowski et al., 1988). Both limitations can be compensated by optimizing the selection and screening process (Icirchhoff et al., 1995; Klehr-Wirth et al., 1997). Further options of this system include a reverse tTA, mediating activation of expression in presence of Tc (Gossen et al., 1995), transactivators with reduced toxicity (Baron et al, 1997) and established transactivator expressing cell lines. While the promoter was designed to permit strict regulation, it can be used even for high-level constitutive expression. Furthermore, since its activity does not require any cell-specific factors it is expected that it is active in virtually all tissues. Since the bidirectional promoter Pbl.l is symmetric, a balanced transcription level of the divergent mRNAs is usually achieved. Using the bidirectional promoter Pb,., in cell lines constitutively expressing the tTA a strictly regulated, coordinate expression of two reporter genes has been shown (Baron et al, 1995; Li et al, 1998). Applications include co-expression of a reporter gene and a gene of interest (Nishizawa et al., 1998; Senner et al, 1999; Keyvani et al., 1999; Kwissa et al., in press), expression of the two protein kinase I1 subunits (Vilk et al, 1999), wild type and mutants of p53 (Aurelio et al., 2000). Finally, the bidirectional promoter was successfully used in a DNA vaccination model to simultaneously express core and surface antigens of hepatitis B virus (Kwissa et al., in press). However, although the bidirectional promoter Pbl.l principally allows a balanced expression of two genes, an equimolar expression will be hard to achieve. One further development of the tTA system is the establishment of autoregulatory expression systems in which both, the gene of interest and the transactivator, are under the control of the tTA dependent promoter. This was realized either on separate transcription units (Shockett et al., 1995) or in a dicistronic transcription unit (Hofmann et al., 1996). Low basal levels of transactivator expression allows self-amplified activation upon withdrawal of Tc. An advantage of autoregulation is that the tTA-expression and - in consequence - also the toxic side effects can be restricted to the time of induction. Furthermore, the autoregulatory cassettes can be expressed in a tissue-independent manner (Shockett et al., 1995). Due to the necessary minimal basal expression level in the repressed state, autoregulatory expression units cannot be applied for expression of highly toxic proteins which are lethal, even if expressed in minimal amounts (e. g., diphteria toxin). However, the low basal expression level is of no concern for most applications (Keyvaniet al., 1999). Autoregulatory expression cassettes based on modified bidirectional tTA dependent promoters have been established (A-Mohammadiand Hawltins,l998; Strathdee et al., 1999). 3.3 Combining polycistronic and bidirectional expression

The potential of the bidirectional promoter can be substantially extended, if it is combined with polycistronic expression cassettes. Although simple cloning vectors are not available yet, the advantages of such a vector design are obvious: two coor-

7 Polyvalent Vectorsfor Coexpression of Multiple Genes

dinately transcribed mRNAs encoding several cistrons can be achieved. Thereby, problems or limitations arising from I RE S-mediated translation of certain genes can be bypassed. Certainly, this approach can also be extended to achieve autoregulatory expression.

4

Perspectives

The polyvalent methodology is of great importance. However, the individual components cannot be used like building bloclts for vector construction with guaranteed success. This is due to the complexity of mammalian gene expression and a number of properties and interactions of the individual components which are not understood in sufficient detail, such as the sensitivity of IRES function to flanking cistrons. Therefore, a reliable prediction of the expression levels is currently not possible. Similarly, the de nouo composition of different cistrons on one mRNA does not guarantee a certain efficiency of expression. Posttranscriptional RNA processing, mRNA half-life time and translational initiation are influenced by the RNA configuration. A better understanding of the mechanisms of posttranscriptional events and translational initiation will be needed to permit a directed design of polyvalent expression. Currently there are no alternatives to polyvalent expression methods in mammalian cells. Nevertheless, homologous recombination into a set of defined chromosomal sites or genes could be used to achieve coordinated expression. Transgenic mouse models might help to pioneer such approaches. Applications of IRES elements and bidirectional promoters as outlined in this article are not restricted to naked DNA vectors. The described expression cassettes may be used to construct viral vectors and by this gene transfer to cells or organisms can be obtained with much higher efficiency. In fact, IRES elements and bidirectional promoters are already used in retroviral and adenoviral vectors.

References AIURI, G., NAHARI,D., FINKELSTEIN, Y., ATTAL,J., THERON,M.C., HOUDEBINE,L.M. LE, S.Y., ELROY-STEIN,O., LEVI,B.-2. (1998), (1999),The optimal use of IRES (internal Regulation of vascular endothelial growth ribosome entry site) in expression vectors, factor (VEGF) expression is mediated by Genet Anal. 15, 161-165. internal initiation of translation and alterna- AURELIO,0.N., KONG, X.T., GUPTA, S., tive initiation of transcription, Oncogene 17, STANBRIDGE,E. J. (2000), p53 mutants have 227-236. selective dominant-negative effects on apopA-MOHAMMADI,S., HAWKINS, R. E. (1998), tosis but not growth arrest in human cancer Efficient transgene regulation from a single cell lines, Mol. Cell. Biol. 20, 770-778. tetracycline - controlled positive feedback BARON,U., FREUNDLIEB, S., GOSSEN,M., regulatory system, Gene Tner. 5, 76-84. BUJARD,H. (1995),Co-regulation of two gene

I

133

134

I

Peter P. Mu//er et a/. activities by tetracycline via a bidirectional promoter, Nucl. Acids Res. 23, 3605-3606. BARON,U., GOSSEN, M., BUJARD,H. (1997), Tetracycline - controlled transcription in eukaryotes: novel transactivators with graded transactivation potential, Nucl. Acids Res. 25, 2723-2729. BELSHAM,G. J., SONENBERG, N. (l996), RNA-protein interactions in regulation of picornavirus RNA translation, Micro-obiol. Rev. 60, 499-511. BERNSTEIN,J., SELLA,O., LE, S.-Y., ELROYSTEIN,0. (1997), PDGF2/c-sis mRNA leader contains a differentiation linked internal ribosomal entry site (D-IRES),J. Biol. Chem. 272, 9356-9362. BORMAN,A.M., BAILLY,J. L., GIRARD, M., ICEAN,I<.M. (1995), Picornavirus internal ribosome entry segments: comparison of translation efficiency and the requirements for optimal internal initiation of translation in vitro, Nucleic Acids Res. 23, 3656-3663. BORMAN,A.M., LE MERCIERP.,GIRARD, M., ICEAN,I<.M. (1997), Comparison of picornaviral-IRES driven internal initiation of translation in cultured cells of different origins, Nucl. Acids Res. 25, 925-932. CHAPPELL, S.A., EDELMAN,G. M., MAURO, V: P. (2000), A 9-nt segment of a cellular mRNA can function as an internal iibosorne entry site (IRES) and when present in linked mu]tiple copies greatly enhances IRES activity, Proc. Natl. Acad. Sci. U S A 97, 1536-1541. DAVIES M.V., KAUFMANR. J. (1992). Internal translation initiation in the design of improved expression vectors, Cum Opin. Biotechnol 3, 512-517. DAY,D. A,, TUITE,M. F. (1998), Post-transcriptional gene regulatory mechanisms in eukaryotes: an overview, Endocrkdogy 157, 361-371. D E ZOYSA, P.A., HELLIWELL, R. J., DUKE,V.M., QUINTON, R., BOULOUX,P. M. (1998), Contrasting expression of I U L in cell-free systems: 5’ UTR and coding region stmctural effects on translation, Protein Expr. Pun$ 13, 235-242. H. (19931, DIRKS,W., WIRTH,M., HAUSER, Dicistronic units for gene expression in mammalian cells, Gene 128, 247-249. DIRKS,W., SCHAPER, F., KIRCHHOFF, S., C., HAUSER, H. (1994), A multiMORELLE, functional vector family for gene expression in mammalian cells, Gene 149, 387-388.

DOYEN, N., DREYFUS, M., ROUGEON,F. (1989), Regulatory elements involved in the bidirectional activity of an immuiloglobulin promoter, Nucl. Acids Res. 17, 1977-1987. FUSSENEGCER M., SCHLATEQS., DAETWYLER, D., MAZUR, X., BAILEY,1. E. (1998a),Controlled proliferation by multigene metabolic engineering enhances the productivity of Chinese hamster ovary cells, Nature Biotechnol. 16, 468-472. FUSSENEGGER, M., MAZUR, X., BAILEY J. E. (1998b),pTRIDENT, a novel vector family for tricistronic gene expression in mammalian cells, Biotechnol, Bioeng. 51, 1-10, GAINES, P., WOJCHOWSICI, D. M. (1999), pIRES-CD4t, a dicistronic expression vector for MACs- or FACS-based selection of transfected cells, Biotechniques 26, 683-688. CAN,W., RHOADS R. E. (1996), Internal initiation of translation directed by the 5‘ a n translated region of the mRNA for eIF4G, a factor involved in the picornavirus-induced switch from cap-dependent to internal initiation, J. Bid. Chem. 271, 623-626. GIDONI,D., KADONAGA, J. T., BARRERASALDANA, H., TAKAHASHI, K., CHAMBON, P., TJIAN,R. (1985), Bidirectional SV40 transcription mediated by tandem Spl. binding interactions, Science 230, 511-517. GOSSEN, M., BUJARD, H. (1992), Tight control of gene expression in mammalian cells by tetracycline-responsive promoters, Proc. Natl. Acad. Sci. U S A 89, 5547-5551. S., BENDER,G., GOSSEN, M., FREUNDLIEB, MULLER, G., HILLEN, W., BUJARD, H. (1995), Transcriptional activation by tetracyclines in mammalian cells, Science 268, 1766-1769. HAVENGA, M. J. E., VOGELS, R., BRAAKMAN,E., KROOS, N., VALERIO,D. et al. (1998), Second gene expression in bicistronic constructs using short synthetic intercistrons and viral IRES sequences, Gene 222, 319-327. S., STRUMPF, N. L., GOLDHENIS-KORENBLIT, STAUB, D., KIMCHI, A. (ZOOO), A novel form of DAP5 protein accumulates in apoptotic cells as a result of caspase cleavage and internal ribosome entry site-mediated translation, Mol. Cell. Bid. 20, 496-506. HOFMANN, A,, NOLAN,G.P., BLAU, H.M. (1996),Rapid retroviral delivery of tetracycline-inducible genes in a single autoregulatory cassette, Proc. Natl. Acad. Sci. U S A 93, 5181-5190.

7 Polyvalent Vectorsf o r Coexpression of Multiple Genes

HOLCIK, M., LEFEBRE,C., YEH, C., CHOW,T., KORNELUK,R. G. (1999), A new internalribosome-entry-site motif potentiates XIAPmediated cytoprotection, Nature Cell Biol. 1, 190-192. C. M., WILLIAMS, M.A., LAMB, R. A. HORVATH, (1990), Eultaiyotic coupled translation of tandem cistrons: identification of the influenza B virus BM2 polypeptide, E M B O ] . 9, 2639-2647. JACKSON, R. J., KAMINSKI,A. (1995), Internal initiation of translation in eultaryotes: the picornavims paradigm and beyond. R N A 1, 985 -1000. JANG, S. I<.,KRAUSSLICH, H. G., NICKLIN,M. J., DUKE,G.M., PALMENBERG, A.C., WIMMEQ E. (1988), A segment of the 5’ nontranslated region of encephalomyocarditis virus RNA directs internal entry of ribosomes during in vitro translation, J. Virol. G2, 2636-2643. J O H N S O N , P., FRIEDMANN, T. (1990), Limited bidirectional activity of two housekeeping gene promoters: human HPRT and PGK, Gene 88, 207-213. KAWFMAN,R. J., DAVIES,M. V., WASLEY, L. C., MICHNICK,D. (1991), Improved vectors for the stable expression of forein genes in mammalian cells by use of the untranslated leader sequence from EMCV vims, Nucl. Acids Res. 19, 4485-4490. KAUFMAN, R. J., MURTHA,P., DAVIES, M.V. (1987) Translational efficiency of polycistronic mRNAs and their utilization to express heterologous genes in mammalian cells. E M B O ] G, 187-193. KEIPER, B.D, GAN,W., RHOADS,R. E. (1999), Protein synthesis initiation factor 4G, Int. J. Biochem. Cell Biol. 1, 37-41. KEYVANI, K., BAUR, I., PAULWS, W. (1999), Tetracycline-controlledexpression but not toxicity of an attenuated diphtheria toxin mutant, LFe Sci. 64, 1719-1724. KIM, J. G., ARMSTRONG,R. C., BERNDT,J.A., KIM,N. W., HUDSON, L. D. (1998), A secreted DNA-binding protein that is translated through an internal ribosome entry site (IRES)and distributed in a discrete pattern in the central nervous system, Mol. Cell Neurosci. 12, 119-140. KIRCHHOFF,S . , K ~ S T E RM., , WIRTH,M., F., GOSSEN, M. et aI. (1995), IdenSCHAPER, tification of mammalian cell clones exhibiting highly regulated expression from inducible promoters, Trends Genet. 11, 219-220.

I

KLEHR-WIRTH, D., KWHNERT, F., UNSINGER, J., HAWSER H. (1997), Generation of mammalian cells with conditinal expression of cre recombinase, Trends Genet. T40067. KOZAK, M. (l987), At least six nucleotides preceding the AUG initiator codon enhance translation in mammalian cells, /. Mol. Bid. 196, 947-950. KOZAK, M. (1989), The scanning model for translation: an update, ].Cell Bid. 108, 229-241. KOZAK, M. (1999), Initiation of translation in proltaryotes and eukaryotes, Gene 234, 187-208. M., UNSINGER, J., SCHIRMBECK, R., KWISSA, HAUSER, H., REIMANN,J. 2000, Polyvalent DNA vaccines with bidirectional promoters,. /. Mol. Med. 78, 495-506. LE, S.-Y. A N D MAIZEL,J R , J.V. (1997), A common RNA structural motif involved in the internal initiation of translation of cellular mRNAs, Nucl. Acids Res. 25, 858-866. LENNARD,A,, GASTON, I<., FRIED,M. (1994), The Surf-1 and Surf-2 genes and their essential bidirectional promoter elements are conserved between mouse and human DNA, Cell B d . 13, 1117-1126. LI, X., NGO, N., Hou, C., CUNNINGHAM, S., ZHANG,R. B. et al. (19983, Screening for positive transfected clones with coexpressed green fluorescent protein, Biotechniques 24, 52, 54-55. LIAO, WC., ASH,J., J O H N S O N , L. F. (1994), Bidirectional promoter of the mouse thymidylate synthase gene, Nucl. Acids Res. 22, 4044-4049. P. (1991), Internal MACEJAIC, D. G., SARNOW, initiation of translation mediated by the 5’ leader of a cellular mRNA, Nature. 353, 90-94. MARTINEZ-SALAS, E. (19991, Internal ribosome entry site biology and its use in expression vectors, Curr. Opin. Biotechnol. 10, 458-464. MCKENDRICIZ,L., PAIN,V.M., MORLEY,S. J. (1999),Translation initiation factor 4E, Int. J. Biochem. Cell Bid. 1, 31-35. MEEROVITCH, I<., SVITICIN, Y. V., LEE, H. S., LEJBKOWICZ,F., KENAN, D. J. et al. (1993), La autoantigen enhances and corrects aberrant translation of poliovirus RNA in reticulocyte lysate, ]. Virol. 7, 3798-807. C., TUMMLER, M., SCHUBELER,D., VON MIELKE, HOEGEN, I., HAUSER, H. 2000, Stabilized, long-term expression of heteromeric proteins from tricistronic mRNA. Gene 254, 1-8.

135

136

I

Peter P. M i i h

et a1

NAORA,H. (19991, Involvement of ribosomal proteins in regulating cell growth and apoptosis: translational modulation or recmitment for extraribosomal activity? Immunol. Cell Bid. 77, 197-205. NATARAJAN, V., MADDEN, M. J , , SALZMAN, N. P. (1984), Proximal and distal domains that control in uitro transcription of the adenovirus IVa2 gene, Proc. Natl. Acad. Sci. USA. 81, 6290-6294. NEGULESCU, D., LEONG,L. E.-C, CHANDY, I<. G., B. L., GUTMAN,G. A. (1998),TransSEMLER, lation initiation of cardiac voltage-gated POtassium channel by internal ribosome entry, /. Bid. Chem. 273, 20109-20113. NISHIZAWA, I<., FREUND, C., LI, J., WAGNER,G., REINHERZ, E. L. (1998), Identification of a proline-binding motif regulating CD2-triggered T lymphocyte activation, Proc. Natl. Acad. Sci. USA. 95, 14897-14902. ORII, I<. E., ORII, K. O., SOURI,M., ORII, T., KONDO, N. et al. (19993, Genes for the human mitochondrial trifunctional protein alphaand beta-subunits are divergently transcribed from a common promoter region, J. B i d Chem. 274, 8077-8084. OUMARD, A., HENNECICE, M., HAUSER,H., NOURBAKHSH, M. (ZOOO), Translation of NRF mRNA is mediated by highly efficient internal ribosome entry, MoZ. Cell. Bid. 20, 2755-2759. PELLETIER, J., SONENBERG, N. (1988), Internal initiation of translation of eukaryotic mRNA directed by a sequence derived from poliovirus RNA, Nature 334, 320-325. PESTOVA, T.V., HELLEN,C. U.T., SHATSKY,1. N. (1996), Canonical eultaryotic initiation factors determine initiation of translation by internal ribosome entry, Mol. Cell. Bid. 16,6859-6869. POZNEQA., GOLDENBERG, D., NEGREANU, V., LE, S., ELROY-STEIN, 0. et al. (ZOOO),Tram scription-coupled translation control of AMLl/RUNXl is mediated by cap- and internal ribosome entry site-dependent mechanisms, Mol. Cell. Biol. 20, 2297-2307. REES,S., COOTE,J,, STABLES, J., GOODSON,S., HARRIS, S., LEE, M. G. (1996), Bicistronic vector for the creation of stable mammalian cell lines that predisposes all antibiotics-resistant cells to express recombinant protein, BioTechniques 20, 48-56. REYNOLDS, J. E., KAMINSKI, A.,. KETTINEN,H.J, GRACE,K., CLARICE, B. E. et al. (1995), Unique

features of internal initiation of translation of hepatitis C virus translation, EMBO J. 14, 6010-6020. ROSS, J. (1995),mRNA stability in mammalian cells, Microbiol. Rev. 59, 423-450. SACHS,A. B., SARNOW, P., HENTZE,M.W (l997), Starting at the beginning, middle, and end: translation initiation in eultaryotes, Cell 89, 831-838. SADOWSKI, I., MA, J., TRIEZENBERG, S., PTASHNE, M. (1988), GAL4-VP16 is an unusually potent transcriptional activator, Nature 335, 563-564. SCHIAVI,A., HUDDER,A., WERNER, R. (1999), Connexin43 mRNA contains a functional internal ribosome entry site, FEBS Lett. 464, 118-122. SCHILLING, L. J.. FARNHAM, P. J. (1995),The bidirectionally transcribed dihydrofolate reductase and rep-3a promoters are growth regulated by distinct mechanisms, Cell Growth Dz$ G, 541-548. SCHILLING, L. J., FARNHAM,P. J. (1994), Transcriptional regulation of the dihydrofolate reductase/rep-3 locus, Crit. Rev. Eukaryot. Gene Expr. 4, 19-53. SCHIRMBECK, R., V O N IGZMPEN,J., METZGER, K., WILD,J., GRUNER,B. et a. (1999), DNAbased vaccination with polycistronic expression plasmids, Methods Mol. Med. 29, 313-322. SCHMIDT,C., FISCHER,G., KADNER,H., GENERSCH, E., KBHN, K., POSCHL,E. (1993), Differential effects of DNA-binding proteins on bidirectional transcription from the common promoter region of human collagen type JV genes COL4A1 and COL4A2, Biochim. Biophys. Acta 1174, 1-10. SENNER,V., STURM,A., BAUR,I., SCHRELL, U.H., DISTEL,L., PAULUS, W. (1999), CD24 promotes invasion of glioma cells in vivo, J. Neuropathol. Exp. Neurol. 58, 795-802. SHATI~IN, A. J. (1976), Capping of eucaryotic mRNAs. Cell. 9, 645-653. SHOCKETT,P. E., DIFILIPPANTONIO, M., HELLMAN, N., SCHATZ,D. G. (1995), A modified tetracycline-regulated system provides autoregulatory, inducible gene expression in cultured cells and transgenic mice, Proc. Natl. Acad. Sci. USA.92, 6522-6526. SONENBERG, N., PELLETIER, J. (1989), Poliovirus translation: a paradigm for a novel initiation mechanism, Bioessays 11, 128-132.

7 Polyvalent Vectors for Coexpression of Multiple Genes

VILIC,G., SAULNIER,R.B., ST.-PIERRE, R., STEIN,I., ITIN A,, EINAT, P., SKALITER,R., GROSSMAN, Z., KESHET E. (1998), Translation LITCHFIELD, D. W (1999), Inducible expression of protein kinase CK2 in mammalian of vascular endothelial growth factor mRNA cells. Evidence for functional specialization of by internal ribosome entry: implications for translation under hypoxia, Mol. Cell. Bid. 18, CI<2 isoforms. /. Bid. Chem. 274, 144063112-3119. 14414. B., METZGER,I<., KUHROBER, STRATHDEE, C.A., MCLEOD,M. R., HALL,J. R. WILD,J., GRUNER, A., PUDOLLEK, H. P. et al. (19981, Polyvalent (1999), Efficient control of tetracycline-revaccination against hepatitis B surface and sponsive gene expression from an autoregucore antigen using a dicistronic expression lated bi-directional expression vector, Gene plasmid, Vnccine 16, 353-360. 229, 21-29. I., OGRIS,E., WINTERSWEICHSELBRAUN, SVITKIN, Y V , GRADI, A,, IMATAIW, H., BERGER, E. (1990), Bidirectional promoter MORINO,S., SONENBERG, N. (1999), activity of the 5’ flanking region of the mouse Eulcaryotic initiation factor 4GII (eIF4GII), thymidine lcinase gene, FEBS Lett. 275, but not eIF4G1, cleavage correlates with in49-52. hibition of host cell protein synthesis after WRIGHT,K. L., WHITE,L. C., KELLY,A., human rhinovims infection, J. Virol. 73, BECK,S., TROWSDALE, J., TING,J. P. (1995), 3467.3472. Coordinate regulation of the human TAP1 TAKAMI, Y., NAKAYAMA, T. (1992), Presence of and LMP2 genes from a shared bidirectional distinct transcriptional regulatory elements promoter, J. Exp.Med. 181, 1459-1471. in the 5’-flanking region shared by the YE, X. P., FONG,P., IIZUKA,N., CHOATE, D., chicken H3 histone gene homopair, Nucl. CAVENER, D. R. (1997), Ultrabitlzorax and Acids Res. 20, 3037-3041. Antennapedia 5’ untranslated regions A. A. TEERINK, H., VOORMA,H. O., THOMAS, (1995),The human insulin-like growth factor promote developmentally regulated internal translation initiation, Mol. Cell. Biol. 17, I1 leader 1 contains an internal ribosomal 1714-1721. entry site, Biochim Biophys. Acta 1264, ZITVOGEL, L., TAHARA, H., CAI,Q., STORKUS, 403-408. VAGNER, S., GENSAC, M. C., MARET,A,, BAYARD, W, J., MULLER,G. et al. (1994), Construction and Characterization of retroviral vectors F., AMALRIC, F. et al. (1995), Alternative translation of human fibroblast growth factor expressing biologically active human interleulin-12, Hum. Gene Ther. 5 , 1493-1506. 2 mRNA occurs by internal entry of ribosomes, Mol. Cell. B i d . 15, 35-44.

I

137

P/asmids for Therapy and kchation by M.Schleef Copyright 0 WILEY-VCH Verlag GmbH, 2001

I139

8 Form Follows Function: The Design o f Minimalistic lmmunogenically Defined Gene Expression (MIDGE') Constructs Claas Junghans, Matthias Schroff, Sven A. Koenig-Merediz, Jens Alflten, Colin Smith, Florian Sack, Reinhold Schirmbeck and Burghardt Wittig"

1

The problem

Practitioners of gene transfer for medical purposes can chose today between virusderived transfer technologies and plasmid DNA. The issues governing the selection of one technology over the other will mainly be safety and efficacy. We propose a technology that combines the safety features of plasmid DNA with some of the mechanisms of viral transfer efficacy. Plasmid DNA of modern bacteria did not evolve to transfer genetic material for gene expression into the nuclei of mammalian cells. Man-made derivatives of plasmids were constructed and refined by molecular biologists for all kinds of genetic engineering. Their shape, size, sequence content, and structural dynamics do not fit, although, into the expression machinery of higher eultaryotes. In the latter, the nuclear processes of transcription and replication are separated and shielded from the cytoplasmic compartment where protein biosynthesis, processing, and degradation take place. Bacteria did arrive at specialized systems to exchange plasmids between their mating cells. No genuine equivalent of this process has been identified in mammalian cells - be it among the various populations of single cells, as in the hematopoietic system, or between the well-ordered cell assemblies of tissues and organs. From the physicochemical properties of DNA, no cell- or tissue-specific targeting can be envisaged in animal or human patients. Likewise, the absence of an active cellular uptake mechanism makes the diffusion of negatively charged plasmid DNA through the lipid bilayer of the mammalian cell membrane an unlikely event. Instead, local temporary disruptions of the fluid membrane mosaic will allow some molecules to sneak in. Here, smaller constructs will have an advantage. The usual size of plasmids recombined for expression of mammalian genes ranges from about 4,000 to 12,000 basepairs (bp). The viruses of mammalian cells, by impressive avenues of molecular evolution, adaptation, and mimicry, transfer nucleic acids fairly efficiently into their target cells to achieve all kinds of replication, recombination, amplification, and gene ex-

140

I

Claasjunghans et al.

pression. However, this process needs virally encoded proteins, which then interact, block, or make use of cellular proteins and other components of mammalian cells to accomplish transfer and expression. The viral proteins employed in this process are in turn recognized as “foreign”, “mutated”, or “expressed in wrong quantity” by the immune system of the host. Compared to plasmids, the genetic make-up of viruses is far more complex, leading in most cases to much longer nucleic acids, although cases of extreme compression of genetic information are known. With the exception of adeno-associated virus (AAV) and Simian virus 40 (SV40),both ofwhich are the size ofplasmids, most viral vectors recombined for mammalian gene expression range from about 30,000 to over 100,000 bp. Plasmids and viral vectors have both been used for preventive applications, such as in prophylactic DNA vaccines, or in curative approaches (see also Chapters 3 and S), as intended by gene therapy or therapeutic DNA vaccines. Depending on the peculiarities of vector design (see also Chapter 9) and construction, the questions of efficacy and safety dominate the discussions about the corresponding experiments. There will be no clinically successful and approved gene therapy or DNA vaccine, as long as the problems summarized and briefly commented in the following are not solved: 1. Many of the currently discussed targets and clinically tested applications require expression of one or several genes for hours, days, or several weeks. Efficacious transfer of constructs for transient expression into target cells and tissues will, therefore, suffice for a great variety of medical applications. Stable integration of the transgene into the genome of the host cell is not needed in the majority of applications to achieve sufficient expression levels. Today, unwanted accidental integration into the genome of the host cell as well as recombination and mobilization of host genetic material are regarded to be the most important delimiters of safety and regulatory approval. The persistence of vector DNA is indirectly related to the problem of integration. High expression levels attained with less vector DNA in a short time make unwanted genetic effects, - of which unwanted recombination with germline DNA would be the worst very unlikely. 2. Plasmid-encoded bacterial genes which are not needed for the desired prophylactic or therapeutic effect, but only required for propagation and production purposes, will also be wealdy expressed in mammalian cells. Many of these are strong antigens and will lead to elimination of the expressing cells, and could cause severe adverse immunological reactions upon repeated applications. Even weak antigens will almost certainly produce less well defined immunological complications, once large healthy populations are vaccinated and observed over longer times. In the context of mass vaccination campaigns, the spread of bacterial antibiotic resistance genes is an important public health issue in itself and will have to be avoided, even if the antibiotic is clinically not relevant. 3. Viral gene transfer by recombinants closely resembling the wild-type virus jointly aggravates problems 1 and 2. Removal of infection-promoting genes

8 Form Fo//ows Function

and of the traits responsible for the recombination and replication properties results in a trade-off between lower immunoreactivity and pathogenicity on one side, and exponential loss of gene transfer efficacy on the other. It is likely that most of the second and third generation viral vector strategies will end up in production schemes extremely difficult to manage pharmaceutically, and their resulting gene expression will come down to the efficacy of plasmids. 4. Size and shape (see also Chapter 2) matter in gene transfer and expression efficacy. Small, rod-shaped expression constructs will much easier surmount, escape, and pass cellular membranes, endosomes, and nuclear pores than wide, circular constructs. Integration into mammalian genomes probably requires DNA lengths of more than 30,000 bp. Thus, concatemerization of many small constructs would be required, but three of the longer expression plasmids or one “non-integrating”viral vector, like adenovirus, could be sufficient for unwanted integration. 5. Plasmids contain sequence motifs often referred to as CpG sequences. They stimulate the release of cytolcines typical for the response of organisms to bacterial infection. This represses transcription from transferred genes and activates the innate immune system. Although such effects can serve as important adjuvants in DNA vaccination schemes, they are to be avoided, if strong gene expression lasting longer than several hours is the therapeutic goal.

2 The solution

The problems described above define the goals to be reached by the design process: transient gene expression at high levels for several hours and at lower levels for several weeks, no other sequence content than the gene expression cassette for therapy or DNA vaccination, no viral genes or genome compromising regulatory sequences involved, much smaller than plasmid, rod-shaped, not supercoiled, no replication origins, unique, specific sites for chemical attachment of targeting signals, no immunomodulatory DNA sequence motifs present per se, but to be added, if desired. Since the ultimate goal of the development discussed here is a key technology for gene therapy and DNA vaccination, the solution will enable pharmaceutical production standards in quality (GMP), quantity, and storage stability (see also Chapter 11).Production has to be adaptable to a one-vessel scheme, where sequential reactions are accomplished by subsequent addition of new reactants, and with no intermediate purification step required.

142

I

CIaasJunghans e t a\.

2.1

MIDGE - the concept

The MIDGE acronym stands for “Minimalistic Immunogenically Defined Gene Expression”. The concept is a synthesis of the developments in vector safety and efficacy from the first laboratory to conduct somatic gene therapy in Germany. The MIDGE vector combines superior safety with the power of ligand-receptor interaction to enhance transfection and expression. MIDGE vectors are linear double-stranded DNA expression cassette molecules with stem-loop structures to protect against exonuclease degradation, and to provide unique, specific sites for targeting signals. As shown in Figure 1, they are made by excision of the expression sequence from plasmids by restriction endonuclease digestion. The resulting fragment of double-stranded DNA is then covalently closed at both ends by ligation of stem-loop-forming DNA oligonucleotides, and the remaining plasmid backbone degraded. This is a one-vessel, two-step process, which is followed by only a simple purification step to yield the MIDGE vector. Preliminary costing analyses have shown that the expense of post-plasmid modification is balanced by increased efficacy of transfer and a resulting lower dose per application. The entire process can be performed in a pilot-scale facility under GMP.

I

-1

/

/

2. Cut out expression cassette and seal it

with ODNs

lasmid backbone

Fig. 1. Production o f MIDGE vectors from plasmid DNA: The expression cassette in pMCVl.2 is situated between two cutting sites o f the rare cutting type I I endonucleases Eco31 I and Esp31. It can be excised by restriction endonuclease digest with either enzyme. The incompatible ends o f the resulting expression cassette fragment are subsequently ligated

t o hairpin oligonucleotides, resulting in covalently closed molecules. The remaining plasmid backbone does not ligate t o the hairpin oligonucleotide, and is digested by the 3‘->5’ exonuclease activity o f T7 DNA polymerase. This i s followed by a simple purification step t o yield the MIDGE@ vector.

8 Form Fo//ows Function

2.2

Simple MIDGE

MIDGE vectors without targeting signals (Simple MIDGE) are linear, not supercoiled, and significantly smaller than plasmid DNA. MIDGE vectors for peptide expression are well below 1,000 bp in size, MIDGE for expression of a typical cytokine is somewhere above 1,200 bp. They can be used for all in vitvo and in vivo gene transfer protocols as a safe and efficacious substitute for plasmid DNA. Simple MIDGE vectors have been used successfully for electroporation, ballistic transfer, microinjection, and lipid- or polymer-mediated gene transfer. Simple MIDGE vectors have been determined to have a transfection efficiency in terms of amount of transgenic protein expressed, of between 30% and 300% compared to the corresponding plasmid. Transfection efficiency seems to vary depending on the cell type and method of transfection in vitvo. For in vivo studies, Schirmbeclc et al. (unpublished results) found a somewhat reduced titer level after immunization with hepatitis B surface antigen expressing Simple MIDGE vectors compared to plasmid. 2.3 Smart MIDGE

One of the most important features of the MIDGE technology lies in the attachment of targeting signals to the nucleotides comprising the single-stranded loops at both ends of the molecule. We named the corresponding MIDGE family “Smart MIDGE”. Signals may be separated into two groups, the first of which targets MIDGE to tissues or specific cell types. Such signals could be sugar moieties to target liver cells or macrophages and dendritic cells, or lipids including steran derivatives for tissues expressing LDL receptors. Mostly, the ever growing population of peptide ligands will be used here to enable selective, high-affinitybinding to cell surface receptors that specifically mark almost every cell type. The second group comprises signals for Smart MIDGE which surmount cellular traps, such as endosomes, and utilize the cytoplasm-to-nucleustransport systems. Here the fusogenic peptides of viral origin and the nuclear localization peptides are typical examples. Most promising - and in certain aspects a unifying concept of both groups of signals - are the protein-transducing peptides, of which the TAT peptide derived from HIV is an eminent example. If it can be utilzed to carry Smart MIDGE through the cellular membrane and into the nucleus, a rich scenario of in vivo applications will become reality. Figure 2 explains the signal coupling reactions which are currently used to explore such ligands in cell culture as well as in animal experimentation: This work is currently ongoing; examples of our own experiments as well as links to providers of peptide-coupled oligonucleotides will be available on our website soon.

I

143

144

I

C/aasJunghans et a/.

CAGGTC

0

NH-ODN

SU Ifo-KMUS

GGGAGTCCAG CAGGTC

NH2-P

K V E N LS

K K K R

NH

GGGAGTCCAG CAGGTC

-

Fig. 2. Where indicated the nuclear location signal (NLS) peptide (PKKKRKVEDPYC) was coupled t o the hairpin O D N in two steps. First, the amino-modified oligonucleotide was activated with sulfo-KMUS (Pierce) i n phosphate-saline at room temperature. After 30 m i n the reaction was quenched with TRlS buffer, and the activated O D N was preciptated

D P Y

I C--aCOM

4

0

S

H*C,

N

with ethanol. The activated oligonucleotide was resuspended and reacted with peptide at room temperature for 1 h. The reaction was monitored by native electrophoresis through 3 % low melt agarose and ethidium bromide staining. The resulting NLS-coupled ODN was purified by HPLC and used t o generate the HBsAY-MIDGE-NLS construct.

8 Form Follows Function 2.4

Applications

We have extensively used the MIDGE vector in vaccination trials with a number of antigens. When discussing the contribution of the MIDGE vector system for the application of genetic vaccination or somatic gene transfer, a distinction has to be made with regard to the method of delivery. Injection of “naked DNA into muscle or skin leads to a cell-mediated immune response in a majority of cases. This appears to be at least partly due to the high amounts of DNA needed to achieve any response at all, and the adjuvant properties of bacterially derived DNA that help bring about such cellular responses. We were able to demonstrate (Schirmbeck et al., unpublished results) that peptides comprising a nuclear localization sequence from SV40 covalently coupled to MIDGE vectors lowered the amount of MIDGE vector necessary to achieve a given titer by more than one order of magnitude, compared to the “naked”MIDGE vector. Similar results have been obtained when injecting transgene expression constructs into muscle or connective tissue. A striking feature of this approach is that in our hands, the higher expression efficacy of peptide-coupled vectors was not paralleled by similar results in vitro. It appears that the efficacy of gene transfer in v i m is governed by a great number of yet unexplored parameters, and that in vitro models need to be adapted to a specific cell and tissue type and delivery situation, if one is to be able to draw meaningful conclusions for a specific problem from in vitro experiments. Ballistic transfer of polynucleotides achieves transfer of the expression constructs into the nucleus of antigen-presenting cells residing in the skin. The resulting immune response is predominantely Th2. We altered the immune response after ballistic transfer of feline immunodeficiency virus antigens to the cellular subtype by adding cytokine-encoding expression vectors and achieved protection from experimental homologous challenge. The antibody titer and cellular immunity attained by ballistic transfer of MIDGE vectors is not significantly different from plasmid immunization, and the rationale for using the minimalistic technology is basically the greater intrinsic safety and the regulatory concerns about excess sequence content and antibiotic resistance genes. 2.5

Practical aspects of vector sequence design

If the expression cassette that will form the MIDGE vector is to be excised from our standard vector pMCV1.2 using the restriction enzyme Eco311, Eco31I sites will not be tolerated by the process. If the gene to be inserted contains Eco31I recognition sites, these need to be removed. Our simple and effective protocol for the removal of such sites can be downloaded from www.midge.com/manuals/ Eco 3 1removal. Our own experience has shown that in the process of inserting a gene into pMCV1.2, it is possible to create, by accident, an unwanted ATG start codon,

I

145

146

I which might result in premature and possibly frame-shifted translation initiation, Claas junghans et a/.

and ultimately in the wrong protein being made. Please check your sequences for possible ATG sites prior to the first codon. For eukaryotic expression, you may want to ensure that your gene includes a Kozak consensus sequence upstream of the start codon. If you use PCR to amplify your insert you may want to include a Kozak sequence in your primer design. For more documentation on the Kozak sequence requirements, please refer to www.midge.com/manuals/kozak. Acknowledgement

The MIDGE' vector design and production process is covered under German patents DE 19753182 and DE19781276D and corresponding foreign patents. The CMV promoter sequence is made under U. S. Patent Nos. 5,168,062 and 5,385,839 licensed from the University of Iowa Research Foundation. MIDGE and Mologen are registered trademarks in Europe and the USA.

References Leutenegger, C.M., Boretti, F.S., Mislin, C., Boretti, F.S., Leutenegger, C.M., Mislin, C., Flynn, J.N., Schroff, M. et al. (2000) ImHofmann-Lehmann, R., Konig, S.A. et al. munization of cats against feline immuno(2000) Protection against FIV challenge deficiency virus (FIV) infection by using infection by genetic vaccination using minimalistic immunogenic defined gene minimalistic DNA constructs for FIV env expression vector vaccines expressing gene and feline IL-12 expression, AIDS FIV gp140 alone or with feline interleukin-12 18;14:1749-1757. (IL-l2), IL-16, or a CpG motif, 1 Virol. CBER (1996) Points to Consider on Plasmid D N A Vaccines for Preventa.tive Infectious 74:10447-10457. Disease Indications, Rockville, MD: Center for WHO (1997) Guidelines for Assuring the Quality of DNA Vaccines, WHO Technical Biologics Evaluation and Research, FDA. CBER (1998) Guidance for Industuy: Guidance Report, Geneva. for H u m a n Somatic Cell Therapy and Gene Wittig, B., et al. (in press) Therapeutic vacciTherapy, Rockville, MD: Center for Biologics nation against metastatic carcinoma by expression-modulated and immuno-modified Evaluation and Research, FDA. Commision of the European Communities autologous tumor cells: a first clinical phase 1/11 trial, Hum. Gene Ther. (1993) Gene Therapy Products - Quality Aspects in the Production of Vectors and Genetically Modified Somatic Cells.

P/asmids for Therapy and kchation by M.Schleef Copyright 0 WILEY-VCH Verlag GmbH, 2001

I

9 Synthetic Genes for Prevention and Therapy: Implications on Safety and Efficacy o f DNA Vaccines and Lentiviral Vectors Kurt Bieler and Ralf Wagner

1 Introduction

Both, design and generation of recombinant bioproducts for research and development critically depends on the availability of naturally occurring templates. Conventional strategies supporting a rational modification of these sequences such as site-directed mutagenesis or established cloning strategies are time-consuming and make strong demands on laboratory personnel. Furthermore, such strategies can be hardly put on automated platforms and therefore render the resulting products fairly expensive. Novel developments in the field of combinatorial biology such as scrambling of epitopes, shuffling of functional protein domains, construction of polygenes and the generation of combinatorial libraries are not sufficiently covered by the currently used technologies (Table 1).Novel approaches that aim towards avoiding biological limitations such as low RNA stabilities, inefficient nuclear export, antitermination or insufficient translational efficiency critically depend on technological innovation (Figure 1). Many of the above limitations may be solved by means of automated, solid-phase and high-throughput gene synthesis. De novo generation of synthetic genes solely relies on the sequence information that is usually available via publically accessible or commercial databases without depending on a physical template. De novo gene synthesis allows a precise adaptation of the desired product to the requirements of the final application without any need to consider technical limitations resulting from complex cloning strategies. In the past, hands on gene synthesis has been primarily used to increase translational activity by adapting the codon usage of the gene to be expressed to the t-RNA frequencies of the heterologous production system. A series of examples are given in Table 2. However, progress in the technology of gene synthesis opens a wide variety of potential applications in different fields of biotechnology. Accordingly, the choice of the appropriate technological platform for automated gene synthesis enables a hitherto almost unachieveable diversity in the area of combinatorial biology. Libraries of single-chain antibodies for, e. g., drug screening programs may be gen-

147

148

1

Kurt Biefer and Ray Wagner

Properties of Synthetic Genes

-

multifunctional immunogenic hypermutated

Fig. 1 . Power of synthetic genes ( I ) . Modulation of of gene expression at the level of transcription, RNA stability, nuclear translocation and translation.

erated by randomizing all three complementarity determining regions (CDRI-3) in a single synthesis by stepwise addition of synthetic and partially randomized oligonucleotides to the nascent gene. Moreover, novel strategies like generating polygenes or scrambling genes for DNA vaccination purposes or proteinaceous therapeutic bioproduct synthesis (Hanke et al., 1999) are greatly facilitated by targeted solid-phase gene synthesis (Schneider et al., 1998). Clinical trials using synthetic genes to fight tumor diseases such as malignant melanoma or to prevent infectious diseases such as malaria and AIDS are already in progress. Table 1. Advantages o f synthetic genes

-

template not required optimized expression yields by

-

increased RNA stability enhanced nuclear export avoiding anti-termination eliminating cryptic splice sites increased translation efficiency

epitope scrambling, domain shuffling combinatorial libraries

. safety issues (proteins) . safety issues (DNA vaccines/gene therapy vectors)

-

imodulate immunogenicity

- destroy unfavourable functions - generate toxoids - avoid unfavourable cis-acting sequences - (packaging sequence) - reduce homology to naturally occuring sequences - reduce "time to market''

9 Synthetic Genes for Prevention and Therapy

Table 2.

Synthetic genes in homologous and heterologous expression systems

Properties

Reference

Protein

Modification

Expression

Edible vaccine protects mice against Escherichia coli heat-labile entero-

1998

cok heat-labile enterotoxin B subunit (LT-B)

mized

tuberosurn L. (potato)

Wallis et al., 1997

antifreeze protein gene

synthetic gene

Solanurn tuberosurn L (potato)

toxin (LT): potatoes expressing a synthetic LT-B gene Expression o f a synthetic antifreeze protein in PO-

(AW

tat0 reduces electrolyte release at freezing temperatures Specific sequence modifications o f a cry3B endotoxin gene result in high levels of expression and insect resistance

Iannacone et al., 1997

Bacillus thuringefisis (Bt) gene Bt43 belonging to the cry3 class

partly modified

Solanurn me1ongena (eggplant)

Enhanced expression in tobacco of the gene encoding green fluorescent protein by modification of its codon usage

Rouwendal et al., 1997

green fluorescent protein (GFP) from jellyfish Aequorea victoria

plant optimized

transgenic tobacco lines

Transgenic elite indica rice plants expressing CiyIAc delta-endotoxin of Bacillus thuringiensis are resistant against yellow stem borer (Scirpophaga incertulas)

Nayak et al., 1997

cryIAc gene

partly plant optimized

indica rice

A recombinant ribosomeinactivating protein from the plant Phytolacca

Del Vecchio Blanco et al., 1998

ribosomeinactivating protein

synthetic gene

Phytolacca dioica L. (tomato)

Escherichia coli optimized

Escherichia coli

~~

~

dioica L. produced from a synthetic gene Gene synthesis by a LCRbased approach: Highlevel production of leptinL54 using synthetic gene in Escherichia coli

Au et al., 1998 recombinant leptin-154

I

I

149

750

I

Kurt Bider and Ray Wagner Table 2.

(continued)

Properties

Reference

Expression and characterization of soluble 1997 human erythropoietin receptor made in Streptowlyces lividans 66

Protein

Modification

Expression

domain of the human erythropoietin receptor (EPO-R)

cetes optimized

lividms

Expression of tetanus toxin fragment C in E. colt high-level expression by removing rare codons

Malcoff et al., 1989

tetanus toxin fragment c from Clostridium tetani

Escherichia coli optimized

Escherichia coli

A “humanized” green fluorescent protein cDNA adapted for highlevel expression in mammalian cells

Zolotukhin et al., 1996

green fluorescent protein (GFP) from jellyfish Aequorea victoria

mammalian cell optimized

human cell line 293, also in neurosensory cells of guinea pig eye

Codon optimization for high-level expression of human erythropoietin (EPO) in mammalian cells

Kim et al., 1997

human erythropoietin (EPO)

adaptation on highly expressed human and yeast genes

different mammalian cells

The consequent implementation of novel findings that are also discussed in this issue - such as the elimination of immunostimulatory DNA motifs in gene therapy vectors or the introduction of stimulatory CpG islets into DNA vaccine constructs (Krieg, 1996) - are greatly facilitated by means of automated gene synthesis. Also, as outlined in this chapter in more detail, profound insight into different aspects of regulated gene expression followed by rational gene design critically contributes to avoiding phenomena known to limit efficient gene expression (Figure 1).Simultaneously, crucial safety issues can be easily met by, e. g., avoiding the occurrence of packaging sequences in the case of vector design or removing toxic domains or unfavorable enzymatic activities in the case of therapeutics and vaccines by means of gene synthesis. Taken together, facilitated production processes, increased yields, improved safety profiles and proven efficacy of bioproducts based on synthetic genes will finally add to reduce the “time to market”. In the following, two examples will be given illustrating that a profound understanding of regulated gene expression combined with a rational gene design may result in innovative bioproducts exhibiting safety and efficacy profiles that are significantly superior to those achieved by the use of naturally occurring genes. Both, improvements in the biological activity and safety profiles of recombinant bioproducts may considerably contribute to fulfill standards set by regulatory

9 Synthetic Genes for Prevention and Therapy

authorities. This is especially true for all cases where bioproducts are deduced from life threatening pathogens such as human immunodeficiency virus (HIV) in order to develop effective candidate vaccines or vector constructs for efficient gene delivery.

2

Paradoxon: HIV-derived vaccines and gene delivery systems

In the vast majority of reported cases infection of humans by HIV-1 leads to immunodeficiency, severe clinical symptoms and death often within less than 10 years. However, a considerable portion of HIV-1 infected people callled long-term nonprogressors is capable of controlling the infection for up to 15-20 years. There is growing evidence from several laboratories that cellular immune responses primarily directed towards highly conserved viral proteins are critically involved in limiting the infection, controlling viral replication and keeping the virus load in the circulation and lymphoid tissues low (Harrer et al., 1998; Wagner et al., 1999). Detailed immunological analysis of symptomatic and asymptomatic HIV infected individuals revealed an inverse correlation between

the cellular immune responses directed towards the HIV-1 group-specific antigens (gag) or polymerase functions (Pol) and the plasma virus load (Rosenberg et al., 1997). Necessary consequences for novel vaccine concepts are the presentation of a large repertoire of antigenic sites as well as the stimulation of different effectors of the immune system including a Thl -dominated cellular immune response directed towards the gugpol gene (Wagner et al., 1999). Despite the life-threatening properties of HIV per se, selected components of this virus including the above named gugpol genes are currently being exploited to construct a novel class of gene delivery systems allowing the transduction of non-dividing, postmitotic and terminally differentiated cells. In contrast to standard retroviral gene transfer as mediated, e.g., by Moloney murine leukemia virus (MoMuLV) based vectors, lentiviral gene transfer can be used for gene delivery into a variety of quiescent cells, such as neuronal, muscle, and liver as well as hematopoietic stem cells (Naldini et al., 1996; Gallichan et al., 1998; Case et al., 1999; Han et al., 1999). Lentiviral vectors might, therefore, considerably contribute to the prevention and treatment of genetic disorders, tumor diseases and infections. However, in the viral context, the expression of the gugpol products as well as the infection of quiescent cells - a property that is among all retroviruses restricted to lentiviruses - critically depends on the assistance of a series of viral trans-acting proteins and cis-active sites, all of which should be avoided in the context of therapeutic or preventive products based on HIV. A better understanding of the regulatory mechanisms controlling the timely coordinated expression of viral gene products has been required in order to construct safe lentiviral vaccines and vector constructs.

I

151

152

I

Kurt Bieler and Ray Wagner

3 Synthetic genes: Novel tools contributing to the understanding of HIV replication

Efficient production of structural or enzymatic proteins from lentiviruses such as HIV and SIV is limited by their complex regulatory mechanism of transcription, nuclear RNA translocation and expression employing the viral proteins Tat and Rev. Whereas the anti-repressor function of Tat can be easily avoided by using heterologous transcriptional control via viral and/or cellular promotor/enhancer units, the expression of late genes coding for Gag, Pol and the envelope protein (Env)depends on the presence of the trans-activeRev protein. This is known to promote the export of unspliced and partially spliced RNAs from the nucleus to the cytoplasm via the interaction with its cognate RNA recognition site Rev responsive element (RRE), a complex of a 351 nucleotide (nt) RNA stem-loop structure located within the env open-reading-frame. Although Rev/RRE action is well accepted as a necessary prerequisite for HIV-1 late gene expression, the critical contribution of different cis-acting elements within the unspliced and singly spliced transcripts to Rev dependency and timely regulated expression is still controversially discussed. Accordingly, Rev dependent and timely regulated export of late singly or unspliced lentiviral mRNAs has been explained in most reports either by inefficient splice site formation or attributed to inhibitory sequences located within the coding region (referred to as INS elements) (Figure 2). 3.1 Construction o f a synthetic, HIV-1 derived gag gene

Although the precise character of the postulated INS elements still remains to be defined, they seemed to involve a noticeably high A/T-content. In order to eliminate proposed repressor elements without having particular knowledge of their nature, we designed a synthetic gag gene employing a codon usage occurring most frequently in highly expressed mammalian genes (Graf et al., 2000). Accordingly, more than 400 substitutions homogeneously distributed throughout the complete gag gene were introduced thereby reducing the A/T-content of the wild-type gag gene from 55.9% down to only 33.9%. Mimicking the situation of genornic RNAs a 103 bp untranslated region (UTR) carrying the highly functional HIV-1 major splice donor (SD) was fused upstream (UTR-syngag, UTR-wtgag) and a 861 nt fragment known to carry the RRE was fused downstream to Gag encoding

Fig. 2. Rev-dependent, timely regulated export of late viral RNAs (see text)

UTR-wfgag-RRE wtgag-RRE UTR-syngag-RRE

SYngag UTR-syngag syngag-RRE

5‘ Major Splice Donor

codon usage adapted to mammalian cells Fig. 3. Schematic representation o f wild-type and synthetic gag encoding expression plasmids. Grey boxes indicate synthetic gag (syngag) encoding genes, whereas open boxes indicate wild-type gag (wtgag) encoding genes. Syngag and wtgag reading frames were fused to the cis-acting sequences 5‘ located non-

I 3’ Splice Acceptor Rev-Responsive-Element

transiated-region (UTR) and Rev responsiueelement (RRE). In upper part, the position ofthe Gag encoding region as well a s the RRE are highlighted within the HIV-l genome. Dots mark the HIV-1 major splice donor (SD), vertical lines indicate splice acceptors (SA).

reading frames (syngag-RRE, UTR-syngag-RRE, wtgag-RRE, UTR-wtgag-RRE).The latter fragment accommodates, in addition to the RRE, the most 3’ located splice acceptor site within the HIV-1 genome that is known to be used very inefficiently a property suggested to contribute to timely regulated gene expression (Dyhr Miltlcelsen and Kjems, 1995; Olsen et al., 1992; Staffa and Cochrane, 1994; Dyhr Mikltelsen and Kjems, 1995; O’Reilly et al., 1995). All synthetic gag gene derivatives and RRE containing wild-type gag sequences were cloned into the pcDNA 3.1 (+)expression vector under the transcriptional control of the immediate-early promoter-enhancer of cytomegalovirus (CMV) (Figure 3). 3.2

Codon usage modification in the gag gene abolishes Rev dependency and increases expression yields

To evaluate the critical contribution of inhibitory elements in the presence of the major SD, Gag expression from the wild-type gag gene reporter was compared to the synthetic gag gene driven expression, in presence and absence of UTR, RRE and Rev, respectively. High-level expression of Pr55g”g was achieved after

1

2

3

4

5

6

7

8

9 1 0 1 1 1 2

+-R e v -

+

-

+

-

+

-

t

Fig. 4. Influence o f various cis acting sequences o n Rev dependent gag expression. Human H1299 lung carcinoma cells were tran. siently transfected with the indicated reporter constructs that were either based on the wildtype gag gene or on the synthetic gag gene. Rev-responsiveness was determined by mock co-transfection ( - ) or co-traiisfcction o f a Rev expression plasmid (+). Expression of the gag

55kDa

- + - + reporter was monitored from cell-lysats by Western blot analysis and quantified by a Gag specific capture ELlSA (DuPont, Boston, MA). Levels of Gag production were expressed as the percentage o f Gag protein obtained after cotransfection o f UTR-wtgag-RRE with Rev. The indicated values each represent the mean of four independent transfection experiments. Standard deviations o f the mean are indicated.

transfection of various syngug encoding plasmids into mammalian cells (Figure 4). Noteworthy, expression levels from the optimized gag gene were neither substantially altered by introducing the Rev/RRE system nor influenced by the presence of U T K and the major SD. By contrast, expression of the wild-type gag gene derived product essentially depended on the presence of' RKE, Rev as well as on the 5'-UTR including the major splice donor (Figure 4) confirming previous observations made by several groups that Rev-dependentexpression of late HIV-1 gene products is influenced by splice site usage (Hammarskjold et al., 1989; Kjems et al., 1991; Chang and Sharp, 1989; Hammarskjold et al., 1989; Lu et al., 1990; Mikaelian et al., 1996). I'rSFg cxpression levels from codon adapted genes exceeded those accomplished by the Rev dependent wild-type gag reporter by 1,5-2 fold. Bascd on these results we conclude that Rev responsiveness of HIV-1 late gene expression critically depends on appropriate wild-type codon usage.

Northern blot analysis of nuclear and cytoplasmatic fractions clearly demonstrated that in the absence of the SD the wild-type transcripts are targeted into a intranuclear degradative pathway (Figure 5B). The addition of the 5'-UTR/SD to wild-type gag ( UTR-wtgag-RRk')lcd to a nuclear accumulation of-Gag encoding messages that

9 Synthetic Genes for Prevention and Therapy

A Cytoplasm

+

R U

g

61 z$n: z

-

i-

-

+

+

250 200 150 100 50

Rev

Nucleus --

1

2

3

Fig. 5. Cellular RNA distribution. Northern Blot analysis of cytoplasmatic and nuclear RNA. H1299 cells were transfected with the indicated constructs and harvested 48 hours post-transfection. Rev-responsiveness was tested by cotransfection o f a Rev expression plasmid (+) or , Cells were parempty vector ( ~ ) respectively. tially lysed and nuclei were separated from the cytoplasm. RNA was prepared from the cytoplasmatic fraction [A) and subjected - together with RNA purified from the nuclei ( 6 ) - t o

4

5

6

2 4nc

7

Northern Blot analysis. Gag encoding transcripts and fl-actin RNAs were detected by a radio-labeled RRE antisense riboprobe and a ,O.actin specific DNA probe, respectively. The position and calculated length of the Gag encoding RNAs is indicated from the right. Intensities o f the gag and 8-actin specific signals were quantified by a Phospho-Imager. Bars in the upper panel represent the relative amounts o f specifically detected gag transcripts following normalization towards the fl-actin conlrol.

were lranslocated into the cytoplasm in presence of Rev (Figure 5A, B). By contrast syngag encoded R NAs, exhibiting markedly differences in wobble positions and calculated RNA secondary structure (not shown),were readily detected both within the nucleus and cytoplasm (Figure SA, B). Nuclear and cytoplsmatic levels of syngag transcripts werc neither influenced by Rev/RRE interaction nor by the prescnce of 5'-UTR/SD (Figure 5 A, B). In accordance with the expression data, cytoplasmatic levels of spngag mRNA exceeded those achieved by UTR-wtgug-RKE and Rev by about 50-70%, by 5- to 8-fold in the absence of Rev and by several orders of magnitude in the absence ofthe 5'-UTR (Figure 5A). Taken together, this clearly demonstrates that the elimination of-proposed inhibitory sequence elements by a consequent codon usage adaptation transforms Gag encoding transcripts into a RNA species with altered characteristics, rendering nuclear export and Gag expression completely independent from the presence of-the 5' -UTR/SD and Rev/RRE interaction.

I

755

156

I

Kurt Bider and Ralf Wagner

3.4 Codon usage modification in the gag gene alters the nuclear export pathway of otherwise C R M l dependent RNAs

To determine whether or not the observed differences in the RNA phenotypes can be correlated to different nuclear export pathways, the influence of leptomycin B (LMB) on the nuclear export of the different RNA species was investigated. LMB has recently been shown to effectively inhibit Rev function due to its ability to directly interfere with CRMl(Exportinl)/Ran-GTPmediated nuclear export (Kudo et al., 1998; Otero et al., 1998; Wolff et al., 1997; Kudo et al., 1998; Otero et al., 1998). In contrast to syngag driven Gag expression, Rev dependent expression of wtGag ( UTR-wtgag-RRE) was extremely sensitive towards LMB treatment resulting in sig-

A no LMB 5 nM LMB

? .

1,5 I n 1 0,5

2

p

O(

--*-

I

B

UTR-wtgag-RRE

+Rev

,

b-actin

55kDa

I____

I

N

C

+

-+

LMB -

b-actin

1

2

3

4

Fig. 6. identification of nuclear export pathways. Influence of codon usage within the gag gene on Leptomycin B (LMB) sensitivity and nuclear export pathway. (A) Influence of LMB on Gag expression from the indicated reporter constructs. H1299 cells were transfected by either syngag or UTR-wtgag-RRE in cornbination with Rev each in absence or prescnce of 5 n M LMB. Cells were harvested 48 h post transfeclion. Synthcsized Gag protein was dctcrmincd from cell lysates by Western blol analysis (lower panrl) and quantified by a capture ELISA (upper panel). The Pr55':"" polyprotein is indicated from the right. Amounts of

N

-

C

N

7

8

+ 5

6

produced Gag protein are expressed ng Pr55gag per hg of total cellular protein. (B) Influence of LMB on subcellular distribution o f Gag encoding RNA. H1299 cells were transfected with the indicated constructs and cultivated with (+) or without (-) 5 n M LMB. Nuclear (N) and cytoplasmic (C) levels of Gag-encoding RNA were determined by Northern blot analysis. Gag encoding transcripts were detected by radio-labeled riboprobes specifically matching the syngag or the wild-type gag RNAs, respectively. 0-actin RNAs were detected by a radio-labeled p-actin specific DNA probe.

9 Synthetic Genes for Prevention and Therapy

nificantly decreased levels of Gag expression (>90%) (Figure 6A) (Wolff et al., 1997; Fischer et al., 1995). Northern blot analysis revealed, that LMB treatment and disruption of CRMl function remarkably decreased both nuclear and cytoplasmic levels of wild-type RNAs (Figure 6B) (Wolff et al., 1997;Askjaer et al., 1998).In sharp contrast, syngag derived transcripts were readily detected in the nucleus and shown to be exported constitutively to the cytoplasm whether or not LMB was present (Figure 6B). We therefore conclude that, by altering the codon usage and thereby eliminating proposed INS elements, gag encoding transcripts are targeted to a different C R M l independent nuclear export pathway. This implicates that the targeting of the gag encoding RNAs to the Rev/CRMl export pathway is critically determined by the sequence composition of the wtgag gene. 3.5 Codon usage modification increases R N A stability, modulates nuclear RNA export and increases translational efficiency

Thus, the construction of a codon optimized gag gene enabled us to truly eliminate so-called A/U-rich repressor sequences. We therefore strongly suggest that the overall A/U-content of the gag RNA rather than previously proposed “INS” elements contribute to stability and nuclear retention of wild-type gag RNAs. This assumption is in accordance with several publications correlating instability of certain cellular mRNAs with their AU-content or the presence of AU-rich elements (reviewed by Chen and Shyu, 1995). Furthermore, we were able to show that increased levels of expression achieved after codon optimization of our gag reporter is due to nuclear stability and constitutive nuclear export of its transcripts rather than increased translational efficiency. Consequently, different nuclear sequestration of wild-type and codon optimized gag-RNAs could be demonstrated by the CRMl independent nuclear export of the optimized transcripts (Figure 7).

Model of Rev-Independent Gag Expression

Fig. 7. Model on the contribution of different cis-acting sequences t o

Rev-dependent, timely regulated expression of late gene products (see text)

I

157

158

I

Kurt Bieler and Ralf Wagner

4

Synthetic genes: Implications on the development of safe and effective DNA vaccines 4.1

Safety issues to be considered for DNA vaccine development

The development of HIV specific DNA vaccine constructs expressing gag or derivatives thereof for vaccination purposes faces several limitations regarding both, safety and efficiency. Gag expression using wild-type genes in the Rev-dependent situation is limited by the complex viral regulatory mechanisms which involve several cis-acting elements (RRE, UTR) and trans-acting proteins (Rev,Tat). Gag expression, therefore, needs simultaneous expression of the Rev protein either from a bicistronic construct or from a separate plasmid. Both strategies limit the efficiency of generating cell lines in vitro or reduce the efficacy of DNA vaccine constructs in viuo either due to an increased plasmid size or the necessity to transfectltransduce one single cell with both plasmids at the same time; and by the presence of the Rev protein itself, acting as an RNA shuttle between nucleus and therefore harboring an intrinsic risk. Moreover, when applied as a therapeutic vaccination in chronically HIV infected individuals, Rev could contribute to reactivating latent viruses, thereby enhancing infection, virus replication and disease. However, as outlined above, the use of codon optimized gag genes or derivatives thereof for DNA vaccination resulted in high levels of Gag expression independent of the RRE/Rev system. Furthermore, this strategy excludes the S’-untranslatedregion (5‘-UTR)including a RNA packaging signal, reduces the homology between the vaccine construct and the wtgag sequence and increases the number of potentially immunostimulatory DNA-motifs (Figure 8),

I

Implications for the Developement of DNA Vaccines and Lentiviral Vectors for Gene Thearpy

-

No Rev

* No 5 -UTR * No RRE * No packaging sequences

-

Facilitated expression

Oncogenic potential

Enhanced irnrnunogenicity

Fig. 8. The power of synthetic genes (11). Improved safety profiles of HIV-specific DNA vaccines and lentiviral vector constructs to be achieved by the use of synthetic gagpol genes.

9 Synthetic Genes for Prevention and Therapy

4.2

Codon optimization of a gag-specific candidate vaccines results in increased antibody responses

We have analyzed the capability of the generated gug expression vectors to induce Gag-specific antibodies in female BALB/c mice (Wagner et al., 2000; Deml et al., unpublished results) (Figure 9). Three groups of five animals each received an intramuscular (i.m.) primary immunization of plasmid DNA (100 pg per dose) followed by two i.m. boosts at weeks 3 and 6 with the same DNA dose. A control group was immunized with PBS. Total Ig titers to purified Gag proteins were determined by ELISA. Vaccination with both the syngag and UTR-wtgug-RREIRev vector system induced Th-1 type antibody responses. However, the titers induced by injections of syngag were more rapid and higher than those induced after co-immunization with the UTR-wtgug-RRE and Rev plasmids (pRev). Mice immunized with the syngag plasmid developed measurable levels of p24(CA)-specificantibodies already three weeks after the primary immunization (Figure 9). Reactive antibodies appeared to increase almost 100-foldtwo weeks after the first booster immunization and reached gag-specific endpoint titers of 1:81,000 one week after the second booster injection. By contrast, gag-specific serum antibody responses raised in mice following repeated co-administration of UTR-wtgug-RRE and pRev were significantly delayed and reached only maximum titers of 15,500 at one week after the second booster 1000000

-.-cn 0

100000

z0

:

d-

10000

-m-

UTRgagRRE + Rev

x

non immunized

1000

a c, 0

c

100

10

1

1

3

5

7

9

weeks post immunization Humoral responses. Kinetics and strength o f humoral responses induced in BALB/c mice immunised by intramuscular (i. m.) (co)injection o f 100 pg o f various Gag expression vector systems. Each point represents the group mean (n = 5) f SD for antiGag Ig antibodies as determined by end-point Fig. 9.

dilution ELISA assay. End-point titers of the immune sera were defined as the reciprocal of the highest plasma dilution that resulted in an adsorbance value (OD 495) three times greater than that o f a preimmune serum with a cut-off value of 0.05. The arrows indicate the time points of booster immunisations.

I

159

160

I

Kurt Bieler and Ray Wagner

immunization. No gag-specificantibody response was detectable at any time point in the sera of non immunized mice and animals co-injected with gagRRE and the Rev expression vector. 4.3

Enhanced in vitro cytokine release of splenocytes from mice immunized with synthetic gag plasmid DNA

The recall antigen-specific cytolcine secretion as a measure of T-helper memory cells was determined from splenocytes of mice, obtained 5 days after the second booster immunization. Splenocytes from mice i. m. immunized with both the syngag and UTR-wtgag-RREIRevvector systems showed a substantial I F N - y production (Table 3 ) upon specific in vitro stimulation with purified Gag proteins. A significantly reduced, but specific IFN-y secretion was observed after p. g. injection of the syngag plasmid. In contrast, no significant amounts of IFN-y were secreted from splenocytes of mice i. d. co-immunized with UTR-wtgag-RRE and Rev expression vector after specific restimulation (Table 3 ) . No cytokine secretion was observed from non stimulated splenocytes of all experimental groups. To assess Th2 differentiation, ELISA was performed from the same cell culture supernatants to quantify the concentrations of secreted IL-4 and IL-5. In all groups of immunized and non-immunized mice, independent of the immunization route, no IL-4 and IL-5 secretion was neither detectable from the supernatants of specifically restimulated nor from non-stimulated splenocytes. Thus, i. m. immunization of vectors containing the improved synthetic gag expression cassettes induced a strong Thl cytokine profile, whereas particle gun injection of these plasmid constructs resulted in a weaker Thl-biased cytoltine response.

Table 3. Cytokine profile of in uifro gag-stimulated splenocytes from mice immunized i. m . by needle injection or i.d. by particle gun with various gag expression vectors

UTRgagRRE

IL-5(pg m P )

IFN-y [pg rnl-’)


3220 -t 840

<8


3520 f 1020

<8


<32

<8

<16

80

11-4 {pg ml-’)

DNA Vaccine

+ Rev

(i.m.) syngag (i. m.) UTRgagRRE

+ Rev

(i.d.) syngag (i.d.)

Means f SD of splenocytes of five mice per experiment

* 32

I

9 Synthetic Genes for Prevention and Therapy

4.4

Induction of CTL responses in mice immunized with the modified Gag expression plasmids

In order to analyze the capacity of syngag plasmids to induce a Gag-specific CTL response, splenic cells, derived from immunized mice 3 weeks after the primary immunization, were specifically restimulated in a 6 day mixed lymphocyte tumor cell culture and tested for cytotoxic activity. The AMQMLKETI 9-mer p24(CA)-derivedpeptide used in this assay is known to constitute a Dd-restricted CTL epitope in BALB/c mice. Gag-specific CTL were detectable after a single i. m. injection of the syngag plasmid DNA, whereas administration of UTR-wtgagRRE/pRev and wtgagRRElpRev combinations was not sufficient to induce detectable CTL responses. Furthermore, no CTL priming was observed following an intradermal injection of the syngag vector as well as the UTRgagRRE/pCsRevsg25GFP and gagRRE/pCsRevsg25-GFP combinations by particle gun immunization (Figure 10). These results showed that i.m. injection of the syngag plasmid was the most efficient vaccination strategy to induce both substantial humoral and cellular Gag-specific immune responses. Talcen together these results strongly suggest that the administration of a synthetic gene based DNA vaccine is clearly superior to wt-derived DNA constructs form the safety, production and immunogenicity point of view.

40

UTR-wtgag-RR+Rev

s

h

I syngag (Lm.) 0 UTR-wtgag-RR+Rev

-

v g a g (w.1 non immunized

P u)

.

L

U ’

20

10

5

2

E/T ratio Fig. 10. CTL activities. Cytotoxic T cell activity i n splenocytes from mice immunized intramuscularly by needle injection or intradermally by particle gun with indicated Gag expression vectors. Lymphoid cells obtained from mice five days after the booster injection were co-cultured with Gag peptide-pulsed syngeneic P815 mastocytoma cells (irradiated with 20,000 rad). Control assays included splenocytes of non immunized mice stimulated in vitro with pep-

tide-pulsed P815 cells. Cytotoxic effector populations were harvested after 5 days o f in vitro culture. The cytotoxic response was read against 9-mer Gag-peptide pulsed A20 cells and untreated A20 negative target cells i n a standard 5’Cr release assay. Data shown were mean values o f triplicate cultures. The standard errors of the means o f triplicate data were always less than 1 5 % of the mean.

I

16’

162

I

Kurt B i d e r and Ra,f Wagner

5 Synthetic genes: Implications on the development o f safe lentiviral vectors for gene delivery into quiescent cells

Another application of lentiviral gag and gagpol based vectors is gene transfer into different types of dividing, resting or postmitotic cells. Lentiviral vector particles are currently prepared by triple-transfection of a gagpol expression plasmid together with a transfer construct and expression plasmid encoding the envelope protein derived from Vesicular stomatitis virus into mammalian cell lines (Figure 11A). Lentiviral vector particles from the supernatant of such transfected cells were able to stably transduce a large variety of different cells even after in vivo gene transfer (Blomer et al., 1997; Case et al., 1999; Slcolnilr et al., 1991; Fontenot et al., 1992; Ford et al., 1992; Ruegg and Strand, 1991).To increase the safety of these vectors, most of the accessory genes of HIV-1 were deleted from the packaging construct and the vector, thereby minimizing the risk for the emergence of potentially pathogenic replication competent recombinants (Zufferey et al., 1997; Schnell et al., 2000). However, the development of lentiviral gag and gagpol based vectors for gene transfer into quiescent or non-dividing cells faces several limitations regarding both safety and efficiency (Figure 8).

A

B

p24 Capture ELISA

1

Packaging constructs

300000 200000

50000

-

gag RRE C Transfer construct - GFP UTR 7 1

CMV UTR-syngp-RRE

~

env

POI

CIMY gagl synw

i

RRE

;

c____l POI

Fig. 11. Lentiviral vector constructs. (A) Different variants o f synthetic and wild-type HIV-1 derived gagpol packaging constructs. (6)Expression profiles o f gagpol packaging functions

~

Envelope V

tat1 0

tat2 OmGFP

following transfection of the constructs depicted i n (A) into H1299 cells. (C) VSV-G derived envelope expression construct and SIV derived, GFP expressing transfer plasmid.

9 Synthetic Cenesfor Prevention and Therapy I 1 6 3

5.1 Safety issues to be considered for lentiviral vector development

The currently used HIV-1 or SIV derived gagpol expression plasmids contain parts of the 5' -untranslated region comprising the RNA packaging signal A. Moreover, if the gagpol expression plasmid can indeed be packaged, homologous or non-homologous recornbination events between the vector RNA and the gugpol RNA during or after reverse transcription can not be excluded. This could lead to the undesired transfer of the gagpol gene (or parts of it) into target cells. Additionally, two regions of homology between conventional wild-type based gagpol expression plasmids and the lentiviral transgene vector constructs might facilitate homologous recombination events during either vector production or reverse transcription: the 5'-untranslated region (5'-UTR) or at least part of it including parts, the gag gene required for efficient gagpol expression and packaging of the transgene construct. Homologous recombination events during either vector production or reverse transcription should be excluded by rational design of packaging constructs (Figure 8). 5.2 Construction and characterization o f synthetic gagpol expression plasmids

Different gagpol expression plasmid were generated by cloning the coding region from gagpol including the 5'untranslated region (UTR),and/or the rev-responsive element (RRE) of HIV-1 downstream of the CMV promoter as desired (Figure 11A) (Wagner et al., in press). In the absence of Rev, expression of the Gag capsid protein (p24) from UTR-wtgp-RREwas rather inefficient (Figure 12A). Co-transfection of a Rev expression plasmid led to p24 antigen levels comparable to those seen after transfection of HIV-1 proviral DNA (HX10).To further minimize HIV-1 sequences present on the HIV-1gugpol expression plasmid, the UTR sequences were removed from UTR-wtgag-RRE resulting in wtgp-RRE. However, even in the presence of Rev, CA expression was undetectable (Figure 11B). In order to generate a Rev-independent gagpol expression plasmid, the codon usage of the 4.3 kb gagpol gene of HIV-1 was optimized for expression in human cells (UTR-syngp-RRE,Figure 11A). p24 capsid antigen expression of UTR-syngp-RRE was independent of Rev (Figure 11B). The UTR and the RRE were not required for efficient expression of CA from the synthetic gagpol gene (syngp; Figure 11B) also expressed p24 capsid at similar levels (Figure 11B). Subviral particles were readily released from the transfected cells and sedimentate at a density of about 1 , l G g mL-', comparable to wild-type HIV-1 virions. The gagpol precursor protein could also be detected by adding the HIV-1 protease inhibitor Saquinavir during particle production. No differences were observed in the processing of the gagpol precursor protein encoded by the wild-type virus and the synthetic gagpol gene (not shown).

164

I

Kurt Bieler and Ralf Wagner

5.3

Production of lentiviral vectors using synthetic gagpol genes

The functional integrity of the gagpol proteins encoded by the synthetic genes was analyzed by co-transfection of the synthetic HIV-1 gugpol expression plasmids and an expression plasmid for VSV-G with the SIV transgene vector, which expresses the green fluorescent protein gene (GFP) under the control of an internal promoter (Figure 11C). The vector titers in the supernatant of the transfected cells were in the range of 1 x lo6 to 5 x lo6 GFP-forming units per mL of supernatant. Similar titers were obtained after co-transfection of the SIV transgene vector with a wildtype gugpol expression plasmid of SIV. Of note, the synthetic HIV-1 gagpol expression plasmid also allowed efficient transfer of the SIV vector. Therefore, HIV-1 gagpol must recognize all cis-acting sequences of SIV required for packaging, reverse transcription, and integration with similar efficiency as SIV. 5.4

Transduction o f non-dividing cells

The transduction efficiency of vectors produced with the synthetic gugpol genes for non-dividing cells was assessed by arresting the target cells in the G1 phase of the cell cycle by aphidicolin treatment. The titer of an MLV-based vector in growtharrested cells was reduced to background levels (Figure 12). In contrast, the SIV vector titer was only slightly reduced by aphidicolin treatment. Non-dividing cells could be transduced with the SIV vector independent of the gugpol expression plasmid used for the production of the vector particle (Figure 12).

Aphidicolin (pginil)

Aphidicolin arrests cells in GI-Phase of the cell cycle

107

Only lentiviral vectors are transducing arrested cells

SgpA2

syngp

MLV

I

Transduction efficiencies of vectors are equal

Fig. 12. Transduction of growth-arrested cells. The 293T cells were transfected with ViCABH, pHIT-C and the indicated lentiviral gagpol expression plasmids. The MLV vector was generated by transfecting plasmids pLECFP-N1, pHIT60, and pHIT-G. Vector titers in the supernatant o f transfected cells were deter-

mined on 293 cells in the presence of the indicated concentrations of aphidicolin. The titration was done in triplicates or quadruplicates. The means and the standard deviations are shown. CFU: green fluorescence forming units.

9 Synthetic Genes for Prevention and Therapy I 1 6 5

5.5 Absence of replication-competent recombinants (RCRs)

The frequency of emergence of RCR was tested in an assay system that monitors homologous recombination between the gagpol expression plasmids and an S IV transfer vector for the generation of RCR. Homologous recombination events between the synthetic gagpol expression plasmids and an SIV vector were undetectable and in comparison to a previously used gagpol expression plasmid at least approximately 100-fold less frequent (Figure 13). In sum these results demonstrate that by eliminating regions of homology and sequences involved in packaging, synthetic gagpol genes greatly improve the safety profile of lentiviral vectors to be used in future clinical trials. 10000000

2

+wild 1000000

j-

0

4

u)

100000 10000 1000

type

-

+synthetic

0

5

10

15

20

25

days post transfection

Fig. 13. Detection of replication competent recombinants, CEMxl74-SIV-SEAP cells were infected with the supernatant o f 293T cells

co-transfected with SIV-GFPADP (A) or SIVGFPABP (B) and the indicated gagpol expression plasmids. RLU: relative light units.

6 Future perspectives

Altogether, these data strongly support the use of synthetic genes for the design of novel generations of candidate DNA vaccines and viral vector constructs for gene therapy. Advantages regarding the safety of DNA vaccines and gene delivery systems based on rationally designed genes are obvious and will certainly contribute to developing novel safety standards for approval of comparable products for clinical trials (Figure 8). In addition, the opportunity of designing the genes of interest

. including the plasmid backbone on a rational basis will allow to increase and to modulate immune responses as desired towards Thl or Th2 type effector functions. On the contrary, sequences encoded by any type of viral vector used for gene therapy purposes my be rendered immunosilent by removing known immunostimulatory CpG sequence motifs and introducing inhibitory sequence elements.

166

I

Kurt Bieler and Ralf Wagner

More generally, the safety profile and yields of proteinaceous bioproducts to be produced in various cell lines can be significantly improved by altering or modifying the coding sequence. The examples presented in this chapter clearly demonstrated that increased yields of gag or gagpol products are rather due to a stabilization of the nuclear RNAs and a re-direction of these RNAs into an alternative nuclear translocation pathway than to an increased translational acticity resulting from an adaptation of the codon usage to tRNA frequencies. The opportunity to rationally design and produce (see also Chapter 11) optimized genes without any need of a “wet” template may also contribute to reducing the time span between the identification of genes by sequencing projects or functional genomic programs (differential display, proteomics) and the biochemical analysis and production of the corresponding products for preventive or therapeutic purposes. Finally, progress in synthesizing genes and ribozyme-encoding sequences on automated, high-throughput solid-phase platforms will certainly speed up the generation and broaden the diversity of combinatorial libraries to be used in future drug screening programs. Acknowledgements

The authors wish to thank Dr. Graf, Mrs. Bojak and Dr. Deml, University of Regensburg, for their skillful1 biochemical and immunological characterization of synthetic genes. We would also like to thank Prof, Uberla, University of Leipzig, for the excellent collaboration on the testing of lentviral vector constructs. This work was supported by BMBF grant No. 01KI9765/3 and DFG grant No Wo227/ 7-4 R. w.

References ASICJAER, P., JENSEN, T. H., NILSSON, J., ENGLMEIER,L., KJEMS,J. (1998), The specificity of the CRM1-Rev nuclear export signal interaction is mediated by RanGTP, /. Biol. Chern. 273, 33414-33422. Au, LC., YANG, FY.,YANG,WJ., Lo, SH., IGo, CF. (1998), Gene synthesis by a LCR-based approach: High-level production of leptin-154 using synthetic gene in Escherichia coli, Biochern. Biophys. Res. Comrnun. 248, 200-203. D., SZABO,A,, BINNIE,C., JENISH, D., COSSAR, TRUDEAU, D. et al. (1997), Expression and characterization of soluble human erythropoietin receptor made in Streptornyces lividans 66, Protein E q r . Purq 11, 271-278.

B L O M EU., ~ NALDINI, L., I ~ F R IT., , TRONO, D., VERMA,I. M., GAGE,F. H. (1997), Highly efficient and sustained gene transfer in adult neurons with a lentivims vector, /. Virol. 71, 6641-6649. CASE,S. S., PRICE,M.A., JORDAN, C.T., Yu, X. J., WANG,L. et al. (1999), Stable transduction of quiescent CD34(+)CD38(-) human hernatopoietic cells by HIV-1-based lentiviral vectors, Proc. Natl. Acad. Sci. USA 96, 2988-2993. CHANG,D. D., SHARP,P.A. (1989), Regulation by HIV Rev depends upon recognition of splice sites, Cell 59, 789-795. CHEN,C.Y., SHYU,A. B. (1995), AU-rich elements: characterization and importance in mRNA degradation, Trends. Biochern. Sci. 20, 465-470.

3 Synthetic Genes for Prevention and Therapy

DELVECCHIOBLANCO,F., CAFARO,V., DI MARO,A., SCOGNAMIGLIO, R., SINISCALCO, G. et al. (1998), A Recombinant ribosome-inactivating protein from the plant Phytolacca dioica L. produced from a synthetic gene, FEBS Lett. 437 (3), 241-245. DYHRMIKKELSEN, H., KJEMS, J. (1995), Inefficient spliceosome assembly and abnormal branch site selection in splicing of an HIV-1 transcript in vitro, /. Biol. Clzem. 270, 24060-24066. FISCHER,U., HUBER, J , , BOELENS,W. C., M A ~ A JI.,W., LUHRMANN,R. (1995), The HIV-1 Rev activation domain is a nuclear export signal that accesses an export pathway used by specific cellular RNAs, Cell 82, 475483. GRAF, M., BOJAK,A,, WOLF,H., DEML,L., R. (2000), Concerted action of WAGNER, multiple cis-acting sequences is requifes for Rev dependence of late human immunodeficiency virus type 1 gene expression, /. Virol. 74, 11367-11376. HAMMARSICJOLD, M. L., H E I M E R , I., B., SANGWAN, I., ALBERT, L., HAMMARSKJOLD, REKOSH,D. (1989), Regulation of human immunodeficiency virus env expression by the rev gene product, /. Virol. G3, 1959-1966. A. N., WAREING, M., HAN,J. J., MHATRE, PE-ITIS,R., GAO,W. Q. et al. (1999),Transgene expression in the guinea pig cochlea mediated by a lentivirus-derived gene transfer vector, Hum. Gene fier. 10, 1867-1873. HANKE, T., NEUMANN,V. C., BLANCHARD, T. J., SWEENEY, P., HILL,A. V et al. (1999),Effective induction of HIV-specific CTL by multiepitope using gene gun in a combined vaccination regime, Vaccine 17, 589-596. HARRER, T., HARRER, E., KALAMS,S.A., ELBEIK, T., STAPRANS,S. I. et al. (199G), Strong cytotoxic T cell and weak neutralizing antibody responses in a subset of persons with stable nonprogressing HIV type 1 infection, A I D S Res. Hum. Retroviruses 12, 585-592. IANNACONE, R., GRIECO, P. D., CELLINI, F. (1997), Specific sequence modifications of a cry3B endotoxin gene result in high levels of expression and insect resistance, Plant Mol. Bid. 34 (3), 485-496. KJEMS, J., FRANKEL, A. D., SHARP,P.A. (1991), Specific regulation of mRNA splicing in vitro by a peptide from HIV-1 Rev, Cell G7, 169- 178.

KIM, C. H., OH, Y., LEE, T. H. (1997), Codon optimization for high-level expression of human erythropoietin (EPO) in mammalian cells, Gene 199 (1-2), 293-301. KRIEG, A.M. (1996), CpG DNA: A novel immunornodulator, Trends Microbial. 7, 64-65. KUDO, N., WOLFF, B., SEKIMOTO, T., E. P., YONEDA,Y. et al. (1998), SCHREINER, Leptomycin B inhibition of signal-mediated nuclear export by direct binding to CRM1, Exp.Cell Res. 242, 540-547. LAURENCE,J., SIEGAL, F. P.,SCHAITNER,E., GELMAN, I. H., MORSE,S. (1992),Acquired immunodeficiency without evidence of infection with human immunodeficiency virus types 1 and 2, Lancet 340, 273-274. J., REKOSH,D., HAMLu, X.B., HEIMER, MARSKJOLD, M. L. (1990), U1 small nuclear RNA plays a direct role in the formation of a rev-regulated human immunodeficiency virus env mRNA that remains unspliced, Proc. Natl. Acad. Sci. U S A 87, 7598-7602. LUBAN, J., ALIN,I<.B., BOSSOLT, K. L., HUMARAN, T., GOFF,S. P. (1992), Genetic assay for multimerization of retroviral gag polyproteins, /. Virol. 6G, 5157-5160. MA, H. S., HAQ,T.A., CLEMENTS, J. D., C. J. (19981, Edible vaccine protects ARNTZEN, mice against Escherichia coli heat-labile enterotoxin (LT):potatoes expressing a synthetic LT-B gene, Vaccine 16 (13), 1336-43. A. J., OXER,M. D., ROMANOS, M.A., MAKOFF, FAIRWEATHER, N. F., BALLANTINE,S. (1989), Expression of tetanus toxin fragment C in E. coli: high level expression by removing rare codons, Nucleic Acids Res. 17 (24), 1019110202. MASON,H. S., HAQ,T. A,, CLEMENTS, J. D., ARNTZEN, C. J. (1998), Edible vaccine protects mice against Eschericha coli heat-labile enterotoxin (LT): potatoes expressing a synthetic LT-B gene, Vaccine 16, 1336-1343. MIKAELIAN, I., KRIEG, M., GAIT,M. J., KARN,J. (1996), Interactions of INS (CRS) elements and the splicing machinery regulate the production of Rev-responsivemRNAs, J . Mol. Bid. 257, 246-264. NALDINI, L., BLOMER, U., GALLAY, P., ORY,D., MULLIGAN, R. et al. (1996). In vivo gene delivery and stable transduction of nondividing cells by a lentiviral vector, Science 272, 263-267.

I

167

168

I

Kurt

Bieler and Rdf Wagner

NAYAK,P., BASU, D., DAS,S., BASU,A,, GHOSH, of suboptimal signals in the 3’ splice site, I . Virol. 68 , 3071-3079. D. et al. (1997),Transgenic elite indica rice

R., LESCHONSICY,B., HARRER E., plants expressing CryIAc delta-endotoxin of WAGNER, PAULUS, C., WEBER, C. eta]. (1999),Molecular Bacillus thuriizgiensis are resistant against and functional analysis of a conserved CTL yellow stem borer (Scirpophaga incertulas), Proc. Natl. Acad. Sci. USA 94 ( 6 ) , 2111-2116. epitope in HIV-1 p24 recognized from a longterm nonprogressor: constraints on immune OLSEN, H. S., COCHRANE, A. W., ROSEN,C. escape associated with targeting a sequence (1992), Interaction of cellular factors with essential for viral replication, J. Imrnunol. intragenic cis-acting repressive sequences 162, 3727-3734. within the HIV genome, Virology 191, WAGNER, R. (ZOOO), Virus-like particle and 709-715. DNA based candidate vaccines: Novel strateJ. E., OTERO, G.C., HARRIS, M. E., DONELLO, gies to induce Thl type immune responses, HOPE,T. J. (1998), Leptomycin B inhibits 5th European Conference on A I D S Research equine infectious anemia virus Rev and ECEAR 2000, June 2000, Spain (Plenary Talk). feline immunodeficiency virus rev function R., GRAF,M., BIELER,I<., WOLF,H., but not the function of the hepatitis B virus WAGNER, GRUNWALD, T. et al. (ZOOO), Revindependent posttranscriptional regulatory element, expression of synthetic gag-pol genes of J. Virol. 72, 7593-7597. HIV-1 and SIV Implications for the safety ROSENBERG,E. S., BILLINGSLEY,J. M., S. L., SAX, P. E. et of lentiviral vectors, Hum. Gene Tner. 11, CALIENDO, A.M., BOSWELL, 2403-2413. a1. (1997), Vigorous HIV-I-specific CD4’ T R., SHAO,Y., WOLF,H. (1999), WAGNER, cell responses associated with control of Correlates of protection, antigen delivery and virernia, Science 278, 1447-1450. ROUWENDAL, G. J., MENDES,O., WOLBERT, E. J,, molecular epidemiology: basics for designing an HIV vaccine, Vaccine 17, 1706-1710. DOUWE DE BOER,A. (1997), Enhanced J.G., WANG,H., GUERRA, D. J. (1997), expression in tobacco of the gene encoding WALLIS, green fluorescent protein by modification of Expression of a synthetic antifreeze protein in potato reduces electrolyte release at its codon usage, Plant Mol. Biol. 33 (6), freezing temperatures, Plant Mol. Biol. 35 (3), 989-99. I., GILBERT,S. C., BLANCHARD, 323-30. SCHNEIDER, J. J., WANG,Y. (1997), T. J., HANKE, T., ROBSON, I<. J. et al. (1998), WOLFF,B., SANGLIER, Leptomycin 8 is an inhibitor of nuclear Enhanced immunogenicity for CD8’ T cell induction and complete protective efficacy of export: inhibition of nucleo-cytoplasmic translocation of the human immunodefimalaria DNA vaccination by boosting with ciency virus type 1 (HIV-1) Rev protein and modified vaccinia virus Ankara, Nature Med. 4, 397-402. Rev-dependent mRNA, Chem. Biol. 4, 139-147. SCHNELL,T., FOLEY, P., WIRTH,M., MUNCH,J., ZOLOTUKHIN, S., PO~TER M., HAUSWIRTH, UBERLA, K. (ZOOO), Development of a selfN. (1996), A inactivating, minimal lentivims vector based W.W., GUY,J., MUZYCZIQA, “humanized green fluorescent protein on simian immunodeficiency virus, Hum. cDNA adapted for high-level expression in Gene Tner. 11, 439-447. mammalian cells, I. ViroE. 70 (7), 4646-4654. SINGH,M., O’HAGAN, D. (1999), Advances R. J., R., NAGY,D., MANDEL, in vaccine adjuvants, Nature Biotechnol. 17, ZUFFEREY, NALDINI,L., TRONO, D. (1997), Multiply 1075-1081. attenuated lentiviral vector achieves efficient STAFFA,A., COCHRANE, A. (1994), The tat/rev gene delivery in vivo, Nature Biotechnol. 15, intron of human immunodeficiency virus 871-875. type 1 is inefficiently spliced because

P/asmids for Therapy and kchation by M.Schleef Copyright 0 WILEY-VCH Verlag GmbH, 2001

10 Plasmids in Fish Vaccination Simon R.M. Jones

1

Introduction

The earliest report of foreign peptide expression in a fish (common carp) following parenteral administration of plasmid DNA, referred to herein as NAV (Hansen et al., 1991),coincided closely with the earliest reports of NAV in mice (MacGregoret al., 1989; Wolff et al., 1990; Tang, 1992; Ulmer et al., 1992). Since that time, however, the growth of knowledge in this field has been distinctly different for mammals and for fish. In 1995, fewer than 100 reports of mammalian or avian DNA vaccines had been published and by June 1999 this number exceeded 900. This intensity of research activity was not realized in fish which is reflected in most recent reviews (Heppell and Davis, 2000; Anderson and Leong, 2000) listing altogether not more than seven fish-related NAV publications. This rather moderate growth is surprising compared to the vigor with which more conventional aspects of fish immunology have recently been described (Pastoret et al., 1998). On the other hand, the relatively late exploration of immunology in fish can perhaps partially explain the lag in NAV-associated work, as we are only now beginning to understand the molecular basis of antigen recognition, lymphocyte heterogeneity and antibody diversity in fish. Similarly, the molecular characterization of many fish pathogens is still in its infancy. A recent surge in the number of fish-related NAV articles to over 20 by April 2000, however, may indicate the end of a lag phase for NAV research in fish. The importance of fish species as research models has been recognized for several reasons: the rapid growth of intensive aquaculture which has given rise to a need to better understand the nutritional, metabolic and disease resistance characteristics of farmed species, comparative studies seeking to understand the evolution of physiological processes taking advantage of certain ancestral characters found in fish, and

170

I

Simon R. M. Jones

the low cost and ease with which statistically adequate sample sizes can be accommodated within study designs. To better understand the role of plasmid DNA in fish vaccination, it is important to first have an appreciation of fish and their defence mechanisms and of the function and efficacy of conventional vaccines.

2

Fish

Fish are a diverse assemblage, encompassing between 20,000 and 30,000 species that inhabit both fresh and salt water environments and represent over 60% of all vertebrates (see Table 1). With only 1,000 species of cartilaginous sharks and rays, the vast majority of fish are teleosts possessing a bony skeleton. In addition to their antiquity (tetrapods diverged from an ancestral sarcopterygian fish approximately 300 million years ago), fish are readily distinguished from most other vertebrates by their adaptation to an aquatic life: gills for gas exchange, fins for locomotion, thin skin that is often protected with scales of dermal origin and with mucus secreted by epidermal mucous cells. Most fish are poikilothermic (coldblooded) and the rate of physiological processes, including those associated with disease resistance, are determined by environmental temperature. The intimate association of fish with their aquatic environment is reflected in an acute sensitivity to alterations in the chemical and thermal composition of water. Despite these apparent vulnerabilities, many fish demonstrate exquisite capacity for physiological acclimatization within a range of environmental extremes.

Common and scientific names o f fish species in which nucleic acid vaccination has been described

Table 1.

Common Name

Scientific Name

I

I

Zebrafish

Danio rerio

Goldfish

Carassius auratus

Rainbow trout

Oncorhynchus mykiss

Atlantic salmon

Salmo salar

Sockeye salmon

Oncorhynchus nerka ~~

Common carp

Cypnnus carpio

Tilapia

Oseochromis niloticus

I

70 Plasmids in Fish Vaccination

3

Fish immunology

As mammalian immunology has been largely a reflection “of mice and men” (Pastoret et al., 19981, so too has fish immunology focussed primarily on a core group of four species: rainbow trout, Atlantic salmon, common carp and channel catfish. Studies on these and on relatively primitive species, including sharks and the agnatha (lampreys and hagfish), have revealed that many of the defence mechanisms of mammals and birds also occur in fish. This suggests compelling evidence of the early appearance for these defence mechanisms during vertebrate evolution. 3.1 Innate defence mechanisms

Non-specific defence mechanisms play a proportionally greater role in disease resistance in fish than do adaptive, specific mechanisms, perhaps because of the relatively slow onset of the latter. Soluble non-specific substances important in fish defences are found constitutively or inducibly in mucus (lysozyme, agglutinins, precipitins, proteases) and in plasma (lysozyme, lectins, bactericidins, C-reactive protein, complement, transferrin, interferon and proteinase inhibitors). Non-specific cellular defences are based on many of the same leucocyte populations seen in mammals: monocytes, granulocytes and non-specific cytotoxic (lymphoid) cells. Collectively, these contribute to acute and chronic inflammation and to the release of pharmaco-active agents via degranulation and/or phagocytosis of foreign particles. Cells of the granulocyte and myeloid lineages are stimulated to higher levels of oxidative metabolism following stimulation by exogenous (LPS, peptidoglycan, ,B-glucan) and endogenous (?+interferon)factors. Soluble mediators secreted during this enhanced metabolic activity that contribute to host defence include vasoactive amines or proteins, eicosanoids and cytolcines (Secombes et al., 1996). In addition, although differences have been observed among some species, killing mechanisms associated with fish phagocytes (neutrophils, macrophages) include reactive oxygen (OF, H,Oz) (Chung and Secombes, 1987) and nitrogen (NO) species (Neumann et al., 2000). Natural (non-specific) cytotoxic cells, capable of selectively ltilling target cells of host or parasitic origin on contact have also been reported in cartilaginous and bony fish (Secombes, 1996). 3.2 Adaptive defence mechanisms

It is beyond the scope of this chapter to detail the teleost immune system; recent reviews serve this purpose very well (Iwama and Nalcanishi, 1996; Pastoret et al., 1998). Rather, it is most useful here to emphasize those features of adaptive immunity in fish that distinguishes it from that of homeotherrnic vertebrates (birds, mammals), particularly those features that arc probably important in contributing to protective immunity.

I

171

172

I

Sirnofl

R. M. Jones

Fish do not possess a lymphohematopoietic bone marrow, rather the kidney, particularly the anterior portion or pronephros, is hematopoietic. Together with the thymus, the pronephros serves as the primary lymphoid organ. Secondary lymph organs include spleen in teleosts and spleen, Leydig's organ and epigonal tissue in chondrichthyes. While the large lymphoid aggregates common in birds and mammals such as the bursa of Fabricius, Peyer's patches and lymph nodes are not present in fish, loosely defined clusters of lymphoid cells occur in the intestine and gills of fish and probably play a similar role in the surveillance of foreign antigens. In particular, pigmented clusters of myeloid-like cells in head kidney and spleen, known as melonomacrophage centers (MMC), are associated with antigen retention and may play a role with antigen processing possibly influencing the development and duration of immunological memory in these species. It is tempting to speculate that MMCs serve to provide a microenvironment conducive to antigen presentation analogous to germinal centra within draining lymph nodes in mammals. That antigen processing in teleosts uses a similar machinery to that in mammals is supported by recent observations. MHC class I and I1 genes and cDNA have been studied in several species of teleosts and chondrichthyes. MHC I and IIP expression was described in Atlantic cod, common carp and Atlantic salmon (Rodriguez et al., 1995; Koppang et al., 1998; Persson et al., 1999) and the sequence of a polymorphic MHC class I1 gene was reported in Pacific salmon (Miller and Withler, 199G).While it is expected that all nucleated cells will express class I molecules, the distribution and factors regulating the expression of class I1 molecules require further investigation. The identification of genes encoding proteasome subunits and TAP1 and TAP2 in several species provides additional support (Ohta et al., 1998; Murray et al., 1999). Fish analogs of the MHC-restricted, antigen-presenting dendritic cells of mammals have not been reported. Lymphocyte heterogeneity in fish is defined primarily by using functional criteria adopted from mammalian immunology. Thus in salmonid, cyprinid and ictalurid species, lymphocytes have been segregated according to expression of surface Ig (sIg). Furthermore, putative B cell (sIg+)and T cell (sIg-) subpopulations proliferate following exposure to LPS or ConA (or PHA), respectively. Co-cultivation of heterologous catfish leucocytes initiate mixed lymphocyte reactions involving Iglymphocytes. Thus, although there is substantial indirect evidence for considerable sophistication in the orchestration of the teleost immune system, surface markers capable of distinguishing lymphocyte subpopulations (CD3, CD4 or CD8, for example) have not yet been described. Genes encoding putative TCR b-chains, however, have been cloned from horned shark (Hawke et al., 199G),rainbow trout (Partula et al., 1995),Atlantic salmon (Hordvik et al., 1996) and channel catfish (Zhou et al., 1997). Similarly, the TCRa-chains have been cloned and sequenced from rainbow trout and other species (Zhou et al., 1997).Although as many as 13 cytoltines have been reported from five teleosts and a chondrichthian species based on cDNA sequence homology, serological cross-reactivity using mammalian-specific reagents and functional assays (Pastoret et al., 1998), there is still insufficient evidence to assign functional dichotomies, such as those equivalent to mammalian

10 Plasmids fn Fish Vaccinatm

Thl and Th2 subpopulations of CD4' lymphocytes. Thus, considerable effort is now being devoted to the identification of lymphocyte markers, to identifying cytolcines and correlating their activities to specific cell fractions and to establishing the significance of the many putative MHC genes (or cDNAs) that have been described. Although direct evidence is lacking, these data collectively support the notion that antigen presentation in fish is mediated in part by T cells and is MHC restricted. The ability to secrete antigen-specific Ig is a hallmark of vertebrates from chondrichthians to mammals. Convincing evidence for these molecules in the Agnatha is lacking. However, a significant divergence from the mammalian paradigm exists in fish with respect to the structure, isotypic variation and epitope-specific repertoire of secreted Ig. Fish Ig is a large molecule (-800 kDa) composed of heavy (H) and light (L) chains; each of which is organized into constant (C) and variable (V) domains, the latter bearing antigen binding sites. Teleost Ig has a potential valency of eight (tetrameric) and that of chondrichthyes is 10 (pentameric).Their large size and multiple valency has led workers to describe these molecules as IgM-like, despite the fact that the amino acid sequence of the four CH domains of catfish shares only 24% homology with that of the mouse .u-chain. Isotypic (H-chain) variants of the IgM-like molecule have been reported within several fish species. In most cases this variation has apparently not been sufficient to justify the recognition of heavy chains other than the .u-type. An ability to switch to C, y , 8, E or a chains, with their corresponding functional capabilities including affinity maturation, has not been reported for fish. However, since structural diversity among Ig molecules can also be influenced by post-translational processes such as glycosylation or variable disulfide bond formation, it has been hypothesized that in fish, functional Ig heterogeneity may rely more heavily on epigenetic processes (Kaattari and Piganelli, 1996). Thus, while immunological memory is described in fish, it is characterized essentially by enhanced secretion of the IgM molecule. The relatively limited sophistication of the fish adaptive immune system undoubtedly reflects the ancestral position of this poililothermic vertebrate class and its adaptation over several hundred million years to the aquatic environment. Another significant factor, one that is shrinking rapidly however, is the lack of sufficient experimental data to fully elucidate the similarities and differences with immune mechanisms of terrestrial vertebrates. Of particular pragmatic interest is the extent to which the unique features of the teleost defence repertoire influences the opportunities for vaccination.

4

Vaccination of fish

Annual global production of farmed salmonids now exceeds 1.2 million metric tonnes with an estimated value of $4 billion. Vaccination, most frequently by injection, but also by direct immersion and infrequently by oral delivery, has contributed significantly to the growth of the salmonid aquaculture industry by reducing

I

173

174

I mortality and morbidity associated with bacterial diseases and has reduced depenSimon R. M. Jones

dency on antibiotics. In contrast, vaccination of salmon against diseases caused by viruses, parasites and other intracellular pathogens has proven to be more challenging, due partly to technological constraints in antigen production and to possible limitations both in the immunological capacity of poikilothermic animals and in the defence mechanisms elicited by conventional formulations of inactivated antigen. Diseases of viral origin are becoming more significant as causes of economic loss in finfish aquaculture. The success of contemporary husbandry practices including the use of highly effective vaccines has reduced the impact of many diseases of bacterial origin. In contrast, management of virus-associated diseases has been largely restricted to avoidance and were necessary, culling of affected stock. Effective vaccines against the majority of viral diseases are not available. Viruses belonging to the Rhabdoviridae are among the most thoroughly studied of pathogens with significance to aquaculture. These include IHNV and VHSV. It is not surprising, therefore, that the feasibility of NAV be established using these well-defined models.

5

Nucleic acid vaccination o f fish

Two parallel and equally compelling forces are driving the application of NAV in fish. First, this method provides opportunities to elucidate fundamental characteristics of fish immunology such as antigen trafficking, MHC restriction of antigen presentation and the identification, distribution and significance of cytokines and their receptors. Secondly, and possibly as a consequence of an improved understanding of fish immunology, vaccination with plasmid DNA may provide a rational method of effectively preventing diseases caused by many of the intractable viral, parasitic and intracellular bacterial agents that cause disease in farmed fish. Since the mid 1990s, it has become clear that the expression of heterologous peptides, coded by parenterally administered plasmid DNA occurs in fish and that immunological responses are elicited by these peptides. In addition, when the foreign peptide is an antigen of an infectious agent, its expression has elicited a protective response. It is still premature to draw conclusions regarding “typical” fish responses as results to date represent data from Atlantic salmon, rainbow trout, gold fish and zebrafish. Furthermore, the reports often describe plasmid DNA of unknown quality (% super-coiled, endotoxin content, etc.) expressing a variety of peptides (see Table 2) usually following intramuscular injection. Despite the inevitable divergence in methodologies and observations that occur early during independent investigations into a new field, some clear trends are beginning to emerge. The remainder of this chapter will review the literature to date (April, 2000) highlighting similarities and differences in the application of plasmid DNA vaccines in fish.

70 Plasmids in Fish Vaccination 6

Plasmid constructs used in fish studies

A total of eight promoter/enhancers have been reported in studies involving intramuscular injection into seven fish species (Table 2). While the vast majority of studies described the efficiency of the CMV immediate early promoter in several fish species, CBAP was similar to CMV in driving expression of luciferase in Atlantic salmon (Gomez-Chiarri and Chiaverini, 1999). Similarly, CBAP-driven expression of luciferase in rainbow trout exceeded that driven by the glucocorticoid-responsive MMTV promoter (Anderson et al., 1996a).A broader variety of promoters, particularly those of fish origin, in the context of various gene constructs, must be studied to properly appreciate their relative strength, particularly given possible concerns of commercial vaccines containing promoters of viral origin (Gomez-Chiarri and Chiaverini, 1999). The expression of relatively few antigens have been studied in fish following NAV (Table 3). These are broadly divided here for convenience into antigens derived from infectious agents and those derived from non-infectious agents. The paucity of ltnowledge regarding genes encoding protective antigens in pathogens of importance in cultured fish species i s reflected in the short list. The RNA genomes of two rhabdoviruses,IHNVand VH SV, have been thoroughlycharacterized. The rhabdoviral transmembrane glycoprotein elicits neutralizing serum antibodies and i s therefore the focus of several NAV studies in fish. The bacterial pathogen Renibacterium salmoninarum, causative agent of bacterial kidney disease in farmed salmon, produces a Table 2.

Promoters tested by injection into muscle of various fish species

Promoter

Fish Species

Glucocorticoid-responsive mouse mammary tumor virus (MMTV)

Oncorhynchus mykiss

Cytomegalovims (CMV) immediate early (alone or with translational promoter)

Oncorhynchus mykiss Oncorhynchus nerka Salmo salar Danio rerio Carassius auratus

Carp /3-actin

0ncorhynchu.s mykiss Oreochromis niloticus

SV40 early

Cyprinus carpio

Rabbit /3-cardiac myosin heavy chain (MHC)

Cypnnus carpi0

Human MxA

Cyprinus carpio

Herpes simplex virus thymidine lcinase (CMV enhancer)

Oncorhynchus mykiss

Killifish lactate dehydrogenase-B

Salmo salar

I

I

175

176

I

Simon R. M. Jones Table 3.

Antigens expressed following intramuscular injection o f fish with plasmid DNA

Fish Species

Antigen

Non-infectious Origin Luciferase

firefly

sea pansy

6

I Oncorhynchus mykiss Danio rerio Carassius auratus Salmo salar

Escherichia coli p-galactosidase

Carassius auratus Oncorhynchus mykiss

Mouse granulocyte-macrophage colony stimulating factor (GM-CSF)

Carassius auratus

~~~

~

~~~

Green fluorescent protein

Oncorhynchus mykiss

Infectious Origin

Infectious hematopoietic necrosis virus (IHNV) glycoprotein

Salmo salar Oncorhynchus mykiss

IHNV nucleocapsid protein IHNV non-virion protein IHNV matrix protein

Oncorhynchus mykiss Oncorhynchus mykiss Oncorhynchus mykiss

Viral hemorrhagic septicemia virus (VHSV) glycoprotein

Oncorhynchus mykiss

VHSV nucleocapsid protein

Oncorhynchzis mykiss

Renibacterium salmoninarum p57 protein

Oncorhynchus mykiss

soluble, immunogenic protein known as p57. Whether peptides encoded for by the few other described genes of fish pathogens elicit a protective immunity has received scant attention. Thus much work is still needed to characterize the genomes of the majority of fish bacterial, viral and parasitic pathogens as a prerequisite for NAV.

7

Routes of plasmid administration 7.1

Intramuscular injection of plasmid DNA

The relatively large muscle mass of teleosts, consisting of dorsal and ventral metameric blocks along the entire axial skeleton, makes this tissue an obvious target for DNA immunization, particularly given the stellar successes of this delivery route in

10 Plasmids in Fish Vaccination

small mammals. The structure of striated, skeletal muscle of fish is microscopically similar to that of other vertebrates: individual muscle cells or fibers are elongate, multinucleate and surrounded by a delicate endomysium. Unlike that in mammals, striated skeletal muscle of fish is arranged in a longitudinal series of metameric myomeres, separated by sheets of connective tissue or myosepta. As the function of this muscle architecture is to confer a sinuous, caudal movement to the axial skeleton, the entire muscle mass undergoes continuous rhythmic contractions as the animal swims. In terrestrial tetrapods, ancestral metamerism of the axial musculature is largely obscured by well-developedmuscle bundles of the appendicular skeleton. In addition, mammalian muscle has levels of organization not seen in fish muscle: individual endomysium-enclosed fibers are contained within fasciculi, each contained by a perimysium. The grossly recognizable muscle is a collection of fasciculi surrounded by a thick layer of connective tissue, the epimysium. The endomysium is the connective tissue framework within which capillaries supply myocytes and is similar in its pattern of vascularization to that of fish muscle. Histiocytes associated with teleost musculature are largely myeloid in origin and are represented by migratory macrophages. The few studies comparing NAV in mice and fish suggest that structural differences between teleost and murine striated muscle may help explain differences in plasmid uptake, expression and/or a subsequent immunological response. Luciferase activity was consistently higher in 0.5-2.0 g rainbow trout muscle and in 0.20.9 g zebrafish following i. m. injection with 0.1-50 pg pCMV-luc compared with that in six to eight week balb/c mouse anterior tibialis (Heppell et al., 1998a). Similarly, four weeks after injection with 35 pg pCMV-lacZ, anti-,&gal antibodies were detected in only five of 10, six week balb/c mice and in all 10, 4-10 g goldfish. By eight weeks, nine of 10 mice had seroconverted (Russell et al., 1998),suggesting that levels of expressed antigen were poorly immunogenic in the mice used in this trial. These authors suggested that the relatively simple muscle structure of fish may have promoted greater accessibility of DNA to myocytes or, conversely, that greater compartmentalization of mammalian muscle restricted movement of plasmid DNA to other tissues, possibly restricting more widespread antigen expression. The relatively small target presented by the anterior tibialis for injection may also influence the volume of DNA reaching myocytes. Studies in mice, however, have shown that surgical removal of the limb receiving plasmid DNA very shortly after injection had no impact on subsequent immunity, suggesting rapid transport of DNA or antigen, perhaps via draining lymph, from the injection site (Lewis and Babiuk, 1999). Thus it is possible that luciferase levels in mouse muscle were lower because of more efficient capacity to transport plasmid or antigen from the site of injection. Clearly, more comparative data are required from a much wider variety of fish and mammalian species to begin to understanding how muscle structure affects DNA uptake and subsequent expression.

I

177

178

I

Simon R. M.Jones

7.2 Other routes of plasmid administration

Limited information is available on administration of plasmid DNA into fish by routes other than i.m. injection. Gomez-Chiarri et al. (199Gb) reported B-gal expression in epidermis, dermis and muscle tissues of rainbow trout following bombardment with plasmid labeled gold particles. The same authors reported luciferase expression in muscle and skin or eye homogenates of rainbow trout following gold particle bombardment. Particle acceleration, as measured by helium pressure, and particle size were important determinants of foreign antigen expression (Gomez-Chiarri, 199Gb). The latter study concluded, however, that injection is rapid, simple, reproducible and results in significantly higher levels of gene expression compared with particle bombardment. More recently, rainbow trout were immunized against infectious hematopoietic necrosis virus (IHNV) by i. p. injection and by particle bombardment with plasmid DNA, but not by immersion, epidermal scarification or buccal gavage with plasmid (Corbeil et al., 2000b). It is interesting to note that while Corbeil et al. (2000b) were able to detect gold particles only as deep as the dermis at the site of particle bombardment, intramuscular expression of b-gd was reported in Atlantic salmon following particle bombardment (Gomez-Chiarri, 1996b). Fernandez-Alonsoet al. (1999) used expression of GFP in fins as evidence for the uptake of DNA during immersion of rainbow trout in 10 pg mL-' of plasmid pQBI,,. However, it is still not clear whether expression by myocytes is a necessary prerequisite for the development of protective immunity in fish or whether antigen expression and recognition by cells associated with the dermis will also serve this purpose.

8 Fate of injected plasmid DNA

Comparatively little is known concerning the fate of plasmid DNA following intramuscular injection in fish. Cell membrane fluidity is dependent on cholesterol content and in salmonids this is strongly dependent on acclimation temperature (Robertson and Hazel, 1995). The extent to which the altered fluidity of salmonid cell membranes affects uptake of naked DNA or polycation-DNA complexes is not known. In addition, the critical importance of DNA stability in the cytosol as a predeterminant of delivery to the nucleus (Lechardeur et al., 1999)has not been examined in fish. There is no evidence, however, that plasmid DNA integrates into fish genomic DNA following in vivo transfection. DNA extracted from rainbow trout muscle was examined 63 days following injection with plasmid pCMV4EAL (expressing firefly luciferase) (Anderson et al., 1996a). Southern blotting of restriction-digested DNA obtained from injection-site muscle, but not from muscle obtained from other sites, revealed a band with the predicted migration pattern of unreplicated, unintegrated plasmid DNA. Similarly, PCR was used to demonstrate the integrity of the VHS-glycoprotein gene for up to 45 days following i. m. injection of

10 Plasmids in Fish Vaccination

plasmid DNA into rainbow trout (Boudinot et al., 1998). In goldfish, super-coiled, circular and linear plasmid DNA was found only in injection-site muscle 70 days following injection (Kanellos et al., 1999~).Using Southern blotting, a pcDNA3.1 probe recognized a DNA fragment of the predicted size in injection-site muscle but not elsewhere (Icanellos et al., 1999a). Studies to date have involved sexually immature fish under laboratory conditions, thus no data has been generated on the occurrence of plasmid DNA in gametes or ovarian fluids. However, it is not unreasonable to expect that unincorporated DNA will rapidly degrade in the presence of nucleases present in the serum and cytosol of fish as they do in mice (Kawabata et al., 1995). A more determined effort is required to fully appreciate the longevity and distribution of plasmid DNA following injection in a wider variety of species.

9 Magnitude, distribution and longevity of expressed antigen

Foreign antigen is readily expressed by fish muscle cells at the site of plasmid DNA injection. While transfection efficiency (percent of cells expressing antigen) has not been explicitly described in fish species, relative frequency of @-gal-positivemyocytes peaked 21 days after injection of goldfish maintained at 22 "C and /I-gal negative fish began to appear on day 28 (ICanellos et al., 1999a).Factors affecting antigen expression are beginning to be understood for fish and include temperature, plasmid concentration and volume of DNA injected. The importance of transcriptional promoters was discussed earlier. The duration of antigen expression in goldfish was inversely correlated with the temperature at which fish were maintained. Thus, the number of muscle cells expressing @-galpeaked one or four weeks after plasmid injection at 25 "C and 15 "C, respectively, and declined thereafter. In fish held at 9 "C or less, the number of D-gal positive cells continued to increase for up to 18 weeks. A dose effect was also evident as luciferase activity in rainbow trout and zebrafish muscle increased in proportion to plasmid concentration between 0.01 pg and 1.0 pg (Heppell et al., 1998a) and in 7-15 cm rainbow trout between doses of 10 pg and 50 pg (Gomez-Chiarri, 199Gb). Luciferase activity was inversely correlated with the volume in which 1 pg plasmid was injected into rainbow trout muscle (Heppell et a]., 1998a). In contrast, Anderson et al. (199Ga) reported greater reproducibility of luciferase expression in rainbow trout following i.m. injection with 25 pg plasmid in 200 pL compared with 100 pL. The former observations suggest several possibilities including disproportionately greater loss of more diluted DNA following injection with larger volumes, physical trauma to tissue following injection with larger volumes or simply that dilution of plasmid DNA in the larger volume reduced its concentration below a threshold suitable for uptake and detectable expression. Luciferase was detected as early as two hours following i. m. injection of 0.2-0.9 g zebrafish with 10 pg DNA and peaked at about five days both in zebrafish and in 0.5-2.0 g trout (Heppell et al., 1998a). Similarly luciferase activity peaked seven

I

179

180

I days after injection of 50 g trout with Simon R. M.Jones

25 pg DNA (Anderson et al., 1996a).The persistence of luciferase activity in muscle was found to be similar in both studies particularly in 1.0 g trout where luciferase was still evident 115 days after injection. In contrast, luciferase activity continued to increase for 60 days following i. m. injection of Atlantic salmon with 25 pg of plasmid pCMVtlduc (Gomez-Chiarri et al., 199613).The presence of both a thymidine ltinase promoter and CMV enhancer in the plasmid used in the latter study may help explain the differences observed in the patterns of luciferase expression. In either case, luciferase is relatively nonimmunogenic in mice (Wolff et al., 1990) and poor iminunogenicity in fish may help explain the duration of expression of this antigen, as suggested by Heppell et al. (1998a). Messenger RNA transcripts of the VHSV glycoprotein (vhsG) were detected within injection site myocytes of two rainbow trout on each of 7, 14 and 21 days after injection with 30 pg plasmid DNA (Boudinot et al., 1998).The latter study reported that inflammatory cells infiltrating the injection site also immunostained positive for VHSV glycoprotein. In another study, the glycoprotein of VHSV was directly detected by immunohistochemistry in 4 of 16 rainbow trout sampled two to three weeks after injection with 10 pg or 50 pg of plasmid pCMV-vhsG (Lorenzen et al., 1998). Antigen expression in tissues not associated with the site of injection has been reported only for luciferase. The enzyme was evident in gills both of 0.2-0.9 g zebrafish and of 0.5-2.0 g rainbow trout (Heppell et al., 1998a) and in non-injection site muscle, kidney, heart, liver, spleen and gill of 50 g rainbow trout (Anderson et al., 1996a). Expression in non-injection site tissues, possibly due to translocation via inflammatory cells, was 200- to 1,000-fold lower than at the site of injection. What is relevant to the development of meaningful vaccination programs is a clear understanding of the factors affecting duration and distribution of antigen expression and whether these are vector-, antigen- or species-specific. Following coilventional vaccination, antigen is retained in lymphoid tissues of fish (Press et al., 1995), thus providing a partial explanation of the duration of the observed protection. It is compelling to speculate that the presence of foreign antigen in lymphoid hematopoietic tissue is a necessary requirement for NAV to elicit long-lasting protection and, therefore, to be commercially successful as the importance of immunological memory in contributing to longevity of protection in fish is particularly poorly understood. Thus, the increasing dependency on vaccination by the aquaculture industry is in spite of a failure to fully appreciate the protective mechanisms that are elicited. This is particularly apparent for those organisms against which the application of NAV technology is most likely to be successful: intracellular and parasitic pathogens. Therefore, the success of candidate nucleic acid vaccines will initially be based on their performance in fish against experimental or natural challenges in much the same way as present-day conventional vaccines are tested, rather than on a well-defined repertoire of immunological mechanisms. Coincidental opportunities to capitalize on the antigen-specificresponse mechanisms elicited by NAV have begun to be explored and these are described below.

70 Plasmids in Fish Vaccination I 1 8 1 10 Responses

of fish to injection with plasmid DNA

10.1 Inflammatory responses

Intramuscular or intraperitoneal injection or epidermal delivery by ballistic bombardment with plasmid DNA results in the expression of foreign antigen by fish as described above (see Table 3). The responses of the fish to these processes can be categorized either as inflammatory or immunological. Inflammation of muscle attributable to the presence of plasmid DNA or to the expressed foreign antigen is rarely observed in fish. Thus, Gomez-Chiarri (1996b) reported inflammatory cell infiltrate and tissue damage associated only with needle track injury in Atlantic salmon injected with pCMVtldacZ. Similar observations were made in rainbow trout muscle following injection with 19 pg pCMV-lac2 in 10 yL (Heppel1 et al., 1998a). Muscle cells expressing P-gal appeared morphologically normal 28 days after injection. Transient tissue damage observed both in saline controls and plasmid-injected fish up to 14 days after injection however, was considered to be injection-associatedtrauma (Heppell et al., 1998a). The absence of plasmidor foreign antigen-associated inflammation suggests that plasmid vaccines delivered by i. m. injection will be safe. However, histological examinations are required following injection with a much greater variety of plasmid constructs, particularly those with proven ability to stimulate immunity, before such generalizations are valid. Specific and non-specific immunological responses have been observed in fish following NAV Furthermore, when the expressed antigen was derived from an infectious agent, protective immunity has been reported. Thus, the responses of fish to expressed foreign antigens will be discussed separately: those derived from nonpathogenic sources (avimlent antigens) and those derived from pathogenic sources (virulent antigens). 10.2 Avirulent antigens

Serum antibodies are produced by fish in response to the expression of certain foreign antigens following injection with plasmid DNA. Thus, while firefly luciferase is apparently as non-immunogenic in fish as it is in mice (Davis et al., 1997), E. coli /?-galis strongly immunogenic and stimulates the production of specific antibodies following i. m. or i. p. injection in goldfish (Russell et al., 1998; Kanellos et al., 1999~;Russell et al., 2000). All 20 goldfish receiving 35 yg plasmid pCMV-lac2 had seroconverted by four weeks and were still seropositive with similar titres by eight weeks after injection (Russell et al., 1998).Titres of /?-gal-specificserum antibodies were in proportion to the dose of plasmid between 1 yg and 125 yg of DNA, but were significantly reduced if plasmid DNA had been linearized prior to injection. Similarly, the titer was significantly reduced if injection site muscle tissue was

182

I

Simon R. M. Jones

disrupted by prior treatment with barium chloride, reinforcing the importance of myocytes in foreign gene expression and possibly in antigen presentation. Feeding level (percent body weight per day) affected the onset of production of P-gal specific antibodies in goldfish i.m. injected with 50 fig pcDNA3.1/His/LacZ. Thus two weeks after injection, fish fed at the 3 % level had significantly higher titers than those fed at the 0.33 % level. By four weeks, however, there was no difference in mean titers despite a 20% loss in body weight among fish fed at the lower level (Russell et al., 2000). Temperature also affected the rate at which goldfish seroconverted. Interestingly, although seroconversion occurred most rapidly in fish held at 25"C, the highest titers were observed in the 15°C group when observed at 18 weeks. At this time, titers were lower but still increasing in the group held at 9°C (Russell et al., 2000). In another study (Kanellos et al., 1999c), p-galspecific antibodies were detected in goldfish following intraperitoneal injection of plasmid, suggesting that DNA vaccines may be successfully delivered to fish by i. p. injection, the same route used to deliver conventional vaccines. Therefore, antibody production in plasmid-injected goldfish appears to show similar relationships with temperature and dose as that seen in antigen-injected fish and this requires confirmatory studies in salmonid species. P-Galactosidase-sensitized lymphocytes were detected in goldfish following NAV (Kanellos et al., 1999a; Kanellos et al., 1999b).Stimulation indices observed in vitro peaked approximately 28 days after injection and were enhanced by coinjection with plasmid encoding murine granulocyte-macrophagecolony-stimulating factor (GM-CSF).The number of P-gal-expressingmuscle fibers was significantly reduced in goldfish coinjected with plasmids encoding P-gal and GM-CSF compared with those coinjected with the pcDNA3 blank, suggesting development of a cytotoxic response in association with expression of the cytoltine (Kanellos et al., 1999b). Clearly, expression of ,&gal in goldfish is a valuable model for understanding some fundamental relationships governing immunological mechanisms following NAY These observations however, while demonstrating potential to drive the development of NAV strategies should be repeated using fish species and antigens of greater significance to aquaculture. 10.3 Virulent antigens

Ultimately, the value of nucleic acid vaccination lies in its ability of confer protection, measured either as increased survival or as reduced morbidity following exposure to an infectious agent under farm conditions. Numerous controlled, experimental studies have reported increased survival in nucleic acid-vaccinatedfish following virulent challenges, the majority with the fish rhabdoviruses IHNV and VHSV. Despite their experimental nature, these studies have provided important observations that will ultimately help define commercial nucleic acid vaccines. Anderson et al. (199Gb) were first to report that NAV elicited protective immunity in fish. In that study, 1 g rainbow trout were i.m. injected with 10 fig pCMV4-G or pCMV4-N (encoding the IHNV glycoprotein or NP, respectively) or with both plas-

10 Plasmids in Fish Vaccination

mids concurrently. Fish were challenged by immersion in 5.0 x lo7 PFU per L of IHNV (Rangen strain) six weeks (-500 degree-days)after vaccination. Significantly more trout that were vaccinated with glycoprotein encoding plasmid (alone or in combination) survived the challenge. No increase in survival was reported among fish vaccinated with the NP-encoding plasmid. Subsequent studies using rainbow trout or Atlantic salmon have reported similar results while examining the effects of DNA dose, routes of administration, onset of immunity, cross-protection and other viral antigens. Thus Lorenzen et al. (1998)reported 97 and 94 relative percent protection (RPS) (RPS = 1-[%vaccinatemortality/% control mortality] x 100)following VHSV challenge of 13 g rainbow trout injected with 50 pg or 10 pg, respectively of pCMV-vhsG (glycoprotein encoding). Protection was lower (RPS = 66) but still significant in fish vaccinated with 5 pg pCMV-vhsN (NP). Subsequently, and conforming to earlier findings, Lorenzen et al. (1999) found no protection against VHSV in trout vaccinated with plasmid encoding viral NP. Similarly, 26 pg of pCMV4-G, expressing IHNV glycoprotein, elicited protection in Atlantic salmon presmolts (RPS = 96 or 100) or smolts (RPS = 90 or 93), each challenged by immersion or cohabitation with infected fish, respectively (Trader et al., 1999). Corbeil et al. (1999) have recently confirmed that i.m. injection with 1.0 pg of IHNV glycoprotein-expressing plasmid protected 2 g rainbow trout. In contrast, neither NP- (1 p g ) , non-virion protein- (1 pg) nor phosphoprotein- (5 pg) expressing plasmids elicited protection when challenged four weeks after vaccination. Similarly, matrix protein-expressing-plasmid (10 pg) failed to elicit protection after six weeks (Corbeil et al., 1999).While the authors cautioned that expression of NP, matrix protein, phosphoprotein or non-virion protein had not been confirmed, the study emphasized the importance of immunizing with an appropriate antigen to maximize protective immunity. The minimal dose of plasmid DNA necessary to elicit protection was independently investigated for VHSV (Lorenzen et al., 1999) and for IHNV (Corbeil et al., 2000a). Both studies demonstrated that significant protection could be elicited in rainbow trout with as little as 10 ng or with 1 ng plasmid DNA (Corbeil et al., 2000a). A common deficiency in many published accounts of NAV in fish however, is a failure to describe precisely how plasmid DNA is purified and quantitated (see also Chapters 2 and 11).Although spectrophotometric determination of nucleic acid concentration is routine, the model of spectrophotometer and diluent used may significantly affect the optical density and make interlaboratory comparisons difficult, particularly where fish are being vaccinated with nanogram quantities of DNA. Despite the above, it is clear that small trout (-1 g) are effectively immunized with minute quantities of plasmid DNA encoding rhabdoviral glycoprotein. Lorenzen et al. (2000) suggested, citing unpublished results of LaPatra et al. (2000),that the dose of plasmid DNA necessary to elicit protective immunity is related to the mass of the fish. In this vein, Corbeil et al. (2000a) reported that a ratio of 1-10ng DNA per 1 g fish compared favorably to ratios used successfully in mammalian studies. This appears to be borne out by Babiuk et al. (1999a) who described vaccination of cattle with from 400-1000 pg DNA (plus booster injections) and of sheep with from 200-1000 pg DNA (plus boosts). While of obvious economic ben-

184

I

Simon R. M. Jones

efit, the practise of vaccinating small fish with nanogram quantities of DNA raises questions concerning the longevity of protection, an important consideration for commercial vaccines. Protection against IHNV elicited by 1 pg pcDNA-G, while significantly greater than controls at all times, declined between 28 (2% mortality) and 80 days (49% mortality) after vaccination (Corbeil et al., 2000a). In contrast, Lorenzen et al. (2000) reported significant protection against VHSV elicited by 1 pg pCDNA3vhsG for as long as 168 days after vaccination. In the latter study protection had begun to wane by 168 days and the authors suggested protection may not be life-long at this dose. The relationship between DNA dose and longevity of protection will need to be properly established for each plasmid construct using standardized challenge methodologies for each species. The capacity of NAV to elicit protection in fish against heterologous strains of virus, and therefore be of practical value under farm conditions, was investigated by Lorenzen et al. (2000) and Corbeil et al. (2000a). In both studies, protection against heterologous strains was significant, but consistently lower than that seen against homologous challenge. Structural variability in the glycoprotein undoubtedly contributed to the reduced protection against heterologous challenge. Whether this is compensated for by the evident importance of non-neutralizing antibodies and/or non-specific immune mechanisms (see below) remains to be evaluated. There is very little information on the efficacy of NAV against bacterial challenge in fish. Gomez-Chiarri et al. (1996a),immunized 7-10 cm rainbow trout with two different DNA-based vaccines: a Renibacterium salmoninarum expression library comprised of 500 bp segments of genomic DNA and pGFPp57, a plasmid encoding the p57 surface protein of R. salmoninarum. Fish received two i. m. injections, spaced 10 days apart, each with 50 pg plasmid DNA. Compared with controls, both DNA vaccines significantly reduced mortality associated with R. salmoninarum challenge. While promising, considerable work is still necessary with this model as protective antigens are not yet known for R. salmoninarum nor are parameters associated with optimization of protection by NAV While NAV is able to induce protection in fish, the mechanisms contributing to protective immunity are still far from clear. Non-specific immunity is induced in response to viral infection (de Kinkelin and Dorson, 1973) and following injection of fish with plasmids encoding viral peptides (Boudinot et al., 1998).One stimulus of non-specific immunity is the occurrence in plasmid (and proltaryotic genomic) DNA of hypomethylated ISS (Pisetsky, 1996). ISS are associated with a PuPu-CpGPyPy oligodeoxynudeotide or CpG motif (Ieieg et al., 1995). The role of CpG in augmenting antibody production has been examined in goldfish. Antibody levels in fish co-injected with the motif AACGTT, present in the ampR gene within the expression vector, pcDNA3.1 (Invitrogen), and either recombinant p-gal or lacZ (in pcDNA3.1) were significantly increased compared to controls. Substituting a G for the C in the ampR motif eliminated the augmenting effect (Kanellos et al., 1999b). In the same study, an increase in antibody obtained with a synthetic ODN (GACGTT) however, was significant in mice but not in goldfish (Kanellos

TO Plasmids in Fish Vaccination

et al., 1999b).Similarly a third motif, present in the CMV promoter (GACGTC) of a different vector (containing the kanR gene),was ineffective in enhancing antibody levels in either mice or goldfish. The importance of 3’ and 5’ sequences flanking the CpG motif is widely recognized as important in affecting the ability of the ODN to stimulate immune cells (Lewis and Babiuk, 1999) and the observations of Kanel10s et al. (1999b) suggest that differences in flanking sequences may help define ODNs with species-specific capabilities to optimally stimulate immune cells. While cultured fish leukocytes display elevated oxidative metabolic activity and elaboration of poorly-defined cytolrines following exposure to mitogens or other surface moieties associated with foreign cells (e. g., LPS, peptidoglycan, b-glucan), similar studies using defined ODNs have yet to be published. Given the relative importance of innate and non-specific defence mechanisms in fish, there may be greater potential in fish compared with mammals for ISS to add significantly to protection afforded by NAY This hypothesis is worthy of timely examination with a wide variety of ODNs in several fish species and work in this direction appears to have been initiated (Heppell et al., 1998b). Meanwhile, a number of observations are suggestive of mechanisms that may be associated with cytoltine production by fish in response to ISS. Expression of Mx genes is initiated by interferon in response to viral infection and the Mx protein is believed to play a role in antiviral defence (Freses et al., 1996). Mx gene transcripts were reported in injection-site muscle 7, 14 and 21 days after injection of rainbow trout with pcDNLgVHS but not in those injected with pcDNAl alone (Boudinot et al., 1998). These observations were subsequently confirmed and Mx transcripts were detected as early as 10 hours after injection (Boudinot et al., 1999). The latter study also described vig2 (VHSV-inducedgene), a novel gene associated with lymphoid tissue of rainbow trout. The expression of vig-1 was upregulated in response to virulent or b-propiolactone-inactivatedVHSV or to injection with plasmid pcDNA1-GVHSV, encoding the VHSV glycoprotein. vig-2 was also induced via interferon-like soluble factors associated with the replication of IPNV, a fish birnavims. Interestingly, the deduced amino acid sequence of VIG-1 was 80 % homologous (over 291 residues) to that of the human cytomegalovirusinduced gene (cig-5).VIG-1 may play a role in regulating NO synthesis (Boudinot et al., 1999). Early expression of vig-1 and Mx genes following NAV may help explain protection against VHSVobserved as early as 8 days after vaccination (Lorenzen et al., 2000). Together these observations suggest a heretofore unappreciated sophistication in non-specific anti-viral defences mounted by fish. Nucleic acid vaccination may therefore provide a valuable tool for better understanding the roles played by ODN ISS in many of these processes. Immunological responses of fish following injection with plasmids encoding viral peptides include elaboration of serum antibody, possible cytotoxic T cell activity and protection against virulent challenge. Levels of anti-VHSV and anti-IHNV neutralizing antibodies (nAbs) were first detected among most fish examined 38 days after injection; higher titers being observed in both groups by 45 days (Boudinot et al., 1998).This study elegantly showed that levels of specific nAbs in rainbow trout co-injected with plasmids encoding IHNV or VHSV glycoproteins were

I

185

186

I

Simon R. M. Jones

similar to levels of the respective antibodies in fish receiving either plasmid alone, suggesting a lack of interference between the expression of these genes or in the processing of their respective antigens. Although the capacity of these sera to cross-neutralize heterologous virus was not demonstrated in vitvo, passive transfer studies demonstrated the protective quality of the serum against homologous challenge. The specificity of these reactions was suggested, however, by the failure of passively transferred serum from trout immunized with plasmid pcDNA-gRV, encoding the glycoprotein of the rabies virus, to protect against either IHNVor VHSV challenge (Boudinot et al., 1998). While protection against rhabdoviral diseases of fish are controlled mainly via nAbs (Lorenzen et al., 1990; Corbeil et al., 1999), that other defence mechanisms, including non-neutralizing antibodies, may play a role is suggested by the absence of detectable nAbs in a significant proportion of immune fish (Lorenzen et al., 1998; Traxler et al., 1999 (see Table 4)). In particular, the latter study demonstrated the ability of Atlantic salmon serum with low neutralizing activity (1:20) to confer protection in rainbow trout against challenge with IHNV Thus, the possibility exists that certain antibodies incapable of neutralizing virus in vitro, for example anti-nucleocapsid protein antibodies, display anti-viral functions in uivo (Lorenzen et al., 1990). A recurring theme in fish immunology is our poor understanding of the specific defence mechanisms that contribute to an observed protective immunity. Antibodies are quite frequently undetected in immune fish leading some authors to conclude that a cell-mediated mechanism is responsible for protection (Traxler et al., 1999; Corbeil et al., 2000b). Co-injection of rainbow trout with pCMV-luc and pCMV-vhsG or with pCMV-luc and pcDNA3 provided indirect evidence of a cytotoxic response (Heppell et al., 1998a). Luciferase activity declined significantly faster in fish receiving pCMV-vhsG leading the authors to suggest the presence of VHSV-glycoprotein-specificcytotoxic T lymphocytes, perhaps similar to that proposed by Kanellos et al. (1999b) in goldfish. More direct evidence of MHC-

Table 4. Survival and serum neutralizing antibodies among DNA-vaccinated rainbow trout following challenge with virulent infectious hematopoietic necrosis virus (Lorenzen et al., 1998)

Treatment Group

Percent Survival (sample size)

% Seropositive (sample size) a f e r vaccination a f e r challenge

50 pg PCMV-vhsG

97 (67)

nd

71 (21)

10 pg PCMV-vhsG

94 (66)

72 (18)

62 (21)

94 (69)

60 (20)

76 (21)

68 (66)

0 (12)

14 (21)

59 (68)

14 (21)

48 (21)

5 pg PCMV-vhsG

+

1.3 pg inactivated V H S V

I

10 Plasmids in Fish Vaccination

restricted T cell mediated immunity is limited due in part to the limited availability of inbred fish lines. Similarly, further work is necessary to confirm whether a poor or absent neutralizing antibody titer reflects total levels of antibody. The possibility that non-neutralizing antibodies contribute to protection via opsonization and enhanced phagocytosis should also be explored.

11

Regulatory issues and future directions

The rapidity with which DNA vaccines are developed and licenced for use in aquaculture will require that manufacturers satisfy the same criteria used to regulate other veterinary biologicals: safety, purity, potency and efficacy (Anonymous/ FDA 1996; Anonymous/CFIA 1998; see also Chapters 11 and 12). While the potential risks associated with NAV have been identified (Griffiths, 1995; Babiuk et al., 1999b), careful investigations, largely in mice and other mammals, have provided little evidence to support the induction of immunotolerance, the integration of plasmid DNA into the host genome (of germ-line or somatic cells), the induction of anti-DNA autoimmune reactions or hypersensitivity reactions associated with the adjuvant effects of certain nucleic acid motifs (Babiuk et al., 1998; Davis and McClusltie, 1999). Furthermore, Babiuk et al. (1998), based on observations of rats fed injection-site muscle from DNA-vaccinated calves, expressed confidence that humans consuming meat from DNA-vaccinated animals would not be at risk. A similar argument could be made for the consumption of DNA-vaccinated fish. The administration of plasmid DNA to organisms inhabiting an aquatic environment may require additional levels of regulation because of the potential for environmental impact, however slight. It is unfortunate that science cannot prove that DNA vaccines will be absolutely safe when applied to fish, however, extrapolation of the available information (e. g., shedding of plasmid DNA and transmission to non-target species, toxicity of plasmid DNA) suggests the environmental risk is extremely low (Babiuk et al., 1999b; J. Heppell, personal communication). Plasmid DNA has often been described as a third-generation vaccine, the result of an evolutionary continuum from the methods of Jenner and Pasteur. However, while our understanding of bacterial physiology and genetics has grown tremendously, the vast majority of licenced vaccines for use in fish are still based on inactivated products of bacterial fermentation. These first-generation products are inexpensive to produce and when used correctly, work very well to prevent many of the diseases present in salmon culture, and are therefore unlikely to be entirely supplanted by vaccines based on plasmid DNA. On the other hand, certain fish diseases of viral or parasitic origin or those with an obligate intracellular bacterial aetiology, for which no vaccines presently exist, are good candidates for NAV. A common practice in vaccination of salmonid fish is to immunize small animals while they are still relatively accessible in freshwater hatcheries. Immunization regimes requiring protection of (10 g fish often involve one or more immersions offish in vaccine. Alternatively, a single i. p. injection with an adjuvanted, multi-valent vac-

I

187

188

I cine will protect fish until harvest, Simon R. M. Jones

12-24 months later. In both cases, vaccination is valued at pennies per dose. Thus the challenges facing the commercial application of NAV in aquaculture are two-fold: first ensuring an economical product, and second, developing the means for integrating both levels of technology without compromising the performance of either. Future work should capitalize on the evident feasibility of i. p. injection of fish with plasmid DNA as a method of immunization (Kanelloset al., 1999c; Corbeil et al., 2000b). The commercial requirement of longevity of protection, preferably conferred by a single immunization, suggests that methods of improving the efficiency of DNA uptake by fish cells will need to be developed. In addition, the cell population most suitable for antigen presentation with respect to the development of protective immunity should be identified and methods for targeting plasmid DNA to those cells identified, possibly requiring alternative routes of delivery. In this regard, co-injection of plasmids encoding antigen and cytokines (Russell and ICanellos, 2000) may prove to be extremely valuable in species of significance to aquaculture. Perhaps the most exciting possibilities that immunization with plasmid DNA offer, however, are the unique opportunities to dissect the various defence mechanisms that contribute to the protective response in fish.

Acknowledgements

The author thanks Drs N. Lorenzen, S. LaPatra and P. Russell for making prepublication materials available. Dr Jeff Lewis, Atlantic Veterinary College, University of Prince Edward Island, provided a critical review of the manuscript.

References Veterinary Nucleic Acid Vaccines, pp. 1-2 ANDERSON, E. D., MOURICH, D.V, LEONG, (www.cfia-acia.agr.ca/english/anima/vetbio/ J. C., (199Ga),Gene expression in rainbow vb323e.pdf). trout (Oncorhynchus mykiss) following intramuscular injection of DNA, Mol. Mar. Biol. ANONYMOUS(1996), US Food and Drug Administration (FDA) Points to Consider on Biotechnol. 5 (2), 105-113. Plasmid D N A Vaccinesfor Preventive Infectious E. D., MOURICH,D.V., FAHRENANDERSON, Disease Indications, pp. 1-13 (www.fda.gov/ ICRUG, S. C., LAPATRA, S., SHEPHERD,J., cber/cberftp.html). LEONG,J. C. (1996b), Genetic immunization BABIUK, L.A., LEWIS, P. J., VAN DRUNEN of rainbow trout (Oncorhynchus mykiss) LITTEL-VAN D E N HURK,S., TIKOO,S., LIANG, against infectious hematopoietic necrosis X. (1998), Nucleic acid vaccines: veterinary virus, Mol. Mar. Biol. Biotechnol. 5 (2), applications, in: Current Topics in Microbiology 114-122. and Immunology Vol. 26. D N A Vaccination/ ANDERSON, E.D., LEONG, J.C., (ZOOO), DevelGenetic Vaccination (Koprowski, H., Weiner, opment of DNA vaccines for salmonid fish, D. B., Eds.), pp. 90-105. Springer-Verlag, in: Methods in Molecular Medicine Vol. 29, D N A Vaccines: Methods and Protocols (Lowrie, Berlin, Heidelberg. LITIEL-VAN D E N BABIUIC, L. A., VAN DRUNEN D. B., Whalen, R. G., Eds.), pp. 105-121. HURK,S., BABIUIC,S. L. (1999a),ImmunizaHumana Press Inc., Totawa, NJ. ANONYMOUS (1998),Canadian Food Inspection tion of animals: from DNA to the dinner plate, Vet.Immunol.Immunopathol. 72,189-202. Agency (CFIA) Guidelinesfor Licensing

10 Plasmids in Fish Vaccination

study fish DNA immersion vaccination by BABIUK,L. A., LEWIS, J., VAN DRUNEN LITTLEVAN D E N HURIC, S., BROWN, R. (1999b),DNA using the green fluorescent protein, J. Fish Dis. 22, 237-241. immunization: present and future, Adu. Vet. Med. 41,163-178. FRESE, M., KOCHS, H., FELDMANN, H., HERTKORN, c., HALLER, 0. (1996), Inhibition of BABIUIC,L.A., VAN DRUNENLIITEL-VAN D E N bunyaviruses, phleboviruses, and hantaHURK,S., LOEHR,B. I., UWIERA, R. (2000), Veterinary applications of DNA vaccines, in: viruses by human MxA protein, J. Virol. 70, 915-923. Development and Clinical Progress of DNA (;OMEZ-CHIARRI, M., BROWN,L. L., LEVINE, Vaccines. Deueloprnents in Biology Vol. 104 (Brown, F., Ciehutek, I<., Robertson, J., Eds.), R. P. (1996a), Protection against Renibacterium sulmoninarum infection by DNA-based pp. 73-81. Karger, Basel. immunization. International Congress on the BOUDINOT,P., BLANCO, M., DEKINXELIN, P., B E N M A N S O UA. ~ (1998), Combined DNA Biology of Fishes, Proceedings ofthe Aquaculture immunization with the glycoprotein gene of Biotechnology Symposium, pp. 155-157. San viral hemorrhagic septicemia virus induces Francisco State University. GOMEZ-CHIARRI, M., LIVINGSTONE,S., double-specific protective immunity and M., SANDERS, S., LEVINE, R. P. nonspecific response in rainbow trout, MURO-CACHO, Virology 249, 297-306. (1996b), Introduction of foreign genes into the tissue of living fish by direct injection and BOUDINOT,P., MASSIN, P., BLANCO,M., RrmAuLT, S., BENMANSOUR,A. (1999), uig-1, particle bombardment, Dis. Aquat. Urga-ganisms a new fish gene induced by the rhabdovirus 27, 5-12. glycoprotein has a virus-induced homologue GOMEZ-CHIARRI, M., CHIAVERINI, L. A. in humans and shares motifs with the MoaA (1999), Evaluation of eukaryotic promoters for the construction of DNA vaccines for family, J. Virol. 73, 1846-1852. C H U N GS., , SECOMBES, C. J. (1987), Activation aquaculture, Genet. Anal.: Bzomol. Eng. 15, of rainbow trout macrophages, J. Fish Biol. 121-124. 31A, 51-56. E. (1995),Assuring the safety and GRIFFITHS, CORBEIL, S., LAPATRA, S. E., ANDERSON, E. D., efficacy of DNA vaccines, Ann. NYAcad. S L ~ . JONES, J., VINCENT, B. et al. (1999), Evalua772, 164-169. tion of the protective immunogenicity of the HANSEN, E., FERNANDES, I<., GOLDSPINK, G., N , P, M, NV and G proteins of infectious BUTTERWORTH, P.,UMEDA,P. I<., CHANG, hematopoietic necrosis virus in rainbow trout I<. C. (1991), Strong expression of foreign Uncorhyndzus mykiss using DNA vaccines, genes following direct injection into fish muscle, FEBS Lett. 290, 307-312. Dis. Aquat. Organisms 39, 29-36. S., LAPATRA, S. E., ANDERSON, E. D., HAWICE, N. A., RAST, J. P., LITMAN, G.W (1996), CORBEIL, KURATH, G. (2000a),Nanogram quantities of Extensive diversity of transcribed TCR-8 a DNA vaccine protect rainbow trout fry in a phylogenetically primitive vertebrate, J. Irnmunol. 156, 2458-2464. against heterologous strains of infectious J., LORENZEN, N., ARMSTRONG, HEPPELL, hematopoietic necrosis virus, Vaccine 18, 2817-2824. N. I<., WU, T., LORENZEN,E. et al. (1998a), Development of DNA vaccines for fish: vector S., KURATH, G., LAPATRA, S. E. CORBEIL, (2000b), Fish DNA vaccine against infectious design, intramuscular injection and antigen hematopoietic necrosis virus: efficacy expression using viral haemorrhagic septiof various routes of immunization, Fish caemia virus genes as model, Fish Shellfish Immunol. 8, 271-286. Shellfish Immunol. 10, 711-723. J., SANCHEZ-DARDON, J., KRIEG, DAVIS,H. L., MCCLUSICIE, M. J., (1999), DNA HEPPELL, A. M., DAVIS,H. L. (1998b), Stimulation of vaccines for viral diseases, Microb. InJect. 1, 7-21. rainbow trout immune cells with oligodeoD E KINKELrN, P., DORSON,M., (1973), Interxynucleotides bearing CpG motifs. Proceedferon production in rainbow trout (Salmo ings ofthe Tnird International Symposium on gairdneri) experimentally infected with Egtved Aquatic Animal Health (Abstract S3-1), virus, J. Gen. Virol. 39, 125.127. pp. 101, Baltimore. FERNANDEZ-ALONSO, M., ALVAREZ, F., ESTAPA, HEPPELL, J., DAVIS,H. L. (2000), Injection of A., BLASCO,R., COLL,J. M. (1999), A model to DNA vaccines in fish, in: Methods in

I

189

190

I

Simon R. M. Jones

Molecular Medicine Vol. 29, DNA Vaccines: LEWIS, P. J., BABIUK,L.A. (1999), DNA Methods and Protocols (Lowrie, D. B., Whalen, Vaccines: a review, Adv. Virus Res. 54, R. G., Eds.), pp. 99-103. Humana Press Inc., 129-188. Totawa NJ. LORENZEN,N., LORENZEN, E., EINER-JENSEN, J., K., HEPPELL, HORDVIK, I., JACOB, A. L. J., CHARLEMANGE, J., DAVIS,H. (1999), Genetic ENDRESEN, C. (1996), Cloning of T-cell antivaccination against viral haemorrhagic septigen receptor beta chain cDNAs fromn cemia virus: Small amounts of plasmid DNA Atlantic salmon, Salmo salar, lmmunogenetics protect against a heterologous serotype, Virus 45, 9-14. Res. 63, 9-25. IWAMA, G., NAKANISHI,T. (1996),The fish LORENZEN,N., LORENZEN,E., EINER-JENSEN, immune system: organism, pathogen, and K., HEPPELL, J., Wu, T., DAVIS,H. (1998), environment. Academic Press, London. Protective immunity to VHS in rainbow trout T. S., SYLVESTER, I. D., AMBALI, A. G., (Oncorhynchusmykiss, Walbaum) following KANELLOS, P. H. (1999a),The HOWARD, C. R., RUSSELL, DNA vaccination, Fish Shellfish rmmunol. 8, safety and longevity of DNA vaccines for fish, 261-270. Immunology 96, 307-313. LORENZEN, N., LORENZEN,E., EINER-JENSEN, KANELLOS,T. S., SYLVESTER, I. D., BUTLER,V. L., J., DAVIS,H. L. (1999), Genetic K., HEPPELL, AMBALI, A. G., PARTIDOS, C. D. et al. (1999b), vaccination of rainbow trout against viral Mammalian granulocyte-macrophage colony- haemorrhagic septicaemia virus: small stimulating factor and some CpG motifs have amounts of plasmid DNA protect against a an effect on the immunogenicity of DNA and heterologous serotype, Virus Res. 63, 19-25. subunit vaccines in fish, Immunology 96, LORENZENE., EINER-JENSEN, K., MARTINUSSEN, 507-510. T., LAPATRA, S. E., LORENZEN, N. (2000),DNA I. D., HOWARD, C. R., KANELLOS, T., SYLVESTER, vaccination of rainbow trout against VHS RUSSELL, P. H. (1999c),DNA is as effective as virus: a dose response and time-course study, protein at inducing antibody in fish, Vaccine J. Aquat. Anim. Health 12, 167-180. 17, 965-972. MURRAY,B.W., SULTMANN, H., KLEIN, J. KAATTARI S. L., PIGANELLI, J. D. (1996),The (1999),Analysis of a 26 kb region linked to specific immune system: humoral defence, the mhc in zebrafish: genomic organization in: The Fish Immune System: Organism, of the proteasome component Bltransporter Pathogen, and Environment (Iwama, G., associated with antigen processing-2 gene Nakanishi, T., Eds), pp. 207-254. Academic cluster and identification of five new proteaPress Inc., London. some $ subunit genes, /. Immunol. 163, B. H., LIE,0., KOPPANG, E.O., DANNEVIG, 2657-2666. RPINNINGEN, K., PRESS, C.MCL. (1999), N. F., STAFFORD. J. L., BELOSEVIC. NEUMANN, Expression of Mhc class I and I1 mRNA in a M. (2000), Biochemical and functional charmacrophage cell line (SHK-1)derived from acterization of macrophage stimulation facAtlantic salmon, Salmo salar L., head kidney, tors secreted by mitogen-induced goldfish Fish Shellfish Immunol. 9, 473-489. kidney leukocytes, Fish Shellfish Immunol. 10, KRIEG,A.M., YI, A.-K., MATSON,S., 167-186. WALDSCHMIDT, T. J.. BISHOP, G. A. et al. D. E., HANSEN, J., OHTA,Y., HALINIEWSKI, (1995), CpG motifs in bacterial DNA trigger FLAJNIK, M. F. (1999),Isolation of transporter direct B-cell activation, Nature (London) 374, associated with antigen processing genes, 546-549. TAP1 and TAP2, from the homed shark LECHARDEUR,D., SOHN,K. J., HAARDT, M., Heterodontus francisci, Immunogenetics 49, JOSHI,P.B., MONCK,M. et al. (1999), Meta981-986. bolic instability of plasmid DNA in the PARTULA, S., FELLAH, J. S . , DE GUERRA, A., cytosol: a potential barrier to gene transfer, CHARLEMANGE, J. (1994), Identification of Gene Ther. 6, 482-497. cDNA clones encoding the T-cell receptor p LI'ITEL-VANDEN LEWIS,P. J., VAN DRUNEN chain in the rainbow trout (Oncorhynchus HURK,S., BABIUK,L. A. (1999),Altering the mykiss), C.R. h a d . Sci. D317, 765-770. cellular location of an antigen expressed by a PARTULA, S., D E GUERRA, A,, FELLAH,J. S., DNA-based vaccine modulates the immune CHARLEMANGE, J. (1995), Structure and response, /. Virol. 73, 10214-10223.

70 Plasmids in Fish Vaccination diversity of the T cell antigen receptor B chain in a teleost fish, /. Immunol. 155, 699-706. PARTULA, S., D E GUERRA, A,, FELLAH,J. S., CHARLEMANGE, J. (1996), Structure and diversity of the T cell receptor a-chain in a teleost fish, 1.Immunol. 157, 207-212. P. P., GRIEBEL, P., BAZIN, H., PASTORET, GOVAERTS, A. (1998), Handbook of Vertebrate Immunology. Academic Press. London. PERSSON, A., STET,R. J.M., PILSTROM, L. (1999), Characerization of MHC class I and $2-microglobulin sequences in Atlantic cod reveals an unusually high number of expressed class I genes, Immunogenetics SO, 49-59. PISETSKY, D. S. (199G), Immune activation by bacterial DNA: A new genetic code, Immunity 5, 303-310. PRESS, C. M., REITAN,L. J., LANDSVERK,T. (1995), Antigen retention and enzyme reactivity in the spleen of Atlantic salmon Salmo salar L., following administration of injectable furunculosis vaccines, /. Fish Dis. 18, 199-210. A., MACLEAN,N. (1992), Fish transRAHMAN, gene expression by direct injection into fish muscle, Mol. Mar. Bid. Biotechnol. 1, 286-289. ROBERTSON,J. C., HAZEL, J. R. (1995), Cholesterol content of trout plasma membranes varies with acclimation temperature, Am. J . Physiol. 269, 1113-1119. RODRICUES,P. N. S., HERMSEN, T T., ROMBOUT,J. H. W. M., EGBERTS,E., STET, R. J. M. (1995), Detection of MHC class 11 transcripts in lymphoid tissues of the

common carp (Cyprinus carpio), Dev. Compar. Immunol. 19, 483-496. RUSSELL,P. H., KANELLOS,T., SYLVESTER, I. D., CHANG, I<. C., HOWARD, C. R. (1998), Nucleic acid immunisation with a reporter gene results in antibody production in goldfish (Carassim auratus L.), Fish Shellfish Immunol. 8,121-128. RUSSELL,P. H., I ~ N E L L O T. SS., (2000),The potential use of cytokines for the development of DNA vaccines for fish (Abstract) Proceedings of Veterinary Cytokine and Vaccine ConJrence, Tsubuka Science City, Japan. P. H., ~(ANELLOS, T., NEGROU,M., RUSSELL, AMBALI,A. G. (ZOOO), Antibody responses of goldfish (Carassius auratus L.) to DNA-irnrnunisation at different temperatures and feeding levels, Vaccine 18, 2331-2336. SECOMBES, C. J. (1996), The nonspecific immune system: cellular defenses, in: The Fish Immune System: Organism, Pathogen, and Environment (Iwama, G., Nakaniihi, T., Eds), pp. 63-103. Academic Press Inc., London. L. J., DANIELS, G. SECOMBES, C. J., HARDIE, (1996), Cytokines in fish: an update, Fish Shellfish Immunol. G, 291-304. G. S., ANDERSON,E., LAPATRA, S. E., TRAXLEQ RICHARD, J., SHEWMAKER, B., KURATH, G. (1999), Naked DNA vaccination of Atlantic salmon Salmo salar against IHNV, Dis. Aquat. Organisms 38, 183-190. ZHOU, H., BENGTEN,E., MILLER,N.W, WAR% G. W., CLEM,L.W, WILSON, M. R. (1997), T cell receptor sequences in the channel catfish, Deu. Comp. Immunol. 21, 238.

P/asmids for Therapy and kchation by M.Schleef Copyright 0 WILEY-VCH Verlag GmbH, 2001

I

11 Plasmid Manufacturing - An Overview Guilherme N. M. Ferreira, Duarte M. F. Prazeres, Joaquim M. S. Cabral and Martin Schleef ''

1

introduction

Gene therapy is a therapeutic strategy in which sense nucleic acids are introduced in human cells, modifying their genetic repertoire for therapeutic purposes (Crystal, 1995). Curing diseases at the level of the specific gene defect rather than at the conventional level (Alton et al., 1993; Caplen et al., 1994),the possibility of preventive or curative vaccination against pathogenes or viruses (Gregoriadis 1998; LeBorgne et al., 1998; Major et al., 1995; Michel et al., 1995; Schleef, 2000 and Chapters 3, 5 and G of this book) or the induction of regenerative processes (Bonadio et al., 1999; Shea et al., 1999; and Chapter 4 of this book) have already been demonstrated. Generally, the respective nucleic acid is double-stranded DNA (dsDNA)encoding for a therapeutic, a destructive or a marker protein. The nucleic acid can also be antisense RNA or single-stranded DNA (ssDNA) binding to targeted sequences in the host cells and inhibiting the expression of a specific gene, either by blocking messenger RNAs or gene promoters (Doux et al., 1995). The first gene therapy products are expected to get market approval by the beginning of the 21th century, and market expectations for this kind of products by the year 2010 exceed US$ 45 billion (Glaser, 1997). A prerequisite for effective gene therapy is that the nucleic acids are delivered efficiently into the targeted cells, and - with the exception of antisense therapy that the nucleic acids must reach the cell nucleus. Both viral and non-viral vectors have been developed to efficiently deliver nucleic acids (Anderson, 1998; Luo et al., 2000). However, in spite of some advantages on using viral vectors, no significant evidence has been presented for the clinical effectiveness of any gene therapy protocol (Anderson, 1998; Luo et al., 2000; Mountain, 2000). The use of viral vectors for gene delivery has raised safety and regulatory concerns owing to their toxicity and immunogenicity as well as to the possible activation or deactivation of oncogenes or tumour repressor genes, respectively (Anderson, 1998; Luo et al., 2000). Half a dozen deaths have been reported recently in sev-

793

194

I

Guilherrne N. M. Ferreira et a1

era1 separate gene therapy trials involving viral vectors (Fox, 1999; Lehrman, 1999), reinforcing the safety and regulatory concerns on using such vectors. For these reasons, non-viral delivery systems have become more and more desirable (Luo et al., 2000) as they are the most attractive gene transfer systems for commercial products (Mountain, 2000). The number of approved gene therapy protocols using plasmid DNA-based delivery vectors has increased exponentially since 1995, representing approximately 25 % of the ongoing gene therapy trials (http://www.wiley.co.uk/ genetherapylclinical). Plasmid DNA has also been used to express specific antigens on cell membranes stimulating and enhancing the immune system response and memory, as a new, safer, generation of vaccines (Tighe et al., 1998). Plasmid DNA-based delivery vectors, however, are less effective in transfecting cells than are viral vectors (Luo et al., 2000). It is estimated that only one in every 1000 plasmid molecules presented to the cells reaches the nucleus and is expressed (Crystal, 1995),with full treatments requiring milligram quantities of plasmid DNA. Therefore, large-scale plasmid DNA manufacturing processes must be developed, as gene therapy and DNA vaccine applications move from the laboratory to clinical trials and market approval (see also www.PlasvnidFacto~.cow).In addition to the requirement of large quantities of therapeutic DNA, the quality of this pharmaceutical is essential. Only little is known about the efficacy and safety of the different plasmid isoforms (Schmidt et al., 1999a; see also Chapter 2 of this book). This requires additional research on the mode of action of plasmid therapeutics in the respective applications with a focus on the structure of the molecule.

L

Structure of nucleic acids

The design of purification processes should always consider the structural properties of the target molecules. For instance, adsorption techniques are based on specific interactions between the target molecules and the adsorbent (Sofer and Hagel, 1997), and the critical binding properties can be either exposed on the surface or hidden in the interior parts of the target molecules. Furthermore, the separation principle of some unit operations is based on the size and shape of the target molecules. The design of purification processes and the selection of downstream processing unit operations is thus dependent on the structure and conformation of the desired product. 2.1 Brief structural description of DNA and RNA structures

Nucleic acids are polymers in which nucleotides are connected by phosphodiester bonds between the hydroxyl group of the 3’ sugar carbon of a nucleotide and the phosphate group of the 5‘ sugar carbon of the adjacent nucleotide. Thus nucleic acids have two distinct ends termed the 3’ and 5’ ends. Both ends are chemically

I 7 Plasmid Manufacturing - An Overview

and biologically different. The 5‘ end (the starting end) is derived with a phosphate group while the 3’ end (the terminal end) is derived with a hydroxyl group. Nucleic acids are polar molecules, with two negative charges at the phosphate end of the molecule. Another important feature of nucleic acids is that at pH > 4 (Stryer, 1995) each nucleotide contributes with one negative charge to the overall net charge of the molecule. Thus, under physiological conditions (pH = 7) RNA and DNA are polyanionic molecules with an overall net charge equal to the number of nucleotides plus two charges per strand located at the 5’ end of each strand of the molecule. The nucleic acid structure is influenced by the conformation of its building blocks. Different sugar conformations (puckers) and variations in the structural relationship of the base-to-sugar bond leads to different nucleic acid conformations (Neidle, 1994; Sinden, 1994; Stryer, 1995; Voet et al., 1990). The structure of DNA, today termed B-DNA, was first deduced by Watson and Crick in 1953 from X-ray diffraction studies of highly hydrated (> 92%) DNA fibers (Stryer, 1995). While B-DNA is the most common DNA conformation, diffraction studies performed in different conditions generate different diffraction patterns, thus leading to different structures termed A-DNA and Z-DNA (Neidle, 1994; Sinden, 1994; Stryer, 1995) (Table 1).DNA is highly polymorphic with the different forms thus far isolated corresponding to distinct, yet interconvertible, molecular structures of the nucleotides (Neidle, 1994; Sinden, 1994). For instance, dehydration of the DNA molecule to less than 65 -75 % relative humidity or reduction of the water activity by addition of alcohols results in DNA helixes of the A type, while Z-DNA is favored at high salt concentrations and high supercoiling densities (Neidle, 1994; Stryer, 1995).A detailed description of DNA polymorphism can be found in Neidle (1994).Table 1 summarizes some of the structural features of the three most abundant DNA polymorphs (A-, B-, and Z-DNA). In the heliptical structure of DNA the aromatic bases are oriented to the interior of the molecule, perpendicularly to the sugar-phosphate backbone, and stacked over each other in the helix axis. Watson and Crick hydrogen bonds are formed between complementary bases located in each DNA strand, therefore holding the two strands together. Some structural features of ideal A-, B-, and 2-DNA (Voet et al., 1990); detailed descriptions of these DNA polymorphs can be found in (Neidle, 1994)

Table 1.

A-DNA

B-DNA

Helical sense Helix diameter (nm)

Z-DNA

left handed 0.26

0.20

0.18

~~~~~

Base pairs per helical turn Helix rise per base pair (nm)

11 0.26

10 0.34

12 0.45

Major grove

narrow and deep

wide and deep

flat

Minor grove

wide and shallow

narrow and deep

narrow and deep

I

I

195

196

I

Guilherme N. M. Ferreira et a/.

A

B M

c1”

--N

\C1’

Fig. 1. Accessibility of the aromatic bases within the major (M) and minor (m) groves in the G-C (A) and A-T (B) pairs. Arrows indicate the available hydrogen bond sites (adapted from Neidle, 1994).

An important consequence of the base arrangement in DNA is that the sugar linked to an individual base is on the same side of the molecule (Neidle, 1994). Thus when the bases are stacked over each other, the gaps between the sugar residues form continuous indentations in the molecule surface. They are termed groves and wind along in parallel to the sugar-phosphate chains (Neidle, 1994; Sinden, 1994; Stryer, 1995; Voet et al., 1990). Within these groves the bases are accessible to solvent and ligand molecules. The accessibility of the aromatic electrons and the available sites for hydrogen bonding (Figure 1) are crucial features for ligand-DNA binding and recognition which is biologically relevant (Neidle, 1994). 2.2

DNA supercoiling

With an average molecular weight of 660 Da per base pair (Sambroolt et al., 1989) DNA molecules generally are heavy and big (Table 2). Thus, in order to fit into the proltaryotic cell limits or into the eultaryotic cell nucleus, DNA molecules must be condensed. In this supercoiling plays an important role (Boles et al., 1990; Vologodslcii et al., 1992). Supercoiling can result from winding around proteins, as in eukaryotic nucleosomes, or from the topological constraints imposed upon underwound or overwound circular DNA molecules (e. g., plasmids), as in proltaryotic cells or in solution (Boles et al., 1990; Sinden, 1994). Supercoiled DNA contains large amounts of free energy (AG) that is used to drive biological reactions, influencing a large range of fundamental processes such as DNA replication, recombination and transcription (Boles et al., 1990; Sinden, 1994; Summers, 1996; Vologodslii et al., 1992). A detailed description of the structure and thermodynamics of supercoiling can be found in Sinden (1994),Vologodskii et al. (1992),and Boles et al. (1990).Briefly, supercoiling is characterized by the linking number (Lk), which is the number of

I 1 Plasmid Manufacturing Table 2.

- An

Overview

Sizes and molecular weights of some DNA molecules (Voet et at., 1990) Number ofBase Pairs

Organism

Contour Length (rrm)”

Wl Polyoma virus (SV40)

Escherichia coli

I

5.1

I

1.7

Molecular Weight (kDalb

I

3,366

4,000

1,360

264,000

Yeast (in 17 haploid chromosomes)

13,500

4,600

8,910,000

Human (in 23 haploid chromosomes)

2,900,000

990,000

1,914,000,000

a

The contour length was calculated considering the 0.34 nm helix rise per base for B-DNA (Table 1). The average molecular weight of DNA is 660 Da per base pair (Sambrook et al., 1989).

times the two DNA strands intenvind when the molecule is made to lie flat on a plain (Sinden, 1994).A relaxed DNA molecule, which is a molecule without supercoils, has a linking number (Lko) equal to the number of base pairs (N) divided by the helical repeat, the number of base pairs per helix turn (Table 1).The degree of supercoiling is usually expressed relatively through the superhelix density (o = (Lk-&)/&). A deficiency in the linking number (Lk < Llco)leads to negative supercoils, i. e., o < 0. Most supercoiled DNA molecules isolated from either prolcaryotes or eukaryotes are negatively supercoiled, with o values between -0.05 and -0.07 (Boles et al., 1990; Vologodskii et al., 1992). In the structural and thermodynamic model developed by Boles et al. (1990)and further complemented by Vologodslcii et al. (1992),circular DNA molecules such as plasmids adopt a structure in which the superhelix diameter, the axis length, the number of branches, and the free energy are correlated with the superhelix density. In these models, the superhelix axis is the line passing through the nodes and bisects the area enclosed between adjacent DNA nodes (Figure 2). The length of the superhelix axis is independent of the superhelix density (u)and about 41 % of the length of the linear or relaxed DNA molecule (Boles et al., 1990) - equal to the number of base pairs multiplied by the helix rise per base pair (Table 1).If supercoiled DNA molecules adopt a branched structure, as presented in Figure 2, the superhelix axis length (0.41 X molecule length) is equal to the sum of the lengths of each segment (Boles et al., 1990). Therefore, the superhelix axis length is the maximum extension of supercoiled DNA molecules and thus can be regarded as the maximum size of the relaxed conformation. If the relaxed DNA molecules were to assume a circular shape, the circle perimeter can also be used for a rough estimation of the diameter of the molecule. Upon supercoiling, the superhelix diameter D in Figure 2, is a function of the supercoiling density and is given by (Boles et al., 1990): D (nm) = 0.2 X (0.00153 - 0.268 o)-l

I

197

198

I

Cuilherme N. M. Ferreira et a/. Fig. 2. Model of supercoiled circular DNA structure. The superhelix has a diameter (D) and an axis (line) with maximum length equal t o the number of base pairs multiplied by the helix rise. For branched molecules, the sum o f the axis lengths of all the branches equals the maximum superhelix axis (Boles et al., 1990).

Therefore, the superhelix diameter of the majority of the supercoiled DNA molecules is between 9.9 nm and 13.4 nm, the limits obtained from CT values of -0.07 and -0.05, respectively. Virtually all isolated supercoiled DNA molecules are branched (Boles et al., 1990; Vologodslcii et al., 1992). The number of branch points of a supercoiled DNA molecule is independent of the superhelix density and increases linearly with the molecule length (Boles et al., 1990; Vologodslcii et al., 1992) (Figure 3A). Based on the free energy of supercoiled DNA molecules, a branching probability can be calculated for specific buffer conditions (Vologodskii et al., 1992).This probability, that can be understood as the frequency of supercoiled DNA molecules adopting a particular structure within the entire population, indicates that in solution supercoiled DNA molecules can adopt different structures, i. e., different branches (Vologodsliiet al., 1992) (Figure 3B). Therefore, derived by the energetic constraints due to DNA length and ionic conditions (Vologodskii et al., 1992) the B

A

4,o

I

ln

.-E

3,O

0 Q 'El

3,O

4,O

5,O

6,O

Size (kb)

7,O

8,O

1,o

2,o

3,O

4,O

Branched points

Fig. 3. Dependence of the expected number of (triangles), 5.25 kb (circles), and 7 kb (squares) branched points with the D N A length (A) and supercoiled DNA (Boles e t al., 1990 Volobranching probability distribution (B) for 3.5 kb godskii et al., 1992).

7 I Plasmid Manufacturing - An Overview

structure of supercoiled DNA i s thought to be highly dynamic with branches forming and retracting rapidly (Boles et al., 1990), The practical consequence of such a dynamic behavior is that the dimensions of the molecule (superhelix maximum axis and diameter) and its shape (number of branch points) can be altered by changing the solvent conditions (Boles et al., 1990; Vologodskii et al., 1992). For instance, supercoiled DNA molecules can be compacted or expanded by increasing or decreasing the solvent ionic strength, respectively (Vologodskii et al., 1992) or by changing the type of cations with regard to size and charge density (Murphy et al., 1999). Plasmid DNA i s a particular type of DNA. Plasmids are extrachromosomal circular DNA molecules responsible for bacterial conjugation. They encode for a variety of functions that are not essential for the survival of the host, but increase its ability to survive in atypical conditions (Summers, 1996). Plasmids replicate autonomously from the cell life cycle and have been used for a variety of applications, including cloning and recombination (Sambrook et al., 1989).Their size ranges from approximately 1.5 kbp to approximately 120 kbp, with quite different copy numbers per bacterial cell (Davis et al., 1980).For small plasmids the copy number can reach more than 1,000 copies per cell (see also Chapters 1 and 2 of this book). The relaxed replication (amplification) does not depend on any plasmid-encoded protein and is not synchronous with the stringent replication of the bacterial host chromosome (Davis et al., 1980). The exact conformation of a plasmid depends on its integrity (“covalently closed circular - ccc”, or “open circular - oc”) and the result of the replication process (e.g., monomeric or oligomeric). The analysis of the different plasmid forms is decribed in Chapter 2 of this book. Not all known plasmids, however, are of use in commercial bioprocessing. Desirable characteristics for useful vectors include high copy numbers, several unique endonuclease cleavage sites, small size, and genetic stability (ICumar et al., 1991). Plasmids with such characteristics have been widely used in the largescale production of recombinant proteins (Sofer and Hagel, 1997), and their utilization in gene therapy and DNA vaccine protocols i s increasing as their efficiency as delivery vectors i s being demonstrated (Robinson et al., 1997; Radnedge and Richards, 1999).

3

Plasmid DNA Manufacturing

As for proteins, process development for manufacturing plasmid DNA generally starts on a bench scale with the construction and selection of appropriate expression vectors and production microorganisms (upstream processing), followed by the selection and optimization of fermentation conditions (fermentation) and of isolation and purification steps (downstream processing). These three stages of process development - upstream processing, fermentation and downstream processing - are integrated and must not be approached on an individual basis (Kelly and Hatton, 1991). Downstream processing of biologicals is in fact greatly

1

199

200

I

Guilherrne N. M. Ferreira

et a/.

Table 3. Average composition o f rapidly growing Escherichia coli cells (Atkinson and Mavituna, 1991) - plasmid D N A is not considered in this table Component

Relative Amount

Average Molecular Weight Pal

("/. w/wl

I H2O

I 70

I

18

Average Number of Molecules per Cell

I 4 x 10'" ~~

Average Number of Di@rent Molecules

I

1

Inorganic ions

1

40

Carbohydrates and precursors

3

150

2 x 108

200

Amino acids and precursors

0.4

120

3

lo7

100

Nucleotides and precursors

0.4

300

1.2

lo7

200

Lipids and precursors

2

750

2.5 x lo7

50

Other small molecules

0.2

150

1.5 x lo7

200

15

Proteins Nudeic acids DNA" RNA

2.5

10'

2.5 x lo8

10'

20

3,000

4

1

6

16.5 rRNA

500,ooo

3

lo4

I

23s rRNA

1,000,ooo

3

104

1

4 x lo5

40

tRNA mRNA a

1

40,000

I

~

25,000 1,000,000

103

1,000

Rapidly growing E. coli cells have on average 4 molecules of genomic DNA (Atkinson and Mavituna, 1991; Lodish et al., 1995). However, only one molecule of gDNA per cell should be considered at the end of the fermentation (Atkinson and Mavituna, 1991).

affected by impurities and contaminants present in the process streams. They strongly depend on the upstream processing and fermentation conditions (Kelly and Hatton, 1991). The concept of impurity and contaminant is different. Impurities are undesired substances that are internal to the system, i. e., substances which are derived from the host cell or are present during growth or purification of the desired product (Briggs and Panfili, 1991; Sofer and Hagel, 1997).Thus, cell components released after lysis (Table 3 ) or remainders of the fermentation broth are examples of impurities. On the other hand, contaminants are undesired substances that are external to the system. Contaminants can be introduced or added to the process, either accidentally or as part of a purification operation. Viruses, residual solvents, and chromatographic leachaes are examples of such substances (Briggs and Panfili, 1991; Sofer and Hagel, 1997).

7 7 Plasmid Manufacturing - An Overview

The major difference between manufacturing processes for classical protein vaccines and DNA vaccines (Giese, 1998) is the potential of saving time during their development (Schleef et al., 2000). Classical approaches require many years and a substantial part of the time is spent for process development. A generic process is not suitable for different protein vaccines, since they are usually different and require different processing technologies. For plasmid DNA a generic process can - in principle - be developed at the beginning of pharmaceutical developments, and only minor modifications may have to be applied for each individual type of plasmid. 3.1

Major impurities and main product specifications

Biopharmaceuticals based on plasmid DNA are chemically highly defined and, therefore, can be analyzed by chemical, biochemical, and physical assays (Schleef, 1999). Assessing the quality of the final plasmid preparation in terms of safety, potency and purity - in order to ensure that the product specifications are met (Table 4) - is a key issue for process validation (Marquet et al., 1997a; Marquet et al., 1997b; Schleef, 1999). Although quality standards can not be established a priori as they are dependent on the intended theraputic use (Prazeres et al., 1999), and therefore defined during the clinical trials, regulatory agencies provide guidelines that are helpful in the early establishment of such standards (Table 4) during process development (Marquet et al., 1995; Prazeres et al., 1999; Schleef,

Table 4. Principal approval specifications and recommended assays for assessing the final plasmid DNA preparation purity, safety and potency for gene therapy and DNA vaccines applications

(USFDA, 1996a,b; see text)

Impurity

Recommended Assay

Approval Specification

Proteins

BCA protein assay

undetectable

RNA

agarose gel electrophoresis

undetectable

gDNA

agarose gel electrophoresis Southern blot

undetectable < 0.01 pg-' plasmid

Endotoxins

LAL assay

< 0.1 EU pg-' plasmid

Plasmid isoforms (linear, relaxed, denatured)

agarose gel electrophoresis

< 5%

Biological activity and identity

restriction endonucleases

Coherent fragments with the plasmid restriction map expected migration from size and supercoiling comparable with plasmid standards

agarose gel electrophoresis transformation efficiency

I

I

201

202

I

Cudherme N. M. Ferreira et a/

1999). However, the authors recommend to apply even more testing than requested by the references of Table 4. For the identity test, sequencing and PCR should be added. The most important test concerns the topology of plasmids used for clinical applications: It is well known within the scientific community, that agarose gel electrophoresis may not be sufficient to detect any type of topology included (see also Schmidt et al., 1999a). It is essential to perform a CGE assay (see Chapter 2 of this book) to replace agarose gel electrophoresis as soon as possible. The general lists of impurities and contaminants that should be measured during the production of recombinant proteins in E. coli also apply for plasmid DNA (Marquet et al., 1997b; Schleef, 1999). In the production of plasmid DNA for clinical applications, host nucleic acids, host proteins, and endotoxins are the impurities of major concern, since some of them co-purify with the plasmids (Prazeres et al., 1999; Schleef, 1999; Wicks et al., 1995) and can cause significant side effects once administered to living organisms (Janson and Pettersson, 1993).

Host nucleic acids Concerns regarding the contamination of the final biopharmaceutical product by genomic DNA (gDNA) and/or RNA fragments have 40 years of history (Briggs and Panfili, 1991; Riggin et al., 1996). They are based on the possibility of an oncogene being encoded by those fragments or on the possible activation of oncogenes or deactivation of tumor repressor genes, once they occur in the recipient cells (Briggs and Panfili, 1991; Riggin et al., 1996). Although these are theoretical risks, as shown by experimental data in which less than frequency of tumor formation was observed when a 1 ng DNA dose from a cell containing 100 copies of oncogenic DNA was administered in living organisms (Riggin et al., 1996), the regulatory agencies impose limits with the sole concern of patient safety (Janson and Pettersson, 1993; Sofer and Hagel, 1997). The principles of gene therapy and DNA vaccination also apply to nucleic acid impurities. Thus, induction of the immune system memory and activation or repression of cellular genes are other relevant issues and concerns regarding these impurities (Briggs and Panfili, 1991). According to the regulatory agencies (USFDA, 1985; USFDA, 1996a; USFDA, 1996b), genomic DNA and RNA must be undetected by ethidium bromide stained agarose gel electrophoresis (Marquet et al., 1995; Prazeres et al., 1999; Schleef, 1999). These impurities must also be assayed using molecular methods, such as Southern and Northern assays, in which case less than 10 ng per dose and 100 ng per dose are imposed by the USFDA and World Health Organization (WHO), respectively (Janson and Pettersson, 1993; Marquet et al., 1995; Riggin et al., 1996; Sofer and Hagel, 1997). The issues of performing validated lot release testing of pharmaceutical-grade plasmid DNA or other biologicals are presented in detail in Chapter 12 of this book. 3.1.1

I 1 Plasmid Manufacturing

3.1.2

- An

Overview

Proteins

The risks associated with the presence of proteins in final plasmid DNA preparations are the possible generation of immune responses and allergies as well as the activation and development of biological reactions on the recipient by promoting the production of cytoltines, hormones and/or antibodies (Briggs and Panfili, 1991). It is difficult to quantify the minimum immunogenic amount of proteins, since this value depends on the specific protein and recipient. However, immune responses were reported with less than 1 ng of contaminating protein (Briggs and Panfili, 1991). Proteins can be monitored by silver staining polyacrylamide gel electrophoresis (PAGE) either under native or denatured conditions. The detection limit of this method is as low as 0.5 ng of analyzed protein (Briggs and Panfili, 1991), which in principle is below the 10 ng per dose limited imposed by the WHO. Standard colorimetric assays are also recommended for the analysis of trace protein in plasmid DNA therapeutics (Marquet et al., 1995). These methods can be used instead of, or as well as, silver stained PAGE. Among them, the bicinchoninic acid (BCA) assay has been used more frequently (Ferreira et al., 1999; Horn et al., 1995; Prazeres et al., 1998; Schleef, 1999). Immunoblot assays, enzyme linked immunosorbent assay (ELISA), and radioimmuno assay (RIA) can also be adapted whenever appropriate (Briggs and Panfili, 1991; Marquet et al., 1995). However, these assays require the generation and production of antibodies, which is costly and time consuming. Furthermore, the U SFDA specifications for the final plasmid DNA preparation regarding contamination by proteins (Table 4) only requires this impurity to be undetectable by silver staining PAGE or BCA assays (USFDA, 1996a; USFDA, 199613). Thus, the necessary investment to develop antibody-based methods may until now prevent to analyze protein traces in the final plasmid DNA preparations for gene therapy and DNA vaccination applications. However, the authors recommend the timely establishment of such analytic systems in case of intended future large-scale manufacturing.

3.1.3

Endotoxins

A pyrogen is defined as any fever causing substance, and the most common pyrogenic substances are endotoxins (Sofer and Hagel, 1997; Wicks et al., 1995). Subnanogram amounts of endotoxins (Montbriand and Malone, 1996) or less than 5 endotoxin units (EU) per kg body weight (Sofer and Hagel, 1997) are enough to cause a pyrogenic response in humans. The maximum allowed endotoxin administration dose for humans required by regulatory authorities is about 0.5 ng per kg body weight per hour (Janson and Pettersson, 1993).Thus it was recommended to keep endotoxin levels below 0.1 EU per pg plasmid in the final plasmid preparation (Schleef, 1999). Endotoxins are highly negatively charged lipopolysaccharides (LPS) from the cell wall of gram-negativebacteria such as E. coli. LPSs are composed by three different chemical regions: an innermost lipidic region (lipid A) that is responsible for most

204

I

CLiiiherme N. M. Ferreira et al

of the biological activity of LPSs (Wicks et al., 1995), an intermediate conserved oligosaccharide core region, and an outermost phosphopolysaccharide side chain that confers immunospecificity to the molecule (Sofer and Hagel, 1997). LPS have, among other effects, profound effects on a variety of cells, including the release of tumor necrosis factor, interleukins, stimulation of T cells to produce interferons, (Wicks et al., 1995; Weber et al., 1996). LPSs monomers have molecular weights between 10,000 Da and 20,000 Da (Janson and Pettersson, 1993: Sofer and Hagel, 1997). However, LPSs have the ability to form spherical-shaped aggregates of up to 0.1 pm diameter through interactions between ionic groups and between individual lipid A moieties (Janson and Pettersson, 1993: Sofer and Hagel, 1997). The aggregates are stabilized by the presence of divalent cations, such as calcium and magnesium. In the absence of such cations the aggregates will dissociate into micelles of 300,000-1,000,000Da (Jansonand Pettersson, 1993: Sofer and Hagel, 1997). Further dissociation into the monomers can be achieved by exposing the LPS micelles to surfactants (Janson and Pettersson, 1993). From a practical point of view, after treatment with surfactants and in the absence of divalent cations, LPS complexes can be regarded as having the monomer average molecular weight (Janson and Pettersson, 1993; Sofer and Hagel, 1997). The most sensitive and widely used method for the measurement of endotoxins in biopharmaceuticals is the Limulus amoebocyte lysate (LAL)assay (Marquet et al., 1997b; Montbriand and Malone, 1996; Schleef, 1999; Wicks et al., 1995). This assay is based on the dose-dependent enzymatic coagulation of an extract of blood cells from horseshoe crab, Limulus polyphemus, by LPS (Wicks et al., 1995). Preparing appropriate endotoxin standards and by performing successive dilutions of the EU per pg plasmid plasmid DNA preparations, specific sensibilities of 1.2 X can be achieved (Horn et al., 1995). It is important to note that LPS molecules give negative LAL test results, if their molecular weight is below 10,000 Da. Furthermore, the lipid A moiety of the LPS molecule, which has a molecular weight close to 2,000 Da, can be pyrogenic (Sofer and Hagel, 1997). Thus, other assays for pyrogenic substances should be performed whenever a sample gives negative LAL tests. The leukocyte pyrogen assays are examples of such assays (Sofer and Hagel, 1997). In these assays, the presence of endogenous pyrogens such as interleukins, tumor necrosis factor, and interferons is checked after the administration of biopharmaceutical formulations to laboratory animals (Sofer and Hagel, 1997). 3.2 Factors influencing the production of plasmid DNA: Some considerations on the upstream processing and fermentation stages 3.2.1

The plasmid vector

The relative amount of plasmid DNA over genomic DNA and other impurities, such as RNA, proteins, and lipopolysaccharides (LPS), is critical for the successful large-scale production of plasmid DNA (Varley et al., 1998). Therefore, the design (see also Chapter 9) and selection of plasmid vectors for gene therapy and DNA vaccination should consider not only the safety and potency demands, but also

1 1 Plasmid Manufacturing

-

An Overview

the inherent requirements for the large-scale production and purification of the vector. For instance, the use of high copy number plasmids and the introduction of specific selection markers in the plasmid vector will largely improve process yield and productivity (Prazeres et al., 1999) by promoting a plasmid excess during cell growth and, therefore, reducing the overall manufacturing costs (Schleef, 1999). An other option is the reduction of the active DNA molecule to the relevant portions for gene expression (see also Chapter 8 of this book). Several promoters should be evaluated to express the gene once injected in vivo, and the plasmid should be free of viral sequences and open reading frames (Marquet et al., 1995; Schleef, 1999). Selective markers are required during the plasmid design phase for the selection of the transformed host E. coli cells, and during the entire manufacturing process, improving the stability of the plasmid by inhibiting the proliferation of plasmidfree cells (Kumar et al., 1991; Prazeres et al., 1999; Schleef, 1999). However, the regulatory agencies recommend to totally avoid the use of antibiotic resistance genes (USFDA, 1996a; USFDA, 199Gb). If the use of such genes cannot be avoided, genes encoding for resistance to ,&lactam derived antibiotics, such as ampicillin, should not be selected (Marquet et al., 1995; Schleef, 1999), since this type of antibiotics can cause potential adverse reactions in patients and manufacturing personnel (Marquet et al., 1995). Furthermore, ampicillin is rapidly degraded and therefore inappropriate for fermentation scale-up (Marquet et al., 1995). A minor improvement is the use of kanamycin resistance elements as the selective marker (Marquet et al., 1995; Schleef, 1999).

3.2.2

The bacterial host strain

The selection of the host strain is another important issue to be considered (Prazeres et al., 1999). Different genetic characteristics of the hosts, hence different strains, may largely improve downstream processing of the product by reducing or eliminating some of the impurities (Kelly and Hatton, 1991; Prazeres et al., 1999; Schleef, 1999). Furthermore, host strain physiology is known to affect plasmid stability (Kumar et al., 1991; Zabriskie and Arcuri, 1986)being therefore a particular feature to which extreme caution should be exercised. Moreover, for safety and ethical reasons, the selected host strain should be chosen and used with respect to the respective biosafety law in Europe (European Council Directive 90/219/EEC) and the regulative requirements within the U. S. (see Federal Register 51:16.958, 1986, No. 88 and Federal Register 56:33.178, 1991, No. 138; for a review, see Schlumberger and Brauer, 1994a, b). Once selected, the plasmid host strain system should be fully characterized according to the guidelines of the regulatory agencies (USFDA, 1985). A Master Cell Bank (MCB) should be produced and stored, and a single vial should be removed from the MCB to produce a Working Cell Bank (WCB) (Marquet et al., 1995; Schleef, 1999). The WCB is thus the starting point to any plasmid manufacturing process, and therefore regular tests for identity and sterility regarding other microorganisms and plasmid stability are required (Marquet et al., 1995; Schleef, 1999).

I

205

206

I

Guilherwre N.M. Ferreira et of. 3.2.3

Plasmid fermentation

The fermentation strategy is designed to achieve high cell densities under conditions that allow plasmid replication (Zabriskie and Arcuri, 1986) increasing the plasmid copy number (Lahijaniet al., 1996; Schmidt et al., 1996)during cell growth. Altough the plasmid copy number is primarily a function of the genetic structure of the plasmid, it can be strongly influenced by growth conditions (Kumar et al., 1991;Zabrisltie and Arcuri, 1986). For instance, the copy number decreases with increasing growth rates (Zabrisltie and Arcuri, 1986),and nutrient limitations as well as oxygen starvation leads to the loss of plasmid due to segregational instability in some plasmid host systems (Kumar et al., 1991). Optimizing the fermentation conditions by selecting appropriate nutrient and oxygen levels as well as the fermentation temperature is, therefore, an important consideration for the successful operation of any bioreactor (Kumar et al., 1991; Schmidt et al., 1999b).While commercial fermentation media can be used for plasmid production media development adapted to the particular plasmid host system is recommended in order to increase plasmid stability and, therefore, plasmid yield (Marquet et al., 1995). Plasmid yield and stability can be further increased by recurring to fed-batch strategies (Kumar et al., 1991; Lahijani et al., 1996).However, the most critical aspect in such modifications of growth conditions for plasmid production is the integrity of the plasmid molecules, since they are the intended product. The fed-batch strategies mentioned so far unfortunately result in low plasmid quality. Recent developments (Schmidt et al., 1999b, c; Schmidt et al., unpublished results) indicate, that such high cell density fed-batch strategies may result in high plasmid yields plus a quality suitable for pharmaceutical applications. The use of E. coli cells for manufacturing of plasmid DNA has the advantage of working with a highly characterized host with a short generation time. Various cultivation processes have been described (Lee, 1996),but they are used and optimized nearly exclusively for the production of biomass to isolate recombinant proteins. For safety reasons and to avoid any discussion on animal derived elements within the process, these should not be used as components of the fermentation broth. No animal derived peptone should be used. Specific fermentation processes - overcoming the disadvantage of low plasmid yields - were developed (Schleef et al., unpublished data). Additionally, the use of antibiotics may cause irritation if contained in the product. This theoretical aspect is of importance, since the detection of this kind of contamination may be difficult. An ideal cultivation medium does not contain any selection substance like this (e.g., www.~lasmidFuctovy.com). Complex growth media contain large amounts of non-defined substances such as peptones or extracts. Large differences in composition and quantity of the respective components negatively influence reproducibility. This fact also inhibits any optimization work on the cultivation process. Defined media usually contain one single carbon source, e. g., glucose or glycerol. By this, direct control of the substrate is possible, and in consequence optimal growth my be achieved. The presence of low molecular weight components is not that disadvantageous as larger components from complex sources are, especially in product recovery. However, in cases were the use of at least some complex components within the cultivation medium may improve the product yield, such “half-defined medium may be more economic.

I I Plasmid Manufacturing - An Overview

The most simple defined medium for the cultivation of E. coli is "M9", which is a minimal medium usually consisting of glucose as the carbon source, a mineral salt mixture and ammonia chloride as the nitrogen source. Glycerol as the carbon source within cultivation media has - compared with glucose - the advantage of being easier to prepare, since it may be heat-sterilized in the presence of free ammonia ions without leading to the Maillard-reaction. Additionally, glycerol may be used in higher concentrations than glucose. Also the generation of acidic acid, which inhibits bacterial growth, is lower compared to glucose (Holms, 1986). In Figure 4 the most important variable parameters of a 30 L cultivation of E. coli cells containing plasmid pUK21CMVf3 (7.6 kbp, ltanamycin resistance) are indicated. A synthetic M9 medium with glycerol as the only carbon source was used. The pH was regulated by automated addition of ammonia, if required. This was the only nitrogen source. Trace elements and vitamin B1 were added and the process started with a 15 L volume. The feeding medium was a concentrated glycerol solution (87 %), containing magnesium sulfate. The regulation was performed by the regulation of oxygen. The cultivation temperature was 37°C at a pH of 7.0. In cases were the oxygen level reached a pre-defined lower limit, the stirrer speed was increased. In cases were an upper limit was reached, feeding substrate was added to the culture (Figure 5).

200 -

-

I

207

50

'A

2

40

--.150-

I

a d

Fig. 4. Plasmid concentration, plasmid mass content and biomass of a 30 L fedbatch cultivation of €. coli K12 DH5a containing the plasmid pUK21CMVg (see also Figure 5; from: T. Schmidt, University of Bielefeld, Germany).

Fig. 5.

Process parameters

for the 30 L fed-batch cultivation o f E. coli K12 DH5a containing the plasmid pUKZlCMVJ3 (Figure 4). Temperature 37"C, pH = 7.0, lower pOz limit 30%, upper PO, limit 50% (from 1 Schmidt, University of Bielefeld, Germany).

C

30,

8 100n

I

2

5 -

.

20 -x a

a 50-

I I

lor>

0-

0

Cultivation time / h

. 8

g 50 25

0

0

20 30 Cultivation time I h

10

40

208

I

Cuilherme N. M. Ferreira et a/.

M 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 M

Fig. 6. Agarose gel electrophoresis analysis of untreated extracted samples from the 30 L fed-batch cultivation o f E. coli K12 DH5a containing the plasmid pUK21CMVJ

t/h

(Figures 4 and 5) to indicate the plasmid quality over the cultivation time (from T. Schmidt, University of Bielefeld, Germany).

Most important to monitor is the quality of the plasmid product within the cultivation. Figure 6 shows the samples taken during the cultivation process discussed above. Finally, special plasmid host system features, such as temperature-induced plasmid amplification during cell growth (Lahijani et al., 1996) and the release of endogenous nucleases during cell lysis (Monteiro et al., 1999) should be identified to improve the overall production performance. It was also shown that plasmid quality and quantity as well as growth rates and hence the overall productivity of a cultivation depends on the host cell used (Schleef, 1999). A judicious plasmid vector and host strain selection, combined with fermentation media and optimization of conditions, can result in plasmid yields as high as 220 mg L-' fermentation broth (Lahijani et al., 1996) and in a 40 % reduction on the RNA content during cell lysis (Monteiro et al., 1999), without the need for introducing any foreign chemical or enzyme. 3.3

Downstream processing of plasmid DNA

Following the fermentation a sequence of unit operations must be set up, essentially aimed at eliminating impurities in order to obtain a final plasmid DNA preparation complying with the approval specifications (Table 4). Figure 7 shows a typical process flow sheet for the production and purification of supercoiled plasmid DNA. The strategy is to select and integrate appropriate purification operations in two distinct regions. In the first region, called clarification and concentration (Figure 7) which is performed following the lysis step, high volumetric capacity and low resolution operations such as salt precipitation are performed in order to remove cell debris and structurally unrelated impurities, such as proteins and small molecular weight nucleic acids, while simultaneously concentrating and conditioning the plasmid DNA preparations for the purification region. In the second region, chromatographic operations are performed to separate supercoiled plasmid DNA from structurally related impurities, such as relaxed and denatured plasmid DNA, genomic DNA (gDNA),high molecular weight RNA and endotoxins (Figure 7).

7 7 Plasmid Manufacturing - An Overview

I

209

Cell lysis Mechanical Chemical

Clarification and concentration Precipitation (salts, PEG, alcohol) Membrane processes (microfiltration, ultrafiltration, and/or diafiltration)

Waste

I

Proteins Cell debris

Waste Host DNA RNA Proteins Salts Endotoxins

k' Purification

Adsorption processes (affinity, ion-exchange, hydrophobic interaction) Size-exclusion

Final product

Waste Host DNA RNA Proteins Salts Endotoxins Plasmid isoforms

Fig. 7. Process flow sheet for the large-scale purification of supercoiled plasmid DNA. Unit operations to be considered during process development are indicated together with the eliminated impurities.

210

I

Cuilherme N. M. Ferreira et a/

3.3.1

Cell [ y s k

The first critical step of downstream processing of plasmid DNA is cell lysis. At this step, all intracellular components including plasmid DNA, RNA, genomic DNA, endotoxins (LPS) and proteins are released. Together with cell wall and membrane fragments and other minor intracelular compounds (Table 3) this constitutes the mixture of molecules from which plasmids must be isolated. The extraction of large plasmid molecules from the cells without damage is a significant challenge for biochemical engineers. Shear sensitivity of plasmid and genomic DNA (gDNA) molecules (Levy et al., 1999b) as well as the high viscosity of the lysates (Ciccolini et al., 1998) are of major concern during plasmid release (Prazeres et al., 1999). The release and recovery of high amounts of intact, supercoiled plasmid DNA is crucial in order to obtain high overall process yields. Thus, the main objective of the lysis step is to efficiently disrupt the cells in order to release the intact supercoiled plasmid. Studies on mechanical disruption of E. coli cells for plasmid release (Carlson et al., 1995) revealed that only microfluidization and bead milling resulted in the recovery of intact plasmid molecules. The best recoveries of plasmid molecules achieved were 3 5 % with 65% cell disruption on a microfluidizer and 74% when 50% of the cells were disrupted on a bead mill. From the total plasmid released, 80 % and 94 % where in the supercoiled conformation after disrupting the cells on the microfluidizer and on the bead mill, respectively (Carlson et al., 1995). Increasing cell disruption yields led to the drop-off of the supercoiled plasmid recovery to less than 10% (Carlson et al., 1995). The low cell disruption yield, necessary to obtain high relative amounts of supercoiled plasmid, makes these lysis methods unsuitable for large-scale production processes. Variations of the alkaline lysis procedure first described by Birnboim and Doly (1979)were the procedure of choice for cell disruption for plasmid recovery (Chandra et al., 1992; Ferreira et al., 1999; Hines et al., 1992; Lahijani et al., 1996; Marquet et al., 1995; McClung and Gonzales, 1989; Prazeres et al., 1999; Prazeres et al., 1998; Schleef and Schorr, 1998). Generally (Sambrook et al., 1989), after resuspending the cells in an appropriate buffer containing RNase and lysozyme the cells are lysed by adding two volumes of a NaOH-SDS solution (e.g., 0.2 M NaOH, 1 % SDS). The extract is mixed gently and neutralized with 0.5 volumes of a high ionic strength solution (e.g., 3 M potassium acetate, acetic acid mixture). Although the use of RNase (Sambrook et al., 1989) or high temperatures (Chandra et al., 1992) during cell lysis is advocated to reduce the content of RNA, the use of RNase may raise considerable difficulties on process validation (USFDA, 1996a). Since process validation and approval is facilitated if no enzymes are used in the manufacturing process (Schleef, 1999), plasmid DNA manufacturing processes for clinical applications should avoid the use of such enzymes (Marquet et al., 1995). On the other hand, the need for large and efficient heat transfer equipment for raising, maintaining and further cooling the reaction temperature have a significant impact on process economics (Schleef, 1999). Moreover, no significant advan-

7 7 Plasmid Manufacturing - An Overview I 2 1 1

tage was found on lysing the cells either using RNase or high temperature (Ferreira et al., 1999). The alkaline lysis method exploits the denaturation of the double-stranded conformation of DNA at high pH values (Figure 8). After rapid neutralization, the complementary DNA strands cannot hybridize. Both inter- and intra-strand association occur, leading to the formation of random coils that precipitate (Rush and Warner, 1970; Sambrook et a]., 1989). If the DNA molecule is linked in a circular conformation, as is the case in plasmid molecules, contact points between the complementary DNA strands are maintained after alkaline denaturation (Rush and Warner, 1970; Sambroolc et al., 1989).These contact points allow for the re-association of the complementary strands after neutralization. The plasmid molecule is thus renatured through complementary hybridization (Figure 8 ) , remaining soluble.

...... ...... ......

...... ...... ...... ...... ......

. lysing solution

neutralisation

------+

(random coils)

neutralisation (pHd2.5)

. lysing solution

neutralisation (pH> 12.5)

(denatured plasmid) Fig. 8. Course o f alkaline lysis: under alkaline conditions hydrogen bridges are destroyed causing the separation of the complementary strands. Rapid neutralization o f the solution promotes intra-strand association and the formation o f insoluble random coils. Contact points between the complementary strands are

maintained in circular DNA, allowing complementary hybridization after neutralization. The exposition o f circular DNA to p H > 12.5 causes intra-strand association, leading to the formation of denatured circular DNA after neutralization.

212

I

Cuilherme

N. M.

Ferreira et a/.

Mixing is crucial during alkaline lysis (Prazeres et al., 1999) to avoid local pH extremes and the consequent irreversible plasmid denaturation (Rush and Warner, 1970) (Figure 8) while promoting efficient lysis of the entire cell population (Marquet et al., 1995). Mixing should be very gentle to avoid shearing the plasmid molecules (Levy et al., 1999b) and maintain gDNA with high molecular weight in order to maximize its precipitation from the plasmid preparation (Marquet et al., 1995; Prazeres et al., 1999). Such mixing conditions may be difficult to accomplish due to the viscosity variations during the lysis reaction (Ciccolini et al., 1998) and the non-Newtonian properties of the lysate (Ciccolini et al., 1999; Stephenson et al., 1993). After adding the neutralization solution, a precipitate containing denatured gDNA and RNA, SDS-protein complexes and cell debris is formed (Schleef, 1999) that aggregates forming a low-density, shear-sensitive matrix (Levy et al., 1999a, b). If high shear forces are applied, disaggregation of the matrix, breaking and resolubilization of gDNA fragments will occur (Levy et al., 199%; Marquet et al., 1995). Since the separation of gDNA fragments from plasmid DNA is extremely difficult in the subsequent downstream processing steps (Levy et al., 1999a), careful handling of the lysate at the end of the lysis stage avoiding the shearing of the precipitate is a very important issue. Removal of the precipitate formed can be achieved with any solid-liquid unit operation, provided that low shear forces are used. Centrifugation on fEed-angle rotors is the most frequently used operation at laboratory and pre-preparative scales (Ferreira et al., 1999; Lahijani et al., 1996; Maniatis et al., 1982; Prazeres et al., 1998). However, such operation is not suitable for the large-scale production of plasmid DNA. Industrial centrifuges usually operate with continuous feed flow to achieve capacity. The centrifugal acceleration of the liquid entering the centrifuge may cause shearing and, consequently, break the precipitate material and gDNA molecules (Theodossiou et al., 1997). Filtration i s thus the recommended operation to be used in large-scale plasmid production processes (Marquet et al., 1995; Theodossiou et al., 1997). Up to 67% plasmid recovery, with 46% purity with respect to the total recovered DNA, was achieved after removing 99% of the precipitate formed at the end of alkaline lysis with a 5 pm pore diameter filter (Theodossiou et al., 1997). Larger pore sizes (> 15 pm) were unable to fully retain the solids, thus not fulfilling the requirements of the operation (Theodossiou et al., 1997). Filter aids should be used to reduce the fdtration pressure in order to avoid shearing of the precipitate and redissolution of gDNA fragments (Theodossiou et al., 1997). Expanded bed adsorption (EBA) chromatography is a recent technique that allows for the early purification of biological molecules (Chase, 1994). By increasing the extraparticle void volume, small particulate material and viscous solutions can be applied to chromatography columns without leading to gel clogging. The adsorbents used in EBA are analogous to conventional fured-bed adsorbents, but with a higher density (approximately 1.2 g ~ m - achieved ~ ) by the inclusion of inert cores in the particles. Consequently, high flow rates are required to generate

7 1 Plasmid Manufacturing - An Overview

stable expanded beds (Chase, 1994).The physical performance of EBA is strongly influenced by the viscosity of the feedstocks and by the flow rates (Batt et al., 1995) which control contact time (Artolozaga et al., 1998), rate of mass transfer (Draeger and Chase, 1991), and the dynamic binding capacity (Chang and Chase, 199Ga, b). Recently, Zhang, et al. (1999)proposed the use of magnetically susceptible particles as a new generation of EBA adsorbents. With these adsorbents bed expansion is controlled with a magnetic field, therefore enhancing the operation performance by enabling the uncoupling of the bed expansion from the feedstock viscosity and flow rates (Batt et al., 1995). An appropriate selection of adsorbents and conditions permits removal of precipitated material, while capturing and concentrating the desired product in a single step (Chase, 1994). Expanded bed chromatography has been mostly used to recover proteins from unclarified yeast suspensions (Chang and Chase, 199613; Chang et al., 1995), E. coli homogenates (Barnfield et al., 1994),and mammalian cells culture broth (Batt et al., 1995). It has been recently applied for capturing plasmid DNA after the lysis step (Ferreira et al., 2000a; Varley et al., 1998). However, clarification steps prior to the expanded bed operation are required in order to enable an efficient capture of plasmid molecules (Ferreira et al., 2000a; Varley et al., 1998).

3.3.2

Pre-chromatography processing: clarification and concentration

Although most gDNA is denatured and precipitated during lysis, plasmid DNA normally represents only 2 % (w/w) of the total nucleic acids in E. coli lysates (Ferreira et al., 1999; Varley et al., 1998). Large amounts of RNA and proteins remain and must be removed in subsequent steps. Clarification and concentration steps designed to remove proteins and host nucleic acids and increase the plasmid mass fraction are recommended operations to prepare plasmid extracts for the subsequent purification steps (Marquet et al., 1995). In the clarification step removal of high molecular weight RNA is of major concern. The presence of endogenous nucleases in plasmid preparations obtained at the end of the alkaline lysis step can be advantageously used to remove high molecular weight RNA. Monteiro et al. (1999) demonstrated that RNA levels can be reduced by 40 % when the lysates are incubated at 37 "C with only 9 % plasmid loss. The proteins remaining in the lysates are usually removed by salting out, using a high concentration of chaotropic salts - a common operation in the biotechnology industry. Some chaotropic salts, such as LiCl (Chalrabarti et al., 1992) and ammonium acetate (Horn et al., 1995),were identified to have the additional advantage of precipitating high molecular weight RNA together with the proteins. However, in a comparative study on the clarification of a 4.8 kb plasmid extract with LiCl, ammonium and potassium acetate no significant difference on the removal of RNA was observed (Ferreira et al., 1999). All these chaotropes were able to reduce both the protein content (> 92 %) and the amount of RNA and other low molecular weight nucleic acids. Up to 75 % (w/w) plasmid purity degrees, corresponding to 100-fold purification factors, were achieved after clarifying the plasmid extracts with the chaotropes (Ferreira et al., 1999).However, the use of such substances may require

214

1 additional investment in applicable process elements for their removal from the Guiiherffle N. M . Ferreira et a/.

process stream as well as for quality assurance and may cause safety concerns. The concentration of plasmid DNA by polyethylene glycol (PEG) precipitation is commonly used after the clarification step to further remove small nucleic acids and to reduce the volume of process streams prior to chromatographic purification. While several polymer salt systems were tested at a small scale (Cole, 1991; Lis and Schleif, 1975; Nicoletti and Codorelli, 1993; Ohlsson et al., 1978), low step yields were reported when processing large volumes (Ferreira et al., 1999; Horn et al., 1995). Although plasmid precipitation with PEG is highly empirical (Prazeres et al., 1999), scaling-up of the PEG8000-NaC1 system is the common method for the concentration of plasmid streams prior to the final purification steps (Horn et al., 1995). PEG precipitation also enables a buffer exchange, preparing the plasmid extracts for the next purification step.

3.3.3

Chromatographic processing: purification o f supercoiled plasmid DNA

Purification of supercoiled plasmid DNA has been traditionally a molecular biology method, with laboratory research as the sole intended application for the final product. Saccharose or caesium chloride/ethidium bromide density gradient ultracentrifugation are standard molecular biology methods for plasmid purification (Sambrook et al., 1989). These methods are time consuming, difficult to scale up, they use toxic and mutagenic reagents and are thus unsuitable for the purification of plasmids for clinical purposes (Schleef, 1999).Table 5 compares standard laboratory and clinical manufacturing processes for producing supercoiled plasmid DNA. Chromatography is the method of choice for the large-scale purification of supercoiled plasmid DNA. The size of the target nucleic acid molecules, and by consequence the chemical properties related to the molecule size, such as charge and hydrophobicity as well as the exposition or the accessibility of the nucleotide bases within single- or double-stranded nucleic acid molecules, respectively, and the topological constraints due to supercoiling, are explored in the interaction of nucleic acids with solid supports. In any case, the objective is to selectively isolate and purify plasmid DNA from impurities, such as gDNA, endotoxins, RNA and proteins as well as removing traces of contaminants, such as isopropanol and PEG 8000, introduced in the previous downstream processing steps. Reverse-phase, hydrophobic interaction and ion-pair chromatography

Reverse-phase chromatography explores the hydrophobicity of the nucleotide bases. It has been used in molecular biology to detect partial denaturation on DNA fragments (Huber and Berti, 1996),to purify oligonucleotides and DNA fragments (Floyd et al., 1986; Huber et al., 1995; Patient et al., 1979),to purify plasmid DNA (Best et al., 1981; Colote et al., 1986) and to separate plasmid topoisomers (Kapp and Langowski, 1992). Reverse-phase chromatography is based on the interaction of hydrophobic, nonpolar molecules or regions of the molecules, with non-polar immobilized ligands. Bound molecules are eluted with decreasing polarity gradients by adding organic

7 7 Plasmid Manufacturing

- An

Overview

Comparison of laboratory methods and large-scale pharmaceutical processes for purifying supercoiled plasmid D N A

Table 5.

Process Stea

Laboratorv Method

Large-Scale Process

Cell lysis

RNase, lysozyme

no enzymes only GRAS" reagents

Removal of cell debris

centrifugation

filtration, centrifugation or expanded bed chromatography

Removal of host contaminants: (e.g., RNA, gDNA, proteins, endotoxins)

RNase, proteinase K, organic solvents (phenol, chloroform)

salting out, PEG precipitation

Plasmid enrichment

alcohol precipitation

alcohol precipitation but PEG precipitation is preferred

Plasmid purification

ultracentrifugation (mutagenic reagents) (ethidium bromide) IEC" (gravity flow columns provided in commercial kits) RPC" (organic, toxic solvents)

IECa and or SEC" (use only GRAS reagents)

a

I

I

Abbreviations: G U S , generally regarded as safe; RPC, reverse phase chromatography;IEC, ion-exchange chromatography: SEC, size-exclusionchromatography

modifiers to the column eluent (Brown and ICrstulovic, 1979; Sofer and Hagel, 1997). Therefore, the solutes are eluted in the order of decreasing polarity or increasing hydrophobicity (Brown and ICrstulovic, 1979; Sofer and Hagel, 1997). When the target molecules are polar, which is the case for nucleic acids, however, reverse-phase principles are achieved in a technique termed ion-pair reverse-phase chromatography or simply ion-pair chromatography (Huber, 1998). In this technique, the solute polarity is reduced by adding amphiphilic organic ions to the buffers. These ions establish ionic interactions with the target molecules, forming hydrophobic non-polar ion pairs that bind to reverse-phase resins (Huber, 1998; Meyer, 1994). The major factors affecting the retention time of nucleic acids in reverse-phase columns are size, base composition and secondary structure (Colote et al., 1986). Nucleic acids of increasing size are expected to elute with increasing retention times in reverse-phase chromatography (Patient et al., 1979). However, the effect of the nucleic acid size is not independent on the base composition (Colote et al., 1986; Huber and Berti, 1996; Kapp and Langowski, 1992; Patient et al.,

I

215

216

I

Cuifherme N. M. Ferrefra et a/.

1979). Experimental data on reverse-phase chromatography of nucleic acids indicate that AT-rich DNA binds more tightly to reverse-phase matrices than DNA of equivalent size which is not AT-rich (Patient et al., 1979). The explanation for this result relies in the thermolability of the AT-rich regions (Patient et al., 1979). Partial denaturation of the double helix at the AT-rich locations forming single stranded regions within the molecule leads to the exposition of the bases to the ligands, and thus to an increase of the hydrophobic interaction strength. This hypothesis is consistent with the observation that single-stranded oligonucleotides bind more tightly to reverse-phase matrices than duplex DNA fragments of the same length (Colote et al., 1986; ICapp and Langowsli, 1992; Patient et al., 1979). In ion-pair reverse-phase chromatography an opposite trend has been reported: Partial denaturation of DNA fragments at AT-rich regions decreases the retention factor when compared with native duplex molecules of similar size (Huber, 1998; Huber and Berti, 1996). This apparent contradiction is explained by the mode of molecule-ligand interaction. In ion-pair chromatography nucleic acid molecules interact with the immobilized ligands through amphiphilic ions. Formation of single-stranded regions at the AT-rich locations implies a decrease of the charge density from two to one charges per helix repeat. Thus, less amphiphilic ions bond to these regions, and, consequentially, lower hydrophobicities of the ionic pair are achieved (Huber and Berti, 1996). In either situation, ion-pair (Huber, 1998; Huber and Berti, 1996) or reversephase chromatography (Colote et al., 1986; Patient et d., 1979), the formation of single-stranded regions within the molecule depends on the denaturation conditions, such as, e. g., temperature or formamide concentration. Thus, by selecting and optimizing such conditions two different separation dimensions can be explored (Huber, 1998): one dimension is the size of the molecules and the other is the base sequence. The secondary structure of nucleic acids also influences the retention time in reverse-phase chromatography (Colote et al., 1986; Huber, 1998; ICapp and Langowski, 1992). As mentioned above, single-stranded nucleic acids are more retained in reverse-phase chromatography than equivalent size duplex molecules. On the other hand, the torsion constraints due to supercoiling render the bases more accessible to interact with the stationary phase. Therefore, the retention times are expected to be higher with increasing supercoiling (Colote et al., 1986). This hypothesis was confirmed by the separation of plasmid topoisomers by reverse-phase chromatography,were the elution profiles of a 2.7 kb plasmid followed the order of increasing superhelix density (Kapp and Langowsli, 1992). Reverse-phase and ion-pair chromatography have been used to purify supercoiled plasmid DNA from crude cell lysates (Colote et al., 1986; Green et al., 1997).In either case, low molecular weight RNA, gDNA fragments and linear plasmid DNA were completely separated from the supercoiled plasmid DNA, with the last being more strongly retained (Colote et al., 1986; Green et al., 1997). However, a second pass in the column, i. e. recycling, is suggested due to the failure in completely removing high molecular weight RNA from the plasmid peak (Colote et al.,

1 7 Plasmid Manufacturing

- An

Overview

1986). Genomic DNA and endotoxins remain bound to the column and are removed after sanitizing the column with 1 M NaOH (Green et al., 1997). The need to use organic solvents to elute the plasmid from reverse-phase columns constitutes a disadvantage of these purification techniques (reverse-phase and ion-pair chromatography). Most of the organic solvents are toxic or mutagenic, volatile, and even explosive. Thus, special safety concerns regarding the manufacturing personnel and facility may be required, such as the design of explosion proof facilities and use of appropriate protection masks (Marquet et al., 1995). Although large-scale plasmid purification schemes based on reverse-phase and ionpair chromatography are scarce, these techniques have been widely used for molecular biology applications (Huber, 1998),and they are suggested techniques for the characterization of final plasmid DNA preparations (Marquet et al., 1997a, b). Ion-exchange chromatography

The polyanionic structure of nucleic acids can be explored in ion-exchangechromatography. The strategy relies upon binding the nucleic acids to positively charged matrices, followed by selective elution with increasing ionic strength. It has been used in molecular biology (Chandra et al., 1992; Hines et al., 1992; Merion and Warren, 1989) to separate DNA restriction fragments (Drager and Regnier, 1985; Muller, 1986; Newton et al., 1983; Yamakawa et al., 1996) and plasmid topoisomers (Onishi et al., 1993),as well as in the purification of plasmid DNA for gene therapy and DNA vaccination applications (Ferreira et al., 1998; Ferreira et al., 1999; Ferreira et al., 2000a; Lahijani et al., 1996; Prazeres et al., 1998; Varley et al., 1998; Schleef, 1999). As mentioned above, nucleic acid molecules have one negative charge per base, i.e. the overall net charge equals the number of bases in the molecule. The expected elution profde thus follows the order of increasing molecule size, with the larger nucleic acid molecules being eluted at higher ionic strengths (Colpan and Riesner, 1984; Yamakawa et al., 1996).The relative retention time between single- (ss) and double-stranded (ds) nucleic acids can also be predicted by this rule depending on the overall molecule charge, therefore, on the number of bases that constitute the molecule (Colpan and Riesner, 1984). Inversions in the predicted retention factors were observed in the anion-exchange purification of dsDNA fragments on a variety of stationary phases (Huber, 1998; Muller, 1986; Yamaltawa et al., 1996). It was suggested that the retardation of some dsDNA fragments i s due to their high AT content (Muller, 1986; Yamakawa et al., 1996), therefore indicating that the ion-exchange elution profile of nucleic acids i s also sequence-dependent. A more detailed analysis revealed that peak retardation i s not always proportional to the high AT content of the dsDNA sequences (Huber, 1998; Yamakawa et al., 1996). It was found that binding of nucleic acid molecules onto anion exchangers i s favored by optimal interaction with the curvature of the particle pores (Colpan and Riesner, 1984) (Figure 9). Therefore, a conformation-dependent separation was proposed (Huber, 1998; Yamalawa et al., 1996). Nudeic acid bending promotes better fits within pore curvatures, enabling more charges to interact with the solid phase (Figure 9A, B), thus

I

217

218

I

Cuilherme

N. M.

Ferreira et a/.

A

B

C

D

Fig. 9. Representation o f possible interactions with a particle pore of linear (A) and bent (B)dsDNA fragments, supercoiled (C)and relaxed (D) plasmid DNA.

leading to higher retention factors. Another physical explanation for the stronger binding of bent dsDNA fragments relies upon the appearance of a dipole character due to the local compression of charges at bent regions (Huber, 1998). Bending increases the local charge density and, by consequence, stronger electrostatic interactions take place at these locations. Nucleic acid bending is mainly the result of poly A runs (Sinden, 1994). Therefore bent dsDNA fragments exhibit an inherently high AT content. Consequently, the observed sequence-dependent elution profile of dsDNA fragments (Miiller, 1986; Yamakawa et al., 1996) is an artifact that hides the phenomena responsible for peak retardation, which is sequence bending. The separation of plasmid topoisomers (Onishi et al., 1993), mainly the relaxed and supercoiled plasmid isoforms (Lahijani et al., 1996; Marquet et al., 1995; Prazeres et al., 1998) with anion-exchangers, confirms the conformation-dependent retention hypothesis outlined above. Supercoiled DNA is more stretched, compacted and bent than its relaxed isoform, therefore presenting higher charge densities and better fits with the particle pores’ curvature (Figure 9C, D). Selectivity and peak resolution are other factors affected by the interaction of nucleic acids with the solid phase (Colpan and Riesner, 1984; Huber, 1998; Thompson, 1986). At least 50 nm pores were needed to completely separate a mixture of RNAs with molecular weights in the range of 25,000 Da to 1,000,000 Da, in which case the elution profile followed the order of increasing size (Colpan and Riesner, 1984). Circular RNA molecules added to the RNA mixture were lastly eluted, at the highest ionic strength (Colpan and Riesner, 1984). Furthermore, a 4.2 kb plasmid was not separated from an E. coli extract with particles of less

1 1 Plasrnid Manufacturing

- An

Overview

than 400 nm pores (Colpan and Riesner, 1984).These experimental results suggest that nucleic acids must fit well into the particle pores, within their maximum extension, while matching the pore inner curvature as well in order to achieve optimal resolution and selectivity (Colpan and Riesner, 1984; Huber, 1998; Thompson, 1986). Elution buffer, gradient former, and flow rate also influence the selectivity and resolution in ion-exchange chromatography (Sofer and Hagel, 1997). While the buffer pH has no apparent effect on the separation of nucleic acids with strong anion exchangers, a decrease in pH usually requires higher ionic strengths to elute the bound nucleic acids in week anion exchange matrices (Colpan and Riesner, 1984; Huber, 1998).This result is attributed to the increase of the charge density of week anion exchangers at lower pH values (Colpan and Riesner, 1984; Sofer and Hagel, 1997), being therefore an obvious corollary when using such supports (Sofer and Hagel, 1997). Variation of the cation (lithium, sodium, potassium and caesium) while keeping chloride as the anion showed that smaller cations eluted the nucleic acids at lower ionic strengths (Huber, 1998). This observation, however, simply reflects the decrease of the activity coefficients with increasing cation radius (Atlcins, 1990; Muller, 1986),with particular activity values and therefore ionic strengths achieved at higher concentrations for larger cations (Muller, 1986). Thus, the variation observed on the retention factors by changing the gradient cation is due to gradient slope alterations which are unrelated to the system selectivity and therefore should not be confused with. It is well known in ion-exchange chromatography that longer gradients usually improve the resolution of two components, which is at maximum with gradients within the range of 5-20 column volumes (Sofer and Hagel, 1997). While the gradient slope has no significant effect in the separation of low molecular weight dsDNA fragments it is very important to maintain shallow gradients, if the separation of high molecular weight nucleic acids is desired (Huber, 1998; Yamalcawa et al., 1996). Regarding the purification of plasmid DNA, removal of impurities from the plasmid preparations is also improved with shallow gradients (Colpan and Riesner, 1984; Huber, 1998). Furthermore, the optimization and selection of shallow NaCl gradients leads to the partial separation of relaxed, supercoiled and denatured plasmid DNA (Prazeres et al., 1998) which is not accomplished with sharp NaCl steps (Ferreira et al., 1998; Ferreira et al., 1999). The optimal gradient, however, should combine acceptable resolutions with relatively fast purification times in order to enable high throughputs and maximize the purification productivity. The flow rate has no significant effect on the resolution of nucleic acids (Colpan and Riesner, 1984),with the sole disadvantage of high flow rates being the dilution of the eluted nucleic acids, However, in process chromatography the flow rate controls the residence time which is a function of target molecule diffusivity, sample viscosity, and particle size (Sofer and Hagel, 1997).The loss of plasmid DNA in the flow through the column (Prazeres et al., 1998) can thus be attributed to high flow rates and, therefore, shorter residence times than required for efficient plasmid binding.

I

219

220

I

Cuilherme N M. Ferreira et al.

A carry-over phenomenon, which i s the contamination of a nucleic acid peak by impurities that normally elute earlier, has been described (Colpan and Riesner, 1984; Huber, 1998).This phenomenon was attributed to the interaction of the nucleic acid with the impurity, either directly or promoted by di- or multivalent metal ions such as Ca2+ or Mg2+ (Colpan and Riesner, 1984; Huber, 1998). Addition of chelating agents such as EDTA to the elution buffers to complex the metal ions, and/or denaturating agents such as urea, formamide or isopropanol, to suppress intermolecular hydrophobic interactions is thus suggested (Huber, 1998). Furthermore, highly charged cationic molecules such as urea, spermidine and spermine are known to selectively compact supercoiled plasmid DNA (Murphy et al., 1999), thus improving its separation from relaxed plasmid DNA and other impurities in anion-exchangechromatography (Colpan and Riesner, 1984; Murphy et al., 1999). Although high protein and nucleic acid clearances can be achieved in the anion exchange purification of supercoiled plasmid DNA, this operation is hampered by the inevitable contamination of plasmid DNA with high molecular weight RNA, gDNA and endotoxins (Ferreira et al., 1999; Prazeres et al., 1999). While carryover phenomena contribute to this fact to some extent, unselectivity of anion exchangers (Huber, 1998; Muller, 1986) and diffusion constraints of the macromolecules in the particle pores (Huber, 1998; Lyddiatt and O'Sullivan, 1998) are better explanations for this. In anion exchange chromatography molecules with higher charge densities are expected to elute at higher ionic strengths. As previously mentioned, the elution profile of linear, unbent nucleic acids i s expected to be a function of the molecule size. If a mass action law (Bellot and Condoret, 1993; Brooks and Cramer, 1992) is considered for the adsorption of nucleic acids onto anion-exchangers, a linear relationship between the elution ionic strength and the nucleic acid size is expected. However, experimental elution profiles of linear dsDNA fragments showed a drastic compression of the elution position for fragment sizes above 180 bp (Huber, 1998; Muller, 1986), suggesting unselectivity of anion exchangers after this limit (Figure 10). Competitive binding between plasmid DNA and impurities present in the feedstocks, such as high molecular weight RNA (Figure ll), as demonstrated by batch adsorption experiments (Ferreira, 2000) provides further evidence for the unselectivity of anion exchangers. Therefore, the selective elution of molecules with high charge density such as plasmid DNA, gDNA, high molecular weight RNA and endotoxins, can thus be difficult to accomplish in a single anion-exchange step. The diffusional constraints of macromolecules in the particle pores also contribute to the recovery of contaminated plasmid DNA. Large plasmid DNA molecules are partially excluded from particle pores smaller than 400 nm (Colpan and Riesner, 1984),being adsorbed at the particle surface (Ferreira et al., 2000b; Ljunglof et al., 1999). On the other hand, RNA i s small enough to be accommodated in the particle pores. Therefore, size exclusion mechanisms can override the ionexchange principles (Thompson, 1986), with the eluted RNA diffusing slowly through the pores of the particles and through the column. It is important to

0,50

Fig. 10. Plot o f the eluting NaCl activity versus dsDNA fragment size n (bp) (Muller, 1986).

- +

0,45 h

5 m

Y

0,40

1 I

0,35

Fig. 11. Adsorption isotherm for the binding o f high molecular weight E. coli W1<6 RNA (open circles, dashed line) and a 4.8 Itb plasmid (closed circles, solid line) onto a strong anion-exchanger (Streamline QXL) at 0.5 M NaCl and 25 "C (Ferreira,

2000).

0

2

4

6

8

10

12

C*(mg/l)

note that the occurrence of this phenomenon can be behind the apparent unselectivity of anion exchangers discussed above. Even if the nucleic acid molecules were selectively eluted with different ionic strengths, retardation due to restricted diffusion within the particle pores can lead to the recovery of contaminated samples. The diffusional constraints hypothesis is consistent with the experimental evidence of complete separation of plasmid DNA from RNA in non-porous supports (Stowers and I<eim, 1988),in supports with pores accessible to plasmid molecules (Colpan and Riesner, 1984), and with the better resolutions obtained when using shallow ionic strength gradients (Huber, 1998; Lahijani et al., 1996) or compacting agents (Colpan and Riesner, 1984; Murphy et al., 1999). Extrapolating to other charged impurities, diffusion restriction also explains the co-purification of endotoxins (Wicks et al., 1995) and small gDNA fragments (Marquet et al., 1995; Prazeres et al., 1999; Schleef, 1999) with plasmid DNA. One of the major limitations of process chromatography of plasmid DNA is the lack of capacity of the majority of the commercial adsorbents (Prazeres et al., 1999). Batch adsorption experiments performed with Q-sepharose media (190 n m pores) showed that the maximum binding capacity increased from 0.7 mg mL-l to 3 mg mL-l by decreasing the mean particle size from 200 pm to 34 pm, respec-

222

I

Cttilherme N. M. Ferreira et a/.

4’0 3,0

2,O

0,o

HP

FF

Fig. 12. Correlation between the maximum binding capacity with the particle radius. Data obtained for a 4.8 kb plasmid contacted with Q-sepharose High Performance (HP), FastFlow (FF), and Big Beads (BB) a t 25°C and 0.5 M NaCl (Ferreira et al., 2000b). The maximum binding capacities were expressed in terms of bead volume (line graph) and in terms o f bead surface area (bar graph).

--E .

-

BB

Media

tively (Ferreira et al., 2000b) (Figure 12). Further evidence of plasmid surface binding is the common saturation limit obtained when expressed in terms of mean bead surface area (Figure 12) and the fact that this limit was achieved when only 4% of the total binding sites were occupied by plasmid molecules (Ferreira et al., 2000b). The dynamic maximum binding capacity also depends on the particle size, as revealed by the decrease from 5.3 mg mL-’ to 3.4 mg mL-’ when the mean particle size of Q-sepharose XL was increased from 55 pm to 90 pm (Ljunglof et al., 1999). Plasmid surface binding was also confirmed by direct visualization of plasmid-saturated anion exchangers in confocal microscopy were small adsorption layers thickness (7.3 pm) were measured (Ljunglof et al., 1999). The major consequence of plasmid surface binding is that large columns have to be used with low capacity utilization in order to achieve high plasmid adsorption yields (Ferreira et al., 200Ob). During scale-up of chromatography, the final column volume depends on the sample load that should be maximized to increase throughput and on the capacity utilization which is dictated by the amount and type of molecules that bind to the anion exchanger (Sofer and Hagel, 1997). As mentioned above, endotoxins, high molecular weight RNA and gDNA also bind to anion exchangers, therefore competing with plasmid molecules for the available binding sites. In conjunction with the relatively low amount of plasmid DNA in E. coli extracts which is about 2 % (w/w) of the total nucleic acids (Ferreira et al., 1999; Varley et al., 1998). Binding of impurities limits the operational sample load. Anion exchange purification of plasmid DNA directly from the E. coli extracts, either in fxed (Ferreira, 2000) or expanded bed mode (Ferreira et al., 2000a), revealed that less than one column volume load improved the operation performance, measured as the best plasmid recovery yield and purity balance (Figure 13). Plasmid loads were increased, however, with the introduction of appropriate clarification operations directed to the removal of impurities prior the the anion-exchange step (Ferreira et al., 2000a).

I I Plasmid Manufacturing - An Overview Fig. 13.

Correlations between the plasmid recovery yield (open circles, dashed line) and purity (closed circles, solid line) with the loaded volume (V). Vo, sedirnented column volume.

8

120

30

100

25

80

20

7 3 95 5

v

9 60 m F

.-> . I .

40

lo

20

5

0

2

0 0.5

1

1.5

2.5

Loaded volume (VNo)

In spite of the limitations described above, anion exchange chromatography has been used as one of the first purification steps to capture and concentrate plasmid DNA (Ferreira et al., 1999; Ferreira et al., 2000a; Lahijaiii et a]., 1996; Prazeres et al., 1998; Varley et al., 1998). Binding of plasmid DNA to anion exchange matrices at high ionic strengths (> 0.5 M NaCI) significantly reduces impurity levels (Ferreira et al., 1999). However, in this operation high molecular weight RNA, gDNA and endotoxins are co-purified with plasmid DNA. Therefore, further purification operations are required (Marquet et al., 1995; Prazeres et al., 1999). Size-exclusion chromatography The co-purification of gDNA and endotoxins with plasmid DNA during anion exchange chromatography is inevitable (Prazeres et al., 1999). However, this problem can be addressed by using size-exclusion chromatography (SEC) instead (Bywater et al., 1983; Ferreira et al., 1997; Horn et al., 1995; McClung and Gonzales, 1989; Raymond et al., 1988; Suominen et al., 1984) or ion-exchange (Ferreira et al., 1998; Ferreira et al., 1999; Marquet et al., 1995) as well as specific reagents (Colpan et al., 1995). The difference in size between molecules is the basic property explored in sizeexclusion chromatography. The separation is governed by the accessibility of the molecules to the particle pores, with the larger molecules being excluded and therefore being less retained in the column. It is estimated that the pore size needs to be at least 2 times greater than the molecule size to allow minimal permeation (Sofer and Hagel, 1997). One of the most important operational parameters regulating SEC is the flow rate. The resolution of macromolecules generally decreases with increasing flow rates (Sofer and Hagel, 1997). On the other hand, low flow rates can also lead to poor resolutions in SEC, particularly for permeated molecules, i. e., small molecular weight molecules to which the media pores are completely accessible (Sofer and Hagel, 1997).There is an optimal range of operating flow rates that depends on the support - mainly the particle diameter - and on the particular loaded mixture and

I

223

224

I objective of the

Guilherwre N. M. Ferreira et a/.

SEC purification step. Even though this range is normally suggested by the support manufacturer (Pharmacia-Biotech, 1997a; Pharmacia-Biotech, 1997b) flow rate optimization for a particular application should be performed. Flow rate is one factor that determines SEC throughput, the other is the sample load. It is well known that in SEC the maximum solute concentration i s only set by physicochemical restrictions and that high solute concentrations can be loaded to the column (Sofer and Hagel, 1997). However, the viscosity of highly concentrated mixtures leads to a decrease of peak resolution (Sofer and Hagel, 1997).\Xihile for proteins this i s only observed with very high concentrations, e. g., > 75 mg BSA per mL-' (Sofer and Hagel, 1997),viscosity effects occur when loading high molecular weight nucleic acid mixtures at concentrations as low as 200 pg mL-l (Suominen et al., 1984). The restrictions on the concentration of the loaded mixture turns the loaded volume the sole productivity determining factor in SEC (Sofer and Hagel, 1997). However, high sample volumes produce large sample zones in the column, therefore leading to a decrease in the resolution (Sofer and Hagel, 1997). As a general rule, the loaded volume in SEC should be kept within 2-6 % of the column volume (Preneta, 1989; Sofer and Hagel, 1997). By exploring the reduction of the plasmid hydrodynamic radius due to supercoiling (Boles et al., 1990; Vologodskii et al., 1992) and by selecting a size-exclusion support with appropriate selectivity it i s possible to fractionate the different DNA molecules, with baseline separation of RNA and other smaller molecules such as endotoxins and proteins (Ferreira et al., 1997) (Figure 14). Two different SEC supports have been used for this purpose: Superose 6 (Ferreira et al., 1997; McClung and Gonzales, 1989) and Sephacryl SlOOO (Bywater et al., 1983; Ferreira et al., 1997; Horn et al., 1995; Raymond et al., 1988). In both cases high molecular weight nucleic acids, such as gDNA and plasmid DNA, are excluded or eluted close to the support exclusion limit, while small molecular weight impurities, such as RNA, proteins, and endotoxins, are retained in the column (Bywater et al., 1983; Ferreira et al., 1997; Horn et a]., 1995; Raymond et al., 1988) (Figure 14). In Superose G columns with 40 X 10' Da and 450 bp linear DNA exclusion limits (Pharmacia-Biotech, 1997b) high molecular weight DNA molecules, including gDNA and plasmid DNA, are excluded from the matrix, being well separated from the low molecular weight impurities (Ferreira et al., 1997; McClung and Gonzales, 1989) (Figure 14A). Even though up to 600 pg of a 7.3 ltb plasmid were purified directly from E. coli extracts in a single cycle (McClung and Gonzales, 1989), the removal of gDNA and the isolation of supercoiled plasmid DNA i s difficult to accomplish with this support (Ferreira et al., 1997). Furthermore, SEC purification of E. coli extracts containing a range of plasmids up to 150 kb showed that these molecules were eluted with the same retention time, always in the column void volume (McClung and Gonzales, 1989). The size-exclusion separation of genomic and plasmid DNA can be accomplished with matrices having larger pores (Figure 14A). With exclusion limits of 100 X 10' Da for proteins, 20 kb for linear DNA, and 400 nm for spherical particles

1 7 Piasmid Manufacturing - An Overview Fig. 14. (A) Size-exclusion chromatography elution profiles in Sephacryl 51000 (closed circles, solid line) and Superose 6 (open circles, dashed line), and (B) ethidium bromide stained 1 % agarose gel o f 50 pL aliquots o f fractions 15 t o 2 7 o f the Sephacryl SlOOO purification (gDNA, genomic DNA; L, linear plasmid; OC, open circular plasmid; SC, supercoiled plasmid). Data refers to 5 m L fractions and 16 x 1,000 m m columns.

A 1.o

0.8

E a

0.6 0.4 0.2

0.0 0

5 10 15 20 25 30 35 40 45 50 55 60 fraction

B 15

16

17

18

19

20

21

22

23

24

25

26

27

gDNA

L; oc

sc

(Pharmacia-Biotech, 1997a), Sephacryl SlOOO enabled such separation (Ferreira et al., 1997; Horn et al., 1995). Within the first peak eluted from the column, host gDNA runs slightly ahead of plasmid molecules in the column void volume followed by relaxed plasmid and then supercoiled topoisomers (Ferreira et al., 1997; Horn et al., 1995) (Figure 14B). A judicious selection of the eluted fractions, after performing an analysis by agarose gel electrophoresis, permits the recovery of almost pure supercoiled plasmid DNA with about 70 % recovery yield (Ferreira et al., 1997; Ferreira et al., 1999). The separation between genomic and plasmid DNA achieved with Sephacryl SlOOO media, however, depends on the plasmid relative concentration. Individual gDNA peaks were observed in SEC elution profiles of unclarified E. coli plasmid

I

225

226

I extracts (Ferreiraet al., 1997)and with up to Gui/herme N. M. ferreira et a/.

1 pg mL-' plasmid concentrations (Bywater et al., 1983).These results are clear evidences that minimum plasmid/gDNA relative concentrations can improve the resolution between plasmid DNA and gDNA, as suggested by Bywater et al. (1983). However, the utilization of tandem size-exclusion columns which is equivalent to an increase of the column length is suggested in order to improve the resolution between gDNA and plasmid DNA (Horn et al., 1995). The sole disadvantage of such a strategy is - especially if desired for manufacturing scales - the high dilution factors achieved, and consequently, the loss of material due to detection limitations and the large volumes to be handled. As for proteins, SEC purification of plasmid DNA is also limited by low throughputs. Apart from the low flow rates, small loaded volumes must be used to improve the separation. Throughput, however, can be increased if concentrated samples with less relative amounts of impurities are loaded. Therefore, SEC should be one of the final purification steps in downstream processing of plasmid DNA (Ferreira et a]., 1997; Marquet et al., 1995).In this situation, SEC can be used to exchange the process buffer for an adequate formulation or storage buffer (Marquet et al., 1995; Prazeres et al., 1999) ensuring rigorous process control over small contaminants introduced during the previous downstream processing operations (Marquet et al., 1995). Furthermore, SEC is probably the only way to remove endotoxins from plasmid DNA preparations without the need for particular purification steps, such as affinity chromatography with histamine or polymixin B (Sofer and Hagel, 1997) targeted to this impurity. Up to 8000-fold reduction of the endotoxin level were achieved with a single SEC step, leading to a final plasmid preparation complying with the regulatory agencies specifications (Horn et al., 1995; Marquet et al., 1995). Triplex affinity chromatography

The formation of triple helices between two pyrimidine strands and one purine strand was discovered soon after the structure of double-stranded DNA was determined (Neidle, 1994). Complementary hydrogen binding interactions are responsible for the specificity of the third strand interaction (Sinden, 1994).As mentioned above, Watson-Crick base pairs still have available sites for hydrogen binding (Figure 1) both in the major and minor groves. However, since the bases are already involved in Watson-Crick bonds. The third strand must bind to another side of the bases in a hydrogen binding pattern termed Hoogsteen base pairing (Sinden, 1994) (Figure 15). In this pattern the third strand thymine recognizes AT base pairs forming T-ATtriplets, and the third strand cytosine that must be protonated at the N3 position (Neidle, 1994) recognizes GC Watson-Crick pairs forming Cf-GC triplets (Neidle, 1994; Sinden, 1994). The requirement for protonation of the third strand cytosine base leads to a stability pH dependence of the triplets formed with this base (Schluep and Cooney, 1998).These triplets are formed at pH values close to cytosine pKa (= 4.8) and destabilized by increasing the pH (Neidle, 1994; Sinden, 1994).The kinetics of triple helix formation is generaly low, and triple helices are stabilized by high ionic strengths (Wils et al., 1997).

H'

Fig. 15. Hoogsteen binding of a third base ( T o r P ) t o a Watson-Crick pair (ATor CC)

forming T-AT and P - G C triplets (Neidle, 1994).

Affinity chromatography based on the formation of triple helices between oligonucleotides linked to solid supports and specific sequences introduced in target plasmid molecules has been developed (Ito et al., 1992; Schluep and Cooney, 1998; Wils et al., 1997) (Figure 16). Apparently, these supports have higher affinities for plasrnids in supercoiled conformations than for the relaxed topoisomers (Wils et al., 1997). Up to 6 2 % recovery yields were achieved with simultaneous reduction of the RNA and gDNA content to undetectable levels by agarose gel electrophoresis (Schluep and Cooney, 1998) and 0.1 % (w/w), respectively (Ito et al., 1992; Schluep and Cooney, 1998; Wils et al., 1997).

-

n a

n

am,

nnnn n

An

n

n

a-+ 1 -

(Wash)

(Loading)

(Eltition)

spacer arm (thin line), being therefore sepaFig. 16. Schematic illustration of triple helix rated from the impurities (triangles) present in affinity purification of plasmid DNA. Plasmid DNA is captured by the oligonucleotides (thick the feedstock. line) bonded to a solid support (S) through a

228

I

Guilherme N. M. Ferreira et al

Limitation of this separation are the small reduction of the endotoxin levels only 2-fold reduction on the endotoxin levels were achieved (Wils et al., 1997) and an enrichment on denatured plasmid DNA (Schluep and Cooney, 1998). These limitations in conjunction with the poor capacity of the supports up to 50 pg mL-l (Wils et al., 1997) constitutes a technological problem that makes triple helix affinity purification of plasmid DNA still economically unfeasible. 3.4

Purification Strategies

Plasmid DNA purification strategies rely on the selection and integration of purification operations mostly based on Chromatography. The choice of using one or more chromatographic steps with contrasting selectivities is determined by the nature and distribution of residual impurities and contaminants as well as by the anticipated plasmid dosage (Marquet et al., 1995). Table 6 describes the most frequently used clinical-grade plasmid DNA purification processes for applications published so far or communicated to the authors. Size-exclusion chromatography using Sephacryl S 1000 works exceptionally well for endotoxin, RNA and protein removal, enabling a mild separation of supercoiled plasmid DNA from its topoisomers (relaxed and denatured) and from gDNA. SEC may be used as the single chromatographic step to purify low amounts of plasmid DNA (10-300 pg) for phase I and phase I1 clinical trials (Marquet et al., 1995). However, clarification and concentration steps before SEC purification are required in order to increase the operation selectivity and throughput (Ferreira et al., 1997; Horn et al., 1995) (see Table 6). Table 6. The purification

of clinical grade plasmid DNA; all described strategies are based on the alkaline lysis method. For further applications see also Chapters 4 and 6 Target disease

Purification Strategy Clarification and concentration

Reference Purification

isopropanol precipitation ammonium acetate precipitation PEG precipitation

SEC

(Horn et al., 1995)

Melanoma

isopropanol precipitation ammonium acetate precipitation PEG precipitation

fixed bed IEC SEC

(Horn et al., 1998)

Cancer

-

Expanded bed IEC SEC

(Varley et al., 1998)

Cystic fibrosis

isopropanol precipitation ammonium sulfate precipitation

HIC desalting

~~~~~~~

(Diogo et al., 2000)

Abbreviations: PEG, polyethylene glycol; IEC, ion-exchange chromatography; SEC, size-exclusion chromatography; HIC, hydrophobic interaction chromatography

1 1 Plasmid Manufacturing

-

An Overview

To purify higher plasmid amounts two chromatographic steps should be performed. Anion exchange chromatography has been used as one of the first purification steps to capture and concentrate plasmid DNA (Ferreira et al., 1999; Prazeres et al., 1998; Varley et al., 1998) (Table 6). Binding the plasmid DNA to anion exchange matrices at high ionic strengths (> 0.5 M NaCl) significantly reduces impurity levels. An approximate 40-fold increase in the plasmid purity leading to approximately 90% (w/w) plasmid purities (Ferreira et al., 1999) can be achieved in a single anion exchange chromatography step. Although performing clarification and concentration steps before anion exchange chromatography is suggested in order to increase the column capacity for plasmid DNA (Horn et al., 1998; Schleef, 1999), these steps can be bypassed. By going directly from cell lysis to anion exchange in fxed (Ferreira et al., 1999; Schleef and Schorr, 1998) or expanded bed mode (Ferreira et al., 2000a; Varley et al., 1998) plasmid preparations of identical purity can be achieved. Co-purificationof gDNA, high molecular weight RNA and endotoxins with plasmid DNA during anion exchange chromatography is inevitable (Prazeres et al., 1999) because these molecules have similar affinities for anion exchange matrices (Marquet et al., 1995). However, this problem can be addressed by performing a final SEC step on Sephacryl S1000. In this case, SEC has the additional advantage of enabling a process buffer exchange for an adequate formulation or storage buffer (Marquet et a]., 1995; Prazeres et al., 1999), ensuring rigorous process control over small contaminants introduced during the previous downstream processing operations (Marquet et al., 1995). Another purification strategy recently developed relies upon using hydrophobic interaction chromatography (HIC) instead of anion exchange chromatography as the first purification step (Diogo et al., 2000) (Table 6). In this strategy, plasmid preparations are fed to the column at high ionic strength after performing a clarification with a chaotropic salt (ammonium sulfate) in order to precipitate the proteins. Supercoiled plasmid DNA does not interact with the HIC support eluting in the flow-through, while the majority of the impurities (RNA, gDNA, denatured and relaxed plasmid DNA, and endotoxins) are retained and further eluted from the column with a decreasing ionic strength gradient (Diogoet a]., 2000). Although a final plasmid preparation that complies with the approval specifications is obtained with this strategy in a single chromatographic step (Diogo et al., 2000) further desalting and buffer exchange operations, such as diafiltration or SEC, are required to obtain a final “clinical”plasmid solution.

4

Concluding remarks The increasing demands on the production of pharmaceutical-grade supercoiled plasmid DNA require the development of plasmid DNA purification processes and their optimization in terms of selectivity, yield, efficiency and productivity in order to provide sufficient material for the ongoing gene therapy and DNA vaccine

I

229

230

I clinical trials. Chromatography plays Guilherme N. M. Ferreira et a/.

a role in the purification of plasmid DNA because it is scaleable, reproducible, it uses only chemicals that are generally regarded as safe and is sufficiently robust to withstand the abrasive cleaning conditions required for process approval and validation (Marquet et al., 1995).Anion exchange and hydrophobic interaction chromatography are the most suitable operations for the initial capture, purification and concentration of supercoiled plasmid DNA. Loading the column at high ionic strength enables the removal of impurities such as gDNA fragments, RNA, proteins and endotoxins. However, a final size-exclusion chromatography may be required in order to separate supercoiled plasmid DNA from other plasmid topoisomers (relaxed, linear and denatured plasmid) and/or to promote a plasmid’s medium exchange to the appropriate formulation or storage buffers. Chromatographic purification of plasmid DNA is hampered by the low perfoi-mance of commercially available sorbents. These sorbents were developed and designed for protein purification and, therefore, have small pore dimensions. Plasmid DNA adsorbs only at the particle surface leading to low capacities; 50-fold less than achieved with proteins. The design of new chromatography supports with higher binding capacities for plasmid DNA is therefore crucial in order to move these technologies forward. Further improvement is also required in the production of the biomass used for the plasmid preparation. No purification technology of today will be able to lead to high quality at economic conditions as long as the biomass used is not of sufficient quality. Within the last few months, certain plasmid manufacturers were focusing on the scale-up of the manufacturing process, rather than on aspects of quality assurance. Most interesting is the fact that over years now, no quantitative analysis was performed on the topologic plasmid variants within this type of pharmaceuticals except for the work presented by Schmidt et al., 1999b and in Chapter 2 of this book. It turned out that a complete set of multimeric variants (for examples see www.CGEsewice.corn) of plasmid DNA was not detected - or even worse, it was misinterpreted as nicked (or in other words “defect”) plasmid DNA, although this was simply a dimer of the product. Manufacturing of plasmids is still at the beginning of its development. While scaling up might not be difficult in the future, validation, quality assurance and quality control will have to be improved drastically and requires more time and investment. Depending on the intended application, plasmid manufacturing processes will be diversified within their final steps with respect to further processing and storage. Contracting out this part of product development - even for research or preclinical plasmid DNA - is the option to get access to recent technology and the complete set of quality assurance for innovative therapeutic or preventive DNA pharmaceuticals.

1 7 Plasmid Manufacturing - An Overview

References ALTON, E. W. F.W., M I D D L E T O N , p. G., BROOKS,C.A., CRAMER, S. M. (1992), Steric CAPLEN, N. J.. SMITH,S. N., STEEL, D. M. et al. mass action ion exchange: Displacement (1993), Non-invasive liposome-mediated gene profdes and induced salt gradients, AIChE J. 38, 1969-1975. delivery can correct the ion transport defect in cystic fibrosis mutant mice, Nature Genet. BROWN,P. R., KRSTULOVIC,A.M. (1979), Practical aspects of reversed-phase liquid 5, 135-142. ANDERSON, W. F. (1998), Human gene therapy, chromatography applied to biochemical and biomedical research, Anal. Biochem. 99, 1-21. Nature 392, 25-30. M.J., JONAS, R., SCHNEIDER, BYWATER,M., BYWATEBR., HELLMAN, L. (1983), ARTOLOZAGA, A. L. et al. (1998), One step partial purificaA novel chromatographic procedure for purtion of b-galactosidase from Ktuyueromyces ification of bacterial piasmids, Anal. Biochem. marxianus CDB 002 using Streamline DEAE, 132, 219-224. CAPLEN, N. J.. GAO,X., HAYERS, P., Bioseparation 7, 137.143. ATICINS, P. W. (1990),Physical Chemistry. Oxford ELASWARAPU,R., FISHER,G. et al. (1994), University Press, Oxford, UK. Gene therapy for cystic fibrosis in humans by liposome-mediated DNA transfer: U. K. regATKINSON,B., MAVITUNA,F. (19911, Biochemical Engineering and Biotechnology Handbook, ulatory process and production of resources, Gene fier. 1, 139-147. Macmillan, New Yorlc. BARNFIELD,A. B., HJORTH, R., HAMMARSTROM, CARLSON, A., SIGNS,M., LIERMANN,..I et al. A. (19943, Pilot scale recovery of recombinant (1995), Mechanical disruption of Escherichia coli for plasmid recovery, Biotechnol. Bioeng. annexin V from unclarified Escherichia coli homogenate using expanded bed adsorption, 48, 303-315. Biotechnol. Bioeng. 44, 922-929. S., OHI, S. (1992), CHAKRABARTI, A,, SITARIC, A procedure for large-scale plasmid isolation BATT, B.C., YABANNAVAR, VM., SINGH,V (1995), Expanded bed adsorption process for without using ultracentrifugation, Biotech.nol. protein recovery from whole mammalian cell Appl. Biochem. 16,211-215. CHANDRA, G., PATEL, P., KOST, T. A. et a\. culttire broth, Bioseparation 5 , 41-52. (1992), Large-scale purification of plasmid BELLOT,1. C., CoNnoRET, 1. s. (1993), Modelling of liquid chromatography equilibria, DNA by fast protein liquid chromatography using Hi-Load Q Sepharose column, Anal. Process Biochem. 28, 365-376. Biochem. 203, 169-172. BEST, A. N., ALLISON, D. P., NOVELLI,G. D. CHANG, (1981), Purification of supercoiled DNA of Y. K., CHASE,H.A. (1996a), Development of operating conditions for protein plasmid Col E l by RPC-5 chromatography, purification using expanded bed techniques: Anal. Biochem. 114, 235-243. The effect of the degree of expansion on adBIRNBOIM, H. C., DOLY, J. (1979), A rapid sorption performance, Biotechnol. Bioeng. 49, alcaline extraction procedure for screening recombinant plasmid DNA, Nucl. Acids Res. 512-526. CHANG, Y. K., CHASE, H. A. (1996b), Ion ex7, 1513-1523. BOLES, T. C., WHITE,J. H., COZZARELLI, change purification of G6PDH from unclariN.-R. (1990), Structure of plectonemically fied yeast cell homogenates using expanded supercoiled DNA, /. Mol. Bid. 213, 931-951. bed adsorption, Biotechnol Bioeng. 49, BONADIO, J., SMILEY, E., PATIL,P., GOLDSTEIN,204-216. S. (1099), Localized, direct plasmid gene CHANG,Y.K., MCCREATH, G.E., CHASE,H.A. delivery in viva prolonged therapy results (1995), Development of an expanded bed technique for an affinity purification of in reproducible tissue regeneration, Nature G6PDH from unclarified yeast cell homoMed. 5, 753-759. BRIGGS, I., PANFILI, P. R. (1991), Quantitation genates, Biotechnol Bioeng. 48, 355-366. CHASE,H.A. (1994), Purification of proteins of DNA and protein impurities in biopharmaceuticals, Anal. Chem. G3, 850-859. by adsorption chromatography in expanded beds, Trends Biotechnol. 12, 296-303.

I

231

232

I

Cui/herrne N. M. Ferreira et a/.

CICCOLINI, L. A., SHAMLOU, P.A., TITCHENER- the contained use of genetically modified microorganisms (90/219/EEC). N. J. et al. (1998), Time course of HOOKER, G. N. M. (ZOOO), Development of SDS-alkaline lysis of recombinant bacterial FERREIRA, plasmid DNA downstream processing stratecells for plasmid release, Biotechnol. Bioeng. gies based on Chromatography, Ph. D. Thesis. GO, 768-770. Centre for Biological and Chemical EngiCICCOLINI, L. A. S., SHAMLOU, P.A., neering, Instituto Superior Tecnico, Lisbon. N. J. et al. (1999), TITCHENER-HOOKER, J. M. S., PRAZERES, G. N. M., CABRAL, Rheological properties of chromosomal and FERREIRA, D. M. F. (1997),A comparison of gel filtration plasmid DNA during alkaline lysis reaction, Bioprocess Eng. 21, 231-237. chromatographic supports for plasmid COLE,K. D. (1991), Purification of plasmid and purification, Biotechnol. Techn. 11, 417-420. J. M. S., PRAZERES, FERREIRA, G.N.M., CABRAL, high molecular mass DNA using PEG-salt two-phase extraction, BioEchniques 11, 18-24. D. M. F. (1998), Purification of supercoiled S., FERRAZ, C., LIAUTARD, J. P. (1986), plasmid DNA using chromatographic COLOTE, Analysis and purification of plasmid DNA by processes, J . Mol. Recognition 11, 250-251. FERREIRA, G.N. M., CABRAL,J.M. S., PRAZERES, reversed-phase high-performance liquid chromatography, Anal. Biochem. 154, 15-20. D. M. F. (1999), Development of process flow sheets for the purification of plasmid vectors COLPAN,M., SCHORR,J., MORITZ,P. (1995), for gene therapy applications, Biotechnol, Process for producing endotoxin-free or Prog. 15, 725-731. endotoxin-poor nucleic acids and/or oligoG.N.M., CABRAL,J.M. S., PRAZERES, nucleotides for gene therapy. WO 95/21 2 77. FERREIRA, D. M. F. (2000a), Anion-exchange purification COLPAN,M., RIESNER, D. (1984), High-perof plasmid DNA using expanded bed formance liquid chromatography of high adsorption, Bioseparation 9, 1-6. molecular weight nucleic acids on the FERREIRA, G.N. M., CABRAL, J.M. S., PRAZERES, macroporous ion exchanger nudeogen, D. M. F. (2000b), Studies o n the batch J. Chromatogr. 296, 339-353. adsorption of plasmid DNA onto anionCOLPAN,M., SCHORR,J., MORITZ,P. (1995), Process for producing endotoxin-free or exchange chromatographic supports, Biotechnol. Prog. 16, 416-424. endotoxin-poor nucleic acids and/or oligonucleotides for gene therapy. WO 95/21 177. FLOYD,T.R., CICERO,S. E., FAZIO,S. D. et al. (1986), Mixed-mode hydrophobic ion CRYSTAL, R. G. (1995), The gene as the drug, Nature Med. 1, 15-17, exchange for separation of oligonucleotides and DNA fragments using HPLC, R., EISEN, H. H. DAVIS,B. D., DULBECCO, Anal. Biochem. 154, 570-577. (1980), Microbiology 3rd. Edn. Harper and Fox, J. L. (1999), Gene therapy safety issues Row, Philadelphia, PA. DIOGO,M. M., QUIROZ,J.A., MONTEIRO, G.A. come to fore, Nature Biotechnol. 17, 1153. et al. (2000), Purification of a cystic fibrosis GIESE,M. (1998), DNA-antiviral vaccines: new developments and approaches - a review, plasmid vector for gene therapy using Vir. Genes 17, 219-232. hydrophobic interaction chromatography, GLASEBV. (1997), Promise of gene therapy for Biotechnol. Bioeng. 68, 576-583. treating disease remains o n trial in the clinic, J.R. YARMUSH,M. L. Doux, J. D., MORGAN, Gen. Eng. News 14, 14, 35, & 38. (1995), Gene therapy, in: The Biomedical A. P., PRIOR,G. M., HELVESTON, N. M. Engineering Handbook (Bronzino, J. D., Ed.), GREEN, et al. (1997), Preparative purification of pp. 1518-15-35. CRC Press and IEEE Press, supercoiled plasmid DNA for therapeutic USA. applications, BioPhann (May), 52-61. DRAEGER, N. M. CHASE,H. A. (1991), Liquid GREGORIADIS, G. (1998), Genetic vaccines: fluidized bed adsorption of protein in strategies for optimization, Phann. Res. 15, the presence of cells, Bioseparation 2, 67-80. 661-670. DRAGER, R. R. REGNIER, F. E. (1985), Highperformance anion-exchange chromatogra- HINES,R. N., O’CONER,I<. C., VELLA,G. et al. (1992), Large-scale purification of plasmid phy of oligonucleotides, Anal. Biochem. 145, DNA by anion exchange high-performance 47-56. liquid Chromatography, Biochrom. 12, EUROPEANCOUNCIL DIRECTIVE 90/219/EEC (1999), Council Directive ofApril 23, 1990 on 430-434.

7 1 Plasmid Manufuctun'ng - An Ovewlew

HOLMS, W. H. (1986), The centraI metaboiic I n uivo induction of specific cytotoxic 'I pathways of Escherichia coli: Relationship lymphocytes in mice and rhesus macaques between flux and control at a branch point, immunized with DNA vector encoding HIV efficiency of conversion to biomass, and epitope fused with hepatitis B surface antiexcretion of acetate, Curr. Top. Cell. Regul. 28, gen, Virology 240, 304-315. 69 - 105. LEE, S.Y. (l996),High cell-density culture of HORN,N., BUDAHAZI,G., MARQUET,M. (1998), Escherichia coli, TIBTECH 14, 98-105 Purification of plasmid DNA during column LEHRMAN, S. (1999),Virus treatment queschromatography, U S Patent 5 707812. tioned after gene therapy death, Nature 401, HORN,N.A., MEEK,J.A., BUDAHAZI, G. et al. 517-518. (1995), Cancer gene therapy using plasmid LEVY,M. S., CICCOLINI, L. A. S., YIM, S. S. et al. DNA: purification of DNA for human clinical (1999a), The effects of material properties and fluid flow intensity on plasmid DNA trials, H u m . Gevle Ther. 6, 565.573. HUBER, C. G. (1998), Micropellicular stationary recovery during cell lysis, Chem. Eng. Sci. 54, phases for high-performance liquid 3171-3178. LEVY, M. S., COLLINS, I. J., YIM, S. S. et al. Chromatography of double-stranded DNA, (1999b), Effect of shear on plasmid DNA /. Chromutogr. 806, 3-30. HUBER, C. G., BERTI, G. N. (1996), Detection of solution, Bioprocess Eag. 20, 7-13. LIS, J. T., SCHLEIF, R. (1975), Size fractionation partial denaturation in AT-rich DNA fragof double-stranded DNA by precipitation ments by ion-pair reversed-phase chromatowith polyethylene glycol, Nucl. Acids Res. 2, graphy, Anal. Chem. 68, 2959-2965. 383-389. HUBEBC. G., OEFNER P. J., BONN, G . K. LJUNGLOF, A., BERGVALL, P., BHIKHABHAI, R. (1995),Rapid and accurate sizing of DNA et al. (1999), Direct visualisation of plasmid fragments by ion-pair chromatography on DNA in individual chromatography adalkylated nonporous poly(styrene-divinylbenzene) particles, Anal. Chem. 67, 578-585. sorbent particles by confocal scanning laser microscopy, /. Chromatogr. 844, 129-135. ITO, T., SMITH,C. L., CANTORC. R. (1992), Sequence-specific DNA purification by triplex LODISH, H., BALTIMORE,D., BERK, A. et al. affinity capture, Proc. Natl. Acad. Sci. U S A 89, (1995), Molecular Cell Biology. Scientific 495-498. American Books, New York. Luo, D., SALTZMAN, W.M. (ZOOO), Synthetic JANSON, J.-C., PETTERSSON, T. (1993), Largescale chromatography of proteins, in: PreDNA delivery systems, Nature Biotechnol. 18, parative and Production Scale Chromatography 33-37. (Ganetsos,G., Barker, P. E., Ed.), pp. 559-590. LYDDIATT,A., O'SULLIVAN, D. A. (1998), BioMarcel Deldcer, New York. chemical recovery and purification of gene therapy vectors, C u r . Opin. BiotechnoL 9, IGPP, U., LANGOWSKI, J. (1992), Preparation 177-185. of DNA topoisomers by RP-18 highMAIOR,M. E., VITVITSKI, L., MINK,M. A,, performance liquid chromatography, Anal. Biochem. 206, 293-299. SCHLEEF, M., WHALEN, R. G. et al. (1995), DNA based immunisation using chimeric KELLY, B. D., HATTON,T. A. (1991), The fermentation downstream processing interface, vectors for the induction of immune Biosepavation 1, 333-349. responses against the hepatitis C virus KUMAR,P. I<. R., MASCHKE, H.-E., FRIEHS,K. et nucleocapsid, /. Virol. 69, 5798-5805. T., FRITSCH, E. F., SAMBROOK, J. al. (1991), Strategies for improving plasmid MANIATIS, (1982), Molecular Cloning: Laboratory Manual, stability in genetically modified bacteria in bioreactors, Trends Biotechnol. 9, 279-284. Cold Spring Harbor Press, New York. LAHIJANI, R., HULLEY, G., SORIANO,G. et al. MARQUET, M., HORN,N.A., MEEK,J.A. (1995), Process development for the manufacture of (1996), High-yield production of pBR322derived plasmids intended for Human gene plasmid DNA vectors for use in gene therapy, therapy by employing a temperature control- BioPharm (September),26-37. lable point mutation, H u m . Gene Ther. 7, MARQUET, M., HORN,N.A., MEEK,J.A. 1971-1980. (1997a), Characterization of plasmid DNA vectors for use in Human gene therapy, LE BORGNE, S., MANCINI, M., LE GRAND, R., Part 1, BioPharm (May), 42-50. SCHLEEF, M., DORMONT, D. et al. (1998),

I

233

234

I

Cuilherme N. M. Ferreira et al MARQUET, M., HORN,N.A., MEEK,J.A. of circular DNA using aqueous two-phase (1997b), Characterization of plasmid DNA partition system, Nucl. Acids Res. 5, 583-590. vectors for use in Human gene therapy, ONISHI,Y., AZUMA,Y., KIZAICI, H. (1993), An assay method for DNA topoisomerase activity Part 2, BioPharm (June),40-45. MCCLUNG, J. K., GONZALES, R.A. (1989), based on separation of relaxed DNA from Purification of plasmid DNA by fast protein supercoiled DNA using high-performance liquid chromatography o n Superose 6 preliquid chromatography, Anal. Biochem. 210, 63 - 68. parative grade, Anal. Biochem. 177, 378-382. MERION,M., WARREN, W (1989), Purification PATIENT,R. K., HARDIES, S. C., LARSON, of supercoiled plasmids from crude cell J. E. et al. (1979), Influence of A-T content on the fractionation of DNA restriction fraglysates using high performance anion ments by RPC-5 column chromatography, exchange chromatography, BioTechniques 7, 60-67. J. Biol. Chem. 254, 5548-5554. V. R. (1994), Practical High-Pefomance PHARMACIA-BIOTECH (1997a), Computer Aided MEYER, Liquid Chromatography, John Wiley & Sons, Technical Service: Sephaciyl SIOOO. Pharmacia Chichester, UI<. Biotech, Uppsala, Sweden. M., MICHEL,M.-L., DAVIS,H. L., SCHLEEF, PHARMACIA-BIOTECH (1997b), Computer Aided MANCINI,M., TIOLLAIS, P., WHALEN, R. G. Technical Seruice: Superose 6 Preparative Grade. (1995), DNA-mediated immunization to the Pharmacia Biotech, Uppsala, Sweden. G. N. M., hepatitis B surface antigen in mice: Aspects PRAZERES, D. M. F., FERREIRA, of the humoral response mimic hepatitis B MONTEIRO, G.A. et al. (1999), Large-scale viral infection in humans, Proc. Natl. Acad. production of pharmaceutical-grade plasmid Sci. USA 92, 5307-5311. DNA for gene therapy: Problems and P. M., MALONE, R. W. (1996), MONTBRIAND, bottlenecks, Trends Biotechnol. 17, 169-174. D. M. F., SCHLUEP, Improved method for the removal of endo- PRAZERES, T., COONEY, C. L. (1998), Preparative purification of supercoiled toxin from DNA, /. Biotechnol. 44, 43-46. MONTEIRO, G.A., FERREIRA, G. N. M., CABRAL, plasmid DNA using anion-exchange J. M. S. et al. (1999), Analysis and use of en- chromatography, J. Chromatogr. 806, 31-45. A. 2. (1989), Separation on the basis donuclease activities in Eschevichia coli lysates PRENETA, during the primary isolation of plasmids for of size: Gel permeation chromatography, in: gene therapy, Biotechnol. Bioeng. GG, 189-194. Protein Puvijcation Methods: A Practical MOUNTAIN, A. (ZOOO), Gene therapy: the first Approach (Harris, E. L.V., Angal, S., Eds.), decade, Trends Biotecnhol. 18, 119-128. pp. 273-3015. IRL Press: Oxford, UK. MULLER,W. (l986), Fractionation of DNA RADNEDGE,L., RICHARDS, H. (1999), The restriction fragments with ion-exchangers for development of plasmid vectors, Methods Microbiol. 29, 51-96. high-performance liquid chromatography, Eur. 1.Biochem. 155, 203-212. RAYMOND, G. J.z BRYANT 111, P. I<., NELSON,A. MURPHY,J. C., WIBBENMEYER, J.A., Fox, G. E. et al. (1988), Large-scale isolation of covalently closed circular DNA using gel filtration et al. (1999), Purification of plasmid DNA using selective precipitation by compactation chromatography, Anal. Biochem. 173, 125-133. agents, Nature Biotechnol. 17, 822-823. NEIDLE,S. (1994), DNA Structure and Recogni- RIGGIN,A., DAVIS,G. C., COPMAN, T. L. (1996), tion. Oxford University Press, New York. Reassessing the control of residual DNA NEWTON,C. R., GREEN,A. R., HEATHCLIFFE, i n biopharmaceuticals, Biopharrn (October), 36-41. G. R. et al. (1983), Ion-exchange highH. L., GINSGERG, H. S., DAVIS,H. L. ROBINSON, performance liquid chromatography of oligodeoxyribonucleotides using formamide, et al. (1997), The Scient$c Future o f D N A Anal. Biochem. 129, 22-30. Immunization. American Academy of MicroV. G., CODORELLI, D. F. (1993), NICOLE-~TI, biology, Washington, DC. Optimized PEG method for rapid plasmid R. C. (1970), Alkali RUSH,M. G., WARNER, denaturation of covalently closed circular DNA purification: high yield from "midi-prep", BioEchniques 14,532-536. duplex deoxyribonucleic acid, /. Biol. Chem. R., HENTSCHEL, C. C., WILLIAMS, 245, 2704-2708. OHLSSON, J. G. (1978),A rapid method for the isolation

1 I Plasmid Manufacturing - An Overview

polymer matrices for tissue engineering, SAMBROOK, I., FRITSCH,E. F., MANJATIS,T. (1989),Molecular Cloning: A Laboratory Hand- Nature Biotechnol. 17, 551-554. book, CSH Press, Cold Spring Harbor, USA. SINDEN, R.R. (1994), DNA Structure and Function, Academic Press, Sail Diego, CA. SCHLEEF, M., SCHMIDT,T., FLASCHEL, E. (ZOOO), Plasmid DNA for pharmaceutical SOFER, G., HAGEL, L. (1997), Handbook ofProapplications, in: Development and Clinical cess Cromatogaphy: A Guide to Optimization, Scale- U p and Validation. Academic Press, Progress of DNA Vaccines, Developmental Biololgy Vol. 104 (Brown, F., Cichutek, I<., San Diego, CA. D., NORMAN,F., CUMMING, Robertson, J., Eds.), pp. 25-31. Karger, Basel. STEPHENSON, M. (1999), Issues of large-scale plas- R. H. (1993), Shear thickening of DNA in SCHLEEF, SDS lysates, Bioseparation 3, 285-289. mid DNA manufacturing, in: Recombinant STOWERS, D. J., KEIM, J. M. B. (1988), High Proteins, Monoclonal Antibodies and Tnerapeutic Genes (Mountain, A., Ney, U., Schom- resolution chromatography of nucleic acids burg, D., Eds.), pp. 443-469. Wiley-VCH, on the GEN-PAK FAX column,/. Chromatogr. Weinheim. 444,47-65. SCHLEEF, M. (ZOOO), Genetische Impfung mit STRYER, L. (1995), BiocheinistT, W H. Freeman and Company: New York. Plasmid-DNA, Bioforum 23, 766-769. J. (1998), Plasmid DNA SUMMERS,D. I<. (19961, f i e Biology ofPlasmids. SCHLEEF, M., SCHORR for clinical phase I and I1 studies: large scale Blackwell Science, Oxford, UI<. cGMP manufacturing and quality assurance, SUOMINEN,A. I., KARP, M. T., MANTsALA,P. I. (1984), Fractionation of DNA with Sephacryl in: Gene Therapy of Caizcer (Walden, P., Trefzer, U., Sterry, W, Farzaneh, F., Eds.), pp. S-1000, Biochem. Int. 8, 209-215. I., COLLINS, I. J., WARD,J.M. 481-486. Plenum Press, New York, London. THEODOSSIOU, SCHLUEP, T., COONEY, C. L. (19981, Purification et al. (1997),The processing of plasmid based gene from Esclzen’chia. coli. Primary recovery of plasmid DNA by triplex affinity interby filtration, Bioprocess Eng. lG, 175-183. actions, Nucl. Acids Res. 26, 4524-4528. THOMPSON, J.A. (1986),A review of high perSCHLUMBERGER, H.D., BRAUER,D. (1994a), formance liquid Chromatography in nudeic Das Regulationssystem der USA zur acids research 111. Isolation, purification, Anwendung der Gentechnilc in Forschung, and analysis of supercoiled plasmid DNA, Entwicklung und Produktion, Teil 1, Phann. Biochrom. 1, 68-80. Ind. 55, 28-35. TIGHE,H., CORR,M., ROMAN,M. et al. (1998), SCHLUMBERGER, H.D., BRAUER,D. (1994131, Gene vaccination: Plasmid DNA is more than Das Regulationssystem der USA zur just a blueprint, Immunol. Today 19, 89-97. Anwendung der Gentechnik in Forschung, Entwiclclung und Produltion, Teil 2, Pharm. USFDA (1985), Points to consider in the Ind. 55, 144-148. production and testing of new drugs and T., FRIEHS,I<., FLASCHEL, E. (1996), biologicals produced by recombinant DNA SCHMIDT, technology, US FDA Center for Biologics Rapid determination of plasmid copy Evaluation and Research: Roclville, MD. number, 1.Biotechnol. 43,219-229. USFDA (1996a), Addendum to the points to T., FRIEHS, IZ., SCHLEEF, M., SCHMIDT, consider in Human somatic cell and gene Voss, I<., FLASCHEL, E. (1999a), Quantitative therapy (draft), US FDA Center for Biologics analysis of plasmid forms by agarose and capillary gel electrophoresis, Analyt. Biochem. Evaluation and Research: Roclville, MD. 274, 235-240. USFDA (1996b), Points to consider on plasmid DNA vaccines for preventive infectious T., SCHLEEF, M., FRIEHS,I<., SCHMIDT, disease indications, US FDA Center FLASCHEL,E. (1999b), Hochzelldichtefermentation zur Gewinnung von Plasmid-DNA for Biologics Evaluation and Research: Rockville, MD. fur Gentherapie und genetische Impfung, BioForulTz 22, 174-177. VARLEY, D. L., HITCHKOCK, A. G., WEISS, A.M. E. et al. (l998), Production of plasmid SCHMIDT,T., FRIEHS,K., FLASCHEL,E., SCHLEEF, M. (1999~1,Method for the isolation DNA for human gene therapy using modified alkaline cell lysis and expanded bed of CCC plasmid DNA, WO 99/61633 anion exchange chromatography, BioseparaSHEA,L.D., SMILEY,E., BONADIO, J., tion 8, 209-217. D. J. (19991, DNA delivery from MOONEY,

I

235

236

I

Guilherrne A.! M. Ferreira et a/.

VOET, D., VOET, J.G. (1990), Biochemistry. WILS,P., ESCRIOU,V., WARNEY, A. et al. (1997), John Wiley & Sons, New York. Efficient purification of plasmid DNA VOLOGODSICII, A.V, LEVENE,S. D., KLENIN, for gene transfer using triple-helix affinity chromatography, Gene n e r . 4, 323-330. I<. V. et al. (1992), Conformational and YAMAICAWA, H., HIGASHINO,ICI., OHARA,0. thermodynamic properties of supercoiled DNA, /. Mol. Biol. 227, 1224-1243. (1996), Sequence-dependent DNA separation WEBER, M., MOLLER,I<., WELZECK, M., SCHORR, by anion-exchange high-performance J. (1996), Effects of lipopolysaccarides on liquid chromatography, Anal. Biochem. 240, liposome-mediated transfection and the 242-250. D. W., ARCURI,E. J. (1986), Factors consequence for DNA preparation for gene ZABRISKIE, influencing productivity of fermentations therapy, in: Artijicial Self-Assembling Systems employing recombinant microorganisms, for Gene Delivery (Felgner, P. L., Heller, M. J., Lehn, P., Behr, J.-P., Szoka, F. C., Eds.), Enzyme Microb. Echnol. 8, 706-717. pp. 56-62. American Chemical Society, ZHANG,Z., O'SULLIVAN, D. A,, LYDDIATT, A. (1999), Magnetically stabilised fluidised bed Washington, DC. WICKS,I. P., HOWELL, M. L., HANCOCK, T. et al. adsorption: pratical benefit of uncoupling bed expansion from fluid velocities in the (1995), Bacterial lipopolysaccharides copurifies with plasmid DNA: Implications for purification of a recombinant protein from Escherichia coli, J . Chem. Technol. Biotechnol. animal models and human gene therapy, 74, 270-274. Hum. Gene Ther. G, 317-323.

P/asmids for Therapy and kchation by M.Schleef Copyright 0 WILEY-VCH Verlag GmbH, 2001

I

12

Quality Control of pDNA Andreas Fels, Ruth Baier and Andreas Richter”

1

Introduction

There are several guidelines (CBER, 1998; WHO, 1998; ICH QSD, 1997; ICH QGB, 1999) and a draft guideline (CPMP, 1999) that address the production and control of pDNA intended for use in humans. The purpose of these guidelines is to assure the consistent safety and efficacy of pDNA. The guidelines should be invoked for the following areas: design of the vector, cell substrate used for the production of pDNA, the manufacturing process itself, characterization of the final product (pDNA). The following analysis is focused on the characterization and the quality control of plasmid DNA (final product; for further information on the other points see Chapters 9 and 11 of this book).

2

Characterization and quality control o f pDNA

According to several guidelines, certain lot-to-lot release tests have to be performed to ensure and control the quality of plasmid DNA (CBER, 1998; WHO, 1998; CPMP, 1999). Plasmid DNA is characterized with respect to identity, purity and stability. As summarized in Table 1 there are different sets of tests for plasmid DNA which is already the final product (the so-called drug product) or for plasmid DNA which still has to undergo further preparation steps (e.g., packaging into liposomes). Besides the analysis of the final product an in-process control (IPC) is performed to ensure the quality of the intermediate product(s) and to gain data for process monitoring. In comparison, the quality control of the final product decides “lot

237

238

I

Andreas Fels et a/. Table 1.

Tests applicable to different types of pDNA (according to CBER, 1998)

Test

Drug Substancea

Drug Productb

J

J

Content

J

J

Homogeneity

J

J

Host DNA

J

J

Proteins

J

J

Non-infectious viruses

J

Toxic materials involved in production

Jd

J d

Endotoxins

-

J

J

J

Sterility

J

J

Mycoplasma

J

C

-

Adventitous viruses

J

C

-

I Sterility Purity

Identity Adventitous agents

Potency

C

J

J

I

C

J

bulk product, not final formulation drug product, final formulation ‘ optional, depending on the type of cell substrate; not necessary for e. g. E. coli host cells optional, depending on the purification of the pDNA

release” or “no release”, the IPC data control and even influence the process. Regarding the overall quality control these influences should, nevertheless, be as little as possible because otherwise a revalidation of the procedure might be necessary. Parameters analyzed for IPC are, e. g., cell growth and vital cell number, plasmid quality and quantity and possible contaminants (for details, see Schleef, 1999 and Chapter 11 of this book). Another parameter of fermentation control is the plasmid copy number, i. e. the number of plasmid molecules per genome equivalent or per host cell. This copy number not only depends on the host strain and the kind of construct, but on several fermentation parameters such as growth rate, media composition, etc. For details on plasmid copy number determination, see Icang et al. (1991), Richter et al. (1997) and Schmidt et al. (1996; see also Chapter 2). In the following, quality control aspects are described in detail for the characterization and release testing of final product plasmid DNA. The IPCs are discussed elsewhere (Schleef, 1999 and Chapters 11 and 12 of this volume).

3 Validation o f test procedures

In principle, all applied test procedures have to be validated according to international guidelines [ICH guideline Q2A (1995) and Q2B (1997)) The objective of validation of an analytical procedure is to demonstrate its suitability for the intended purpose. Typical validation parameters are: precision (repeatabilityand intermediate precision), accuracy, specificity, detection limit and quantitation limit linearity. In the following, typical validation experiments are described in detail for the quantitation of host/genomic DNA in pDNA samples (see Sect. 5.2.3): Precision - repeatability and intermediate precision

Repeatability and intermediate precision are calculated as a criterion for the distribution of single measured values. Repeatability and intermediate precision analyses are performed as sixfold spiking experiments, e. g., well-defined amounts of pDNA are spiked with genomic DNA in order to determine the repeatability and the intermediate precision of the test procedure. The analysis of repeatability is characterized by as little variation as possible regarding the time of analysis, the used reference items, instruments and materials. Furthermore, the determinations are performed by a single person. The intermediate precision analysis is characterized by as much variation as possible regarding those parameters. Additionally, the analyses are performed by several persons within the same laboratory. The content of genomic DNA of each analysis is determined, and the amount of recovered spike is calculated. Afterwards, repeatability and intermediate precision are evaluated as standard deviation and coefficient of variation of the recovered spike amounts. Accuracy

The accuracy of a test procedure is determined by recoveiy experiments. To determine the accuracy of the test procedure “quantitation of host DNA in pDNA”, pDNA is spiked threefold with four different amounts of genomic DNA. After determination of the DNA amount of each sample the amount of the recovered spike is calculated. The accuracy is calculated as the deviation (in %) of the mean recovered spike from the nominal spike. Uncertainty o f measurement

Combination of accuracy and precision yields the uncertainty of measurement for a certain test procedure.

240

I

Andreas Fels e t a/.

Linearity, quantitation limits and detection limit

To determine the lower and upper quantitation limit and the linearity of the test procedure, pDNA is spiked in triplicate with different amounts of genomic DNA. The measured mean signal intensities of the spiked pDNA are determined and plotted vs. the spiked DNA amount. Afterwards, a linear regression function is calculated. The smallest amount of this correlation function is defined as the lower, the largest amount as the upper quantitation limit. Both limits define the quantitation range. Within this range, the test procedure is valid to quantitate genomic DNA within pDNA samples. The detection limit is defined for each single analysis (e. g., for each hybridization filter, see Sect. 5.2.3) as the smallest signal of the calibration row that is at least twice as high as the background signal. The determined validation parameters could be used directly to specify criteria for the rejection of a production batch of plasmid DNA, because it is possible to decide whether a certain deviation was caused by the test item itself or if this deviation is due to the uncertainty of measurement. Furthermore, a validation covers as many variations of the method as possible. Thus, the validation parameters are very helpful to analyze the robustness of a test procedure.

4

GLP

GLP (good laboratory practice) is a quality system concerned with the organizational process and the conditions under which non-clinical health and environmental safety studies are planned, performed, monitored, recorded, archived and reported. The first GLP regulation of international scope was that of the FDA which came into force in 1979 (Anonymous, 1979). The purposes of GLP are to ensure consistent standards of safety testing worldwide, to promote the quality and integrity of test data and to improve human and environmental safety. Since 1990, GLP is provided by law for testing of medicinal products in Germany (Anonymous, 1990; Christ et al., 1998).NewLab BioQuality AG as an example of a typical German contract service company obtained the GLP certificate from the responsible state ministry in Nordrhein-Westfalen. The certification was preceded by a one-year-period of GLP-like work in order to demonstrate the appropriate application of the quality control system. During this time, a complete reorganization of the laboratory was performed including the following steps : 1. determination of the current (pre-GLP)state, including: - personnel, - facility, - equipment, - methods;

72 Quality Control ofpDNA

2. determination/recruitment and training of the GLP key personnel:

test facility management, board of study directors, - quality assuraiice unit; 3 . GLP training of the whole staft 4. construction of a standard operation procedures(SOP) system covering: test facility management, - quality assurance system, - training, laboratories, - equipment, - test procedures, - validation, handling of test items, - study plans, - final reports and other topics; 5. performance of all the lab work according to the SOPS. -

-

Reorganization phases of one year or more are common. After this phase, certification will be achieved by passing the inspection (NewLab example: one day of preinspection, three days of main inspection). The GLP quality system strongly regulates every-day life in the laboratory, but on the long run will lead to a substantial improvement of the analytical work. GLP is a very reasonable basis of analytical work with regard to the acceptance and release of gene therapy products.

5 Detailed description of the characterization of pDNA (final product) 5.1 Sterility

The absence of any contaminating microorganisms or spores is tested by USP23 (1995) or conforming sterility tests according to Anonymous (1998), or the load with microorganisms is determined by bioburden assay (Microbial Limit Tests). For plasmid DNA which is the final container product (the so-called drug product) sterility testing must be performed. For plasmid DNA which still has to undergo further preparation steps (e.g., encapsulation, formulation, drug substance) a bioburden assay is sufficient. Bioburden assay (microbial limit tests)

Microbial limit tests are procedures for the estimation of the number of viable aerobic microorganisms present and for the proof of absence from designated microbial species in pharmaceutical articles of all kinds - from raw material to the

I

241

242

I final product. For the total aerobic microbial count Andreas Fels et a/.

two different examination methods can be used: membrane filtration or plate method. For diagnostic biological products not intended for injection

-

-

soybean casein digest medium test incubated for not less than five days at 30-35 “C for the detection of bacteria and Sabouraud dextrose agar medium test incubated for a minimum of five days at 20-25 “C for the detection of fungi and yeast is required.

The volume of material for the bulk test should be no < 2.0 mL, and the sample for the final container test should not be less than three final containers, if the total number filled is 100 or less, and, if higher, one additional container for each additional 50 containers or fraction thereof, but the samples need be no more than 10 containers [Code of Federal Regulations 21CFR610.12 (1999), 5 81. The evaluation occurs by counting the number of colonies on the incubated Petri dishes and express the average for the plates as the number of microorganisms per g or mL of specimen. Sterility test

pDNA intended for injection (parenteralia) are required to be assayed for the presence of bacterial and fungal contaminants. Therefore, testing protocols were initiated for the detection of bacterial and fungal organisms in compliance with standard sterility testing procedures as described in “Titel 21” of the “Code of Federal Regulations” (1999), Section 610.12. These testing procedures require the use of media with the capability to detect aerobic and anaerobic bacteria and also fungal organisms. The test procedures include -

direct inoculation (transfer) into test vessels and membrane filtration techniques.

The membrane filtration procedure is particularly useful for liquids (pDNA in low concentrations) and for soluble powders (freeze-dried products). In cases, where the liquid is highly viscous (pDNA in high concentration) and not directly filterable the direct transfer should be used. The membrane or the test material directly will aseptically transfer to vessels of two culture media (fluid thioglycollate medium and soybean-casein digest). The test mixture is incubated for a minimum of 14 days with fluid thioglycollate medium at 30-35 “C for the detection of contamination with bacteria (e. g., Staphylococcus aureus, Bacillus subtilis, Pseudomonas aeruginosa) and with soybean-casein digest medium at 20-25 “C for the detection of contamination with fungi (e. g., Candida albicans, Aspergdlus niger.) The above mentioned microorganisms should be used to test the growth promoting qualities of each lot of test medium. The necessary amounts of the sample used for each test medium or each incubation temperature of a test medium are summarized in Table 2 (according to Seyfarth, 1998). Subsequently the interpretation of test results occurs by visual inspection of all inoculated vessels. A sample meets sterility testing requirements when no growth

72 Quality Control ofpDNA

Sterility testing of pDNA - requirements

Table 2.

Number of Final Containers Number of containers per lot 5

100

number of containers per medium”

I

lo%, minimum 4

101-500

10

> 500

2 % , maximum 20

Amount of Material

Container content (mL)

minimum quantity to be used for each medium

5 1 mL

total amount of the container

> 1 mL and < 40 mL

one half of the container

2 40 a

mL

I

20 mL

If the content of one container is sufficient for the inoculation of both media, the column describes the required number of containers for both media.

was observed, i.e. the inoculum did not render the media turbid, in any culture vessels inoculated with the test article during a valid assay. 5.2

Purity

According to the Code of Federal Regulations Title 21CFRG10.13 (1999) the product must be free of extraneous material except for that which is unavoidable in the manufacturing process. The CBER guideline (1998) defines in detail the necessary parameters that at least have to be analyzed for pDNA (see Table 1).

5.2.1

Content

According to CBER (1998) the DNA content of the pDNA has to be determined. In general, this test is performed by spectrophotometric analysis and calculation of the DNA concentration according to the absorbance at 260 nm (Cantor and Schimmel, 1980a; Sauer et al., 1998). Furthermore, the A,,o/Azso ratio is determined in order to identify protein impurities (highly purified pDNA exhibits a ratio of 1.80-1.95).Additional information is gained by spectrophotometric scans between 220 nm and 320 nm, because the extinction parameters as maximum extinction and shape of the extinction curve are modified by salt and/or inorganic substances (Wilfinger et al., 1997). Some problems may arise by significant impurities with either low molecular weight or size-inhomogenous DNA, because the spectrophotometric analysis is not able to differentiate between degraded or non-degraded

I

243

244

I

Andreas Fels et a/.

DNA. Nevertheless, a distinctive quantitation of the pDNA is possible using a quantitative agarose gel electrophoresis with commercially available quantitation standards. These standards can be adapted to the size of the linearized pDNA and serve as a calibration row to generate a calibration function. According to this function, it is possible to quantitate the pDNA without influence of degraded DNA. Furthermore, the content of the degraded DNA can be evaluated by comparison of the spectrophotometric and the gel results.

5.2.2

Homogeneity

Tests for structural homogeneity have to be performed to characterize the plasmid DNA. Plasmid DNA exists in several isoforms, e. g., supercoiled monomer (cccform), open circle monomeric and dimeric, supercoiled dimeric, linear (see Chapter 2 of this book). The distribution of these isoforms strongly depends on the E. coli host strain (Schleef, 1999). In order to determine the structural homogeneity ofthe plasmid DNA the different isoforms have to be quantified. There are different methods to fulfill this requirement, e. g. CGE (see Chapter 2) or agarose gel electrophoresis (Schmidt et al., 1999, Schleef et al., 2000). CGE is a highly sophisticated procedure requiring a great deal of technology, whereas agarose gel electrophoresis is less cost intensive and is still the standard procedure for molecular biology laboratories. Using CGE, it is possible to differentiate between the mentioned plasmid isoforms, whereas for agarose gel electophoresis further investigations have to be performed because supercoiled dimeric and open circle monomer might exhibit an identical migration behavior, depending on the gel conditions. Here, a GLP-validated agarose gel electrophoresis based procedure is presented because this procedure is proposed by the relevant guideline (CBER, 1998). A typical agarose gel analysis is presented in Figure 1.The indicated isoforms are clearly separated. In order to optimize these conditions and to designate the different isoforms, experiments have to be performed in advance, e. g., digestion with enzymes that do not cut supercoiled structures. Furthermore, one has to analyze the migration behavior as a function of the ethidium bromide concentration in order to differentiate the isoforms (Cantor and Schimmel, 198Ob).In order to quantitate the different isoforms, the plasmid DNA is linearized with a single-cutting restriction enzyme. The linearized plasmid is used to generate a calibration row by loading different amounts in double determinations onto the gel. The mean signal intensities are used to generate a calibration function. All linear type isoforms contents (linear, open circle and dimeric open circle) of the test items are calculated according to this function. The supercoiled types cannot be quantitated in the way described, because they exhibit a different ethidium bromide intercalation behavior. In order to quantitate these isoforms, the shares of the linear type isoforms are summarized. The summarized shares of the supercoiled type isoforms are calculated as 100% minus summarized share of linear type isoforms. Afterwards, the supercoiled type isoform amounts are calculated according to their share of signal intensity. In order to prove the sensitivity of the test procedure, a sample consisting of the test item spiked with linearized plasmid DNA is always included. The pres-

72 Quality Control ofpDNA

1

2

3 4

5 6

7

8 9 10 S1 1 1 1 2 1 3 14 52

4f-

f=

c

Iin

c ccc

Fig. 1. Homogeneity analysis o f plasmid DNA. Gel conditions: 0.5 % Agarose 1 xTAE, Et hidiu rnb rom ide Lanes: S1: DNA-standard (10,8, 6, 5, 4, 3.5, 3, 2.5,2,1.5,1.2,l.Okbp) 52: DNA-standard (supercoiled ladder, 10,9, 8, 7, 6, 5, 4,3, 2 kbp) 1.10: linearized plasmid DNA (calibration row), 100,50, 20, 10, 5 n g (analysis in duplicates) 11,12:300 n g plasmid DNA (analysis in duplicates) 13: sensitivity control, 300 n g plasmid DNA 10 n g linearized plasmid DNA 14: negative control ccc: covalently closed circle isoform, lin.: linear isoform, oc.: open circle isoform. ccc-di: dimeric ccc isoform, oc-di: dimeric oc isoform, gen: genomic DNA

+

ented example exhibits an amount of ca. 85 % ccc isoform, 11% open circle isoform. All other isoforms including the linear one are lower than 4%. It is not possible to quantitate the residual genomic DNA using agarose gel electrophoresis, because the genomic DNA does not only consist of high molecular fragments but of inhomogenous low molecular weight molecules as well. It is not possible to quantitate such a "smear" by gel electrophoresis. Instead, a hybridization assay is performed (see Sect. 5.2.3). The parameters of a GLP validation of the presented plasmid isoform quantitation are presented in Table 3 . First, the results of the precision analyses (repeatability and intermediate precision) have been summarized. For both parameters the coefficient of variation (CV) is denoted. CV is a criterion of the distribution of the measured values. In order to determine the precision of the test procedure, the test item was spiked in six replicates with a defined amount of linear isoform. The CV is determined as the variation of the single recovered spikes from the mean value. In order to determine the accuracy of the test procedure, the test item was spiked with different amounts of linear isoform plasmid DNA and the recovered spike was quantified. Afterwards, the accuracy was determined as the deviation (in %)

I

245

246

I

Andreas Fels et a/. Table 3.

Validation parameters

-

homogeneity analysis of plasmid DNA

Validation Parameter

I Precision

Repeatability

Results

Comment

I CV 5 10% I valid for linear and supercoiled type I isoforms

Intermediate precision

CV

5

10 %

80%

valid for 0.5 ng to SO ng linear isoform

Accuracy

2

Linearity

0.5-50 ng

amount of linear isoforni

Specificity

proven

proven according to band intensity vs. background analyses

of the mean recovered spike from the nominal spike. For the test procedure “homogeneity of pDNA”, an accuracy of at least 80% was determined for the range of 0.5-50 ng linear isoform within 300 ng of pDNA. Furthermore, the test procedure yields signal intensities for the mentioned range 0.5-50 ng linear isoform that follows a linear regression with a regression coefficient > 0.990. Thus, 0.5 ng is the lower quantitation limit and 50 ng is the upper quantitation limit for the quantitation of linear isoform within 300 ng plasmid DNA. The specificity of the test procedure was proven with different negative controls. These controls did not exhibit a signal larger than the background signal, whereas all mentioned plasmid isoforms did. Note that it was only possible to evaluate the validation parameters only for the linear isoforms, because these forms can be quantified directly. Such a quantitation i s essential for spiking experiments. Nevertheless, the results gained for the linear form can be transferred to supercoiled forms because the latter are calculated according to the amount of linear forms.

5.2.3 Host DNA According to CBER (1998) pDNA contamination with host DNA has to be tested. The CBER proposes an electrophoretic analysis including tests with bacterial host specific probes. In practice, there are different other methods to quantify contaminating host DNA, e. g., a host DNA specific Southern hybridization and a kinetic PCR. In comparison with quantitative agarose gel electrophoresis, a host genomic DNA specific Southern hybridization i s much more sensitive and quantitative than a electrophoretic procedure, because the latter only allows quantitation of high molecular non-degraded DNA, whereas degraded genomic DNA (visible as a smear on the agarose gel) cannot be quantified. However, a host DNA specific hybridization is able to quantify even size-inhomogenous DNA. Recently, a quantitation method was published that i s based on real-time PCR (Smith et al., 1999).The procedure utilizes the amplification of the highly conserved 2 3 s rDNA. Thus, in comparison with the hybridization, a single gene i s used to quantitate the genomic DNA in-

I stead of using a probe prepared from the complete genome. Up to now hybridiza72 Quality Cofltrol of pDNA

tion assays have been more convenient because the complete genome is used for quantitation instead of a single representative gene. In the following, a hybridization method is described that is already validated according to ICH and that has already been used for different types of plasmid DNA. In general, samples are analyzed in duplicates with and without a genomic DNA spike. The spike is necessary in order to quantify a possible loss of DNA during preparation of the samples. Spiked and non-spiked test items are dotted onto a membrane and afterwards hybridized with a genomic DNA specific probe. This probe is prepared by random priming of the host genomic DNA. Furthermore, a calibration row and several controls are also dotted onto the membrane. In order to quantitate the contaminating genomic E. coli DNA in plasmid DNA samples, a highly genomic DNA specific hybridization has to be to applied, because small amounts of genomic DNA have to be quantified within a large excess of plasmid DNA (up to 100-fold).Therefore, hybridization stringency and thus hybridization conditions must be optimized for each test item. Furthermore, one has to use genomic DNA of the host E. coli strain without plasmid in order to prepare reference genomic DNA. This reference DNA is used as template for the synthesis of a radioactively labeled probe using a random priming procedure and as spike DNA. Additional controls must be included in the analysis. Firstly, a negative control containing the applied preparation buffers is necessary. The signal intensity of this control must be lower than a certain threshold, because otherwise the signal is caused by exogeneous contamination and cannot be used. Furthermore, a control for the hybridization stringency has to be found. We used HPLC-treated pDNA test samples as hybridization controls (control items). Because the HPLC treatment significantly removes genomic DNA, the signal intensities of the control items and thus the genomic DNA content must be lower than that of the test items. Thus, hybridization conditions can be optimized until the determined genomic DNA

Table 4.

Validation parameters

-

quantitation o f genomic DNA in p D N A

Validation Parameter

I Precision

Repeatability Inteimediate precision

Results

Icvs

Comment 5 % I -

CV 5 10%

valid for contents of 0.5-8.0%a

Accuracy

2

Linearity

0.5-8.0%

-

Specificity

proven

signal negative control < 2 X background signal, DNA content of test items > 25 % larger than DNA content control item

a

75%

-

Minimum value for the denoted range, most contents exhibit an accuracy > 85 %.

I

247

248

I content of the control items remains constant, because a decreasing content with Andreas Fels et af.

increasing stringency indicates decreasing hybridization with plasmid DNA, thus false-positive hybridization signals are reduced. In order to quantitate the genomic DNA, the mean signal of each step of the calibration row is calculated and a calibration function (linear regression) is evaluated. The parameters of a GLP validation of the test procedure “quantitation of genomic DNA in plasmid D N A are presented in Table 4.

5.2.4

Host cell protein impurities

The guidelines demand the testing of host cell derived proteins (HCP) with appropriate methods. The presence of proteins can be monitored by colorimetric protein assays (like Bradford) or by SDS gel electrophoresis. There is no exact value given by the guidelines for the acceptance limits regarding HCP content. For bulk plasmid DNA intended for use in gene therapy, limits of 5 % are published (Schleef, 1999).This value is very high compared to the limits usually given to recombinant protein drugs. Those drugs have to be proven to be free from protein impurities in the range of 10 ppm to 100 ppm. Therefore, special immunological tests have to be developed (Wolter und Richter, 1999; Hoffmann, 2000). The limit of host cell proteins which will be fmed in the specification of a biopharmaceutical drug might vary from product to product and from process to process. Therefore, it is not excluded, that also for pDNA products in some cases lower limits will be demanded. In any case, the production process should be optimized for an almost complete elimination of host cell proteins.

5.2.5

Endotoxins

The determination of bacterial endotoxins is performed according to USP23 (1995). The test is performed using Limulus amoebocyte lysate (LAL) obtained from aqueous extracts of the circulating amoebocytes of Limulus polyphenus and prepared and characterized for use as an LAL reagent for gel clot formation. The determination of the reaction endpoint is made with a dilution from the test item in direct comparison with parallel dilution of a reference endotoxin, and quantities of endotoxin are expressed in defined endotoxin units. For each test item, no less than two replicates have two be performed for each level of the dilution series. This number of replicates is also necessary for the standard endotoxin dilution. The determination of endotoxins must be validated for the specific test item in order to determine the maximum valid dilution (MVD) and to track down possible interfering factors.

72 Quality Control ofpDNA I 2 4 9

5.3 Identity

5.3.1

Restriction analysis

According to CBER (1998) and the Code of Federal Regulations Title 21CFRG10.14 (1999) one container of each lot of the pDNA should be tested for identity. The used test procedure should be able to distinguish the pDNA from any other DNA being processed in the same laboratory. The method of choice to fulfill these requirements is the restriction enzyme mapping of the plasmid DNA. An example of such a restriction characterization is presented in Figure 2; the restriction fragments are analyzed by agarose gel electrophoresis. In our experience, at least three different restriction enzymes should be used for the characterization and analysis of each plasmid DNA. One of these enzymes

S 1 2 N S 3 4 N S

Fig. 2. Restriction analysis o f plasmid DNA. (A) multiple-cutting enzymes, (B) single-cutting enzyme. Gel conditions: (A) 1.5 % Agarose, IxTAE, Ethidiumbromide (B) 0.8% Agarose, IxTAE, Ethidiumbromide Lanes: Gel A: S: DNAstandard (lo,& 6, 5,4,3.5, 3,2.5,2,1.5, 1.2,1, 0.9,0.8,0.7,0.6,0.5, 0.4,0.3, 0.2,0.1 kbp) 1-2and 3-4:two different multiple-cutting restriction enzymes (analysis i n duplicates) N: negative control Gel B: 5: DNA-standard ( 10,8, 6,5, 4,3.5, 3, 2.5,2, 1.5, 1.2,1, 0.9,0.8,0.7,0.6,0.5,0.4,0.3 kbp) 1-2:Single-cutting restriction enzyme (analysis in duplicate) N: negative control

250

I should linearize the plasmid Andreas Fels et a/

DNA to be able to analyze the complete number of bp, the other enzymes should generate each up to 10 fragments, regarding especially the coding and the flanking regions as well as functional regions like promoter elements. The selection of the used enzymes can be done with different commercially available computer programs. Evaluation of the restriction analysis focuses on two points: firstly the restriction fragment pattern (e.g., the number of bands and the existence of additional bands) and secondly the comparison of the determined fragment lengths with the theoretical ones. The intensity of the band is a third parameter that is important for quantitative analyses, e.g. determination of copy numbers of a certain coding region of the plasmid. Deviations ofthe determined restriction pattern from the calculated restriction pattern are either caused by additional bands or by missing bands. If there are additional bands, onehas toensure thatthesebands arenot causedbyincomplete digestion.Thus, the incubation parameters must be optimized very seriously, because incomplete digestion might be due to a tertiary structure of certain DNA motifs or to mutated restriction sites. If there are still additional or even missing bands after optimization it might even be necessary to sequence the pDNA (see Sect. 5.3.2). The comparison of determined and theoretical restriction fragment length i s the main criterion for the identity analysis. In order to decide whether an observed difference falls within the uncertainty of the test procedure itself or is due to a change of the pDNA-sequence (e.g., deletion or insertion), the uncertainty of measurement has to be known. The determination of this parameter is part of a test procedure validation that i s performed for the restriction characterization according to GLP (see Sects. 3 and 4). Using modern sophisticated gel documentation devices and software such a validation yields results as summarized in Table 5. The parameter uncertainty of measurement can be used for the decision whether a restriction pattern resembles the theoretical one or not. Deviation of more than the uncertainty of measurement can be an indication for a change of the pDNA sequence or structure. In this case, it i s essential to determine at least the DNA sequence of the fragment(s) that lead to the deviation. Table 5.

Validation parameters - restriction characterization o f p D N A

Validation Parameter

I Precision

Repeatability Intermediate precision

Results

Comment

I CV 5 6 % I valid for a range of 300 bp to 7 kbp CV 5 6%

Accuracy

2 90%

valid for a range of 300 bp to 7 ltbp

Uncertainty of measurement

5

10%

valid for a range of 300 bp to 7 lcbp

Linearity

0,25-4 pg

specified amounts of plasmid-DNA yield an identical restriction pattern

Specifcity

proven

different plasmids yield different restriction pattern with the same enzyme

I

12 Quality Control of pDNA

5.3.2

Determination of the DNA sequence

The complete nucleotide sequence of both DNA strands has to be determined by DNA sequencing. All steps are performed following standard operation procedures and the data are documented in a sequencing report. The amount of sequencing work performed depends on the specific vector construct. The Center of Biologics Evaluation and Research (CBER, 1998) recommends complete sequencing of the plasmid directly isolated from the MCB, and sequencing of the finished product at least once within each manufactured lot. Further productions require this at least for the relevant enconding sequences of the construct including flanking regions. With any changes in sequence, the resulting plasmid is considered to be a new product and, therefore, must be fully re-characterized. If point mutations have changed the relevant sequence portions without negative or with even positive effects, this may be evaluated as a modification. In such cases, complete documentation demonstrating evidence for improvement is required. For these reasons, the FDA evaluates single events such as these on a case-by-casebasis. (Schleef, 1999).

6 Conclusion

There are different quality control tests necessary for bulk/final product pDNA. In general, the analyses must be performed with validated test procedures. Actually, although a guideline already exists prescribing the necessary quality control tests (CBER, 1998), this guideline does not set distinct specifications. Thus, lot release criteria are still part of discussions. As an example, such criteria. that are relevant for clinical phase I and I1 trials have been summarized in Table G.

Quality control tests on bulk plasmid DNA (according to Schleef, 1999; see also Chapter 11)

Table 6.

Test

Specification

Identity of the plasmid

restriction fragment size conforms, sequencing identity

~~~~

DNA homogeneity

> 90% ccc form

AZ60IA280

1.80-1.95

Scan 220-320 nm

conforms (peak at 260 nm)

Bacterial host DNA

< 5%

Protein

< 5%

Endotoxins

<0.1 EU pg-'

Absence of micororganisms

< 1 micoorganism

I

251

252

I

Andreas Fels e t a/.

In the future, the use of pDNA as medical products will increase and new applications will be developed. Thus, quality control of pDNA is very important. Guidelines that set distinct specifications could be expected within the next years. Furthermore, there will be more advanced analysis techniques that reduce the period of time that is still necessary for the above mentioned tests, e. g. CGE, micro arrays (DNA chips, proteins chips), mass spectrometry, real-time detection, e. g. real-time PCR. These forthcoming techniques will be more sensitive than the actual ones and will certainly influence the release criteria and thus the quality control of pDNA.

References 21CFR610.13 (1999), Code of Federal Regulations, Title 21 Food and Drugs, Department of Health and Human Services, Part 610 General Biological Products Standards Sec. 610.12 Sterility, 610.13 Purity. 21CFR610.14 (1999), Code of Federal Regulations, Title 21 Food and Drugs, Department of Health and Human Services, Part 610 General Biological Products Standards Sec. 610.14 Identity. Anonymous (1979),Guidancefor Industry: Good Laboratory Practice Regulations, Management Briefings, Post Conference Report. U. S. Department of Health and Human Services, Food and Drug Administration. Anonymous (1990), Allgemeine Verwaltungsvorschrift zum Verfahren der behordlichen Ubenvachung der Einhaltung der Grundsatze der Guten Laborpraxis (ChemVwVGLP). Bundesanzeiger Jahrgang 42 Nummer 204a, 31. Oktober 1990. Anonymous (1998), Europliisches Arzneibuch, Nachtrag 1998. Deutscher Apotheker Verlag, Stuttgart; Govi-Verlag-Pharmazeutischer Verlag, Eschborn. CBER (Centre for Biologics Evaluation and Reserach) (1998), Guidance for Human Somatic Cell Therapy and Gene Therapy. U. S. Department of Health and Human Services, Food and Drug Administration. Cantor, C. R., Schimmel, P. R. (1980a), Biophysical Chemistry, Part 11: Techniquesfor the Study of Biological Structure and Function. W. H. Freeman and Company, New York. Cantor, C. R., Schimmel, P. R. (1980b), Biophysical Chemistry, Part III: The Behaviour of

Biological Macromolecules. W. H. Freeman and Company, New York. Christ, G. A,, Harston, S. J., Hembeck, H.-W., Opfer, K:A. (1998), GLP, Handbuchfur Praktiker. GIT Verlag, Darmstadt. CPMP (Committee for Proprietary Medicinal Products) (1999), Note for Guidance on the Quality, Preclinical and Clinical Aspects of Gene Transfer Medicinal Products. EMEA, CPMP/ BWP/3088/99 draft. Hoffman, I<. (ZOOO), Strategies for host cell protein analysis, BioPham (May), 38-45 ICH Guideline Q 2 A (1995),Text on Validation of Analytical Procedures, CPMP/lCH/381/95, Federal Register Vol. 60, pp. 11260. ICH Guideline Q 2 B (1997), Validation of Analytical Procedures: Methodology, CPMP/ ICH/281/95; Federal Register Vol. 62, No. 96, pp. 27463-27467. ICH Guideline QSD (1997), Quality of Biotechnological Products: Derivation and Characterisation of Cell Substrates Used for Production of Biotechnological/Biological Products. CPMP/ICH/294/295; Federal Register Vol. 62, No. 85, pp. 24311-24317. ICH Guideline Q6B (1999), Specifications: Test Procedures and Acceptance Criteria for Biotechnological/BiologicalProducts. CPMP/ ICH/365/96; Federal Register Vol. 64, p. 44928. Kang, J., Immelmann, A., Welters, S., Henco, I< (1991), . Quality control in the fermentation of recombinant cells: Rapid and quantitative determination of plasmid copy numbers in E. coli, Biotech Forum Europe 10, 590-593. Richter, A., Plobner, L., Schumacher, J. (1997), Quantitatives PCR-Verfahren zur Bestimmung der Plasmidkopienzahl in re-

12 Quality Control ofpDNA

kombinanten Expressionssystemen, Bioforum 20, 545-547. Sauer, P.,Muller, M., Kang, J. (1998), Quantification of DNA, Qiageiz News 2, 23-26. Schleef, M. (1999), Issues of large-scale plasmid DNA manufacturing, in: Biotechnology 2nd Edn. Vol. 5a: Recombinant Proteins, Monoclonal Antibodies and Therapeutic Genes (Rehm, H:)., Reed, G., Piihler, A,, Stadler, P. Eds.), pp. 443-469. Wiley-VCH, Weinheim. Schleef, M., Schmidt, T., Flaschel., E. (2000), Plasmid DNA for pharmaceutical applications, in: Development and Clinical Progress of DNA Vaccines (Brown, F., Cichutelt, K., Robertson, J., Eds.). Dev. Biol. Vol. 104, pp. 25-31. Karger, Basel. Schmidt, T., Friehs, I<., Flaschel, E. (1996), Rapid determination of plasmid copy number, /. Biotechnol, 49, 219-229. Seyfarth, H. (1998), Priifung und Sterilitat nach den Vorschriften des Europaischen Arzneibuches/Nachtrag, Pharm. Ind. GO (12), 1073-1083.

Schmidt, T., Friehs, I<., Schleef, M., Voss, C., Flaschel, E. (1999), Quantitative analysis of plasmid forms by agarose and capillary gel electrophoresis, Anal. Biochem. 274, 235 -240. Smith 111, G. J., Helf, M., Nesbet, C., Betita, H. A,, Meek, J., Ferre, F. (1999), Fast and accurate method for quantitating E. coli host-cell DNA contamination in plasmid DNA preparations, BioTechniques 2 6 518-526. USP23 (1995),The United States Pharmacopoeia 23. United States Pharmacopoica Convention, Rockville, MD. WHO (World Health Organisation) (1998), Guidelines for assuring the quality of DNA vaccines. WHO Technical Report Series No. 878, pp. 77-90. Wilfinger, W. W, Macltey, I<., Chomczynslti, P. (1997),Effect of pH and ionic strength on the spectrophotometric assesment of nucleic acid purification, Biotechniques 22, 474-481. Wolter, T., Richter, A. (1999), Proteinverunreinigungen in Biopharmazeutilta, Bioforum 22, 706-707.

I

253

P/asmids for Therapy and kchation by M.Schleef Copyright 0 WILEY-VCH Verlag GmbH, 2001

I

13 From Research Data to Clinical Trials James S. Robertson” and Klaus Cichutek

1

Introduction

It is important for researchers to be aware of the regulatory aspects of taking a laboratory observation through to clinical trials and there are several guidance documents which provide advice on the requirements for the quality, safety, and efficacy of a DNA vaccine. These include a WHO Guideline (1998), a Points to Consider document from the U.S. FDA (1996), and a draft guideline from the CPMP of the European Union (1999). The latter document also covers viral gene therapy products under the umbrella of gene transfer products. The documents are intended to assist those submitting applications for marketing approval. Within Europe, no document is available specifically to provide advice for clinical trial submission. However, from the experience gained by regulators and industry in the production of other vaccines, and from available guidance in the above documents, requirements for clinical trials can be determined. This chapter will outline the information required for licensing, with special emphasis on the area of clinical trials, and an example of applying for clinical trial authorization within an EU member state (Germany) will be set out. The entire process of licensing a novel vaccine can be broken down to the following steps: a laboratory demonstration of the “proof of principle” of the vaccine, the development of a manufacturing process, preclinical safety studies, clinical trials, 5. application for marketing approval (licensing), 6. post-marketing surveillance.

1. 2. 3. 4.

Several DNA vaccines are, or have been, in the initial stages of clinical trial, but it is likely to be some time before an application for marketing approval is made. This is primarily due to the length of time it takes for clinical trials of a vaccine, in general, but also may reflect the novelty of DNA vaccines.

255

I

256 James 5. Robertson and Klaus Cichutek 2

Approaching regulators

While an approach to an official regulatory body needs not be made until step 4 (above) for permission to begin clinical trials, it is highly recommended to open communications with the appropriate authorities as early as possible during development. Early informal communication with the authorities is beneficial to both sides in that the authorities become familiar with the candidate novel vaccine and also in that those involved in developing the vaccine can discuss the necessary requirements and testing which may have to be implemented. This is especially important when the vaccine or its method of production is quite novel or when the manufacturers have little or no previous experience of vaccine manufacture. Within the European Union (EU), a DNA vaccine will ultimately be licensed through a centralized procedure whereby the licence application will be submitted to the European Medicines Evaluation Agency (EMEA) for assessment by the Committee for Proprietary Medicinal Products (CPMP). However, for clinical trial authorization, there is no Europe-wide procedure and each individual state has its own set of committees, rules, and regulations to be followed. Plans are afoot to harmonize clinical trial requirements throughout the EU but until this is achieved, vaccine developers must obtain regulatory permission from each individual state within which they wish to conduct a trial. The regulatory hurdles and the various bodies involved within each state vary widely.

3

Vaccine manufacture

Development of a novel vaccine will begin with a laboratory demonstration in a suitable animal model of the potential efficacy of the vaccine. Beyond these initial basic laboratory experiments, the development of a novel vaccine becomes an expensive and tightly regulated process. For a DNA vaccine, manufacture of a batch of DNA for clinical trials is possibly less troublesome than for other novel vaccines due to the generic nature of the manufacturing process (see also Chapter 11).The first DNA vaccines are based on plasmid DNA derived from a bacterial culture of E. COG. Fermentation of large batches of E. coli is not a novel process and the biotech industry has gained experience from the manufacture of recombinant biotherapeutics. Purification of plasmid DNA was less well advanced, although various companies have now developed appropriate technologies for large-scale purification of plasmid DNA. Because growth of E. coli and purification of the plasmid is a generic process, contract manufacture of plasmid DNA suitable for clinical use in humans is widely available. There may also be unique aspects to the manufacture of a DNA vaccine, especially if the final formulation contains additional material such as a specialized adjuvant or carrier. The guidelines mentioned above issued by the WHO, the U. S. FDA and the EU CPMP outline quality, safety, and efficacy aspects of a DNA vaccine. Prior to the

73 From Research Data to Clinical Trials

first clinical trial, not all tests will necessarily be required, although adequate information on the quality and preclinical safety of the vaccine will be required to justify the first carefully supervised inoculation of a novel vaccine into humans. The manufacture of any vaccine involves a complex biological production system; there is a well-tried and tested approach to assuring the quality, safety, and efficacy of a vaccine (and of biological medicinal products in general) which involves not only the application of tests to the final product, but also careful control of the entire manufacturing process. This approach is known as “in-process”control and is essential to ensure consistency of the final product from batch to batch. In-process control is encompassed within Good Manufacturing Practice (GMP)which includes a review of the background to - and the laboratory development of - the vaccine, characterization of plasmidlcell banks and starting materials, a description of the complete manufacturing process including production and purification, and a detailed characterization of the final purified vaccine (see also Chapters 11 and 12). This latter point will include an appropriate potency assay, ideally an in vitro assay, but more likely an in vivo assay. Within the EU, plasmid DNA incorporating expression constructs intended to be used as DNA vaccines should be constructed and analyzed in laboratories approved according to Council Directive 90/219/EEC.

4

Preclinical safety testing

An initial batch of GMP material will be used in preclinical safety studies. These are designed to ensure that the vaccine, in its final formulated form, will be as safe as can be experimentally determined to a reasonable degree, prior to its first administration to humans. There are several aspects of a DNA vaccine which raise hypothetical concerns which require investigation during preclinical safety testing. The extent to which these concerns are to be addressed prior to the first clinical trial in humans may vary from authority to authority. One concern is that the plasmid DNA may integrate into the chromosomes of the recipient and transform the phenotype of the cell. In a worst case scenario, this may lead to tumor or cancer formation. It is likely that approaches designed to increase the efficacy of a plasmid vaccine, e.g., by increasing its uptake by cells, will lead to an increased risk of integration. As yet, the true extent of the risk associated with this is unknown. However, so far resuIts of tests indicate that integration, if it happens, occurs at a very low frequency, possibly lower than that due to natural mutation. Another aspect which raises some concern is that of adverse immunopathology such as chronic inflammation or generalized immunosuppression. Again this is a somewhat hypothetical issue based on our poor understanding of the mechanism of the immune response to an antigen expressed from a plasmid DNA molecule and the duration of expression. An issue which is becoming of less concern is that of anti-DNA antibodies. There has been concern that the inoculation of quantities of plasmid DNA may lead to the induction of antibodies against DNA and

I

257

258

I

James 5. Robertson and Klaus Cichutek

possibly result in an autoimmune disease such as systemic lupus erythematosus. Experimentally,it has been difficult to induce such antibodies in laboratory models and often the antibodies induced become denatured and are not of the same type as those found in autoimmune disease states. Some DNA vaccine trials involve the co-administration of plasmids expressing cytokines in order to boost or drive a specific immune response. There is some experimental evidence that this should proceed with care since stimulation of the immune system in a particular direction may diminish its ability to respond appropriately to other infectious agents or antigens. Overexpression may also result in toxic levels of the cytokine. Preclinical data, including analyses of the pharmacological-toxicologicalcharacteristics as outlined in the guidelines mentioned above, should be generated in approved laboratories or facilities and experiments involving challenge with relevant pathogens should be performed under safety levels necessary for the pathogen used.

5

Clinical trials

Clinical trials proceed through several progressive stages. The first stage, phase 1, involves a small number of volunteers. The objective of a phase 1 trial is to provide information on short-term clinical tolerance; however, it is also likely to provide the first set of data on the immunogenicity of a novel vaccine in humans. The second stage of the trial (phase 2) involves increased numbers of volunteers and is generally designed to provide information on dosage and vaccination schedules. On some occasions, a combined phase l/phase 2 trial is performed. The third stage (phase 3) will involve many thousands of volunteers and provide information on the protective efficacy of the vaccine. This will ultimately prove or disprove the value of the vaccine. Both phase 2 and phase 3 will also provide more comprehensive data on the safety of the vaccine. Vaccines must have as low a risk of adverse events as possible which is in the order less than one per million, and studies of such events will continue even after licensing through phase 4,the purpose of which is to detect and assess the rate of any rare adverse events. Material for the phase 1 trial should be of adequate quality and as safe as can be reasonably assessed from initial preclinical studies. As clinical trials proceed into phase 2 and phase 3 , the manufacturing procedures may be expanded or modified to cope with a requirement for larger quantities of material. However, if modifications to the manufacturing procedure occur during clinical trials, it is important that the quality of the material used throughout the trials must be comparable or the data obtained may be compromised. Full preclinical testing will be completed, if not already done, during the trials. As trials proceed and manufacturing is refined, information on the consistency of production will also accrue. Since assuring the quality of a vaccine cannot solely be based on testing the final formulated vaccine, consistency in manufacture from batch to batch will be an important aspect of the overall quality assurance package.

13 From Research Data t o C h i c a / Triak

6

Approval for clinical trials

There are no Europe-wideregulations governing clinical trials, and regulations vary between individual member states. In order to promote harmonization, a European Directive on Good Clinical Practice (GCP) is currently under discussion. This Directive also calls for the establishment of a central European database of ongoing clinical trials and adverse events reported. In addition, the current draft of the Directive calls for a better harmonized and possibly also a more centralized approach for obtaining approval for clinical trials in European member states. Additional discussion will be needed to foster the harmonization of clinical trial regulations for DNA vaccines in Europe. The manufacture and use of plasmid DNA in a clinical trial may have to be authorized by the competent authority of the European member state concerned, or notification may be all that is required. Therapeutic DNA vaccines may be prepared for broad or individual usage. In either case, patients may be treated within a clinical trial or under a “compassionate use” regimen. DNA vaccines for preventive use will most likely be used only in clinical trials and authorization is likely to be required.

7

Clinical trial applications in Germany

The following is an example of the regulatory procedure to be followed in order to obtain authorization to conduct a clinical trial of a DNA vaccine within an EU member state. In Germany, the competent authority of the federal state (Land) where a production facility is located is responsible for authorizing the manufacture of a plasmid DNA vaccine according to 0 13 of the German Drug Law (Arzneimittelgesetz, AMG) and in consultation with the Paul-Ehrlich-Institut (PEI). The German Directive “Operation Ordinance for Pharmaceutical Entrepreneurs” (“Betriebsverordnung fur pharmazeutische Unternehmer”) and GMP requirements are also pertinent documents which should be consulted. The competent authority of the federal state (Land) may also review the details of the production procedure prior to or after giving approval for the trial, inspect facilities, take samples for further analysis, or terminate manufacturing. However, it is the manufacturer who has the legal responsibility for the safety and quality of the DNA produced. The competent authority also has to be notified prior to the initiation of any clinical study according to g 40 of the AMG. This states that the scientific background of a proposed trial of a DNA vaccine, ethical considerations, and a description of the quality and safety testing of the DNA should be provided in the clinical trial protocol (“Prtifplan”).This should also contain appropriate information to assure the conduct of the trial according to current and proven scientific standards. The information provided should correspond to descriptions given in the German Di-

I

259

260

I rective “Official Drug Testing Guidelines” James 5. Robertson and Klaus Cichutek

(“Arzneimittelprufrichtlinien”). While this directive outlines specific issues pertinent to applications for marketing authorization of medicinal products for the German Drug Agencies, it may also be considered as a guideline for protocols describing clinical trials of all gene transfer products, regardless of whether or not they are undertaken in view of a later application for marketing authorization. The clinical trial protocol has also to be filed with the competent local ethics committee, formed according to the law of each federal land. The appraisal by local ethics committees is also made in collaboration with the central Commission for Somatic Gene Therapy which was formed following the publication of directives for somatic gene therapy (“Richtlinien zum Gentransfer in menschliche Korperzellen”) by the Scientific Board of the German Medical Association (“Bundesarztekammer”).These directives define ethical standards which would improve the safety and quality of the gene therapy products used in patients and also require the competent local ethics committee advising the investigators of clinical trials to seek advice from the central committee, as its members are selected experts in gene therapy. Occasionally, it may be necessary for the investigator to provide additional oral or written information before a positive appraisal is given with regard to a trial. The competent higher federal authority in Germany (“Bundesoberbehorde”)for vaccines is the Paul-Ehrlich-Institt (Federal Institute for Sera and Vaccines) in Langen. As noted above, the PEI has to be notified, prior to the beginning of a clinical trial. This notification will include documentation providing the pharmacological-toxicological data of the drug, the clinical protocol (“Prufplan”)with the names and affiliations of the investigator, the location(s) where the trial is going to take place and the vote of the competent local ethics committee. Forms for this submission are available by Internet (www.dimdi.de/gewn/awtg/bekannt-txt.htm). Further information on clinical trial regulations in Germany or discussion of data relevant for obtaining marketing authorization as well as scientific advice can be obtained directly from the author K. Cichutek.

Note: Any views expressed in this paper are those of the authors and do not necessarily represent the policy of the NIBSC or of the PEL

References Biologics Evaluation and Research, Food and European Union Guideline (December 1999), Note for Guidance on the Quality, Preclinical Drug Administration, USA. WHO Guideline (1998), Guidelines for and Clinical Aspects of Gene Transfer assuring the quality of DNA vaccines, in: Medicinal Products. CPMP/BWP/3088/99 WHO Expert Committee on Biological Stan(released for consultation). dardization, Forty-seventh Report. Technical FDA (USA) Guideline (1996), Points to Consider on Plasmid DNA Vaccines for Pre- Report Series No. 878. World Health Orgaventive Infectious Disease Indications. Office nization, Geneva. of Vaccine Research and Review. Center for

P/asmids for Therapy and kchation by M.Schleef Copyright 0 WILEY-VCH Verlag GmbH, 2001

I

14 Market Potential for DNA Therapeutics Jurgen M. Dobmeyer and Rita Dobmeyer“’

1

Definition o f biotechnology

In its contemporary sense, “biotechnology”refers to a set of molecular biology techniques that employ living organisms or parts of organisms to make a range of products useful to humans. The power of bio-technologies such as cell fusion, genetic recombination, and polymerase chain reaction (PCR) lies in their collective ability to manipulate and modify the genes of organisms in a targeted fashion. As scientists perfected how to identify, transfer, and express specific genes over the course of the 1970s and 1980s it became possible to genetically “engineer” the entire spectrum of organisms - be they microbes, plants, or animals - to express particularly useful traits. The possibilities were exciting, even mind-bending: defective genes could be replaced with healthy ones to cure disease, with anti-viral genes from bacteria could resist common blights, and microorganisms could be tricked into producing human growth hormones as miniature “drug factories.” By the 1970s, the commercial potential of transgenic organisms was apparent and industry observers touted biotechnology as a revolution for both science and industry. Though bio products steadily enter the market, the pace is a trot rather than the stampede initially expected. The impediments are scientific and social. First, discovering and developing affordable but commercially useful bio products was more difficult than early products would seem to indicate. Second, questions of safety and ethics put a brake on biotechnology’s quick social acceptance in the non-medical fields. Since its inception, biotechnology has been bathed in an aura of both awe and suspicion. Consumer advocates, environmentalists, religious leaders, and even some scientists are critical of the potential dangers of and the ethical questions surrounding genetic engineering. Therefore, marketing of such products has become the most exciting challenge the pharmaceutical industry was faced with during the last few decades. Initially, the sheer novelty of transgenic organisms made it difficult to determine what effects they would have on ecological stability and public health. The NIH

262

I (National Institutes of Health) thus created a regulatory framework for research Jiirgen M. Dobmeyer and Rita Dobmeyer

and development as a precautionary measure to allay public fears. The guidelines for rDNA (recombinant DNA) progressively have relaxed as confidence in safety mounted. Nevertheless, regulations have contributed to the uncertainty surrounding recombinant products, and slowed down their development. A public distaste for “playing G o d also engendered a certain distrust of biotechnology-derivedproducts, especially in areas in which non-recombinant products are easily available and considered “more natural.” Commercial progress for marketing biotechnology in many fields has, for the above reasons, been slower than initially expected.

L

History

This overview shows infectious diseases and their historic incidence. Most of these diseases now have almost disappeared or at least treatment or prophylactic procedures exist due to the impact of modem medicine, e. g., vaccination (Table 1). The speactrum of biotechnology, nevertheless, is deep and wide. In 1994 the biotech industry had sales of $7.7 billion in the United States, and world sales are expected to reach $50 billion by 2010. While pharmaceuticals and diagnostics account for the bulk of the biotechnology products on the market, agricultural products are now growing fastest at a rate of 30 % per year. In bio-electronics,the sensitivity of biological components (e. g., cells or monoclonal antibodies) and the information processing ability of microprocessors are combined to create sensors. The greatest promise in bio-electronics is in the sorting of complex biological information (as done in combinatorial chemistry). Finally, other sectors such as energy and mining have interests in biotechnology, but have been slow to bring competitive alternatives to market as of yet. Overview o f historic incidences o f infections diseases (The Scientific Future of D N A for Immunization, Report of the America Academy of Microbiology, 1997)

Table 1.

Disease

Maximum Cases

Year

Diphteria

206,939

1921

Measles

894,134

1941

Mumps

152,209

1968

Pertussis

265,269

1934

Polio(para1ytic)

21,269

1952

Rubella

57,686

1969

CRS

20,000

1965165

Hepatitis B

26,611

1985

I

14 Market Potentialfor DNA Therapeutics I 2 6 3

Fig. 1. From research t o market: developmental phases of a pharmaceutical including their support activities.

Biotechnology is thus an umbrella term, gathering an impressive range of end products. All applications of biotech are united by the concept of manipulating DNA for human use. In practice, however, the steps necessary for finding, transferring, and expressing useful genes are far from routine. Each organism, or each end product, presents a unique set of challenges. Companies must specialize in niche applications - even within pharmaceutical - simply to master the science and technology and subsequent marketing essential for new bio product development. Different diseases, different end markets, and different research techniques become the fault lines that fragment the hundreds of firms in the biotech “industry” into specialized subgroupings that are not easily interchangeable.

3

Process of pharmaceutical development

A new therapeutic that is launched has passed an average development period of 12-15 years within a pharmaceutical company. The process that is needed to register and market a product has become more and more sophisticated within the last few decades (Figure 1).

264

I

Jurgen M. Dobmeyer and Rita Dobmeyer

Whereas some years ago in the 1970s a new therapeutic only needed to show that it was not toxic in man, today besides safety efficacy has also become a more crucial point. Huge clinical trials mostly designed as global projects to register New Chemical Entities (NCE) have become a real challenge for the pharmaceutical industry. However, this challenge is even greater for new therapeutic approaches where products are of biological rather than of chemical origin. Particularly those products using active ingredients derived from a human being or even more genetically modified products from human beings have caused substantial discussions on their stability and efficacy and, therefore, suitability as a standardized pharmaceutical therapeutic agent. New guidelines referring to the product shape have to be determined for a registration procedure for such a product (see also Chapter 13). Apart from the methodology in a rapidly developing scientific environment only a few scientists are able to fully understand these products and are able to define the methods for large-scale production (see also Chapter 11). Whereas a classical NCE relies on an understandable mechanism of action, such as inhibition of certain biochemical pathways, e. g., inhibition of cyclooxygenase for aspirin-type products or interaction within renal functions such as ACE inhibitors or &blockers to lower blood pressure, plasmid DNA products need more insight into a complex mode of action than these classical products. Moreover, to finally market such a pDNA product needs more effort and explanation to their use than any other of the products registered so far. On the other hand, gene-derived products have more than any other item become a pivotal source for investors and brokers. In the past, no product with a more than several hundred-fold loss of production over years would have been able to generate exponential enhancement of share value. Marketing of a gene technology product i s beginning when the molecule starts its first reactions in vitro. Smaller companies still in their development stage and equipped with a range of promising molecules or even only staffed with the idea of generating such molecules, may leave the big pharmaceutical industry far behind when evaluated for stock value and share prices. No wonder that such challenging business opportunities have forced scientists and managers to develop and enter marketing strategies much earlier than during the last decades of pharmaceutical developments.

4

Human society and technical revolution

Human society is currently passing from the end of the industrial revolution to the first stages of the technological revolution. One clear expression of this transformation is the leap in electronics and computer technology which revolutionizes both production and communication. A second expression i s the emergence of biotechnology as a major factor in pharmaceutical and agricultural production. Biotechnology represents

74 Market Potential for DNA Therapeutics I 2 6 5

.

new manufacturing technology for commodities previously available only by isolation from plants, animals or microbes, and the means to modify the genomes of living organisms so that the changes are propagated in subsequent generations.

This vast acceleration of organic evolution has the potential to generate qualitative changes in the Earth’s organisms and their interactions between each other and the ecosystem. The same technology applied to the human species can be used to modify the genome of humans, an aspect that frequently dominates news stories and popular imagination. The evolution of the extraordinary diversity of living creatures involved their segregation and separation into distinct species. Genes are the blueprints which specify the structures of large molecules which form the living cells of organisms. Until recently, the exchange of genes between unrelated species has been limited by the biologically evolved barriers to gene exchange that are necessary for the formation of species. Thus, in the normal course of evolution, the genes of corn do not mix with the genes of cows since these organisms neither mate with each other nor have they other efficient means of recombining their genetic material. This separation or reproductive isolation means that corn plants produce kernels and cows produce milk proteins. Prior to the biotechnology revolution, it was not possible to tap the capacity of either organism to produce other foods or materials. Genetic engineering technology has now made it possible to cross these barriers so that the proteins of animals can be produced in plants and plant proteins in animals. Genetic engineering permits the isolation of genes from almost any organism - humans, clams, oak trees, rattlesnakes - and their splicing or transfer into the genetic apparatus of other organisms. Human insulin is now produced in bacteria. Cows, goats, sheep and pigs are genetically modified to produce a whole variety of human proteins. (Jonathan King and Doreen Stabinsky: Race and Class, 1999, 40, 2-3).

5

From sequence to product: Applications of biotechnology

The information stored in genes is in the form of a linear sequence of the nucleotides linked together as long polymers to form DNA or its cousin, RNA. The sequence along the gene specifies the sequence of amino acids that will be linked to form the protein molecules that are the building blocks and machinery of cells. The identification of the nucleotide sequences of genes and the amino acid sequences of their encoded proteins is proceeding at an extraordinarily rapid pace. This has been sharply accelerated by the human and other genome projects. The application of biotechnology follows two stages. The first is the isolation of DNA from the cells of an organism and the determination of the nucleotide sequence of particular regions of this DNA that represent genes. US and European patent offices have been granting patents on such sequences since the Chakrabarty

266

I decision (describedbelow).The second stage is the introduction of these genes into Jurgen M. Dobmeyer and Rita Dobmeyer

cells of different species or the modification of the genes and their re-introduction back into the original species. In these cases, the living organisms have been genetically modified, providing one of the legal arguments for patentability: novelty. If the germ cells that provide the sperm or eggs for the next generation are modified, then the progeny of the original individual will also carry the modification. These are referred to as transgenic organisms. For pharmaceutical production technology, the most useful hosts are presently single-cell organisms which can be grown in a vat, such as bacteria or yeast, or cells isolated from higher organisms, but grown outside the body. Such cells reproduce by relatively simple cell division or budding processes. In agribusiness the hosts for foreign genes are usually crop plants. In biomedical research and pharmaceutical development transgenic mice, hamsters, goats, pigs and other animals are generated. The production of human insulin provides a useful example of the early application of biotechnology to the pharmaceutical industry. From 1930-1985, the pancreases of steers and hogs were cut in the slaughterhouses out of the carcasses to provide enough insulin for the millions of insulin-dependent diabetics. From hundreds of thousands of pancreases, the islet cells representing less than 1% of the tissue mass were dissected out and then diced up. The insulin was isolated in a multistep process that, nationwide, required thousands of workers. Through the application of genetic engineering technology (often called recombinant DNA technology), the gene for human insulin has been spliced into bacteria. These cells are grown in 10,000 L vats and produce 20 % of their mass as insulin. One Eli Lilly plant in Indianapolis produces sufficient human insulin for most of the US diabetics population. Harnessing the intrinsic reproductive capacity of living organisms means there is no scarcity of insulin in the US or in the world. However, prices have not dropped significantly to reflect the new abundance. High prices have been maintained by a variety of barriers giving monopoly like control of the market; one of these barriers is patents on the gene sequences (Table 2).

Distribution o f market control for medical products between USA, EU and Japan (US figures, Ernst and Young, 1992, p. 45. EC figures, Jurgen Drews, SACB, 1994. Japan figures, adapted from Mark D. Dibner and R. Steven White: Biotechnology japan, McCraw Hill Publishing Company, New York, 1989, p. 201)

Table 2.

Sector

USA

EC

lapan

Therapeutic

38 %

20 %

26 %

Diagnostic

28 %

23 %

4%

Suppliers

16 %

19 %

11%

Ag-bio

10%

20 %

16%

8%

17 %

40 %

~~~

Chemical Sr Environmental

I

74 Market Potentialfor DNA Therapeutics I 2 6 7

5.1

Milestones in biotechnology: The Human Genome Project

With the announcement of the completion of the Human Genome Project comes the end of the sequencing stage of the genomic revolution and the start of the postgenomic era. The 3 billion letters of DNA coding the human genome have been fully sequenced, well before the initial target date of 2005. This is an important milestone in the efforts to translate this knowledge into practical uses for the benefit of mankind. For all of its importance, however, the sequencing of the genome is but a humble first step into the genomic era. It provides the letters of a new alphabet and a tantalizing glimpse of the future, but the ultimate payoff will be the successful application of genomics to improve human health and quality of life. kowledge of the sequence of the human genome is virtually useless in that pages in a foreign script are of little value without a way to decipher the meaning and the intelligence to evaluate that meaning. There exists a tremendous opportunity for companies to intercede in this process at multiple levels with technologies and information that will catalyze the realization of the genomics promise in its ultimate incarnation: novel approaches to improving human health. Normal genes and variations known as polymorphisms will be pieced together with their corresponding proteins like a giant jigsaw puzzle, identifying hidden biochemical pathways at work in health and disease. Some of these proteins will directly serve as drugs. Others will serve as new targets for drug intervention. Protein, antibody, and small-molecule (chemical) drugs will be developed to act on these targets with much more selectivity and potency than seen today. Still other genes will serve as diagnostic markers for disease, response to treatment, and risk of disease. These powerful tools will enable a new age of personalized medicine. Individuals at risk for disease will avoid or forestall disease by changes in diet or other environmental exposures, just as most of us avoid cigarette smoking to protect our health. Carriers of disease will receive drugs targeted to their specific form of disease, rather than the “one size fits all” approach of today. With the start of the post-genomic era, efforts will now focus on genetic variability, functional genomics, and lead generation and optimization. The start of the post-genomic era heralds a change in the focus of genomics research to three subsectors: 1. Genetic variability or SNPs: This subsector is concerned with how the single nucleotide polymorphisms (SNPs) that constitute most of the genetic differences between individuals influence susceptibility to disease and response to therapeutics. 2. Functional genomics: With a catalog of the entire genome gene sequences can be more rapidly isolated, but functional genomics - the process of determining the function of genes to find those that will make good targets for drug discovery - is going to be a growing field of research. 3. Lead generation and optimization: Identifying targets for drug discovery from the total pool of genes will allow companies to produce and optimize drug

I

268 Jurgen M. Dobmeyer and Rita Dobmeyer

leads using drug discovery methods such as antibody generation (the most direct route to find an inhibitor of a given a target), combinatorial chemistry (for generating small-moleculecompounds) and high-throughput screening (to find active small molecule compounds among the millions available in libraries). 5.2 The future is now: Examples for existing therapeutic approaches using gene products

There certainly are some therapeutic areas which have always been a major target for new therapeutic developments. One of them has been HIV research and vaccination against the human immunodeficiency virus. Diseases that still produce high mortality rates due to the lack of curative treatment such as oncology have also been under substantial focus of biotechnological intervention, directly followed by cardiovascular diseases still leading the list of highest mortality. Of course we can name all examples, but in order to provide some insight in recent developments and their use in daily life, we have chosen some headlines reporting on the use of gene transfer in cardiovascular diseases to show how much this technology is about to revolutionize cardiovascular treatment:

5.2.1

Gene therapy in cardiovascular diseases

Cardiovascular diseases are still the major causes of morbidity and mortality and have become a pivotal target for gene therapeutic approaches. The following case studies and conference headlines may alert on how the complexity of gene therapeutic approaches may target therapeutic intervention of tomorrow (R&Focus, IMS HEALTH, 2000). Gene therapy is growing bypasses in ischaemic hearts. Three teams say they’re stimulating new collaterals by introducing the gene for vascular endothelial growth factor directly into the myocardium or coronary arteries. And though the phase-1 studies that two U. S. teams and another from Finland reported at the American Heart Association meeting here weren’t designed to measure clinical outcomes, all reported at least a trend toward improvement among patients. At St. Elizabeth’s in Boston, Dr. Jeffrey Isner’s Tufts group gave 16 patients in Canadian functional Class IV, who have had one to three Myocardial Infarctions, low doses of VEGF naked DNA. It was delivered by direct intramuscular injection via a mini-thoracotomy as the sole treatment for refractory angina. Average nitroglycerin use was GO tablets per week. After gene transfer, 40 % reverted to Class I and GO % to Class 11, while nitroglycerin use dropped to an average of 2.5 tablets a week. Of 11 patients followed for at least 90 days, six were completely free of angina. Coronary angiography showed improved filling, and nuclear scans revealed significant improvement in left-ventricular function. Gene expression was documented by ELISA. Injection sites were based on angiographically demonstrated arterial blockages. Average operative time was an hour, and most patients were extubated and

74 Market Potentialfor D N A Therapeutics

awake before leaving the OR, and they were discharged within three to four days. There were no MIS or other adverse events. Before treatment, 15 of the 16 patients had an average of two previous coronary artery bypass grafting (CABG) procedures, and 11 had one to three previous percutaneous transluminal coronary angioplasty (PTCA). Those first 16 patients were given only 125 pg of VEGF DNA, but a dose-escalation protocol will take the last 10 patients in the trial up to 500 pg of DNA. In contrast, patients treated for peripheral vascular disease got 4 mg in three separate intramuscular injections. So far the angiogenesis effects appear to be local. Investigators have seen no sign of tumorigenesis, and none of the patients treated has developed diabetic retinopathy. At the University of Kuopio in Finland, Dr. Seppo Yla-Herttuala’s group delivering the VEGF gene plasmid via catheter during standard angioplasty. I n one double-blind study, the Finns infused 1 mg of DNA into coronaries of 10 patients immediately after balloons had opened the stenosed vessels. Five controls got postangioplasty infusions of either galactosidase or Ringer’s lactate. All but one of the gene-therapypatients moved from NYHA Class I1 or 111 to Class I in the six-month follow-up. But the trial was too small to detect any differences in restenosis rates between the groups. The Finnish researchers are now recruiting patients for a phase-2 trial, and they have seven active gene-transfer studies going. Dr. Ronald Crystal and colleagues at New York Hospital-Cornell Medical Center in New York are putting the VEGF gene into a genetically disabled adenovirus vector and injecting it directly into regions of ischaemic myocardium not amenable to bypass. Gene transfer via mini-thoracotomy was the sole therapy for six patients in their phase-1 trial, and was coupled with standard coronary artery bypass grafting (CABG) in 15 others. They saw no evidence of pericardial effusion, myocarditis, or MI in the region of vector administration, and postop evaluation shows a trend toward improvement. Patients all say they’re doing better. In related work, Dr. Michael Mann’s group at Brigham and Women’s in Boston has seen clinical success in another form of genetic engineering designed to make fragile vein bypass grafts behave like more durable arterial grafts in peripheralartery disease. The team i s poised to use the technology to bypass coronaries. The Haward researchers had shown earlier that neointima formation and arteriosclerosis could be blocked in animals by bathing the vein-graft segments for 10 minutes with E2F decoy oligodeoxynucleotide.Genetic activity was documented in small specimens of the human grafts. After at least seven months’ follow-up, they saw graft failure in only five of 17 patients whose grafts were treated, vs. 10 of IG controls whose grafts weren’t treated. Among patients at highest risk for graft stenosis, the long-term failure rate was cut from five of six to one of seven. Dr. Nicholas Kipshidze’s group at the Medical College of Wisconsin i s taking a more direct approach to angiogenesis and coronary revascularization, injecting VEGF - coupled with fibrin glue to sustain protein release for seven to 10 days - directly into the myocardium.

270

I

Jurgen M. Dobmeyer and Rita Dobmeyer

The team has given those agents to patients in India, five through intracoronary catheters and two more via a transthoracic approach. Dr. Kipshidze reports significant clinical improvement at three months’ follow-up with no complications. But there’s been no angiographic confirmation of new collaterals. Further examples of plasmid therapeutics and vaccines are presented in Chapters 3-6 and 11).

6 Legal aspects of gene technology and pDNA derived products 6.1 The extension of patent law to living creatures and their components

A patent allows the owner to prevent others from using or benefiting from the patented invention, process or construct of matter without permission and without compensating the holder. When the patent is over a form of information, such as is encoded in human genes, the holder can prevent others from using this information. US patent law, by granting a monopoly for 20 years to the patent holders, allows a company to prevent other efforts to produce or utilise the “invention”,even if for medical purposes or human welfare. Most people associate patents with true inventions of machines, novel processes or chemical reactions. Over the last decade, private corporations have moved to extend the monopoly of ownership and use inherent in patents to genes, proteins, cell lines and even strains of organisms. US patent laws were originally written by Thomas Jefferson. Jefferson was an active plant breeder and corresponded with leading breeders in Europe. Nonetheless, the patent laws as developed by him excluded animals and plants from their coverage. Jefferson was clear that patents were a form of monopoly. The new nation had come into existence in part because of the recognition of the noxious aspects of monopoly practices. Jefferson believed the importance of the patent laws did not lay in the generalized protection of private property, but in the limited and specific purpose of ensuring that creative and inventive individuals were able to make a living from their work, and thus continue to contribute to society. He wrote that whenever this monopoly was contrary to the public interest, the public interest would take precedence (Donna H. Smith and Jonathan Kin: The legal and legislative background, Environement: Vol. 24, no.6, 1982; Writings of Thomas Jefferson V and VI; in Graham vs. John Deere Co. of Kansas City et al., 383 US 11065). This 200 year-old legacy was breached in 1980 with the granting of patent protection for a genetically engineered bacterium by the US Supreme Court in Chakrabarty vs. US Patent and Trademark Office. The decision was very close (fiveto four) and was narrowly constructed with respect to genetically modified microorganisms. But it opened the gates, and soon Harvard Medical School scientists applied for, and were granted, a patent on a genetically modified mouse. Since then, US, British and European patent offices have issued thousands of patents on genes,

14 Market Potential for DNA Therapeutics

plants, animals and even human cell lines. In the US, this process was carried on outside of public perception through the procedures of the Patent and Trademark Office. The transformation has never been addressed by the US Congress. Opposition has been much more developed in Europe, South-East Asia and South America The ability to modify genes of organisms, together with legal changes that extend patent protection to such modified organisms, has opened up the possibility of the ownership of entire species - not merely the individuals, but all their progeny. Drug companies own patents on human cell lines and on thousands of human genes. Genentech owns patents on a variety of important human growth factors; Sequana Therapeutics has filed for patents on the cells and genes of indigenous tribes in New Guinea. Eli Lilly company owns the patent on the human insulin gene. They have the legal right to prevent other institutions (including nonprofit organizations) from using the information in these genes for producing the corresponding proteins. Thus, although the technology for isolating the human insulin gene from human cells (skin, blood, etc.) is widely distributed, anyone trying to produce insulin would be subject to infringement suits. Hundreds of millions of dollars of such suits are brought regularly within the biotechnology industry as companies jockey for control of the patents. Perusal of issues of Biotechnology or other trade journals reveals the aggressive nature of the industry in protecting its intellectual property patents (Sweeping patents put biotech companies on the warpath’, Science 1985, 208, 658; Gene therapy patent challenge round one, Nature Biotechnology 1996, 14,428). 6.2 impacts of biomedical patents

The initial extension of patent law to gene sequences and cell lines was driven by the pharmaceutical market. Since competition is suppressed through patents, new drugs and diagnostics can be sold at inflated prices. This is one of the reasons that the new therapies derived from biotechnology continue to be expensive, despite the leaps in productivity available through biotechnology (Jonathan King: Gene patents retard the protection of human health, Genewatch 1996, 10, 10-11). An additional, but less visible effect following the patenting of genes and cell lines is the distortion in the priorities of biomedical research and the health care system. Profitability depends on selling a product that is protected by intellectual property laws. However, in many cases the key public health step is to identify the aetiology of the disease and prevent it; for example, by identifying carcinogens and removing them from the ecosystem. But this is not compatible with securing super-profits by selling people with cancer a palliative or therapeutic drug. Thus, Myriad Pharmaceutical could use its breast cancer gene technology to identify environmental or occupational carcinogens that cause the damage found in the genes of individuals with breast cancer. Instead, it is using its technology and patent rights to sell a diagnostic test which informs the individual how much damage has already occurred in their genome. Exploiting patents requires selling consumers a product, not keeping them from contracting disease.

I

271

272

I

Jurgen

M. Dobmeyer and

Rita Dobrneyer

The profitability of the insulin market depends on millions of people affected by the disease. Despite the very large biomedical research community focused on diabetes, research on what causes the damage to the insulin-producing or sensing cells is limited. Yet there are a flood of studies on managing the patient in terms of diet factors, genetics and behavior. These distortions represent insidious long-term economic influences on the medical community, rather than the limitations of biomedical science. (Roman Saliwanchik: Legal Protection for Microbiological and Genetic Engineering Inventions, Addison-Wesley, Reading, 1982). Corporations claim that, without patent protection, important technologies will not be developed. In fact, patent protection does not ensure technological development, but super-profits. Patents are as often used to prevent the development of new technologies as to exploit them. The role of patents as a mechanism of monopoly pricing rather than technology development has been described in a number of sectors of industry, including antibiotics. Corporate interests argue that venture capital is harder to raise in the absence of patent protection. In those cases when a therapeutic development is not going forward because of problems in raising capital, the capital can be raised through public agencies. For example, vaccines for humans have often been avoided by the pharmaceutical industry, since a few doses provide protection for a lifetime, and liability issues can be serious. In contrast, many firms produce agricultural vaccines for cattle, sheep and pigs since these are slaughtered each year, and new populations are grown up requiring vaccination in turn - a much more profitable market. In these cases, the public sector is fully capable of filling the gap. The US Center for Disease Control oversees the production of a number of vaccines, and the National Institute of Health brought the anti-tumor agent taxol into production as an experimental drug, before the pharmaceutical industry recognized its market potential. 6.3

Resisting corporate ownership o f life forms

In Europe, South-East Asia and South America, there are significant social movements opposing life patents. Dramatic public demonstrations occurred in India in response to W R. Grace’s obtaining patents on the Neem tree, and these were followed by a vigorous battle in the India’s upper parliament to resist the GATT intellectual property requirements. Peasant farmers are struggling to maintain control over the material basis of their livelihood, the agricultural crop plants on which they depend. They are also fighting to maintain control over their culture, as represented in the knowledge of producing and using rice. It is from these farming communities of the world, in particular in the developing world, that much of the resistance is emanating. At the Second Ministerial Conference of the World Trade Organisation in Geneva, peasants from a global movement called Via Campesina were in attendance. Their press release, issued at the Ministerial Conference, stated: “The Via Campesina is demanding ... governments and ... international institutions prohibit biopiracy and patents on life (ani-

14 Market Potential for D N A Therapeutics

mal, plants, parts of human body) including the development of sterile varieties through genetic engineering.” Some of the strongest resistance to life patents has come from indigenous peoples. Because of their potential genetic uniqueness, they have been the brunt of much poking and swabbing, for blood and cells, and the cells of a number of indigenous individuals have been patented by the United States Department of Commerce. The “Declaration of Indigenous Peoples of the Western Hemisphere Regarding the Human Genome Diversity Project” has a view clearly different from that of the officials in the US Patent and Trademark Office: “We oppose the patenting of all natural genetic materials. We hold that life cannot be bought, owned, sold, discovered or patented, even in its smallest form ... We denounce and identify the instruments of intellectual property rights patent law ... as tools of legalized Western deception and theft. We denounce all instruments of economic apparatus such as NAFTA, GATT and the World Trade Organization (WTO) which continue to exploit people and natural resources to profit powerful corporations assisted by governments and military forces of developed countries.” The European Parliament, responding to the initiatives of the Greens, refused to accept patents on genes. The first vote on such legislation a few years ago ended in a decision against patents on life. Only the full mobilization of pharmaceutical and biotechnology forces in 1997 led to the defeat of this effort. As noted above, in 1999 there was a review of the Trade-Related Aspects of Intelectual Property Rights (TRIPS) agreement within the %TO, specifically to revisit the provisions that exempt plants from utility patent protection. Developed countries, led by the United States, will attempt to insert a form of words requiring member states to grant utility patents on plants. There will be a significant amount of resistance to this attempt, at both the governmental and non-governmental levels. In the United States and Canada both religious and secular coalitions have taken clear positions in opposition to life patents. The Blue Mountain Declaration of June 1995 speaks for many: “No individual, institution, or corporation should be able to claim ownership over species or varieties of living organisms. Nor should they be able to hold patents on organs, cells, genes or proteins, whether naturally occurring, genetically altered or otherwise modified. Indigenous peoples, their knowledge and resources are the primary target for the commodification of genetic resources. We call upon all individuals and organisations to recognise these peoples’ sovereign rights to self-determination and territorial rights, and to support their efforts to protect themselves, their lands and genetic resources from commodification and manipulation. Life patents are not necessary for the conduct of science and technology, and may in fact retard or limit any benefits which could result from new information, treatments or products ... As part of a world movement to protect our common living heritage, we call upon the world and the congress of the United States to enact legislation to exclude living organisms and their component parts from the patent system. We encourage all peoples to oppose this attack on the value of life” (Blue Mountain Declaration, 3 June 1995).

I

273

274

I

Jiirgen M. Dobmeyer and Rita Dobmeyer

7 Health care in the light of biotech is different in Europe and US

The application of biotechnology will permit significant increases in new strategies for major advances in the diagnosis of disease and the development of new therapeutic regimes. The impact of biotechnology in the globalized economy is to realize this potential. Pharmaceutical, agricultural and biotech firms are moving rapidly, under the guise of free trade agreements, to privatize human animal and plant genomes, useful cell lines and agriculturally important crop plants and livestock. 7.1 Biotech in the

US

from an economical point o f view

More than 20 years ago, the first private biotechnology industry emerged in the USA. The new companies were founded by scientists with strong ties to university biotechnology. For their survival these new companies functioned as R&D contractors of established companies. However, from the mid 1980s onwards, the vertical division of labor between the different companies has been replaced by an ongoing vertical integration of the biotechnology sector in the USA. In no other country has the number of companies and the amount of capital invested been as high as in the USA. The main US biotechnology trade organization, Biotechnology Industry Organization (BIO), using a broad definition of biotechnology states that in 1994 there were a total of 1,311 biotechnology firms in the US. The 20 years of US biotechnology industry can be seen as a two-phase development. The first phase has been characterized by the establishment of new biotechnologyfinns (NBFs) and a strong division of labor between the new NBFs and established firms. During the second phase, starting in the mid 1980s, NBFs and established firms were involved in an integration process. 7.2 Emergence o f new companies

The first phase occurred approximately between the mid 1970s and 1987, in which NBFs were established. The NBFs generally started as research organizations, selling scientific and technological knowledge, but no products. They did not undertake R&D in the broad range at which established companies were active. Instead they focused on specific technologies and niche markets. Two discoveries really triggered the development of biotechnology and NBFs: Firstly, the discovery of a technique to transfer specific genes from one organism to another by the US scientists Boyer and Cohen in 1973; secondly, the invention of the cell fusion or “hybridoma” technique by the British scientists Milstein and ICohler in 1975. Recognizing the commercial potential of these discoveries, many NBFs were founded by university scientists in collaboration with entrepreneurs and suppliers of venture capital. A strong industry-university relationship is one

74 Market Potential for DNA Therapeutics I275

of the remarkable aspects of the development of the US biotechnology industry. In the 1970s most of the expertise in genetic engineering was found at the universities. Many of the NBFs were started by university faculties interested in simultaneously retaining their professorship and participating in the development of a company. In the close vicinity of universities, special science parks were newly established, where high-tech firms could build or rent facilities under beneficial conditions. Most of the biotechnology companies are now located in the states of California and New England, close to universities such as Harvard, MIT, Stanford and UCLA in Davis. In the beginning these companies had to fund the cost of infrastructural development, without the benefit of internally generating revenues. The NBFs, therefore, had to depend on venture capital, stock offerings, and relationships with established companies for their financing. Although venture capital and stock offerings were an important source of funds for the NBFs, contract research for established firms has always been the most important. Between 1977 and 1985 established enterprises, mostly in the pharmaceutical and chemical industry, provided the mayority of the total funds invested in NBFs. Apart from the need for capital, NBFs benefited from their research and development (R&D) supplyldemand relationship with established firms to get access to downstream capabilities in manufacturing, clinical testing, regulatory processes, and distribution. A last reason was the scope economics in basic biotechnology R&D. Because different commercial products were based on similar basic technologies, the costs (and risks) of developing these technologies could be shared by clients with different commercial interests. For established firms, buying biotechnology R&D from the outside and directing the main focus at commercialization had advantages. Control over manufacturing, testing and distribution facilities could be used to acquire access to the technology on attractive terms, without the large and risky investment in an in-house biotechnology unit. In addition, collaboration allows established firms to tap the specialized expertise of the NBFs. 7.3

6iotech in Germany

Since the 1970s, the German Federal Ministry of Research had been interested in biotechnology. (The ministry was variously named in the past three decades. It is currently called the Federal Ministry of Education, Science, Research and Technology - Bundesministerium fur Bildung, Wissenschaft, Forschung und Technologie.) Long before the German pharmaceutical companies, which were comfortable with their strong performance as “pharmacy to the world,” the Research Ministry identified biotechnology as a growth sector and began spending money to develop it. By the 1980s, after spectacular success in the US, German industry had become interested in biotechnology, but remained moribund or off-shored research to the United States due to German regulatory hurdles and strong anti-biotechnology pressure mobilized by the Greens. Although the Research Ministry doubled its spending on biotechnology by the end of the 1980s, the sector fell farther behind the US, Japan, and even Great Britain. By the early 1990s, biotechnology outside

276

I Germany had clearly become a “future industry” with predicted growth rates of up Jurgen M. Dobmeyer and Rita Dobmeyer

to 25 % in certain sectors. The German government determined, not the least with an eye to unemployment, to make Germany the number one in biotechnology in Europe. The Research Ministry defined its technical strategy by looking to the United States. The approach has been surprisingly successful: regions throughout Germany have mobilized organizations supporting small start-ups and linking industry and universities, banks have become involved in providing pools of venture capital, and the number of small biotech start-ups has doubled twice in the past three years. Prominent industry analysts conclude that Germany has entered a biotech boom (Schitag, Ernst, and Young, 1998). Containing left and Green party opposition to biotechnology is not the only story, despite the publicity it has received. Anti-technology sentiment, at its zenith in the late 1980s, has been contained by national regulations and muted by party leadership. Indeed, the current Social Democratic (SPD)-Green government approved a lukewarm embrace of the sector. The SPD, the party closest to the trade unions, modified its stance after realizing the biotechnology sector is creating hundreds of new highly paid jobs. The Greens, born in opposition to nuclear power, decided that biotechnology opposition need not be a holy cow. Their 1998 agreement with the SPD to form Germany’s government aimed to phase out nuclear power, but to support pharmaceutical biotechnology (human health) and bioremediation (environmental health), to preserve freedom of research despite a re-iteration of the programmatic opposition to genetic engineering, and to accept the existing EU and national regulations. Nods to the fundamentalist position included only longterm monitoring of transgenics, and additional spending for research on the social and ethical consequences of genetic engineering.

8 Who is the health care industry and their clients, a paradigm?

In general, it is relatively easy to identify the consumer, predominantly because the consumer and the buyer is most likely identical, but not so in the healthcare industry. Consumers are patients, but buyers are predominantly health insurance companies and similar sometimes governmentally organized organizations (Figure 2). However, marketing a pharmaceutical compound is rather complex. The following topics might give an idea of the difficult interactions in the health care market: shareholder value, managed care organizations, governmental regulation, uninsured people.

74 Market Potential for DNA Therapeutics

Fig. 2.

Relationship between the parties interacting in health care and society.

8.1 Health care industry share prices

Health care industry share prices through the third quarter of 1999 dropped 2.8 % closing share prices on December 31, 1998, according to the nine-month share price report by Healthcare Markets Group, Hilton Head, South Carolina. The report analyses share price performance through the third quarter of 1999 for 1,000 public health care companies in 36 market segments. By way of comparison, health care industry share prices increased by 1.1% in the first six months of 1999 and dropped by 4.7 % in the first quarter of 1999, they increased by 1.9 % calendar year 1998, and rose by 10% calendar year 1997. Biopharmaceutical group share prices rose by 4.4%, medical devices and supplies group share prices rose by 5.3%, and providers and services group share prices dropped by 21.5 % during the third quarter of 1999. The nine-month results reflect lost enthusiasm for health care stocks by institutional investors and the market in general, excluding some of the large-cap companies, companies with consistently strong earnings growth, and select biopharmaceutical/genome companies that have promising drug and gene therapy products in development or production.

I

277

278

I

Jurgen

M. Dobmeyer and Rita

Dobmeyer

Since January 1, 1999, significant a share price erosion in the long-term care (165.7 %), assisted living (154.8 %), hospital companies (146.7%), drug distribution (137.9%), and practice management (135.8%) segments, coupled with minimal gains for the historically strong pharmaceutical segment (1.6 %), have retarded industry share price appreciation. The Internet segment, which was added to the provider group as a stand-alone segment in the second quarter of 1999, saw its share price rise of 91.9 %. Segments that saw large share price growth were biotechnology (37.1%), lasers (102%). and surgical devices (28.9 %). 8.2 HMO enrolment rose

Health Management Organisations (HMOs) enrolment rose from 94.8 million in December 1997 to 105.3 million in December 1998, an increase of more than 11% according to the SMG Marketing Group Inc., Chicago, IL. California HMO plans continue to lead the market in total enrolment, although Rhode Island has the greatest HMO penetration (more than 64 %), followed by California, Delaware, Connecticut, and Utah. Point-of-serviceplans showed a 34% enrolment increase and account for more than 26 % of total HMO enrolees. In addition, preferred provider organizations (PPO) plans experienced an enrolment increase of more than 10% in 1998. The growth of these two types of plans indicates an effort by employers and employees to gain increased provider choice within their managed care plan. 8.3 Limitations to the access

of health care services

The US Supreme Court will use an Illinois case to decide whether patients have an unlimited right to sue HMOs. Cynthia Herdrich of Bloomington, IL, sued her physician and the Carle Clinic Association for malpractice in 1992, alleging her appendix ruptured and peritonitis set in because physicians delayed diagnostic tests until they could be performed at their own testing facility. Herdrich was awarded $35,000 under state law. Herdrich also sued under the Federal Employee Retirement Income Security Act, alleging the HMO violated its management duty to patients by depriving them from proper medical care and keeping the cost savings. The physician, health plan, and insurance company appealed to the Supreme Court, arguing that the decision creates a new form of liability for physicians and managed care plans. The Supreme Court decision will have far-reaching implications for all managed care plans.

14 Market Potential for DNA Therapeutics

8.4

US uninsured population rose

The nation’s uninsured population increased by about 1 million since 1997, although the proportion of the uninsured population remains statistically unchanged from the previous year, according to the March 1999 Current Population Survey by the US Commerce Department’s Census Bureau. More than 44 million people, or 16.3% of the US population, had no health insurance in 1998. Most of those lacking health insurance were young adults 18-24 years old, people with less education, Hispanics, part-time workers, and foreign-born people. Children’s healthcare coverage remained much the same from 1997-1998,with 11.1 million, or 15.4%, of all children under the age of 18 uninsured. The proportion of people without health insurance ranged from 8.3 % in households with annual incomes of $75,000 or more, to 25.2 % in households with less than $ 25,000 in income. More than 70 % of people who had health insurance were covered by a private insurance plan received through an employer or a union. The government also provided health care coverage, including Medicare (13.2 %), Medicaid (10.3%). and military health care (3.2 %). Comparisons of two-year averages (1997-1998versus 1996-1997) showed that the proportion of the population without health insurance dropped in eight states (Arkansas, Florida, Iowa, Massachusetts, Missouri, Nebraska, Ohio, and Tennessee) and rose in 16 states (Alabama, Alaska, California, Illinois, Indiana, Maryland, Michigan, Montana, Nevada, North Dakota, Pennsylvania, South Dakota, Utah, West Virginia, Wisconsin, and Wyoming).

9 Economical evaluation

of the biotech marked in the future

With a series of recent product development disappointments faced by biotech companies, the market has lost some of its enthusiasm for biotech stocks. Since the market value of biotech stocks is heavily influenced by investor sentiment, it may not be an accurate reflection of the true value of these companies. Also, many biotech companies are not traded on an exchange. As a result, financial models must be used to arrive at a reasonable range of values for the company. Existing financial models, such as discounted cash flow and option pricing models, can be used for this purpose, but they must be tailored to reflect the unique features and risks associated with the biotech industry. While experience and skill are required to distil the vast amount of information required into accurate assumptions, biotech valuation remains a practice somewhere between art and science. PricewaterhouseCoopers explored this topic in a recent paper (Journal of Biotechnology in Healthcare Research and Regulation, 2000, 4,10-15).

I

279

280

I

Jiirgen M. Dobmeyer and Rita Dobmeyer

10 Conclusion

Despite all obstacles of political and economic origin biotech will no doubt have piqued the interest of established pharmaceutical giants as well as small biotechnological companies. Even if early marketing and struggle for money through share holders will be make it somewhat difficult for some of the companies to survive, it will also generate an early selection that may generate survival of the fittest companies and products. This survival is potentially not only based on the biological competence of a compound, but also on the management and ability to finetuned resources of a respective senior management within a company. Both challenges - medical and economical - will make sure that the future market will provide best biotech value for money which will lead us into the next generation of health care.

P/asmids for Therapy and kchation by M.Schleef Copyright 0 WILEY-VCH Verlag GmbH, 2001

I281

Index

A A-DNA 195 active ingredients 264 adenoviral vectors 76 adjuvant 48f, 105, 141 - effects 187 agarose gel electrophoresis (AGE) 35, 201 agricultural vaccines 272 AIDS 148 alkaline lysis ZlOf, 213 allergy 57, 203 ampicillin 205 angiogenesis 269 animal model 76 anti-DNA antibody 98, 257 anti-DNA autoimmune 187 antibiotic resistance 18, 21, 24, 29, 140, 145, 205 antibodies 203 antibody response 50 antigen presenting cells (APC) 93 antigen trafficking 174 antigens expressed in fish 176 antisense RNA 193 aquaculture 169 aquaculture industry 173 AT-rich region 216 attenuated live vector 51 autoimmune disease 57, 98, 258 autoimmune response 59 B

B cell tolerance 97 B-DNA 195 BAC 10, 27 bacterial conjugation bacterial gene 140

199

bacterial host strain 205 balanced transcription 132 batch adsorption experiment 220 BCA assay 201, 203 bidirectional expression 132 bidirectional promoter 130 bioburden assay 241 biomass 230 biopharmaceutical product 202 biosafety law 205 biotechnology industry 274 biotechnology market size 262 boost 55, 94f, 115, 258 buffer condition 198 bulk 242 C cancer 228 cancer formation 257 capillary gel electrophoresis (CGE) 38 cardiovascular diseases 268 cationic lipids 84 cattle vaccination 183 cell growth 3 CGE 244 chaotropic salts 213 chimpanzees 95 chromatographic leachate 200 chromatographic processing 214 - hydrophobic interaction 214 - ion-pair 214 - reverse-phase 214 chromatography - afinity 227 - anion-exchange 220, 229 - hydrophobic interaction 229 - matrice 219

282

I

lndex

-process 219 - scale-up 222 - size-exclusion 223 - triplex affinity 226 cis acting sequences 154 clarification operation 222 clinical monitoring 263 clinical phases (overview) 263 clinical trial 59, 258 - application 259 - approval 259 - authorization 255 -phase 1 258 - phase 2 258 -phase 3 258 clinical trials of DNA vaccines, review on 98 cloning 199 cloning strategies 147 cloning vector 4, 9, 24, 29 CMV enhancer 180 CMV promoter 77, 90, 106, 131, 153 co-purification of endotoxin 221 codon optimization 150, 157 codon usage 153 colorimetric protein assay 248 combinatorial biology 147 Committee for Proprietary Medicinal Products (CPMP) 256 competent bacteria 27 complex growth media 206 conjugation 8-9 contaminating protein 203 contract manufacture of plasmid DNA 256 contract research organization (CRO) 263 coronary artery bypass grafting (CABG) 269 covalently closed circles (ccc) 30 CpG 52, 141, 150, 165, 184 CpG motifs 92, 97 CPMP guideline 255 cross-priming 51 CRS 262 CTL response 51 CTX 114 cultivation processes 206 cystic fibrosis 228 cytolcine 48f, 52, 54, 88, 141, 171f, 185, 203, 258 cytotoxic responses 94

density gradient ultracentrifugation 214 detection limit 240 diafiltration 229 diphteria 262 DNA concentration 243 DNA delivery gene transfer 49 - geiiomic 33 - gold particle 49 - injection 49 - liposome 49 - naked DNA 49, 145 - transfection 49 - virosome 49 DNA intercalating dyes, ethidium bromide 32 DNA production, industrial scale 33 DNA quantitation 239 DNA topoisomerases 31 dose 81, 84 dose-escalation protocol 269 downstream processing 208 drug - large-scale production 264 - marketing 264 - safety 264 - stability 264 - toxicity 264 drug delivery 77, 85 drug efficacy 264 drug factories 261 drug screening 147 duck 97 duck hepatitis B virus 89

E E. coli strain 111 ecological stability 261 economical evaluation 279 electron microscopy 33 endolysosome 53 endoplasmic reticulum 53 endotoxin administration dose 203 endotoxins 33, 201, 204, 220, 228, 248 enzyme linked immunosorbent assay (ELISA) 203 Ethics Committee 114, 260 ethidium bromide 202 European Medicines Evaluation Agency (EMEA) 256

D DDX 114 death 193 deletion 250 dendritic cells 51, 91, 172

F fed-batch mode 206 fermentation 238 fermentation broth 200

Index

fermentation condition 199 fermentation media 206 fish immunology 171 fish vaccination 169 Food and Drug Administration (FDA) 255 food chaim 187 formulation 78, 84f - of plasmid based therapeutics 40 fragile vein bypass grafts 269 functional genomics 166, 267 G gene delivery systems 151 gene design 147 gene expersion - distribution 180 - duration 180 gene gun 52f, 98, 111, 113, 115 gene synthesis 147 gene transfer 139 - ballistic transfer 143 - efficiency 143 - electroporation 143 - injection 48 - lipid- or polymer-mediated 143 - microinjection 143 generic process 201 genome of humans 265 genome size 197 germ-line 187 Good Laboratory Practice (GLP) 240 GMP 141 gold beads 99, 111, 178 Good Manufacturing Practice (GMP) 33, 257 - GMP production 84 - GMP requirement 259 green fluorescent protein 176 ground squirrel hepatitis virus 89 growing bypasses in ischemic hearts 268 growth conditions 206 GTAC 114 guideline 201 gyrases 32

H hepatitis B 98, 143, 262 - acute 88 - B cell epitopes 89 - chronic 87, 96 - chronic carriers 96 - DNA vaccines 89 - HBsAg transgenic mice 96

- immune response to infection 88 - non-responders to HBV vaccines 91 - prevention 90, 97 - surface antigen 89 - T cell epitopes 89 - treatment 87 - virion structure 89 - virus genome 90, 96 hepatitis B virus (HBV) SO hepatitis C virus (HCV) 51 herpes simplex-2 98 high cell density 206 high-throughput screening 268 histochemical analysis 79 HIV-1 98, 151, 164 HIVjSIV 51 HLAtype 107 hormone 203 host 2 host cell 200 host cell protein 248 host DNA 246 host strain 238 human genome project 267 human society 264 husbandry 174 I immune response - cellular 50 - humoral 50 immunization -DNA 45 - genetic 45 - nucleic acid 45 immunoblot assays 203 immunological memory 172 immunoreactivhy 141 in-process control 41 industrial manufacturing processes 41 inflammatory responses 181 influenza 98 influenza virus 55 inhibitory sequence elements 165 injection site 179 insertion 250 insertional carcinogenesis 98 insulin 266 interferon-alpha 87 internal ribosomal entry sites (IRES) 120 intracellular pathogens 174 intramuscular injection 176 ion-exchange chromatography 217 ionic strength 199

I

283

284

I

Index I< kanamycin 205 kanamycin resistance 106 kinetic PCR 246 Kozak sequence 146 L laboratory animals 204 laboratory experiment 256 LAL assay 201, 204 large-scale plasmid DNA manufacturing 194 legal aspects 270 lentiviral vectors 151 leukocyte pyrogen assay 204 licensing a novel vaccine 255 ligand-DNA binding 196 Lilnulus amoebocyte lysate (LAL) 248 linking number 196 lipoplexes 77, 79 lipopolysaccharides (LPS) 203 liposomes 77, 237 long-term memory 94 lot release criteria 251 lot release testing 202 luciferase 175, 180-181 lymph nodes 94 lysis 200, 208, 210 - of bacterial cells 27 lysozyme 210

M malaria 55, 98, 103, 148 - endemic area 105 - infection 103 - life cycle 103 - parasite 103 - symptom 103 - vaccination 104 manufacturing cost 205 manufacturing processes 201 market expectation 193 market potential 261 marketing approval 255 Master Cell Bank (MCB) 205, 251 MCA 114 measles 262 nielanoma 228 microinjection 77 MIDGE 142 migratory macrophages 177 mini-thoracotomy 269 mobilization 8 mode of action 194

model, cap-dependent translation molecular weight 196 mumps 262 muscle fibers 93 muscle structure 177 mutation 257 N necrotizing agents 94 neonatal immunization 96 nitric oxide (NO) 75 non-responder 54 nuclear export pathways 156 nuclear translocation 154 nuclease 213

0 oncogene 193, 202 open circular form 30 open reading frame 205 origin of replication (ori) 29, 78 P parasites 174 parenteralia 242 patents 265, 270 pathogenicity 141 patient safety 202 pCI-system 128 percutaneous transluminal coronary angioplasty (PTCA) 269 personalized medicine 267 pertussis 262 pharmaceutical development 263 pharmacological-toxicological characteristics 258 plasmid - activity 201 - adsorbent 221 - backbone 4, 78, 106, 165 - binding capacity 221 f - biodistribution 112 - ccc-form 112, 244 - CGE assay 202 - characterization 237 - clarification 213 - clinical batch 111 - column capacity 229 - concatemers 31, 39 - concatenates 31 - concentration 179, 213 - conformation 195 - contaminant 200, 202 - content 243

122

- copy number 1f, 4, 29, 199, 206, 238 - cryptic 11, 27 - defect 230 - degradation 114 - dehydration 195 - demands 229 - denatured 208, 219, 228 - design 141, 204 - dimension 197 - dimers 31, 34, 36, 39, 230 - distribution 178 - DNA manufacturing 199 - DNA sequence 251 - dosage 183f, 228 - downstream processing 199 - efficacy 187, 237, 257 - efficiency 229 - fermentation 204, 206 - form 199 - formulation 111, 226 - genetic stability 199 - high copy number 24, 205 - homogeneity 244 - hydrodynamic radius 224 - identity 112, Z O l f , 205, 237, 249 - impurities 200 ff - in vitro assay 257 - in v i m assay 257 - injection 89 - integration 47, 59, 187 - isofoms 194, 201, 218 - large 39, 48, 220 - large-scale production 199 - largescale purification 256 - linear 30, 181 - linearized 40 - low-copy number 4 - manufacturing 193, 230, 237 - manufacturing process 230, 256 - mobilization 2 - mode of action 264 - molecule surface 196 - monomer 34, 36, 39, 244 - multimeric variant 230 - natural plasmids 1, 29 - nick 230 - nomenclature 1 - PCR 202 - potency 112, 187, 201, 204, 257 - potency assay 111 - preparation 212 - product specification 201 - production 111, 237 - productivity 219, 229

- purification 24, 229 - purity 112, 187, 201, 237, 243 -quality 257 - quality assurance 230 - quality control 230, 237 - quality specification 112 - quality standard 201 - quantification 41 - re-association 211 - recovery 212, 220 - relax 219 - release test 237 - replication 2, 29, 31 - rolling circle 4 - routes of administration 176 - safety 187, 201, 204, 237, 257 - segregation 2 - selectivity 229 - sequencing 202 -shape 141 - shearing 212 - size 139, 141, 199 - stability 29, 41, 112, 205, 237 - sterility 205 - storage 114, 226, 230 - structure 194 - structure analysis 33 - supercoil 112, 179, 210, 216, 219c 227, 244 - topoisomer 218 - topologic variant 230 - topology 30, 202 - toxicity 112, 187 - translocation 180 - transport 177 - upstream processing 199, 204 - uptake 177, 257 - vector 204 -volume 179 - yield 206, 208, 229 poiltilothermic animals 174 Points to Consider document 255 polio 262 poIycistronic expression vector 121 polyvalent vector 119 posttranslational modification 51 potency assay 110 preclinical testing 112, 257 primate model 94 process approval 230 process development 199 process flow sheet 208 process yield 210 product launch 263

286

1

Index

product recovery 206 production microorganism 199 promoter 175 proof-oficoncept 76 proteomics 166 protocols 79 pTRIDENT-based plasmid 128 public health 261 purification process 194

Q quality assurance 40 214 quality assurance assay 201 quality control 38, 41 quality control assay 201 R radioimmunoassay (RIA) 203 reactions 187 recombinant proteins 202 regenerative processes 193 registration procedure 264 regulatory aspects 255 regulatory framework 262 removal of impurity 222 replication 2 f, 20 - origin 3 restenosis 75 restriction analysis 249 rhesus macaques 94 ribozyme 166 RNA 33 -export 157 - stability 147, 154, 157 RNase 210 rubella 262 S safety 115, 193 safety profile 84 safety risk 59 scale-up 205 SDS gel electrophoresis 248 selection marker 205 Semlilci forest virus 47 sequence design 145 shallow gradient 219 shear sensitivity 210 sheep vaccination 183 shuttle vector 27 side effect 202 single nucleotide polymorphisms (SNPs) 267 site-directed niutagenesis 147

SIV 164 social acceptance 261 Southern blot 201 Southern hybridization 246 spectrophotometric analysis 243 spiking 239 standard operation procedures (SOP) 241 sterility 241 study designs 170 submitting applications 255 supercoil 32, 195-196 superhelix density 197 surgical procedures 79 SV40 early gene promoter 90 SV40 promoter 131 synthetic genes 147, 152, 158 synthetic oligonucleotides 184

T T cell priming 54 T cell response 50-51 T cell tolerance 97 tet-promoter 131 theoretical risk 202 therapeutic efficacy 82 thymidine lcinase 131 thymidine kinase promoter 180 Ti plasmids 1 2 tissue damage 181 transformation 24 transgenic organisms 266 translational efficiency 157 trauma 181 tuberculosis 51 tumor formation 202 tumor repressor gene 193, 202 tumor vaccine 51, 57 V vaccination - animal model 56 -DNA 45 - genetic 45 - nucleic acid 45 - of infants 91 - polyepitope 47 - polyvalent 53 - prophylactic 54 - therapeutic 54 vaccine - formulation 55 - manufacture 256 - safety 158 - storage condition 55

Index

validation 230 vector - comniercially available 47 - construction 47, 140 - design 59, 132, 140, 237 - integration 140 - persistence 140 - polycistronic 49, 52 veterinav medicine 89 viral challenge 96 viral sequence 205 viral vector 82 virosome gene transfer 76 vim5 production 164 viruses 174

W WHO Guideline 255 woodchuck hepatitis virus 89 Working Cell Bank (WCB) 205 worst case scenario 257

X X-ray diffraction L

2 - D N A 195

195

I

28J