THE MOLECULAR EPIDEMIOLOGY OF HUMAN VIRUSES THE MOLECULAR EPIDEMIOLOGY OF HUMAN VIRUSES Edited by Thomas Leitner Swedish Institute for Infectious Disease Control 17182 Solna, Sweden .... " Springer Science+Business Media, LLC Library of Congress Cataloging-in-Publication Data The molecular epidemiology ofhuman viruses / edited by Thomas Leitner. p. cm. Inc1udes bibliographical references and index. ISBN 978-1-4613-5420-8 ISBN 978-1-4615-1157-1 (eBook) DOI 10.1007/978-1-4615-1157-1 1. Virus diseases -Epidemiology. 2. Molecular epidemiology. 3. Viral genetics. 1. Leitner, Thomas RA644.V55 M65 2002 615.5'75-dc21 2002073053 Copyright © 2002 by Springer Science+Business Media New York Originally published by Kluwer Academic Publishers in 2002 Softcover reprint ofthe hardcover lst edition 2002 AII rights reserved. No part ofthis work may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilm ing, recording, or otherwise, without the written permission from the Publisher, with the exception of any material supplied specificalIy for the purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the work Permission for books published in Europe: [email protected] Permissions for books published in the United States of America: [email protected] Printed an acid-free paper. The Publisher offers discounts on this bookfor course use and bulk purchases. For further information, send email to<[email protected]> . Table of Contents Preface VII Part One: Introduction to Molecular Epidemiology 1. The Use of Molecular Epidemiology Thomas Leitner 2. DNA Technology for Molecular Analysis of Viruses 11 Deirdre 0' Meara and Joakim Lundeberg 3. Phylogenetic Approaches to Molecular Epidemiology 25 Keith A. Crandall and David Posada 4. The HIV Databases: History, Design and Function 41 Bette Korber and Carla Kuiken Part Two: Application of Genetic Methods to Investigate Virus Spread 5. The Evolution of Primate Lentiviruses and the Origins of 65 AIDS Elizabeth Bailes, Roy R. Chaudhuri, Mario L. Santiago, Frederic Bibollet-Ruche, Beatrice H Hahn and Paul M Sharp 6. Recombination and Molecular Epidemiology of HIV -1 and 97 Enteroviruses Mika Salminen 7. Molecular Epidemiology of Human T Cell Leukemia! 121 Lymphoma Viruses Type 1 and Type 2 (HTLV-1/2) and Related Simian Retroviruses (STLV-1, STLV-2 and STLV-L/3) Antoine Gessain, Laurent Meertens and Renaud Mahieux 8. Molecular Epidemiology, Evolution and Dispersal of the 167 Genus Flavivirus Paolo M de A. Zanotto and Ernie A. Gould 9. Molecular Epidemiology of Hepatitis C Virus 197 Alexandra Cochrane and Peter Simmonds 10. Seeking a Pale Horse: The 1918 Pandemic Influenza Virus 217 Thomas G. Fanning, Ann H. Reid, Thomas A. Janczewski and Jeffery K. Taubenberger 11. Molecular Epidemiology in Measles Control 237 Claude P. Muller and Mick N. Mulders 12. Public Health Surveillance and the Molecular Epidemiology 273 of Rabies James E. Childs, John W Krebs and Jean S. Smith 13. Molecular Epidemiology of Rotavirus 313 Christian Mittelholzer and Lennart Svensson 14. Respiratory Syncytial Virus 329 Patricia A. Cane 15. Molecular Epidemiology ofHantavirus Infections 351 Ake Lundkvist and Alexander Plyusnin 16. Molecular Epidemiology of Arenaviruses 385 Remi N. Charrel and Xavier de Lamballerie 17. Molecular Epidemiology of Hepatitis B Virus 405 Jane N. Zuckerman and Arie J. Zuckerman 18. Genomic Diversity of Human Papillomaviruses and its 419 Impact on Molecular Epidemiological Research Hans-Ulrich Bernard Index 439 Preface The field of molecular epidemiology spans many new sciences, including bioinformatics and molecular biology as well as established disciplines such as medicine and mathematics. This makes the use of molecular epidemiological methods both exciting and solid at the same time. In the last decade the use of molecular investigations in virology have avalanched, but a comprehensive overview has been lacking even though almost every journal in medicine, biology, mathematics, evolution, computer science and epidemiology can display papers on this topic. It is, therefore, the intension of this book to collect some central information from this ocean of knowledge and perhaps provide a springboard from which both those who are new to the field and the experts in one niche can jump off. This book contains eighteen chapters covering fundamental knowledge, the high-tech nuts and bolts in the researcher's toolbox, as well as examples from the majority of human viruses, a review of the fascinating recent history and a state of the art snap-shot. Molecular epidemiology means different things to different people. To some, it is an exiting opportunity to apply modem techniques, be it biochemical or computational, on almost anything, even viruses. For others, it means an objective and careful method to extract important information from a virus system that otherwise would be difficult, if not impossible, to retrieve. Then there are those who think molecular epidemiology is just another new fluke in the fast molecular biology and bioinformatics rush, and that it, at best, is another way of telling what they already knew. In addition, to define what is and is not molecular epidemiology can be rather difficult. But surely, with the increased speed at which DNA sequencing improves, computing powers escalate and new emerging viruses are discovered, molecular epidemiology is going to become even more important and interesting in the future. In this endeavor, I wish to sincerely thank all the authors who have contributed to this book. The expertise and active research of the contributing authors has given this book in-depth descriptions and updated views of a vast and diverse field. Their careful and extensive reviews have together created an impressive collection of knowledge on the molecular epidemiology of human viruses. I am also grateful to Laura Leitner for her indefatigable help in various phases of the manuscript preparations and to Joanne Tracy and Dianne Wuori at Kluwer for their encouragement and patience during this project. Thomas Leitner, May 2002 The Molecular Epidemiology of Human Viruses. 2002. Thomas Leitner, ed. Kluwer Academic Publishers, Boston. Chapter 1 The Use of Molecular Epidemiology Thomas Leitner Department of Virology Swedish Institute for Infectious Disease Control 17182 Solna, Sweden 1. INTRODUCTION The use of molecular methods to assess the spread of infectious diseases has become an important way to follow epidemiological patterns. While traditional epidemiology concentrates on the host, the molecular methods focus on the etiological agent, which in this book is the virus that causes the disease one wants to monitor. Naturally, many modern epidemiological studies include data from both hosts and pathogens, and a study without one of these two legs will suffer in its value. As we will see throughout this book, there are many advantages with the use of molecular methods. As an example of the strength of molecular methods, phylogenetic analyses of human immunodeficiency virus type 1 (HIV -1) have proven to be able to reconstruct transmission chains on the individual level to solve criminal investigations. Traditional epidemiology, where one expects all information given by the subjects to be true, could obviously not solve such situations. Other examples where molecular epidemiology excels include the situation when animals are involved and when the subjects are unavailable for questioning (e.g., Chapters 12 and 15). The definition of the term molecular epidemiology is somewhat unclear from the literature. In older literature the term refers to any epidemiological tracking done with any method that uses molecular analyses. Thus, this is a very wide definition. It will include the analysis of both inorganic chemicals and organic molecules such as sugars, proteins, antibodies and nucleic acids. In more modern literature the term, however, most often refers to the use of information derived from nucleic acids or proteins of the pathogen. In its more simple form the analysis can be a molecular typing of the organism that causes a certain disease. For instance, we may perform a 2 Leitner polymerase chain reaction (peR) to conclude that a recent outbreak somewhere was caused by a certain type of bacteria. In more advanced studies the DNA sequence of a gene fragment from the pathogen is analyzed by phylogenetic inference methods. For organisms that evolve more rapidly, like many viruses, this analysis will not only tell us which type is spreading but also how it is spreading. As mentioned previously, in the case of HIV- 1 we are able to link individuals to common sources of infection through analysis of DNA sequences from viral genes with phylogenetic reconstruction methods. 2. RECONSTRUCTION OF A TRUE TRANSMISSION HISTORY Phylogenetic reconstruction attempts to reconstruct the evolutionary history of the sequences under study. If the sequence of a certain gene from a virus is evolving during its spread from one human to another, then the reconstructed phylogeny should tell us how the virus was spread. There are a number of assumptions that have to be made to reconstruct a phylogeny as will be discussed later in this book, but the critical question is whether a transmission chain can or cannot be reconstructed by phylogenetics at all. To answer this question one can refer to many studies on the performance of various phylogenetic methods to reconstruct predefined phylogenetic trees (see for instance Nei, 1991; Hillis et at., 1994; Swofford et at., 1996). While these studies give important information on the powers and pitfalls of the methods, they still deal with hypothetical situations. To study real evolution is difficult for most biological systems because the evolutionary changes are so slow. In contrast, for many viruses we have the opportunity to follow evolution at real time due to high mutation rates. For HIV-l, which evolves at a rate of 6.7 ± 2.1 x 10-3 nucleotide substitutions site-1 year-1 in its hypervariable region V3 (Leitner and Albert, 1999) as compared to most mammal chromosomal genes that evolve at a rate of approximately 10-9, knowledge from a true transmission history has been evaluated by phylogenetic reconstruction. The Swedish transmission chain has shown that phylogenetic reconstruction of sequences from viral genes indeed can reconstruct epidemiological patterns in great detail (Leitner et at., 1996). Figure 1 shows the reconstructed phylogeny using env V3 region sequences of HIV-l compared to the true history. Encouragingly, the reconstruction succeeded to infer the true history at 94% accuracy. The only error in branching pattern that was made was in indicating the direction of the transmission between a mother and her child. All transmission interactions were, however, completely correct. The reason why the phylogenetic tree did not agree with the transmission history in this single case does not necessarily mean that the phylogenetic tree was wrong. Actually, both the tree and the history may be correct due to the The Use ofM olecular Epidemiology 3 fact that the variant that infected the child had gone extinct in the mother at the time point of sampling (Leitner and Fitch, 1999). A B f"I-.."..,.,,.,,....--- p10.9939 L.---:~!"I!!""_ p1 0.9939 r----------- p7.6760 L~rI====~'3i7----p6.6767 L.----l:..:.......:.:.._ p1.719 1980 1982 19841986 19881990 19921994 o 2 3 4 5 6 7 8 9 10 Year Genetic distance ("!o) Fig. 1. (A) The true transmission tree of the Swedish transmission chain. The index case, a Swedish male (pI) who became HIV-l infected in Haiti 1980, infected several females (p2, 5, 7, 8, and 11) between 1981 and 1983. In addition, samples from a later male sexual partner (p6) and two children (p3 and 9) of the females were included in the phylogeny. Blood samples were obtained at different time points between 1986 and 1993. From some individuals more than one sample was available. The information about when the transmissions had occurred (internal nodes) and when the samples were obtained (tips) was compiled into a tree that shows the evolutionary history of the transmitted virus populations. At each tip patient and sample number is indicated. (8) A reconstructed phylogenetic tree based on the HIV -1 env V3 region. The reconstruction was done using maximum likelihood analysis under a general-time-reversible nucleotide substitution model. The discrepancy between branch lengths of the phylogenetic tree and the transmission history is obvious (Fig. 1). Especially short branches tend to be estimated too long. Even though we use substitution models that are capable of giving a linear relationship between genetic distance and time, this discrepancy remains (Leitner and Albert, 1999; Leitner and Fitch, 1999). Once again the explanation lies in the fact that there is a large degree of variation in the virus within one infected person. The discrepancy can be explained by the available genetic variation at the time point of transmission, i.e. the effective population size, and the relationship of the transmitted variant in the recipient to the variants that continue to evolve in the donor. The genetic divergence that exists between these two virus populations is expressed by their pretransmission interval, a unit that will explain the discrepancy between branch lengths of a phylogenetic tree and a transmission history (Fig. 2). Other types of "time errors" caused by, for instance, uncertainties in sampling times and ancestral time points will require other measures such as generalized regression analyses and log likelihood ratio tests (Huelsenbeck and Rannala, 1997; Korber et al., 2000; Rambaut, 2000). The results from the Swedish transmission chain lend ample support to the use of phylogenetic reconstruction to infer epidemiological spread at a 4 Leitner very detailed level. As long as the researcher is careful to apply a fairly realistic model of evolution and enough sequence information, one can be pretty confident in the results when investigating any viral system. Evolutionary Time T1 T2 ) C C BC -LlBC- c Dc B B ABC D abc LlAB A A a I I tl t2 Transmission Time Fig. 2. The pretransmission interval (8) describes the difference between the time of transmission and the most recent common ancestor (MRCA) of the transmitted lineage and the donor lineage. The thin line tree is the transmission history and the bold line tree is the evolutionary history of the transmitted virus. At tl patient A infects patient B, and at t2 B infects C. The virus that infects B shares its MRCA with A at n, and the virus that infects C shares its MRCA with B at T2. The pretransmission interval when A infects B is tl - n = 8AB and t2 - T2 = 8B c when B infects C. Because of the pretransmission interval, the transmitted lineage and the donor lineage will be separated by a genetic distance at the time of transmission. 3. INFORMATION FLOW Investigating the spread of a human virus by molecular methods involves several steps. Naturally, we need epidemiological background information about the samples. Depending on the questions we want to answer we may be interested in geographical data relating to the sample, time point of sampling, disease stage of the patient, gender, age, risk group, anti viral treatment, etc. The type of material (serum, plasma, type of tissue, etc) and genetic regions (polymerase gene, envelope gene, etc) to be analyzed will be important for what type of answer we will get. The number of samples that we will be able to analyze will often depend on the cost of time involved in the laboratory method. Therefore, there will be a relationship between the number of samples and the method of analysis.