Neurobehavioral Genetics - Methods and Applications 2nd Ed - B. Jones, P. Mormede (CRC, 2007) WW
Neurobehavioral Genetics - Methods and Applications 2nd Ed - B. Jones, P. Mormede (CRC, 2007) WW
Neurobehavioral Genetics - Methods and Applications 2nd Ed - B. Jones, P. Mormede (CRC, 2007) WW
Neurobehavioral
Genetics
Methods and Applications
1903_Prelims.fm Page ii Friday, July 21, 2006 7:34 PM
Second Edition
Neurobehavioral
Genetics
Methods and Applications
Edited by
Byron C. Jones, Ph.D.
Department of Biobehavioral Health
The Pennsylvania State University
University Park, Pennsylvania
This book contains information obtained from authentic and highly regarded sources. Reprinted
material is quoted with permission, and sources are indicated. A wide variety of references are
listed. Reasonable efforts have been made to publish reliable data and information, but the author
and the publisher cannot assume responsibility for the validity of all materials or for the conse-
quences of their use.
No part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any
electronic, mechanical, or other means, now known or hereafter invented, including photocopying,
microfilming, and recording, or in any information storage or retrieval system, without written
permission from the publishers.
For permission to photocopy or use material electronically from this work, please access www.
copyright.com (http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC)
222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that
provides licenses and registration for a variety of users. For organizations that have been granted a
photocopy license by the CCC, a separate system of payment has been arranged.
Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and
are used only for identification and explanation without intent to infringe.
Visit the Taylor & Francis Web site at
http://www.taylorandfrancis.com
and the CRC Press Web site at
http://www.crcpress.com
1903_C000a.fm Page v Thursday, July 27, 2006 8:32 PM
Dedication
Between the publishing of our first and second editions, we lost two of our colleagues
in the Behavioral Neurogenetics Initiative, Donald J. Nash and Lorraine Flaherty.
Both Don and Lori had long, distinguished careers. Don was professor of biology at
Colorado State University and Lori was director of the Genomics Institute, chief of
the Laboratory of Mammalian Genomics, and director of the Histocompatibility and
Paternity Testing Laboratory, Wadsworth Center, New York State Department of
Health. Don was president of the Southwestern and Rocky Mountain Division of the
American Association for the Advancement of Science for many years. In 2005, Lori
received the first Lifetime Achievement Award from the International Behavioral and
Neurogenetics Society. Both Don and Lori were active participants in organizing and
teaching at our International Summer School on Behavioral Neurogenetics in Europe
and North America. In fact, Don was a founding member of the Behavioral Neuro-
genetics Initiative (BNI). Both had great and infectious joie de vivre. We even miss
Dons outrageous puns. Above all, was their passion for neurobehavioral genetics
and bringing this emerging discipline to students and researchers worldwide. The
efforts of the BNI have been greatly enriched by the contributions of Lori and Don.
In Memoriam
Gilbert Gottlieb
19292006
Dr. Gilbert Gottlieb was born in Brooklyn, NY in 1929. He earned his Bachelors
(1955) and Masters (1956) degrees from the University of Miami, and his Ph.D. in
clinical psychology at Duke University in 1960. This was no ordinary clinical
psychology degree because his dissertation was entitled, The Following Response
of Wild and Domestic Ducklings of the Same Species (Anas platyrhynchos). Thus
began an amazing career that spanned several interrelated disciplines (e.g., psychology,
zoology, embryology, genetics), the common thread being developmental science
and a unique theoretical perspective that he referred to as simply, the developmental
point of view. Instead of reducing development to a simple interaction of the
organism and its environment, Dr. Gottliebs view embraced the complexity of organ-
ismic development by considering the intricate transactional nature of events that
occur among all levels of organization (genetic, neural, behavioral, environmental)
at every point in time, such that the organism is a new organism throughout its
transactional developmental trajectory.
This view of development emerged from his early career as a research scientist
at the North Carolina Division of Mental Health (Raleigh, NC), where, in 1961, he
established the Psychology Laboratory at Dorothea Dix Hospital and pursued both
field and laboratory investigations on the acoustic basis of species identification in
ducklings. Obtaining baseline data in naturalistic contexts on the kinds of stimuli
that organisms normally encounter, as well as the behaviors in which they typically
engage, provided a foundation for his elegant and innovative experimental analyses
of developmental mechanisms. These studies, which continued for over three decades,
revealed the importance of embryonic (as well as early postnatal) experience on the
development of species-typical behavior. The empirical and conceptual foundations
of this body of work culminated in Dr. Gottliebs volume, Synthesizing Nature-Nurture
(1997, Erlbaum), for which he won the Eleanor Maccoby Book Award of the Devel-
opmental Psychology Division of the American Psychological Association.
In 1982, Dr. Gottlieb became the head of the psychology department at the
University of North Carolina at Greensboro, where he also held an Excellence
Foundation Professorship until his retirement from academia in 1995. That retire-
ment, however, was not a retirement from developmental science, because in that
year, he became a research professor and member of the Executive Committee of
the Center for Developmental Science at the University of North Carolina at Chapel
Hill. In this stimulating multidisciplinary environment, he was able to further develop
his theoretical perspectives on the transactional nature of genetic and experiential
influences on behavioral development.
Throughout his distinguished career, Dr. Gottlieb received recognition and honors
from his peers. He was the guest of the Czechoslovak Academy of Sciences in Prague
1903_C000a.fm Page viii Thursday, July 27, 2006 8:32 PM
(1967) and the USSR Academy of Sciences in Moscow (1989), as well as a Visiting
Fellow at The Neurosciences Institute in San Diego (1996). He was also elected
Fellow of the Animal Behavior Society and the American Association for the Advance-
ment of Science; and in 1997, the Society for Research in Child Development honored
him with the Distinguished Scientific Contributions to Child Development Award. He
was supported from 1962 to 2006 by grants from the National Institute of Child
Health and Human Development, the National Science Foundation, and the National
Institute of Mental Health, all of which enabled him to pursue research questions
and theoretical approaches that he recently described to me as off-the-beaten-track.
But he maintained that it is important for scientists to follow their noses when so
inspired because at the end of the trail, discoveries will be made.
Dr. Gilbert Gottlieb died on July 13, 2006. Although he did not live to see the
publication of this volume, in a taped interview that I conducted with him at his
home in Raleigh 1 month before his death, he proclaimed that the chapter contained
herein is his most important theoretical contribution to science.
David B. Miller
Department of Psychology
University of Connecticut
Gilbert Gottlieb
19292006
Photo Credit: Dr. Marc S. Gottlieb
1903_C000.fm Page ix Wednesday, July 26, 2006 7:58 PM
Preface
This second edition of Neurobehavioral Genetics: Methods and Applications is
issued about 7 years after the first edition. Since its appearance in 1999, the book
has been used as text material for the International Behavioral Neurogenetics Sum-
mer School (formerly the FrenchAmerican Summer School). In September 2005,
we held our 12th school in Moscow. This is particularly remarkable because the
initial funding, provided by the Scientific Mission of the French Embassy to the
U.S. in 1995, was the spark that launched this effort. We will always be thankful to
Dr. Jean-Marie Guastavino, attach for Science and Technology at that time, for
arranging for the generous gift from the French government. An alumna of our third
school, Dr. Yamima Osher, is author of the chapter on genegene interactions.
When we issued the first edition, we knew that the field was developing rapidly,
and, in fact, the advances in methods and our knowledge have outstripped even our
most optimistic expectations. Methods to examine gene expression, new approaches
in bioinformatics, and the rapid growth in knowledge that has followed are remark-
able. Major developments in the field of behavioral neurogenetics are the foundation
of the Complex Traits Consortium (CTC) and the equally important International
Behavioural and Neural Genetics Society (IBANGS). These new societies have
provided much needed forums for examination of brain and behavior from a systems
genetics perspective. There have been remarkable advances also in single-gene
techniques. This second edition has been produced to highlight some of these
important advances. We are particularly pleased to include a contribution by Dr.
Gilbert Gottlieb, who has been an important and tough critic of behavioral genetics.
Fortunately, he has always been kind to behavior geneticists!
Special thanks to our managing editor, Barbara Norwitz, at Taylor & Francis,
for her encouragement and guidance.
Editors
Contributors
Irmgard Amrein Valerie Bolivar
Institute of Anatomy Wadsworth Center
University of Zrich-Irchel Troy, New York
Zrich, Switzerland
Guntram Borck
Rachel Bachner-Melman Institut National de la Sant et de la
Sarah Herzog Memorial Hospital, Givat Recherche Mdicale (INSERM) U781
Shaul Hpital Necker-Enfants Malades
Jerusalem, Israel Paris, France
Andrew Canastar
Amsale T. Belay
Department of Psychiatry
Department of Biology
University of Colorado Health Sciences
University of Toronto
Center
Mississauga, Canada
Denver, Colorado
Florence Molinari
Andr Ramos
INSERM
Departmento de Biologia Celular,
Hpital Necker-Enfants Malades
Embriologia e Gentica
Paris, France
Universidade Federal de Santa
Catarina
Pierre Mormde
Florianpolis, Brazil
Laboratoire de Neurogntique
et Stress INRA
Universit Victor Segalen Nicolas Ramoz
Bordeaux, France INSERM
Paris, France
Tracy L. Nelson
Department of Health and Exercise Laurent Sgalat
Science Centre de Gntique Molculaire et
Colorado State University Cellulaire CNRS
Longmont, Colorado Universit Lyon 1 Claude Bernard
Villeurbanne, France
Yamima Osher
Beer Sheva Mental Health Center Lutz Slomianka
Ben Gurion University of the Negev Institute of Anatomy
Beer Sheva, Israel University of Zrich-Irchel
Zrich, Switzerland
Jeremy L. Peirce
Department of Anatomy and Marla B. Sokolowski
Neurobiology Department of Biology
University of Tennessee University of Toronto
Memphis, Tennessee Mississauga, Canada
1903_C000.fm Page xvi Wednesday, July 26, 2006 7:58 PM
Table of Contents
Chapter 1
A History of Behavior Genetics................................................................................ 1
Stephen C. Maxson
Chapter 2
Developmental Neurobehavioral Genetics: Development as Explanation ............. 17
Gilbert Gottlieb
Chapter 3
Some Basics, Mendelian Traits, Polygenic Traits, Complex Traits....................... 29
Byron C. Jones
Chapter 4
An Introduction to Quantitative Genetics ............................................................... 37
Wim E. Crusio
Chapter 5
From QTL Detection to Gene Identification .......................................................... 55
Marie-Pierre Moisan
Chapter 6
Gene Expression ...................................................................................................... 73
Richard A. Radcliffe
Chapter 7
Bioinformatics of Behavior ..................................................................................... 95
Elissa J. Chesler
Chapter 8
Congenic and Consomic Strains ........................................................................... 115
Lorraine Flaherty and Valerie Bolivar
Chapter 9
Animal Resources in Behavioral Neurogenetics .................................................. 129
Jean-Michel Lassalle
Chapter 10
Sample Size Requirements for Experiments on Laboratory Animals ................. 149
Douglas Wahlsten
1903_bookTOC.fm Page xviii Wednesday, July 26, 2006 8:05 PM
Chapter 11
The Role of Association Studies in Psychiatric Disorders................................... 169
Nicolas Ramoz and Philip Gorwood
Chapter 12
Family and Twin Methods..................................................................................... 183
Keith E. Whitfield and Tracy L. Nelson
Chapter 13
GeneEnvironment Interactions ............................................................................ 189
Byron C. Jones and Leslie C. Jones
Chapter 14
And Now it Starts to Get Interesting: GeneGene Interactions........................... 201
Yamima Osher
Chapter 15
Schizophrenia: Study of a Genetically Complex Phenotype ................................ 209
Michael F. Pogue-Geile and Irving I. Gottesman
Chapter 16
Genetics of Major Affective Disorders ................................................................. 227
Wade Berrettini
Chapter 17
Pedigree Analyses and the Study of Chimpanzee
(Pan troglodytes) Personality and Subjective Well-Being.................................... 247
Alexander Weiss and James E. King
Chapter 18
The Elusive World of Personality Genes: Cherchez le Phenotype ...................... 263
Richard P. Ebstein, Rachel Bachner-Melman,
Jonathan Benjamin, and Robert H. Belmaker
Chapter 19
Aggression: Concepts and Methods Relevant to
Genetic Analyses in Mice and Humans................................................................ 281
Stephen C. Maxson and Andrew Canastar
Chapter 20
Genetic Analysis of Emotional Behaviors
Using Animal Models ........................................................................................... 291
Andr Ramos and Pierre Mormde
Chapter 21
Genetic Analysis of Food Search Behavior in the Fruit Fly (Drosophila
melanogaster) ........................................................................................................ 307
Amsale T. Belay and Marla B. Sokolowski
1903_bookTOC.fm Page xix Wednesday, July 26, 2006 8:05 PM
Chapter 22
Genetic and Molecular Analyses of Drosophila
Courtship Behavior................................................................................................ 319
Jean-Marc Jallon
Chapter 23
A New Era for Drosophila
Learning and Memory Studies .............................................................................. 333
Daniel Comas, Guillaume Isabel, and Thomas Prat
Chapter 24
Behavioral Genetics in the Nematode
Caenorhabditis elegans ......................................................................................... 353
Gert Jansen and Laurent Sgalat
Chapter 25
Genetics, Behavior, and Brain Dopamine Systems .............................................. 371
Robert Hitzemann, Shannon McWeeney, and John Belknap
Chapter 26
Natural Genetic Variation of Hippocampal
Structures and Behavioran Update.................................................................... 389
Hans-Peter Lipp, Irmgard Amrein, Lutz Slomianka, and David P. Wolfer
Chapter 27
Expression and Brain Structure:
Black Boxes between Genes and Behaviors......................................................... 411
Jeremy L. Peirce and Robert W. Williams
Chapter 28
Synaptic Mechanisms Involved in Cognitive Function: Cues from Mental
Retardation Genes.................................................................................................. 435
Guntram Borck, Florence Molinari, Birgit Dreier,
Peter Sonderegger, and Laurence Colleaux
Chapter 29
Pharmacogenetics .................................................................................................. 449
Byron C. Jones
Chapter 30
Alcohol Psychopharmacogenetics ......................................................................... 457
John C. Crabbe
Index...................................................................................................................... 469
1903_bookTOC.fm Page xx Wednesday, July 26, 2006 8:05 PM
1903_C001.fm Page 1 Wednesday, July 26, 2006 6:08 PM
1 A History of Behavior
Genetics
Stephen C. Maxson
CONTENTS
Summary .................................................................................................................... 2
1.1 Introduction ...................................................................................................... 2
1.2 Preliterate and Ancient History ....................................................................... 3
1.3 Two Victorian Cousins..................................................................................... 3
1.4 The First Century of Behavior Genetics ......................................................... 4
1.4.1 1910 to 1952 ........................................................................................ 4
1.4.1.1 Fruit Flies.............................................................................. 4
1.4.1.2 Rodents ................................................................................. 4
1.4.1.3 Humans ................................................................................. 5
1.4.2 1953 to 1970 ........................................................................................ 5
1.4.2.1 Fruit Flies.............................................................................. 6
1.4.2.2 Rodents ................................................................................. 7
1.4.2.3 Dogs ...................................................................................... 7
1.4.2.4 Humans ................................................................................. 7
1.4.3 1971 to 1985 ........................................................................................ 8
1.4.3.1 Nematodes............................................................................. 8
1.4.3.2 Fruit Flies.............................................................................. 8
1.4.3.3 Rodents ................................................................................. 8
1.4.3.4 Canids ................................................................................. 10
1.4.3.5 Humans ............................................................................... 10
1.4.4 1986 to 2005 ...................................................................................... 10
1.4.4.1 Nematodes........................................................................... 10
1.4.4.2 Fruit Flies............................................................................ 11
1.4.4.3 Bees..................................................................................... 11
1.4.4.4 Rodents ............................................................................... 11
1.4.4.5 Humans ............................................................................... 12
1.5 Janus: The Past and Future of Behavior Genetics ........................................ 13
1.6 A Highly Personal Note................................................................................. 14
References................................................................................................................ 14
1
1903_C001.fm Page 2 Wednesday, July 26, 2006 6:08 PM
2 Neurobehavioral Genetics
SUMMARY
This is a broad brush and highly selective history of behavior genetics. I begin it
long ago in prehistoric and ancient times to indicate that even then there was
knowledge of and speculation about the inheritance of behavior in animals and
humans. I then turn to the contributions of two Victorian cousins, Charles Darwin
and Francis Galton. They set the stage for the first century of behavior genetics. I
divide the history of first century of behavior genetics into four phases: 1910 to
1952, 1953 to 1970, 1971 to 1985, 1986 to 2005. Within each, I consider the history
of behavior genetics by key organisms: nematodes, fruit flies, rodents, other critters,
and humans. Throughout, individual contributors to behavior genetics are mentioned.
I conclude with some reflections on past conceptual streams of behavior genetics,
and on how these streams can and should come together in a comparative genetics
of behavior.
1.1 INTRODUCTION
The history of behavior genetics begins long before Darwins work on evolution or
Mendels work on inheritance. It ends in the very recent past. I hope that this history
will be of value in: (1) recognizing who the many contributors to the field were and
are, (2) tracing the origins and developments of conceptual themes in behavior
genetics, and (3) relating how new methods have contributed again and again to
raising questions and enabling answers to them in behavior genetics.
suggested that well-bred parents have well-bred offspring and that poorly bred
parents have poorly bred offspring. Similarly, Plato (427346 B.C.E.) argued in
The Republic that matings of the best man to the best woman would produce
the best children and that matings of the worst man to the worst woman would
produce the worst children. For this reason, both philosophers proposed that
humans, like animals, should be selectively bredwith only the well bred or
the best having children. This was eugenics long before there was a science of
genetics.
There was also an interest in the ancient world in whether the development
of behavior was innate or acquired. For example, Aristotle (384322 B.C.E.) in
Greece argued that the brood parasitism of the cuckoo did not depend on post-
hatching experience. Similarly, Galen (129200 C.E.), a Roman physician, pro-
posed, based on an experiment, that the preference of newborn goats for mothers
milk did not depend on postnatal experience. Some of the ancient arguments for
eugenics may have depended on the belief that if a behavior is inherited, its
development is independent of experience. This mistaken idea, which confounds
inheritance and innateness, has been a persistent problem in behavior genetics,
even until today.
We now leap over almost 2000 years of history to the beginning of a science of
behavior genetics.
4 Neurobehavioral Genetics
There was a great interest in looking for single-gene effects on fly behavior. This
includes the studies of George Wald and Werner Reichard on visual acuity, of Paul
Scott on phototropism, of R. S. McEwen on geotaxis and phototaxis, and of
A. Sturtevant on mate preference and choice. The objects of study, in many cases,
were genes with visible morphological effects on flies, including effects on eye color
and shape, body color, and bristle number. There were also studies by H. T. Spieth
and Dobzhansky on genetics, mating success, sexual isolation, and evolution. These
were both within- and between-species studies.
1.4.1.2 Rodents17
In the 1930s and 1940s, R. C. Tryon reported on the selective breeding of two strains
of rats that differed in maze learning as indexed by trial errors. The two strains or
lines were the maze brights, with few errors, and the maze dulls, with many
errors. Also, C. S. Hall reported in 1938 on the selective breeding of two lines of
rats that differed in emotionality as measured by open-field activity and defecation.
Subsequently, selective breeding would be widely used to study the genetics of rat
and mouse behaviors.
In 1942, two papers were published showing differences in aggressive behavior
among three inbred strains of mice. One was published by John Paul Scott, and the
other by Benson Ginsburg and W. C. Allele. In the 1940s, another focus of strain
difference in mice was sound-induced, or audiogenic, seizures. Subsequently, there
have been a great many studies of strain differences in mouse behaviors. Such strain
differences are considered to be evidence for genetic effects on phenotypic variation.
Not just mice but rabbits and dogs too were among the organisms studied at Hamilton
Station for Behavioral Research, established in 1946 at the Jackson Laboratory. Its
1903_C001.fm Page 5 Wednesday, July 26, 2006 6:08 PM
full-time or summer staff included John Paul Scott, John Fuller, Benson Ginsburg,
Betty Beaman, Calvin Hall, and many others.
For the maze dull and maze bright rats, there were also genotype-by-environment
interactions observed for the maze errors. In enriched environments, the error scores
of the maze dulls but not the maze brights were decreased, whereas in impoverished
environments, the error scores of the maze brights but not the maze dulls were
increased.
A genotype-by-environment interaction was also observed for aggression in
mice. Scotts most aggressive strain was Ginsburg and Alleles most pacific strain
C57BL/10. When mice of the C57BL/10 strain were transferred from cage to cage
by forceps, as Scott did, they were aggressive; and when mice of the C57BL/10
were transferred from cage to cage in a small box, as Ginsburg did, they were pacific.
These treatments had no effect on the aggression of the other two strains. Such
genotype-by-environment interactions would be found again and again for rodent
behaviors and would be an active subject of research up to the present day (see
Chapter 20 in this volume).
In 1951, Calvin S. Hall13 reviewed the literature on the genetics of rat and mouse
behaviors and proposed there were three goals for behavior genetics research: (1)
to determine whether individual differences in behavior were heritable and to what
extent, (2) to determine for each heritable behavior the number of variant genes
involved and the chromosomal location of each, and (3) to determine how each of
these genes affects the heritable behaviors. These are still among the goals of
behavior genetics in animals and humans.
1.4.1.3 Humans12
Family and twin studies were conducted for IQ (H. D. Carter, H. H. Neuman, F. N.
Freeman, and K. J. Holzinger), personality (M. N. Crook, H. D. Carter, and
L. Portenier), schizophrenia (F. J. Kallman), and depression (E. Slater). There were
also some adoption studies. The emphasis was on ascertaining whether individual
differences were due to genes, environment, or both. Although many studies were
consistent with effects of genes on trait variation, definite conclusions were limited
by sample size.
Pedigree analysis for single genes with large effects was also developed at this
time. Using pedigrees, phenylketonuria was shown by L. Penrose in 1933 to be a
single gene recessive trait. It was believed that this was an inborn error of metabolism
similar to alkaptonuria as described by A. E. Garrod in 1909.
6 Neurobehavioral Genetics
now set for Halls third goal for behavior genetics, namely, identifying and charac-
terizing the effect of DNA and DNA variants on behavior.
In 1958, Benson Ginsburg published Genes as a Tool in the Study of Behavior.11
The first part of this paper was a reflection on the implications of molecular genetics
for the study of behavior, and the second part argued that genetics should be used
as a tool to analyze animal and human behavior. It could be used as a tool in an
evolutionary context to determine natural units of behavior, to study the development
of behavior and the respective roles of genes and environment, and to analyze the
biological mechanisms of behaviors. This approach would become one of the main
tributaries of modern behavior genetics.
The other main tributary of modern behavior genetics focused on the contribution
of genetic and environmental factors in individual differences in animal and human
behaviors and on identifying and characterizing each of the genes with effects on
these individual differences. Also, a prime interest of the research with animals was
to develop genetic models relevant to the study of individual differences in human
behaviors. In 1960, the literature on this tributary of behavior genetic research on
human and animals was reviewed and summarized in the text by J. L. Fuller and R.
Thompson. The text was simply titled, Behavior Genetics.12 Developments in quan-
titative genetics as described in D. S. Falconers Introduction to Quantitative
Genetics11 and in K. Mathers Biometrical Genetics24 were a basis for further
advances in this area of behavior genetics.
In 1961 and 1962, there were important meetings on behavior genetics at the
Center for Advanced Studies in the Behavioral Sciences at Stanford CA. These were
attended by many leaders in the field and in many ways set the agendas for the next
decade of behavior genetics research. Some of the presentations at these meetings
were published in 1967.20
During this period, there were several selective breeding projects for behaviors in
fruit flies. Jerry Hirsch and Niki Erlenmeyer-Kimling initiated bi-directional selec-
tion for geotaxis. Selective breeding of these lines has continued to the present day.
Now, crosses of the two lines cannot have fertile progeny. Hirsch and Erlenmeyer-
Kimling also tested for the contributions of individual chromosome regions to the
high and low scores. Similarly, Dobzhansky selectively bred for geotaxis and pho-
totaxis in another species of fruit fly. He reported that very little of the individual
differences in these traits was due to genetic variation. About the same time, Aubrey
Manning carried out a bi-directional study of mating speed.
During this period, Lee Ehrman initiated her studies of rare male mating advan-
tage in several species of fruit flies. Also, Speiss, Ehrman and Dobzhansky among
others continued the research on mating preference and on sexual isolation within
and between species.
About 1967 Seymour Benzer first proposed that single gene mutants with large
effects could be used to dissect the neural and other mechanisms of behaviors in
flies. Later Sydney Brenner made a similar suggestion for the nematode, C. elegans.
These contributions are described more fully in the next period (1970 to 1986) of
1903_C001.fm Page 7 Wednesday, July 26, 2006 6:08 PM
behavior genetics. In some ways, this was to be a large substream of the tributary
that would use genetics as a tool to analyze behavior as originally proposed by
Ginsburg in 1958.
1.4.2.2 Rodents
During this period selective breeding was used to establish new phenotypic lines of
rats and mice. These included the Maudsley Emotional and Non-emotional strains
(Peter Broadhurst) and the Roman High and Low Avoidance strains (C. Bignami)
in rats. Also, mice were selectively bred for high and low activity in an open field
(John Defries) and aggression (Geert van Oortmerssen and K. Lagerspetz).
About this time, Jan Bruel suggested in several articles that adaptive traits of
mice would have low additive genetic variation and high directional dominance
variation. This set the stage for research with diallel F1 crosses and F2 hybrids of
inbred strains to determine the additive and dominance variance of many mouse
behaviors. This program of research will be further discussed in the next period
(19711985) of behavior genetics.
Now strain differences were also being used to investigate the neural basis of
behaviors. For example, M. R. Rosenzweig, D. Krech, and E. L. Bennett had shown
maze dull and maze bright rats differed in brain cholinesterase levels. Subsequently,
Thomas Roderick selectively bred for cholinesterase levels in the brains of rats and
showed that these strains differed in learning performance.
Also, at this time, research was begun on strain differences in consumption and
effects of alcohol (Gerald McClearn). Research on genetics, alcohol, other drugs,
and behavior would become a major area of mouse behavior genetics.
1.4.2.3 Dogs
In 1965, John Paul Scott and John L. Fuller published the book, The Genetics and
Social Behavior of the Dog.34 This book detailed the findings of nearly 20 years of
research at the Jackson Laboratory on the roles of genes and environment in the
behavior of five breeds of dogs and their derived crosses. This work remains a
critically acclaimed classic on dog behavior and on methods for genetic analysis of
social behaviors.
1.4.2.4 Humans
During this period, there were more extensive adoption studies of IQ (M. Skodak
and H. M. Skeels, M. Honzig, and J. Sheilds) and schizophrenia (L. Heston,
D. Rosenthal, S. S. Kety). About this time, the large Louisville Twin Study was used
to assess the contribution of genes and environment to variation in many traits (S.
Vandenberg).
Also, during this time, many single genes with large effects on mental retardation
were identified, and the behavioral effects of many chromosomal anomalies were
also identified. This included the chromosomal basis for Downs and Turners syn-
dromes. Development of staining techniques enabled the identification of each
human chromosome.
1903_C001.fm Page 8 Wednesday, July 26, 2006 6:08 PM
8 Neurobehavioral Genetics
1.4.3.1 Nematodes3
The research on nematodes began in 1973. The nematode has several advantages
for neurobehavioral genetics. These are a small genome (100 Mb), small nervous
system of 302 identified and mapped neurons, fast reproductive cycle (three days),
and self-fertilization. The major approach has been to expose nematodes to chemical
mutagens, to select for single gene mutants with behavioral effects and then to
characterize the DNA of the mutant gene. Initially mutants were detected that showed
sensory and motor effects.
1.4.3.3 Rodents40
ones previously described. Others were new. These later included in rats the Syracuse
high and low avoidance lines (Bob Brush). Also, David Blizard and others continued
the study of the Maudsley Emotional and Non-emotional rats. Blizard also studied
behaviors of mice selected for thyroid function. In mice, Carol Lynch selected for
lines differing in the size of thermal nests and related her work to the ecology and
evolution of mice. Also, Bob Cairns and Kathryn Hood selected mice for male
aggression, and Janet Hyde and Patricia Ebert selected mice for female aggression.
There was also selection in rats and mice for effects of alcohol and drugs on behavior
such as the post-alcohol injection long sleep and short sleep lines. Much of this work
in mice was done by John Crabbe, Andrew Smolen, Toni Smolen, Thomas Johnson,
Bruce Dudek, as well as many others. Another study selected for high and low brain
weights and related these to mouse behaviors (John Fuller and Martin Hahn).
During this period investigators continued to using F1 diallels and F2 dihybrids
to assess the adaptive significance of behaviors. The behaviors studied included
ultrasounds, open field activity, aggression, mating, and learning. Contributors
included Norman Henderson, John Hewitt, Martin Hahn, Tom McGill, and many
others.
There was also an interest in finding the individual genes and chromosomes that
contributed to variation in mouse behaviors. One approach by Del Theissen and
colleagues was to see whether coat color and morphological variants had effects on
behavior. They did, but the mechanisms for the effects seemed trivial. More relevant
was research described by D. L. Coleman on two spontaneous mutants with effects
on feeding, satiety, metabolism, and obesity. These are the Obese and Diabetes
mutant mouse lines. Another approach was to try to use the newly developed
recombinant inbred strains to chromosomally map genes with behavioral effects
(Basil Eleftheriou and Thomas Seyfried). Because there were not enough chromo-
somal markers and too few recombinant inbred strains, these initial attempts to map
mouse variants with behavioral effects were often not replicated. However, Benson
Ginsburg, Stephen Maxson, and their doctoral students successfully used reciprocal
F1s and congenic strains to show that the male specific part of the Y chromosome
had effects on aggression and mating. The findings on the Y chromosome have been
replicated many times (Pierre Roubertoux, Michele Carlier, Frans Sluyter, and Geert
van Oortmerssen).
Also, Michele Carlier, Pierre Roubertoux, and their colleagues also began a
long-term research program on the interactions of maternal factors and genotype in
the development of mouse behaviors.
During this period there were also programs focusing on pharmacology and
neuroanatomy to find the connections between genes and behaviors. For example,
Kurt Schlesinger and colleagues used drugs to manipulate neurotransmitter systems
in relation to audiogenic seizures, and Richard and Cynthia Weimer measured strain
differences in cell number and other morphological parameters of the hippocampus
in relation to learning and memory. Douglas Whalsten also explored strain differ-
ences in corpus colossal morphology in relation to behavior. There were also studies
on strain difference in the endocrine systems and behavior such as those by Thomas
McGill on male mating, by Jack Vale on male aggression, and by Bruce Svare on
female aggression.
1903_C001.fm Page 10 Wednesday, July 26, 2006 6:08 PM
10 Neurobehavioral Genetics
1.4.3.4 Canids
In 1970, a study of the evolution and genetics of canids was established at the
University of Chicago by Benson Ginsburg. This included research on the social
behavior and organization of a pack of wolves and on the inheritance of threat
behaviors in dogs (beagles) and coyotes.
1.4.3.5 Humans19
1.4.4.1 Nematodes2
The genome (100 Mb) of C. elegans was sequenced in 1998. During this period,
there was continued screening for mutants with effects on olfactory and touch
sensory systems, chemotaxis, thermotaxis, biological rhythms, social behavior,
learning and memory and behavioral effects of alcohol, cocaine, and nicotine. Many
genes with effects on these systems were identified and characterized. In many cases,
1903_C001.fm Page 11 Wednesday, July 26, 2006 6:08 PM
the effects of the gene can be or are being traced from DNA to protein to neural
organization and function to behavior.
The genome (160 Mb) of the fruit fly was sequenced in the year 2000. During this
period single gene mutants were used to dissect the molecular and neural basis on
the circadian clock (Jeff Hall), mating (Jeff Hall, Charalambos P. Kyriacou, and
Jean-Marc Jallon), learning and memory (Tim Tully), and effects of alcohol and
drugs on behavior (F. E. Wolf and U. Huberlein). There was also an interest in
natural genetic variation in fly behavior. One example of this is Marla Sokolowskis
studies of the forager gene. Another example is the identification of the genes
involved in Hirschs long-term selection for the high and low geotaxis lines (Ralph
Greenspan). This gene identification used a combination of gene expression chips
and mutants. Greenspan has also investigated and reviewed the complex interactions
of genes with effects on fruit fly behaviors.
1.4.4.3 Bees33
The genome (200 Mb) of honeybees is being sequenced. Recent research (Gene
Robinson and others) on honeybees provides an excellent example of the association
between behavioral development and changes in gene expression and neural struc-
ture. Worker bees engage in nursing behavior from birth to about 2 to 3 weeks of
age. After that they engage in foraging behavior. This behavioral change is associated
with the increase or decrease in expression of many genes in the brain. Two of these
genes are period and forager, which were first found in fruit flies.
1.4.4.4 Rodents7,16,37,38
The mouse genome (3500 Mb) was sequenced in the year 2002. During this period,
the main focus of research has been to identify the genes that cause variation in
behavior. There have been four approaches for this. The first is to chromosomally
map quantitative trait loci (QTLs). One such study used crosses of the lines selected
for open field activity (Jonathan Flint and others.) Another approach used crosses
of inbred strains to find QTLs for aggression (Pierre Roubertoux and others). Some
studies combined recombinant inbred strains and strain crosses to find QTLs for
effects of alcohol and other drugs (John Crabbe, John Belknap, Keri Buck, Thomas
Johnson, Byron Jones, and others). In 1991, John C. Crabbe and R. Adron Harris
edited The Genetic Basis of Alcohol and Drug Actions.6 There has also been devel-
opment of large, recombinant, inbred strain sets that have been used to find QTLS
for brain structures and expressions of genes in mouse brains (Robert Williams and
others). Another approach has used specific gene knockouts and/or transgenics
(Crawley). This has been used with many behaviors including sensation, motor skills,
emotionality (Rene Hen), feeding (C. H. Vaughn and many others), aggression
(Randy Nelson), mating (Sonoko Ogawa and Emilie Rissman), learning and memory
(J. Z. Tsien and Jeanne Wehner), and effects of alcohol and drugs (Tamara Philips,
1903_C001.fm Page 12 Wednesday, July 26, 2006 6:08 PM
12 Neurobehavioral Genetics
Kari Buck, and many others). Also, there are now tissue or temporal specific
knockouts and transgenics.
A related approach is based on spontaneous or induced mutations. As previously
described, Obese and Diabetes are two spontaneous mutations with effects on feed-
ing, satiety and metabolism. Recently, it was shown that the obese mouse has a
coding mutation for the polypeptide hormone leptin, and the diabetic mouse has a
coding mutation in a leptin receptor (J. M. Friedman and others). The study of these
two spontaneous mutants detailing the pathway form DNA to behavior opened wide
the door to research on the hormonal and neural mechanisms of feeding and satiety.
Also, male mice have been exposed to chemical mutagen and the F1 or F2 progeny
of these mice screened to detect mutations with neural or behavioral effects. One
of the first induced mouse mutants to be detected was Clock which affects circadian
rhythms (Joseph Takahashi and others). Other genes with effects on mouse circadian
rhythms were found to have fruit fly homologs, such as Period and Timeless. The
study of the Clock mutant and variants for Period and Timeless from DNA to behavior
opened up further research on the neural mechanism of the circadian clock in
mammals. There are now centers for mouse mutagenesis and behavior at the Jackson
Laboratory, Northwestern University, and the University of Tennessee Health Sci-
ences Center. Each has high throughput screens for a range of mouse neural and
behavioral traits.
Another approach is to look for differences in gene expression among selected
strains of mice. This was done for hippocampal gene expression in SAL (short attack
latency) and LAL (long attack latency) strains (D. E. Feldker and others).
Other research in the period has used inbred or selected strain-based genetic
correlation to relate brain and behavioral phenotypes. One example is the study of
variation in the size of the hippocampal mossy fibers and related cognitive or
emotional behaviors (Wim E. Crusio and Hans-Peter Lipp). Another interest has
been in the standardization of behavioral tests (Norman Henderson, Martin Hahn,
and Stephen Maxson) and phenotypic stability across laboratories (John Crabbe,
Douglas Wahlsten, and Bruce Dudek). Furthermore, during this period mouse models
for Turners, Klinefelters, and Downs syndromes were developed.
One series of fascinating studies has related the genetics of species differences
in the behavior of voles focused on oxytocin and vasopressin systems and aggression,
parenting and monogamy (Larry J. Young).
1.4.4.5 Humans1,5,28,29
The human genome (3500 Mb) was sequenced in the year 2000. During this period
path analysis and model fitting were fully developed to analyze the components of
individual differences in behavior (David Fulker, John DeFries, Lindon Eaves,
Nicholas Martin, Stacy Cherny, and many others). This approach allowed the com-
bined used of family, twin and adoption studies. It was widely applied to cognition
(Robert Plomin, Dorett Boomsma, and Sandra Scarr), personality (John Loehlin),
personality disorders (Richard Rose, Irwin Waldman, Laura Baker, and Liz DiLalla),
sexual orientation (Michael Bailey and Dean Hamer), psychopathology (Peter
McGuffin and Irving Gottesman), and addictions (C. R. Cloninger, Kenneth Kendler,
1903_C001.fm Page 13 Wednesday, July 26, 2006 6:08 PM
Matt McGue and Andrew Heath). There was also an interest in the relative roles of
within family environments, between family environments, genotypeenvironment
correlations and genotypeenvironment interactions (Robert Plomin and David
Rowe) as applied to the same range of human behaviors.
A goal of this period was also to chromosomally map the genes that are involved
in the individual differences for a behavioral domain such as cognition. The two
approaches to this are linkage analysis and association analysis (Eric Lander and
Pak Sham).
The former often uses the affected sib pair method and the latter often uses the
transmission disequilibrium test. Both methods make use of nongenic DNA markers
and association analysis uses not only nongenic DNA variants, but also DNA variants
of known genes. The main problem to date has been replication of findings. For this
reason, meta-analysis across studies is often used.
Animal models may also be used in the search for genes with behavioral effects
in humans. Animal models can indicate chromosome regions or genes to focus on.
14 Neurobehavioral Genetics
taxa. This now includes not only nematodes, flies, mice and humans but also jellyfish,
bees, zebrafish, chickens, pigs, cattle, dogs, cats, monkeys, chimpanzee, and others.
Such a comparative genetics should also be gene centered comparing the behavioral
effects of homologous genes across species. For example, variants of the gene for
MAOA (monoamine oxidase A) have been shown to have similar effects on aggres-
sion in mice, monkeys and humans and depend for their effects on specific aspects
of the social environment during development.
REFERENCES
1. Benjamin, J., Ebstein, R. P., and Belmaker, R. H. Molecular Genetics and Human
Personality. American Psychiatric Publishing Inc., Washington, DC, 2002.
2. Bonn, M. and Villu, Maricq A. Neuronal substrates of complex behaviors in C.
elegans. Ann. Rev. Neurosci., 28: 451501, 2005.
3. Brenner, S. The genetics of behaviour. Br. Med. Bull., 29: 269271, 1973.
4. Broadhurst, P. L. and Fulker, D. W. Behavioral genetics. Ann. Rev. Psychol., 25:
389415, 1974.
5. Carey, G. Human Genetics for the Social Sciences, Sage Publication, London, 2003.
6. Crabbe, J. C. and Adron Harris, R., Eds., The Genetic Basis of Alcohol and Drug
Actions, Plenum Press, New York, 1991.
7. Crawley, J. N. What is Wrong with my Mouse: Behavioral Phenotyping of Transgenic
and Knockout Mice. John Wiley & Sons, New York, 2000.
8. Crusio, W. E. and Gerlai, R. T. Handbook of Molecular-Genetic Techniques for Brain
and Behavior Research. Elsevier, Amsterdam, 1999.
9. Defries, J. C. and Plomin, R. Behavioral genetics. Ann. Rev. Psychol., 29: 473515,
1978.
10. Ehrman, L. and Parsons, P. The Genetics of Behavior. Sinauer Associates, Sunderland,
MA, 1976.
1903_C001.fm Page 15 Wednesday, July 26, 2006 6:08 PM
11. Falconer, D. S. Introduction to Quantitative Genetics. The Ronald Press, New York,
1960.
12. Fuller, J. and Thompson, W. R. Behavior Genetics. John Wiley & Sons, New York,
1960.
13. Fuller, J. and Thompson, W. R. Foundation of Behaviour Genetics, The C.V. Mosby
Co., St. Louis, 1978.
14. Fuller, J. and Simmel, E. C. Behavior Genetics: Principles and Applications.
Lawrence Erlbaum Associates Publishers, Hillsdale, NJ, 1983.
15. Ginsburg, B. E. Genetics as a tool in the study of behavior. Perspectives in Biology
and Medicine, 1: 397424, 1958.
16. Goldowitz, D., Wahlsten, D. and Wimer, R. E. Techniques for the Genetic Analysis
of Brain and Behavior: Focus on the Mouse. Elsevier, Amsterdam, 1992.
17. Hall, C. S. The genetics of behavior. In: Handbook of Experimental Psychology,
Stevens S. S., Ed. John Wiley & Sons, New York, 1951, chap. 9.
18. Hall, J. C., Greenspan, R. J. and Harris, W. A. Genetic Neurobiology. MIT Press,
Cambridge, MA, 1982.
19. Henderson, N. D. Human behavior genetics. Ann Rev. Psychol., 33: 401440, 1982.
20. Hirsch, J., Ed. Behavior-Genetic Analysis. McGraw-Hill, New York, 1967.
21. Kyriacou, C. P. Single gene mutations in Drosophila: What can they tell us about the
evolution of sexual behavior? Genetica, 116: 197203, 2002.
22. Lindzey, G. and Thiessen, D. D. Eds., Contributions to Behavior-Genetic Analaysis:
The Mouse as a Prototype, Appelton Century Crofts, New York, 1970.
23. Lindzey, G., Loehlin, J., Manosevitz, M., and Theissen, D. Behavioral genetics. Ann.
Rev. Psychol., 22: 3994, 1971.
24. Mather, K. Biometrical Genetics, Methuen, London, 1949.
25. McClearn, G. E. Behavioral genetics. Ann. Rev. Genet., 4: 417468, 1970.
26. McClearn, G. E. and DeFries, J. C. Introduction to Behavioral Genetics. W.H. Free-
man, San Francisco, 1973.
27. McClearn, G. E. and Meredith, W. Behavioral genetics. Ann. Rev. Psychol., 17:
515550, 1966.
28. McGue, M. and Bouchard Jr., T. J. Genetic and environmental influences on human
behavioral differences. Ann. Rev. Neurosci., 21: 124, 1998.
29. McGuffin, P., Owen, M. J., and Gottesman, I. J. Psychiatric Genetics and Genomics.
Oxford University Press, New York, 2002.
30. Parsons, P. A. Behavioral and Ecological Genetics: A study in Drosophila, Oxford
University Press, New York, 1973.
31. Pfaff, D. W., Berrettini, W., Joh, T. H. and Maxson, S. C. Genetic Influences on
Neural and Behavioral Function, CRC Press, Boca Raton, FL, 2000.
32. Plomin, R., DeFries, J. C., Craig, I. W., and McGuffin, P. Behavioral Genetics in the
Postgenomic Era. American Psychological Association, Washington, DC, 2002.
33. Robinson, G. E. and Ben-Shahar, Y. Social behavior and comparative genomics: new
genes or new gene regulation. Genes, Brain, Behav., 1: 187203, 2003.
34. Scott, J. P. and Fuller, J. L. The Genetics and Social Behavior of the Dog, University
of Chicago Press, 1965.
35. Sokowlowski, M. B. Drosophila genetics meets behavior. Nat. Genet., 2: 879880,
2001.
36. van Abeelen, J. H. F., Ed. The Genetics of Behavior, North-Holland, Amsterdam, 1974.
37. Wahlsten, D. Single-gene influences on brain and behavior. Ann. Rev. Psychol. 50:
599624, 1999.
1903_C001.fm Page 16 Wednesday, July 26, 2006 6:08 PM
16 Neurobehavioral Genetics
38. Weiner, J. Time, Lover, and Memory: A Great Biologist and His Quest for the Origins
of Behavior. Alfred A. Knopf, New York, 1999.
39. Whener, J. M., Radcliffe, R. A., and Bowers B. J. Quantitative genetics and mouse
behavior. Ann. Rev. Neurosci., 24: 845867, 2001.
40. Wimer, R. E. and Wimer, C. E. animal behavior genetics: a search for the biological
foundations of behavior. Ann. Rev. Psychol. 36: 171218, 1985.
1903_C002.fm Page 17 Wednesday, July 26, 2006 6:09 PM
2 Developmental
Neurobehavioral Genetics:
Development as
Explanation
Gilbert Gottlieb
CONTENTS
Introduction.............................................................................................................. 17
2.1 Preformation and Epigenesis ......................................................................... 17
2.2 Development as Explanation ......................................................................... 19
2.3 A Currently Acceptable Developmental Neurobehavioral
Genetics: Interim Solution ............................................................................. 24
2.4 Summary and Conclusions ............................................................................ 26
Acknowledgments.................................................................................................... 26
References................................................................................................................ 26
INTRODUCTION
From the very dawn of human history, there must have been people who wondered
how we come to benot just in the grand religious sense that ancient texts like the
Hebrew bible attempt to answer, but also in the more mundane and practical sense
of wondering about the mechanisms involved in human (and all animal) development
from egg and sperm to full-grown adult. By the time of Aristotle in the fourth century
B.C.E., there were two main schools of thought on how we become. In fact, the
proper scientific name for the study of individual development was derived from
the Greek language: ontogeny.
17
1903_C002.fm Page 18 Wednesday, July 26, 2006 6:09 PM
18 Neurobehavioral Genetics
ovist or spermistall the parts and organs that an individual will ever have are present
from the outset; development is merely the growth of these preexisting parts until
they reach their full, adult size. Aristotles view, on the other hand, gained from his
personal observations of the developing chick embryo, was that individual develop-
ment happens through epigenesis; that is, through a series of successive transforma-
tions of some sort of early, homogeneous mass that brings the parts and organs of
an organism into being. In this view, what happens during development is not just
the quantitative growth of preexisting parts, but the gradual qualitative differentiation
of parts that then grow as well. Today we still do not completely understand all of
the mechanisms that make development happen, but we do know that it is correct to
say that individual developmentwhether in terms of psychology, behavior, physi-
ology, or anatomyis epigenetic and not preformative.
Nevertheless, some scientists continue to operate under a modern-day version
of preformationist thinking when they say that genes govern individual development.
The only difference between the ancient idea of the unfolding of preformed parts
and the modern idea of the unfolding of the genetic code is that the unfolding, itself,
is now understood to be a transformative processin other words, epigenetic. All
biological scientists understand that successive transformations of the forms and
functions of tissues occur during development. This is what is meant by epigenetic
development. But many scientists seem to think that these transformations are
controlled by little, preformed, inherited, autonomous packets of information that
we know today as genes. This is a modern-day version of preformationism. As the
noted biologist J. S. Haldane13 observed, the idea of genes as the controllers of
individual development only substitutes an extremely complicated molecular struc-
ture for the ancient idea of the original miniature adult(p. 147). I call the idea of
epigenetic transformation that is controlled by genes, predetermined epigenesis.
According to this way of thinking, further developmental processes may occur due
to environmental influences, but genes and the environment are viewed as separate
causal systems rather than as parts of one, integrated system. This idea that there
are separate causal systems is the crux of what has become known as the naturenur-
ture dichotomy.
The advocates of a genetically predetermined epigenesis have often recognized
that the naturenurture dichotomy is a shaky construct because, although based on
a recognition that environment does make a difference in development, it lacks an
adequate theory of how, when, and where this happens.22 In the end, only lip service
is paid to the idea of geneenvironment interactions, and genetic predeterminists
say it is genes primarily that make development happen and that the environment
plays only a passive, permissive, or supportive role. The formative drive, they say,
is governed by genes acting as autonomous agents.
However, as several excellent popular books have already described16,18 recent
research demonstrates that genes are not really unchanging little packets of unfold-
ing information standing on the sidelines of epigenetic processes and stoically
directing the developmental game. Instead, as M.-W. Ho has written,14 they get
variously chopped, rearranged, transposed, and amplified in different cells at
different times (p. 285)and themselves must be activated in order to make any
contribution at all to development. With the supposed role of genes as independent
1903_C002.fm Page 19 Wednesday, July 26, 2006 6:09 PM
20 Neurobehavioral Genetics
DEVELOPMENTAL ANALYSIS
ENVIRONMENT
(Physical, Social, Cultural)
BEHAVIOR
NEURAL ACTIVITY
GENETIC ACTIVITY
Individual Development
FIGURE 2.1 A developmentalpsychobiological systems model to illustrate the comprehensive
analysis required for developmental explanation. (Modified from Gottlieb8 with permission.)
BEHAVIOR
SENSORY
Patterned
STIMU- Neural Activity
LATION
Neural Non-neural
Connectivity Structures
Individual
Nerve Cell
Activity Neural Non-neural
Growth Growth
Extracellular
Cell Membrane Biochemistry
Intracellular
Biochemistry
Protein PHYSICAL
Synthesis INFLUENCES
GENETIC
ACTIVITY
22 Neurobehavioral Genetics
coactions that follow from protein synthesis and its consequences for events at the
cell membrane, coactions among cells, and so on. The model treats genes as an
integral part of the developing system, rather than placing them outside the system,
and sees genes as influencing behavior indirectly, not directly. When a particular
gene has been implicated in the development of some behavior, the model would
accommodate the identification of various roles that the genes activity might play
in development. For example, in a review of the effects of single-gene mutations on
the development of touch receptors in nematodes, Chalfie5 proposed four develop-
mental roles (generation, specification, function, and maintenance) for the 18 genes
that have been implicated so far in touch receptor development. Although Chalfies
taxonomy deals with anatomy rather than behavior, similar taxonomies might be
proposed for the development of behavior (e.g., Gottliebs delineation of the three
roles of experience in behavioral and neural development: induction, facilitation,
and maintenance10). The JohnstonEdwards model (Figure 2.2) has the advantage
that it provides an explicit representation of the intervening coactions implied by
such taxonomies, even though their coactions may not always be specified. Thus, it
indicates the kinds of developmental questions raised by findings that might other-
wise be taken as evidence of a direct link between genes and behavior.
Third, when experience has more than a transient effect on behavior, the effect
is almost certainly mediated through changes in genetic activity. The model
implies that all instances in which experience has been shown to affect behavioral
development must involve some change in genetic activity. Developmental theory
holds that there can be no genetic effects on behavior independent of the envi-
ronment and there are probably no environmental effects on behavior independent
of genetic activity. The model (Figure 2.2) helps to make this statement more
precise by showing the pathway by which experience activates genes through the
agency of neural activity, once again supplying a mediating pathway for findings
that might otherwise be interpreted as evidence of a direct link between genes
and behavior.
Fourth, Johnston and Edwardss model in Figure 2.2 recognizes that nervous
system activity needs to be considered at two levelsin terms of neural activity,
often involving networks of cells in different anatomical regions, and in terms of
the activity of individual nerve cells within which the genes are located. The dotted
lines connecting these two boxes in Figure 2.2 indicate that the activity of individual
cells is nested within the patterns of activity of cell networks. Individual cell activity
neither causes nor is caused by the patterns of activity in cell networks: rather, there
are two levels at which neural activity must be analyzed. Such a dual level of analysis
does not mean that an account of genetic activity must be given individual cell by
individual cell. Rather, ways of describing the developing nervous system both in
terms of populations of cells with similar patterns of genetic activity and in terms
of networks of cells that participate in behavior must be found. For example, Bren-
nan, Hancock, and Keverne have shown that the immediate-early genes c-fos and
zif-268 (but not c-jun) show distinct patterns of both transient induction and persistent
induction in the accessory olfactory bulb (AOB) of female mice immediately after
mating.2 The induction of c-fos is seen only in the granule cells of the AOB, whereas
the zif-268 induction occurs in both the granule and mitral cells. The AOB is known
1903_C002.fm Page 23 Wednesday, July 26, 2006 6:09 PM
24 Neurobehavioral Genetics
attachment with their mothers are also the most likely to have deficits in their central
serotonin metabolism. Because there is a positive correlation between maternal and
infant serotonin level, a genetic deficit could be involved, but it is possible that
aberrant maternal care may make a necessary contribution to the serotonin deficit.
To shed light on the genetic and interactive aspect, Bennett et al.1 genotyped the
monkeys in Suomis laboratory for a known polymorphism (long and short allele) in
the serotonin transporter gene (5-HTT). The short allele confers low transcriptional
efficiency to the 5-HTT gene promoter (relative to the long allele), so low 5-HTT
expression may result in lower serotonergic function. However, evidence for this in
humans is inconsistent because the necessary life experience correlates have not been
examined. In the case of rhesus monkeys, when attempting to correlate the genetic
polymorphism to serotonin metabolism, serotonin concentration did not differ as a
function of long or short 5-HTT status for mother-reared monkeys, whereas, among
peer-reared monkeys, individuals with the short allele had significantly lower sero-
tonin concentrations than those with the long allele.1 Thus, the lowered serotonin
metabolism was not simply a consequence of having the short allele but required the
life experience of peer rearing in this instance. This result supports my idea that the
inconsistencies in the human literature are likely due to unknown but influential
differences in the experiential histories of the populations under study.
Thus, the notion that the short allele of the 5-HTT gene is inevitably associated
with a central nervous system (CNS) deficit or defect is not true: the neural outcome
depends on the developmental rearing history of the animal, as well as the particular
genotype of the animal itself, what has elsewhere been termed relational causality.11
The present finding most likely also explains why there are inconsistencies in the
human literature in finding anxiety-, depression-, and aggression-related personality
traits associated with variations in the serotonin transporter gene.1 The association,
or lack thereof, does not simply reflect genetic causality but developmental-relational
causality.
In sum, the chances of linking genotypes to behavioral (and other) outcomes
will be vastly improved when crucial intervening life experiences are routinely
included in developmental behavioral genetic investigations. This follows from the
empirical work reviewed earlier indicating that gene-environment coactions are the
rule in developmental investigations.9
While gene-environment coactions are a step up from simple single-gene
outcome associations, single-genesingle-experience coactions are going to continue
to be prone to lack of replication because complex behavioral outcomes are no doubt
backed by multiple genes (epistasis) and possibly more than one crucial life
experience. The ultimate solution will be to actually employ more than one relevant
gene and more than one relevant life experience if we are eventually to achieve
highly replicable findings. For example, the Caspi et al.4 study of the low MAOA
genotype and maltreatment alluded to earliera single-gene life experience study
has been replicated at least once7 and has not been replicated at least once.12 So, in
my opinion, the interim solution proposed in the present section of this essay will
be best implemented using more than one genetic marker and, whenever possible,
more than one life experience if we aspire to a mature developmental neurobehavioral
genetics in which replication is the measure of that maturity.
1903_C002.fm Page 26 Wednesday, July 26, 2006 6:09 PM
26 Neurobehavioral Genetics
ACKNOWLEDGMENTS
The authors research and scholarly activities are supported, in part, by National
Institute of Mental Health Grant P50-MH-52429 and National Science Foundation
Grant BCS-0126475. Some material from my 2003 Human Development article18
is included in this chapter. I thank Brenda Denzler for editorial assistance with the
introduction of this chapter.
REFERENCES
1. Bennett, A.J. et al. Early experience and serotonin transporter gene variation interact
to influence primate CNS function, Molecular Psychiatry, 7: 118122, 2002.
2. Brennan, P.A., Hancock, D., & Keverne, E.B. The expression of the immediate-early
genes c-fos, egr-1, and c-jun in the accessory olfactory bulb during the formation of
the olfactory memory in mice. Neuroscience, 49: 277284, 1992.
3. Brodsky, M., & Lombroso, P.J. Molecular mechanisms of developmental disorders.
Development and Psychopathology, 10: 120, 1998.
4. Caspi, A. et al. Role of genotype in the cycle of violence in maltreated children.
Science, 297: 851854, 2002.
5. Chalfie, M. Touch receptor development and function in Caenorhabditis elegans,
J. Neurobiology, 24: 14331441, 1993.
6. Chargraff, E., Hericlitian Fire: Sketches from a Life Before Nature, Rockefeller
University Press, New York, 1978, pp. 106107.
7. Foley, D.L. et al. Childhood adversity, monoamine oxidase a genotype, and risk for
conduct disorder. Archives of General Psychiatry, 61: 738744, 2004.
8. Gottlieb, G. Individual Development and Evolution: The Genesis of Novel Behavior,
Erlbaum, Mahwah, New Jersey, 2002.
9. Gottlieb, G. On making behavioral genetics truly developmental, Human Develop-
ment, 46: 337355, 2003.
10. Gottlieb, G. The roles of experience in the development of behavior and the nervous
system, in Neural and Behavioral Specificity, G. Gottlieb, Ed., Academic Press, New
York, chap. 2, 1976.
11. Gottlieb, G., & Halpern, C.T. A relational view of causality in normal and abnormal
development. Development and Psychopathology, 14: 421435, 2002.
12. Haberstick, B.C. et al. Monoamine oxidase A (MAOA) and antisocial behavior in
the presence of childhood and adolescent maltreatment. Neuropsychiatric Genetics,
135B:5964, 2005.
13. Haldane, J.S. The Philosophical Basis of Biology, Doubleday, Garden City, New York,
1931.
1903_C002.fm Page 27 Wednesday, July 26, 2006 6:09 PM
14. Ho, M.-W. Environment and heredity in development and evolution, in Beyond Neo-
Darwinism: An Introduction to the New Evolutionary Paradigm, M.-W. Ho & P.T.
Saunders, Eds., Academic Press, London, 267289, 1984.
15. Johnston, T.D., & Edwards, L. Genes, interactions, and the development of behavior.
Psychological Review, 109: 2634, 2002.
16. Keller, E.F. The Century of the Gene, Harvard University Press, Cambridge, 2000.
17. Merikangas, K.R., & Risch, N. Genomic priorities and public health, Science, 302:
599601, 2003.
18. Moore, D. The Dependent Gene: The Fallacy of Nature vs. Nurture, Henry Holt,
New York, 2002.
19. Ordovas, J.M. et al. Dietary fat intake determines the effect of a common polymor-
phism in the hepatic lipase gene promoter on high-density lipoprotein metabolism:
Evidence of a strong dose effect in the genenutrition interaction in the Framingham
study, Circulation, 106: 23152321, 2002.
20. Peters, R.J.G., & Boekholdt, S.M. Gene polymorphisms and the risk of myocar-
dial infarctionan emerging relation, New England Journal of Medicine, 347:
19631965, 2002.
21. Rampon, C. et al. Enrichment induces structural changes and recovery from nonspatial
memory deficits in CA1 NMDAR1-knockout mice, Nature Neuroscience, 3: 238244,
2000.
22. Rutter, M. et al. Testing hypotheses on specific environmental causal effects on
behavior, Psychological Bulletin, 127: 291324, 2001.
23. Schaffner, K.F. Genes, behavior, and developmental emergentism: One process indi-
visible? International Journal of Behavioral Development, 24: 514, 1998.
24. Suomi, S.J. A biobehavioral perspective on developmental psychopathology, in Hand-
book of Developmental Psychopathology, A.J. Sameroff, M. Lewis, & S.M. Miller,
Eds., Kluwer Academic/Plenum, New York, chap. 13, 2000.
25. Thelen, E. Motor development: A new synthesis, American Psychologist, 50: 7995,
1995.
26. Venter, J.C. et al. The sequence of the human genome, Science, 291: 3041351, 2001.
27. Wong, A.H.C., Buckle, C.E., & Van Tol, H.H.M. Polymorphisms in dopamine recep-
tors: What do they tell us? European J. Pharm., 410: 183203, 2000.
1903_C002.fm Page 28 Wednesday, July 26, 2006 6:09 PM
1903_C003.fm Page 29 Thursday, July 27, 2006 9:04 PM
CONTENTS
29
1903_C003.fm Page 30 Thursday, July 27, 2006 9:04 PM
30 Neurobehavioral Genetics
32 Neurobehavioral Genetics
short vs. long pea vine length, etc. Moreover, these factors come in pairs that may
be of the same type or different types (what we call homogeneous or heterogeneous).
Finally, during the time of gamete formation, these factors separate to be recombined
with like or different types at fertilization. Some of the factor types (what we call
alleles) would dominate over the alternative form, so that if an individual plant
inherited the factor for wrinkled peas from one parent and the factor for smooth
peas from the other parent, the observed type would be smooth. Mendel then termed
smooth the dominant trait. From these observations, Mendel produced the familiar
mathematical rules to predict outcomes when homogeneous or heterogeneous
matings were performed. Thus, mating pure wrinkled with pure smooth would, in
the first filial generation, produce all smooth. Mating these first filial (F1) plants
among one another would produce a mix of smooth and wrinkled in the ratio of 3
smooth to 1 wrinkled for the phenotype, whereas the genotypes would be
1(SS):2(SW):1(WW).
3.6 LINKAGE
During gametogenesis, or more accurately, meiosis, both of the chromosome
homologs replicate to give 4 copies of the chromosomethe tetrad. The inner
homologous chromosomes can break from physical pressure and reassemble such
that a piece of a maternal homolog, containing several genes, splices with the
remainder of the paternal homolog and vice versa. So, genes for various character-
istics can be linked. The lineup of genes (alleles) on a single strand homolog is
referred to as the haplotype. Now, under certain conditions, a phenotype may inform
the researcher about the genetics of a linked phenotype.
34 Neurobehavioral Genetics
Response
0 1 2 3 . . . . . . . . . . . . . . . . . . . . . . . . .n
Generation Number
FIGURE 3.1 Theoretical representation of progression of phenotype-based bi-directional
selection.
we would expect to see a bimodal distribution. More than 30 years ago, McClearn
and Kakihana produced lines of mice selected for high sensitivity and low sensitivity
to the hypnotic effects of ethanol.7 These lines were named long-sleep and short-
sleep, respectively. Figure 3.1, is a theoretical representation of how bi-directional
selection of a complex trait should progress across generations. Thus alleles that
confer increased or decreased sensitivity would accumulate across generations. At
the end of more than 40 generations of bi-directional selection, it was concluded
that at least four polymorphic genes are involved in sensitivity to the hypnotic effects
of ethanol.5 Thus the genes work together in additive fashion in their influence on
the phenotype.
3.9 INTERACTIONS
Genes do not operate in isolation. As Waddington9 proposed, how genes operate in
embryogenesis depends in large part on the extra- (or epi-) genetic milieu, e.g.,
nutrients, toxins, etc. We also know that genes can influence one another to enhance
or suppress expression.
3.10 SUMMARY
Basic knowledge about what genes are, what they do, and how they influence
neurobiology and behavior has grown exponentially over the past few years. Com-
plex traits analysis is the new look in genetics and holds great promise for helping
us to see brainbehavior relationships from a systems approach.
REFERENCES
1. Cloninger, C.R. Neurogenetic adaptive mechanisms in alcoholism, Science, 1987,
236:410416.
2. Crawley, J.N. Unusual behavioral phenotypes of inbred mouse strains, Trends.
Neurosci., 1996, 19:181182.
3. Griffiths, P., Paterson, L., and Harvie, A. Neuropsychological effects of subsequent
exposure to phenylalanine in adolescents and young adults with early-treated phe-
nylketonuria, J. Intellect. Disabil. Res., 1995, 39:365372.
4. Jones, B.C., Connell, J.M. and Erwin V.G. Isolate housing alters ethanol sensitivity
in long-sleep and short-sleep mice, Pharmacol. Biochem. Behav., 1990, 35:469472.
5. Markel, P.D., Fulker, D.W., Bennett, B., Corley, R.P., DeFries, J.C., Erwin, V.G., and
Johnson, T.E. Quantitative trait loci for ethanol sensitivity in the LS x SS recombinant
inbred strains: interval mapping, Behav. Genet., 1996, 26:447458.
1903_C003.fm Page 36 Thursday, July 27, 2006 9:04 PM
36 Neurobehavioral Genetics
4 An Introduction to
Quantitative Genetics
Wim E. Crusio
CONTENTS
Summary .................................................................................................................. 37
4.1 Introduction .................................................................................................... 38
4.2 Phylogenetic Aspects of Causation ............................................................... 38
4.2.1 Natural Selection and Genetic Architecture...................................... 38
4.2.2 Theoretical Background..................................................................... 39
4.2.3 Examples ............................................................................................ 43
4.3 Phenogenetic Aspects of Causation............................................................... 43
4.3.1 The Correlational Approach: Brain Lesions
and the Locality Assumption............................................................. 43
4.3.2 Genetic Correlations .......................................................................... 44
4.3.3 Theoretical Background..................................................................... 45
4.3.4 Examples ............................................................................................ 48
4.3.4.1 Rearing Behavior in an Open Field and Hippocampal
Mossy Fibers in Mice......................................................... 48
4.3.4.2 Nerve Conduction Velocity and IQ in Man....................... 48
4.3.4.3 Localization of QTL........................................................... 49
4.4 Some Common Misapplications.................................................................... 49
4.5 Conclusion...................................................................................................... 51
Acknowledgments.................................................................................................... 52
References................................................................................................................ 52
Appendix.................................................................................................................. 54
SUMMARY
This chapter provides a brief overview of quantitative-genetic theory. Quantitative
genetics provides important tools to help elucidate the genetic underpinnings of
behavioral and neural phenotypes. This information can then provide substantial
insights into the previous evolutionary history of a phenotype, as well as into
brainbehavior relationships.
The most often employed crossbreeding designs are the classical Mendelian
cross and the diallel cross. The information rendered by the former is limited to the
37
1903_C004.fm Page 38 Wednesday, July 26, 2006 6:33 PM
38 Neurobehavioral Genetics
two parental strains used and cannot be broadly generalized. The principal usefulness
of this design is for testing whether a given phenotype is influenced by either one
gene or by more genes. The diallel cross renders more generalizable information,
the more so if many different strains are used, such as estimates of genetic correla-
tions. To estimate the latter, correlations between inbred strain means may provide
a helpful shortcut.
Some commonly encountered mistakes in the interpretation of the results of
quantitative-genetic studies are presented and explained.
4.1 INTRODUCTION
Behavior is an animals way of interacting with its environment and is therefore a
prime target for natural selection. Furthermore, as behavior is the output of an animals
nervous system, this indirectly leads to selection pressures on neuronal structures. In
consequence, each species behavior and nervous system have co-evolved in the
context of their natural habitat and can be properly comprehended only when their
interrelationships are regarded against that background.4 To arrive at a profound under-
standing of neurobehavioral traits, one will therefore have to consider problems of
causation. Van Abeelen39 distinguished between the phenogenetic and the phylogenetic
aspects of causation. Both aspects deal with the genetic correlates of neurobehavioral
traits, the first in a gene-physiological, the second in an evolutionary sense. In other
words, neurobehavioral geneticists try to uncover the physiological pathways under-
lying the expression of a trait and to evaluate its adaptive value for the organism. As
I have shown before,9 quantitative-genetic methods may be employed with profit to
address problems related to both aspects of causation.
40 Neurobehavioral Genetics
phenotype can then be divided into two main sources: additive-genetic effects and
dominance deviations. In the case of one single gene, with alleles A and a, we may
denote the phenotypical values of the three possible genotypes as follows*:
AA = m + da Aa = m + ha aa = m da (4.1)
The parameter da is used to represent half the difference between the homozy-
gotes, ha designates the deviation of the heterozygote from the midparental point m.
Note that in quantitative genetics capital letters are used to indicate increaser alleles,
which are not necessarily also the dominant ones. Hence, da is positive by definition,
whereas ha may attain all possible values. If we now consider two inbred strains A
and B in a situation where many genes affect the phenotype, we may denote the
average phenotype of strain A by
(shortened to m + [d] for ease of representation), where S(d+) indicates the summed
effects of those genes that are represented by their increaser alleles and S(d) indicates
the same for decreaser alleles. Parameter m is a constant, reflecting the average
environmental effects both strains have in common as well as genetic effects at loci
where the strains are fixed for the same alleles. The average phenotype of strain B
will then equal m [d]. Similarly, the phenotypic value of an F1 hybrid between A
and B may be written as
(shortened to m + [h]). It must be noted that [h] is the sum of the dominance
deviations of many genes. If these effects are balanced in opposite directions, [h]
can be low or zero, even with dominance present. The same applies to [d], of course.
Because variations in a phenotype can be thought of as the summed effects of
variations in genotype and environment, plus the interaction and covariation between
these two factors, we may express the phenotypic variance P of a population as:
* In this chapter, we will follow the notation of Mather and Jinks.31 In the appendix I present a table
comparing this notation with the one followed by Falconer and Mackay.16
1903_C004.fm Page 41 Wednesday, July 26, 2006 6:33 PM
P=G+E (4.5)
The genetic component of the variance (G) can, of course, be divided into
components due to additive-genetic variation (D) and dominance deviations (H).
We may demonstrate the partitioning of genetic variance into its additive-genetic
and dominance components by the example of an F2 cross between two inbred
strains. When only one gene with two alleles influences the phenotype, the expected
genetic composition of the F2 population will be AA, 25%; Aa, 50%; and aa, 25%.
From the foregoing, this leads to a phenotypic mean of
da + ha da = ha (4.6)
(m is set at zero by a simple shift of the measurement scale). The sum of squares
of deviations from the mean then equals
k k
i =1
di2 +
i =1
hi2 (4.8)
shortened to
D+H (4.9)
VP = D + H + E (4.10)
and the proportion of the phenotypical variance due to all genetic effects, the
heritability in the broad sense
42 Neurobehavioral Genetics
D=
i =1
4uividi2 (4.13)
and
H=
i =1
4uivihi2 (4.14)
generalized picture of the genetic architecture of a trait. For instance, if one cross-
breeding experiment indicates dominance in the direction of, say, high expression
of the trait, but another cross indicates dominance in the opposite direction, then
this constitutes prima facie evidence for ambidirectional dominance. In fact, the
presence of directional dominance may only be inferred if all available evidence
indicates that dominance is acting in the same direction.
4.2.3 EXAMPLES
Crusio and van Abeelen14 addressed the question of what exactly is the adaptive
value of various mouse exploratory behaviors carried out in novel surroundings. As
one result of exploration is the collection of new, or the updating of previously
acquired, information, we argued that, if an animal enters a completely novel envi-
ronment, it is obviously of prime interest to collect as much information as possible
in a short time. On the other hand, high exploration levels will render the animal
more vulnerable to predation. Taking all together, we hypothesized an evolutionary
history of stabilizing selection for exploration. This hypothesis was subsequently
confirmed by the results of several crossbreeding experiments11,14 that revealed
genetic architectures comprising additive genetic variation and/or ambidirectional
dominance for most behaviors displayed in an open field. The above reasoning is,
of course, not specific for the mouse species. Indeed, Gerlai et al.21 found similar
genetic architectures for exploration in an open field in a diallel cross between inbred
strains of paradise fish, Macropodus operculatus.
44 Neurobehavioral Genetics
Dxy =
i =1
4uividxidyi (4.15)
and
Hxy =
i =1
4uivihxihyi (4.16)
where dxi and dyi are the additive-genetic effects of the ith gene on characters x and
y, respectively, and hxi and hyi are the respective dominance deviations due to the ith
gene. Evidently, only genes that have effects on both of the characters x and y
contribute to the genetic covariance terms, whereas all genes that affect either x or
y contribute only to the respective genetic variance terms. Combining these compo-
nents of the covariance with the components of the variance obtained in the univariate
analyses we may estimate genetic correlations as follows:
and
46 Neurobehavioral Genetics
cross is very problematic in this respect and although many examples exist in the
literature in which authors claim to have analyzed genetic correlations with this
design, none really have done so. The problem appears to be due mainly to the
frequent occurrence of the phenomenon, first observed by Tryon,38 that the variance
of a segregating F2 population is not significantly larger than those of nonsegregating
populations (in extreme cases, it will even be smaller). Several possible explanations
have been brought forward. First, Hall23 attributed the Tryon effect to an insufficient
degree of inbreeding of Tryons selected (but not inbred) strains. Of course, this
would enlarge the genetic variation within the parental and F1 generations, but one
would still expect the F2 to have a somewhat larger variance. In addition, the Tryon
effect has since also been observed in crosses between highly inbred strains. A
second explanation was presented by Hirsch.26 He argued that most phenotypes are
influenced by more than one gene. If we take the rat, with a karyotype of 21
chromosome pairs, as an example and, for simplicity, treat these chromosomes as
major indivisible genes, one can see that this organism can produce 221 different
kinds of gametes, leading to 321 (= 1.05 1010) different possible genotypes. In
reality, this number will be even larger because chromosomes are not indivisible.
Obviously, no experiment can take from an F2 generation a sample large enough to
have all these genotypes represented and Hirsch assumed this sampling effect to
lower the observed F2 variances below expected levels. Tellegen36 quite correctly
countered that as long as the sampling from the F2 is random, an unbiased estimate
of the population variance should be obtained. In addition, it can easily be seen that
even if Hirschs reasoning were correct, the F2 variance would still be expected to
significantly exceed that of nonsegregating generations, being the sum of environ-
mentally induced variation and, in his reasoning, at least some genetic variation.
Bruell3 had observed that the amount of variance caused by segregation in the
F2 increases if gene effects are larger and decreases if more genes influence
the phenotype studied (cf. Equations 4.13 and 4.14). If environmental influences
on the phenotype are large, an extremely large sample would be needed to detect
the difference in variance between the F2 and F1 populations at a sufficient level of
significance in situations with many genes and relatively small gene effects. As
sample sizes are limited by considerations of time, money, and space, while envi-
ronmental influences on behavioral characters are usually very pronounced, one
should normally expect F2 variances not to differ significantly from F1 variances.
Due to sampling error, they may then even be smaller, although usually not signif-
icantly so. Homeostatic processes may be responsible if the latter situation occurs.27
The problem therefore boils down to one of statistical power. It should be recalled
here that larger sample sizes are needed to obtain accurate estimates of the variance
of a population than for estimating its mean. By analogy, this also goes for covari-
ances. In sum, unless rather huge sample sizes are used, classical crosses will gen-
erally lack the statistical power needed to accurately estimate genetical correlations.
Another reason that the classical cross is less suited for bi- and multivariate studies
is the fact that results are not generalizable, but based on a restricted sample of two
inbred strains only: even if genetic correlations are estimated correctly with this design
(which happens only rarely; see Hayman24 for appropriate statistical methods), there
exists a non-negligible probability that they would be due to a linkage disequilibrium
1903_C004.fm Page 47 Wednesday, July 26, 2006 6:33 PM
instead of pleiotropy. Of course, it is exactly the latter property that researchers employ
when localizing genes. Note, however, that this is a special case in which one
character, the molecular marker, is completely determined by the genotype (see also
the following section).
A more suitable method is the diallel cross, for which a bivariate extension is
available,8 whereas an interesting shortcut is offered by using a panel of inbred
strains. Correlations between strain means either permit the estimation of additive-
genetic correlations,25 or provide a direct lower-bound estimate of additive-genetic
correlations (if the traits to be correlated have been measured in different individuals
from these strains).
From the foregoing, it may easily be seen that the variance of the means of a
set of inbred strains equals
D + E/n (4.19)
where n is the harmonic mean of the number of subjects per strain. The covariance
between the means obtained for two characters x and y can then be expressed as
Dxy (4.20b)
in case characters x and y are being measured on different individuals from the same
strains. The correlation between the strain means in these two situations will now
equal
or
48 Neurobehavioral Genetics
Recombinant inbred strains (RISs) have sometimes also been used to estimate
genetic correlations between phenotypes, using the above-described methods for
correlations using ordinary inbred strains. It should be realized, however, that there
is a considerable risk that any genetic correlations thus found will be due to a linkage
disequilibrium because RISs have been derived from an F2 between two inbred
strains. As was the case with the classical cross itself, this property is of course of
interest for researchers hoping to localize quantitative trait loci (QTL; but see Section
4.3.4.3). In fact, a genetic correlation obtained with RISs implies that both characters
under study are influenced by linked genes, i.e., that map to the same chromosomal
segments, at least in part. Only if previous evidence of the existence of a genetic
correlation has been obtained with other methods (such as a screen of normal,
unrelated inbred strains), can a genetic correlation obtained from an RI study be
considered evidence of the localization of a QTL caused by a common gene.
Except in the case of correlations between inbred strain means, testing the
significance of genetic correlations is often problematic and the power of available
tests is not well known. Fortunately, when the environmental and genetic correlations
are used as input for further, multivariate, analyses, the possible significance or lack
thereof of an individual correlation is no longer very important.
4.3.4 EXAMPLES
4.3.4.1 Rearing Behavior in an Open Field and Hippocampal
Mossy Fibers in Mice
Crusio et al.12 carried out a diallel cross study in which 5 different inbred strains
were crossed in all possible ways. In 150 male mice from the 25 resulting crosses
they measured the rearing-up frequency during a 20 min session in an open field
and the extent of the hippocampal intra- and infrapyramidal mossy fiber (IIPMF)
projection. They obtained a marginally significant phenotypical correlation of 0.138
(df = 148, 0.05 < P < 0.10). Ordinarily, one would take such a result as evidence
that variation in the size of the IIPMF is not related to behavioral variation. However,
a quantitative-genetic partitioning of the covariation showed that the genetic corre-
lation was quite sizable: 0.479. The low phenotypical correlation was explained by
the modest heritability for rearing (0.25 vs. 0.53 for the IIPMF10) and by the fact
that the (low) environmental correlation had a sign opposite to that of the genetic
correlation. The genetic relationship between rearing, on the one hand, and the
IIPMF, on the other hand, was confirmed by the finding that a line selected for high
rearing frequency had larger IIPMF projections than a line selected for low rearing
frequencies,13 exactly as would be predicted from a positive genetic correlation
between these characters. From these data, it was concluded that the IIPMF plays
an important role in the regulation of open-field rearing.9
Lately, human behavior geneticists have also started to use genetic correlations to
uncover brainbehavior relationships, especially the very active group around Dorret
Boomsma at the Free University of Amsterdam (the Netherlands). In a recent
1903_C004.fm Page 49 Wednesday, July 26, 2006 6:33 PM
experiment they used the twin method (see Chapter 12) to examine the possible
existence of a genetic correlation between speed-of-information processing (SIP)
and IQ.35 It was postulated that SIP, as derived from reaction times on experimental
tasks, measures the efficiency with which subjects can perform basic cognitive
operations underlying a wide range of intellectual abilities. Phenotypic correlations
generally range from 0.2 to 0.4. Rijsdijk et al.35 showed that genetic correlations
also fell in this range at ages 16 and 18 years, whereas environmental correlations
were essentially zero. A common, heritable biological basis underlying the SIPIQ
relationship is thus very probable.
Currently, RISs are widely used as a tool to localize QTL. Unfortunately, problems
of statistical power (often also due to multiple testing) lead to many false positives
and negatives, as illustrated by the studies of Mathis et al.32 and Gershenfeld et al.22
In the first study, a number of QTLs associated with open-field behavior were
identified using a large set of RIS between the inbred mouse strains C57BL/6J
and A/J. In the second study, an F2 generation between these same inbred strains
was studied, again leading to the identification of a number of QTLs for the same
behavioral phenotypes. None of the QTL found was common to both studies,
illustrating the problem of both types of statistical errors inherent with the use of
recombinant inbred strains. Nevertheless, there have been a few instances where
QTL have been replicated within and across laboratories.28,33 Recently, recombinant
strain sets have become much larger through the development of new, large sets as
well as of additional strains for existing sets. This development may expect to lead
to a vast increase of power and a sizeable reduction of type I and type II errors.
1. IF SOME GENETIC EFFECTS ARE FOUND FOR ONE CHARACTER BUT NOT FOR ANOTHER,
THIS IMPLIES THAT THESE CHARACTERS ARE INFLUENCED BY DIFFERENT GENES.
WRONG! THE EQUATIONS GIVEN ABOVE SHOULD ALREADY MAKE IT ABUNDANTLY
CLEAR THAT THIS IS NOT TRUE. AN EXAMPLE MAY ILLUSTRATE THIS.
Albinism in mice, for instance, is a character that is completely recessive as far as coat
color is concerned. A quantitative-genetic analysis would indicate the significant pres-
ence of both d and h of equal size, in such a way that the heterozygote would completely
resemble the non-albino homozygote. However, if we now would perform a quantita-
tive-genetic analysis of the phenotype activity of the enzyme tyrosinase, we would
1903_C004.fm Page 50 Wednesday, July 26, 2006 6:33 PM
50 Neurobehavioral Genetics
find that dominance is completely absent for this character, the heterozygote being
completely intermediate between the two homozygotes. Still, as we know very well,
only one and the same gene is involved here, which acts as a recessive on the level of
coat color, but shows intermediate inheritance on the level of the activity of the
responsible enzyme.
In fact, only the presence of a genetic correlation between two phenotypes provides
evidence that (a) gene(s) is (are) simultaneously influencing both. As was pointed out
above, the reverse need not be true.
2. IF TWO CHARACTERS ARE CORRELATED BETWEEN TWO PARENTAL STRAINS BUT NOT
IN THEIR F2 , THEY SEGREGATE INDEPENDENTLY AND, HENCE, ARE INFLUENCED BY
DIFFERENT GENES (IN OTHER WORDS, THERE IS NO GENETIC CORRELATION BETWEEN
THEM). WRONG! IN FACT, THIS OBSERVATION MAY BE TRUE, BUT IN ONLY ONE
EXCEPTIONAL SITUATION: IF AT LEAST ONE OF THE CHARACTERS WE ARE DEALING
WITH (FOR INSTANCE, A MOLECULAR-GENETIC MARKER) IS COMPLETELY DETERMINED
BY THE GENOTYPE. THIS LATTER CONDITION IS CALLED COMPLETE PENETRANCE OR,
IN QUANTITATIVE-GENETIC TERMS, THE HERITABILITY IN THE BROAD SENSE IS SAID
TO EQUAL 1. FOR BEHAVIORAL AND NEURAL PHENOTYPES, THIS CONDITION ALMOST
NEVER OCCURS. A FEW FURTHER OBSERVATIONS SHOULD BE MADE.
First, in all situations (including the above one), the phenotypical correlation within
an F2 generation will be a function of the heritabilities of both characters and the sizes
and signs of the genetic and environmental correlations between them. If, to simplify
the equation, we suppose dominance effects to be absent for both characters x and y, then
where rP is the phenotypical correlation, hx and hy are the square roots of the narrow-
sense heritabilities of characters x and y, ex and ey are the square roots of the environ-
mentalities (the proportion of the phenotypical variance due to the environment;
ex2 = 1 hx2), and rE is the environmental correlation.
Equation 4.22 has a number of important implications. For instance, the size and sign
of rP evidently do not render any information at all about the size and sign of rD. It
can easily be seen that a significant phenotypical correlation may even be completely
absent (rP not significantly different from zero) in case rD and rE have opposite signs
and the absolute value of hxhyrD comes close to that of exeyrE. In recent years, experi-
ments attempting to determine the locations of polygenes, or QTL, have become ever
more popular. In such experiments one character, the molecular-genetic marker whose
possible linkage to a putative QTL is being tested, will have a heritability of 1 and
zero environmentality. In such a case, Equation 4.22 reduces to
rP = hxrD (4.23)
In this case, the additive-genetic correlation rD is equivalent to the square root of the
proportion of the genetic variance explained by the QTL under study. This in turn, is
a function of the distance between the genetic marker and the QTL and the relative
1903_C004.fm Page 51 Wednesday, July 26, 2006 6:33 PM
effect of the QTL compared to other genes. An important implication of this equation
is that there exists an upper bound to the correlation between a behavioral or neural
phenotype, on the one hand, and a marker locus, on the other hand, even if this marker
locus would not just be linked to the hypothetical QTL but actually be identical with
it and even if there is only one single gene influencing the phenotype. In the latter case,
rD would reduce to 1 and the phenotypical correlation would be expected to equal the
heritability. Equation 4.23 explains why, especially when heritabilities are low or
numbers of QTL are large (and also because of the lack of statistical power of estimates
of variance and covariance in F2 populations referred to in Section 4.3.3), the power
to reliably detect QTL is often very low, leading to many false positives and negatives.34
Second, note that in the above erroneous statement the words between two parental
strains were used. The correlation within such strains has obviously no bearing at all
on the eventual presence or absence of genetic correlations. As all individuals within
inbred strains have the same genotype, any correlations occurring within such a strain
are of environmental origin, of course. This is not to say, of course, that at any given
point in time, two individuals belonging to the same inbred strain may not have different
profiles of gene expression. If such is the case, however, then such differences in gene
expression itself must be due to environmental influences (assuming, of course, that
both individuals are at similar stages of development).
This proposition is perhaps formally not incorrect (heritabilities can be derived from
selection studies), but it contains in fact two conceptual errors. The first one is that
there is some information to be gained from a heritability coefficient. Actually, except
as an intermediate step in estimating genetic correlations, heritabilities do not have any
intrinsic value. The only interesting facts about heritabilities are whether they differ
significantly from zero (meaning that there is significant genetic variation) or from
unity (meaning that there is significant environmental variation). There is one further
use of heritability estimates, which is that their size predicts the eventual effects of
selection pressures (whether artificial or natural16). The second conceptual error derives
from this fact: once a selection study has been carried out and selection has led to the
successful establishment of divergent lines, knowledge about heritability become more
or less useless: it is not interesting any more to determine whether selection might be
successful! Thus, there is only a single situation in which one would perform a selection
experiment to estimate heritabilities: when one wishes to estimate genetic correlations
and for some reason inbred strains are not available.
4.5 CONCLUSION
In conclusion, if the above pitfalls are avoided, quantitative-genetic experiments can
render valuable information on the genetic architecture of a trait. In addition, they can
provide information about the multivariate genetic structure of complexes of traits.
Because of this last property, quantitative genetics may serve as a valuable additional
tool in the neuroscientists arsenal and may greatly enhance our understanding of the
genetic and neural mechanisms underlying individual differences in behavior.
1903_C004.fm Page 52 Wednesday, July 26, 2006 6:33 PM
52 Neurobehavioral Genetics
ACKNOWLEDGMENTS
This chapter is dedicated to the memory of my teacher, mentor, and dear friend,
Hans van Abeelen (19361998).
REFERENCES
1. AMMASSARI-TEULE M, FAGIOLI S, ROSSI-ARNAUD C: Genotype-dependent
involvement of limbic areas in spatial learning and postlesion recovery. Physiol Behav,
1992, 52:505510.
2. BROADHURST PL, JINKS JL: What genetical architecture can tell us about the
natural selection of behavioural traits. In JHF van Abeelen (Ed): The Genetics of
Behaviour. North-Holland, Amsterdam, 1974, pp 4363.
3. BRUELL JH: Dominance and segregation in the inheritance of quantitative behavior
in mice. In EL Bliss (Ed): Roots of behavior. Hoeber and Harper, New York, 1962,
pp 4867.
4. CAMHI JM: Neural mechanisms of behavior. Curr Opinion Neurobiol, 1993,
3:10111019.
5. CAREY G: Inference about genetic correlations. Behav Genet, 1988, 18:329338.
6. CRUSIO WE: HOMAL: A computer program for selecting adequate data transfor-
mations, J Hered, 1990, 81:173.
7. CRUSIO WE: Quantitative Genetics. In D Goldowitz, D Wahlsten, R Wimer (Eds):
Techniques for the Genetic Analysis of Brain and Behavior: Focus on the Mouse.
Techniques in the Behavioral and Neural Sciences, Volume 8, Elsevier, Amsterdam,
1992, pp 231250.
8. CRUSIO WE: Bi- and multivariate analyses of diallel crosses: A tool for the genetic
dissection of neurobehavioral phenotypes. Behav Genet, 1993, 23:5967.
9. CRUSIO WE: Natural selection on hippocampal circuitry underlying exploratory
behaviour in mice: Quantitative-genetic analysis. In E Alleva, A Fasolo, H-P
Lipp, L Nadel, L Ricceri (Eds): Behavioural Brain Research in Naturalistic and
Seminaturalistic Settings. NATO Advanced Study Institutes Series D, Behavioural
and Social Sciences, Kluwer Academic Press, Dordrecht, Pays-Bas, 1995,
pp 323342.
10. CRUSIO WE, GENTHNER-GRIMM G, SCHWEGLER H: A quantitative-genetic
analysis of hippocampal variation in the mouse. J Neurogenet, 1986, 3:203214.
11. CRUSIO WE, SCHWEGLER H, VAN ABEELEN JHF: Behavioral responses to
novelty and structural variation of the hippocampus in mice. I. Quantitative-genetic
analysis of behavior in the open-field. Behav Brain Res, 1989, 32:7580.
12. CRUSIO WE, SCHWEGLER H, VAN ABEELEN JHF: Behavioral responses to
novelty and structural variation of the hippocampus in mice. II. Multivariate genetic
analysis. Behav Brain Res, 1989, 32:8188.
13. CRUSIO WE, SCHWEGLER H, BRUST I, VAN ABEELEN JHF: Genetic selection
for novelty-induced rearing behavior in mice produces changes in hippocampal mossy
fiber distributions. J Neurogenet, 1989, 5:8793.
14. CRUSIO WE, VAN ABEELEN JHF: The genetic architecture of behavioural
responses to novelty in mice. Heredity, 1986, 56:5563.
15. DONOVICK PJ, BURRIGHT RG, FANELLI RJ, ENGELLENNER WJ: Septal
lesions and avoidance behavior: Genetic, neurochemical and behavioral consider-
ations. Physiol Behav, 1981, 26:495507.
1903_C004.fm Page 53 Wednesday, July 26, 2006 6:33 PM
16. FALCONER DS, MACKAY TFC: Introduction to Quantitative Genetics. 4th Edition.
Longman Group Ltd, Harlow, Essex, 1996.
17. FANELLI RJ, BURRIGHT RJ, DONOVICK PJ: A multivariate approach to the
analysis of genetic and septal lesion effects on maze performance in mice. Behav
Neurosci, 1983, 97:354369.
18. FARAH MJ: Neuropsychological inference with an interactive brain: A critique of
the "locality" assumption. Behav Brain Sci, 1994, 17:43104.
19. FISHER RA: The correlation between relatives on the supposition of Mendelian
inheritance. Trans R Soc Edinb, 1918, 52:399433.
20. FRYER G, ILES TD: The Cichlid Fishes of the Great Lakes of Africa: Their Biology
and Evolution. Oliver and Boyd, Edinburgh, 1972.
21. GERLAI R, CRUSIO WE, CSNYI, V: Inheritance of species-specific behaviors in
the paradise fish (Macropodus opercularis): A diallel study. Behav Genet, 1990,
20:487498.
22. GERSHENFELD HK, NEUMANN PE, MATHIS C, CRAWLEY JN, LI X, PAUL
SM: Mapping quantitative trait loci for open-field behavior in mice. Behav Genet,
1997, 27:201210.
23. HALL CS: The genetics of behavior. In SS Stevens (Ed): Handbook of Experimental
Psychology, John Wiley & Sons, New York, 1951, pp 304329.
24. HAYMAN BI: Maximum likelihood estimation of genetic components of variation.
Biometrics, 1960, 16:369381.
25. HEGMANN JP, POSSIDENTE B: Estimating genetic correlations from inbred
strains. Behav Genet, 1981, 11:103114.
26. HIRSCH J: Behavior-genetic, or "experimental," analysis: The challenge of science
versus the lure of technology. Am Psychol, 1967, 22:118130.
27. HYDE JS: Genetic homeostasis and behavior: Analysis, data, and theory. Behav
Genet, 1973, 3:233245.
28. JONES BC, TARANTINO LM, RODRIGUEZ, LA, REED CL, MCCLEARN GE,
PLOMIN R, ERWIN VG: Quantitative-trait loci analysis of cocaine-related behav-
iours and neurochemistry. Pharmacogenet, 1999, 9:607617.
29. LIPP HP, SCHWEGLER H, CRUSIO WE, WOLFER D, LEISINGER-TRIGONA
M C, HEIMRICH B, DRISCOLL P: Using genetically-defined rodent strains for the
identification of hippocampal traits relevant for two way avoidance learning: A non-
invasive approach. Experientia, 1989, 45:845859.
30. MATHER K: The Genetical Structure of Populations. Chapman & Hall, London,
1973.
31. MATHER K, JINKS JL: Biometrical Genetics, 3rd ed., Chapman & Hall, London,
1982.
32. MATHIS C, NEUMANN PE, GERSHENFELD H, PAUL SM, CRAWLEY JN:
Genetic analysis of anxiety-related behaviors and responses to benzodiazepine-related
drugs in AXB and BXA recombinant inbred mouse strains. Behav Genet, 1995,
25:557568.
33. MCCLEARN GE, TARANTINO LM, RODRIGUEZ LA, JONES BC, BLIZARD
DA, PLOMIN R: Genotypic selection provides experimental confirmation for an
alcohol consumption quantitative trait locus in mouse. Molec Psychiat, 1997,
2:486489.
34. MELCHINGER AE, UTZ FH, SCHN CC: Quantitative trait locus (QTL) mapping
using different testers and independent population samples in maize reveals low power
of QTL detection and large bias in estimates of QTL effects. Genetics, 1998,
149:383403.
1903_C004.fm Page 54 Wednesday, July 26, 2006 6:33 PM
54 Neurobehavioral Genetics
35. RIJSDIJK FV, VERNON PA, BOOMSMA DI: The genetic basis of the relation
between speed-or-information-processing and IQ. Behav Brain Res, 1998, 95:7784.
36. TELLEGEN A: Note on the Tryon effect. Am Psychol, 23:585587.
37. THODAY JM: Disruptive selection. Proc R Soc Lond B, 1972, 182:109143.
38. TRYON RC: Genetic differences in maze learning ability in rats. Yrbk Natl Soc Stud
Educ, 1940, 39:111119.
39. VAN ABEELEN JHF: Ethology and the genetic foundations of animal behavior. In
JR Royce, LP Mos (Eds): Theoretical Advances in Behavior Genetics. Sijthoff &
Noordhoff, Alphen aan den Rijn, 1979, pp 101112.
40. VAN ABEELEN JHF: Genetic control of hippocampal cholinergic and dynorphiner-
gic mechanisms regulating novelty-induced exploratory behavior in house mice.
Experientia, 1989, 45:839845.
APPENDIX
CONTENTS
Summary .................................................................................................................. 55
5.1 Introduction .................................................................................................... 56
5.2 High-Resolution Mapping.............................................................................. 56
5.2.1 Congenic Strains and Substrains ....................................................... 56
5.2.2 Alternative Strategies ......................................................................... 59
5.3 Positional Cloning.......................................................................................... 62
5.3.1 Pure Positional Cloning ..................................................................... 63
5.3.2 Positional Candidate Approach ......................................................... 65
5.3.2.1 Mapping of Candidate Genes in the QTL ......................... 65
5.3.2.2 Combining QTL Mapping and Expression Profiling ........ 66
5.3.2.3 Haplotype Mapping ............................................................ 66
5.4 Conclusion...................................................................................................... 67
References................................................................................................................ 67
SUMMARY
Although quantitative trait loci (QTL) mapping analysis has become very popular,
the final goal to positionally clone and identify the relevant gene(s) remains difficult
for the genetic dissection of complex traits. This chapter reviews the different
approaches that have been used thus far. First, various breeding strategies and
haplotype mapping have been developed in order to narrow down the QTL interval.
Once high resolution mapping of the QTL is achieved, positional cloning of the
relevant gene can be performed. Until a few years ago, a contig of DNA clones
covering the QTL interval was constructed and several techniques were used in
order to extract the coding sequences of the contig. This is now obsolete in human,
mouse and rat species for which genomes are sequenced. Progress in genome
annotation and techniques based on microarrays gene expression analysis are used
to pull out candidate genes from the fine-mapped QTL interval. Finally, sequence
analysis provides candidate mutations that then need to be validated by functional
studies.
55
1903_C005.fm Page 56 Wednesday, July 26, 2006 6:34 PM
56 Neurobehavioral Genetics
5.1 INTRODUCTION
Recent progress in genetic strategies and technologies has allowed for the dissection
of complex traits in mice, rats and humans. Behavior geneticists took part in this
new research field and a high number of QTL for behavioral traits has now been
reported in mice10,12,13,24,26,40,54,61,62 and in rats.23,42,43,52,57
Many of these studies are conducted in inbred animals. The advantages are
multiple: (1) the environment is controlled, thus its effect on the phenotype is
minimized, (2) problems of genetic heterogeneity (different genes causing the same
phenotype) are avoided, (3) an adequate number of progeny and type of crosses are
available permitting robust statistical testing. The first behavioral QTL mapping
studies used recombinant inbred strains (RISs). The main advantage of using these
strains is that behavioral measures are conducted on groups of animals of each strain
thus strain means and variances become the unit of analysis. This is important for
phenotypes such as behavioral traits that are easily influenced by environmental
noise. However, because only a limited number of RISs are available from the same
panel, mapping resolution is often very low (unless the QTL effect is strong);
therefore, QTL detection using RISs is usually replicated on an F2 population for
validation. Although many successful F2 and backcross QTL mapping analyses have
been reported, very few causative gene variants have been identified (see Korstanje34
for a 2002 report), as the route from QTL detection to gene identification is long
and uncertain. This review will attempt to report various strategies that have been
used thus far for complex traits with their successes and limitations and the future
prospects of this continuously evolving discipline.
m1 m1 M1 M1
m2 m2
X M2 M2
selection m1 M1
m2 M2
F1
M1
selection M2 N2
m1 M1
m2 M2 N3
>6
m1 M1 heterozygous congenic
m2 M2
intercross
M1 M1
homozygous congenic
M2 M2
FIGURE 5.1 Congenic strains. A QTL contained in the interval delimited by the markers
M1 and M2 is introgressed from a donor strain into a recipient strain by marker-assisted
selection. N2 is the first backcross generation.
except at the QTL because at each backcross generation animals are genotyped with
genetic markers flanking the QTL. Heterozygous individuals that have retained the
donor strain alleles (for example M1 and M2) at the QTL are selected and mated
to the recipient strain thus producing the next backcross generation. To speed up the
congenic process, individuals of each generation are also genotyped with a set of
genetic markers covering the rest of the genome. This time one looks for animals
that are homozygous for the recipient alleles at every loci except the QTL. This
strategy is referred to as marker-assisted selection (MAS). Using this strategy, full
congenics could be obtained in 4 to 6 generations of backcrosses instead of eight
backcrosses if no counter-selection on the other chromosome is done.29 Once het-
erozygous full congenics (M1m1, M2m2) are obtained, they are intercrossed in order
to obtain homozygous animals (M1M1, M2M2), theoretically one fourth of the
progeny. Finally, two homozygous congenic animals are selected, based on their
1903_C005.fm Page 58 Wednesday, July 26, 2006 6:34 PM
58 Neurobehavioral Genetics
phenotypic differences with the recipient strains, and bred brothersister to fix the
donor strain alleles in the recipient strain background. Alternatively, a heterozygous
congenic male is mated to several females of the recipient strain in order to obtain
identical congenic animals to be then intercrossed. This procedure takes about two
years to complete in a mouse or rat.
Theoretically, a pure congenic strain differs from the parental strain only in the
QTL area; thus, rigorous phenotypic comparisons between congenic and parental
strains can be made such that: (1) the existence of the QTL is verified (congruent
variation in phenotype), (2) the impact of a given QTL on the phenotype can be
measured and, (3) interaction between various QTLs can be tested.
Once the existence of a QTL has been verified by the development of a congenic
strain, the QTL interval may be further refined by constructing substrains
(Figure 5.2). A congenic strain that shows the expected difference in phenotype from
the recipient strain is crossed to the latter and the offsprings are backcrossed to the
recipient strain, producing a large segregating population. Using a dense genetic
map, individuals recombinant at various places throughout the QTL region are then
selected. Again, homozygous animals for the desired crossover chromosome must
be obtained (by mating 2 heterozygotes) and these homozygotes are bred to fix the
genotype in the congenic substrain. Theoretically phenotypic analysis of these var-
ious congenic substrains should lead to a refined localization of the QTL.
There are, however, several limitations to this strategy:
M1 M1 M1 M1 M1 . m1 . m1
.
M3 . m3 M3 . m3 M3 M3 . m3
. . .
. . .
M4 . m4 . m4 . m4 M4 M4 . m4
. . .
M5 M5 . m5 . m5 M5 . m5 . m5
. .
QTL
.
. . . .
M6 M6 . m6 M6 . m6 . m6 . m6
. . .
M2 M2 M2 M2 m2 . m2 . m2
Phenotype + + - + + - -
FIGURE 5.2 Fine mapping using congenic substrains. A QTL was localized between markers
M1 and M2 in a congenic strain. Phenotypic analysis of congenic substrains recombinant
within the M1-M2 interval has allowed to fine-map the QTL to the M5-M6 chromosomal
segment.
1903_C005.fm Page 59 Wednesday, July 26, 2006 6:34 PM
60 Neurobehavioral Genetics
strain A strain B
F1 F1
F2
n intercrosses
Fn
FIGURE 5.3 Advanced intercross lines (AIL). A standard F2 is produced, then each gener-
ation is intercrossed with itself. Inbreeding is avoided as much as possible.
62 Neurobehavioral Genetics
and cumulative, integrated data from the various participating laboratories.58 Funding
for this endeavor has been obtained from the Ellison Foundation, and progress
of the Collaborative Cross and other CTC projects can be monitored at
www.complextrait.org.
Should one of the aforementioned strategies work, the next step is positional
cloning of the gene(s) involved.
Genetic map
0.28 cM
M1 M2 M3 M4 M5 M6 M7 M8 M9
Physical map
S1 S2 S3 S4 S5 S6
M4 M5
YACs
BACs
10kb
FIGURE 5.4 Construction of a contig. M1-M9 are genetic markers. S1-S6 are sequence tag
sites.
64 Neurobehavioral Genetics
In a 1995 review, Francis Collins14 made the now verified prediction that the posi-
tional candidate strategy would overtake pure positional cloning. This involves
combining fine scale mapping and survey of candidate genes within the interval. In
species not yet sequenced, efforts are underway to improve the transcript maps in
which coding sequences are localized precisely. Besides cDNA sequences already
identified and mapped, short, PCR-amplifiable partial cDNA sequences called EST
(expressed sequence tag, a subtype of STS) have been generated for this purpose
and stored in a public database (dbEST, NCBI). Fine mapping of these coding
sequences has been possible with the advent of radiation hybrid panels. Radiation
hybrid mapping is a somatic cell genetic approach in which diploid cells of the
species studied are exposed to a lethal dose of x-rays, breaking each chromosome
into small fragments. These irradiated cells are then rescued by fusion to non-
irradiated hamster cells. Each independent radiation hybrid clone retains between
15 and 20% of the irradiated cell and the retention of these random fragments of
the species chromosomes is stable without selection. The mapping strategy employs
the frequency of x-ray breakage between two markers as a statistical measure of
distance between these markers. Therefore any sequence that can be distinguished
from hamster DNA can be mapped using this approach without need for polymor-
phisms.16 These tools have been widely used in humans, mice and rats before their
genome was sequenced and remains very useful for other species.
Thus, the first step after establishing a contig, is a computer exercise in which
a query is made to select ESTs mapped in a given chromosomal subregion. These
ESTs are then placed on the contig simply by PCR amplification of the contig DNA
clones using primers specific for a given EST. A considerable amount of information
can be extracted from an EST; for example, homologies with known cDNA
sequences can be obtained (for example using the BLAST program). In turn, these
data may point to a gene already cloned or to a gene family, providing important
clues about the function of the gene and thus its relevance to the disease or mutant
phenotype. If an EST does not display homology to any known sequence, it could
be used as a hybridization probe to examine its tissue expression and to clone a
longer fragment of the cDNA in question. As outlined by Ballabio,7 the features of
candidate genesincluding sequence domains, expression pattern, imprinted expres-
sion, sequence instability (such as nucleotide triplet expansion) and developmental
expressionare compared to the disease or phenotype features in order to select
the strongest positional candidates. Finally, variants of that gene are screened for by
comparing affected individuals to controls.
An example of this approach is provided by the cloning of genes for early-onset
familial Alzheimers disease. The AD3 locus was mapped to chromosome 14 by
linkage analysis and a contig of >5 Mb was constructed around this region. Expressed
sequences encoded within the critical interval were then isolated by hybridization
of brain cDNA to the contig YAC clones. After selection based upon map location
and evolutionary conservation (see above), 19 cDNA clones were kept. The
1903_C005.fm Page 66 Wednesday, July 26, 2006 6:34 PM
66 Neurobehavioral Genetics
sequences of these cDNA clones were compared in affected and normal post-mortem
brain tissue by reverse-transcriptase PCR, leading to the identification of a novel
gene, presenilin-1 (PS-1) within the AD3 locus.53 Shortly after this study, the pres-
enilin-2 gene was identified on chromosome 1. This finding was based on the
location of an EST within the candidate region, defined by linkage studies, that has
a high sequence homology with PS-1.37
The availability of genome sequences, together with the advent of DNA microarrays
or chips, has favored the search for positional candidates through differential
expression analysis. The first time this strategy combining QTL mapping and large-
scale expression analysis was used, it was in a rat model of insulin-resistance.2 cDNA
molecules obtained from reverse-transcribed RNA extracted from adipose tissue (the
target tissue) of the two parental strains and a congenic strain for the detected QTL
were hybridized pairwise to a DNA chip of 10,000 arrayed rat cDNA. Around 70
genes were found to be differentially regulated between the parental and congenic
strains. Among them, there was a biological candidate that was then mapped to the
expected chromosome region. Further analysis of this candidate showed that it maps
at the peak of the QTL and displays functional sequence variation between the
parental strains. Similarly, -synuclein has been proposed as the causative gene of
a QTL associated to alcohol preference in a rat model following QTL mapping
combined to microarray expression analysis in various brain regions.39 In addition
to its strengths, this approach has several limitations: (1) the causative gene may not
necessarily vary in expression between the parental strains (or subject and control
in human studies); (2) the difference in expression may be too small to be detected
by microarray analysis. Indeed, this is of particular concern for behavioral traits in
which the brain is the target tissue and a small difference in gene expression may
have an important functional impact; (3) it is not always obvious which target tissue
to choose in which the causative gene will be differentially regulated; and (4) the
gene under investigation is not represented on the chip; this problem should be
resolved with progress in genome annotation and availability of pangenomic chips.
blocks mosaics derived from few original strains. This genome structure is exploited
to reduce the number of high-priority candidates genes in a QTL interval by con-
centrating on genes that are in divergent haplotype blocks between the two strains
that have served to detect the QTL. Even more rewarding is to compare haplotypes
of a QTL region detected in various crosses using different strains. Such a multiple
crossmapping strategy has been applied and combined to expression profiling and
sequence analysis, leading to the identification of a strong candidate gene for a QTL
associated with basal locomotor activity.30
5.4 CONCLUSION
Although still challenging, identification of genes underlying QTL is becoming
easier. Completion of genome sequences in human and model organisms, advances
in genomic and statistical techniques, and production of new genetic resources in
mice and rats are the main components of this progress. Nowadays, the tendency is
to set up a large-scale project integrating advanced genetic resources, in-depth
phenotyping and high-throughput genomics, such as in the Physgen project in rat
or the Collaborative Cross* in mice. Two tasks demanding a great deal of effort
remain to complete a QTL gene identification study. The first is to prove the
involvement of the candidate gene and then to find the causative mutation(s). The
best approach in model organisms is to introduce the variant allele in the control
strain by means of knock-in technology. So far, however, this is only possible in one
strain of mice and not yet in rats. Thus, as stated and detailed by others,1,28 accu-
mulation of evidence converging to a candidate gene, including functional in vitro
and in vivo studies, is generally accepted.
REFERENCES
1. Abiola O, Angel JM, Avner P, Bachmanov AA, Belknap JK, Bennett B, Blankenhorn
EP, Blizard DA, Bolivar V, Brockmann GA, Buck KJ, Bureau JF, Casley WL, Chesler
EJ, Cheverud JM, Churchill GA, Cook M, Crabbe JC, Crusio WE, Darvasi A, de
Haan G, Dermant P, Doerge RW, Elliot RW, Farber CR, Flaherty L, Flint J, Gersh-
enfeld H, Gibson JP, Gu J, Gu W, Himmelbauer H, Hitzemann R, Hsu HC, Hunter
K, Iraqi FF, Jansen RC, Johnson TE, Jones BC, Kempermann G, Lammert F, Lu L,
Manly KF, Matthews DB, Medrano JF, Mehrabian M, Mittlemann G, Mock BA,
Mogil JS, Montagutelli X, Morahan G, Mountz JD, Nagase H, Nowakowski RS,
O'Hara BF, Osadchuk AV, Paigen B, Palmer AA, Peirce JL, Pomp D, Rosemann M,
Rosen GD, Schalkwyk LC, Seltzer Z, Settle S, Shimomura K, Shou S, Sikela JM,
Siracusa LD, Spearow JL, Teuscher C, Threadgill DW, Toth LA, Toye AA, Vadasz
C, Van Zant G, Wakeland E, Williams RW, Zhang HG, Zou F. The nature and
identification of quantitative trait loci: a community's view. Nat Rev Genet 4:911916,
2003.
* At printing, the Collaborative Cross is operative and had received generous support from the Ellison
Foundation. For further information consult http://atlas.utmem.edu/pipermail/ctc/2005-October/
000284.html.
1903_C005.fm Page 68 Wednesday, July 26, 2006 6:34 PM
68 Neurobehavioral Genetics
2. Aitman TJ, Glazier AM, Wallace CA, Cooper LD, Norsworthy PJ, Wahid FN, al
Majali KM, Trembling PM, Mann CJ, Shoulders CC, Graf D, St. Lezin E, Kurtz TW,
Kren V, Pravenec M, Ibrahimi A, Abumrad NA, Stanton LW, Scott J. Identification
of Cd36 (Fat) as an insulin-resistance gene causing defective fatty acid and glucose
metabolism in hypertensive rats [see comments]. Nat Genet 21:7683, 1999.
3. Altschul SF, Lipman DJ. Protein database searches for multiple alignments. Proced
Nat Acad Sci USA 87[14], 55095513, 1990.
4. Antoch MP, Song EJ, Chang AM, Vitaterna MH, Zhao Y, Wilsbacher LD, Sangoram
AM, King DP, Pinto LH, Takahashi JS. Functional identification of the mouse circa-
dian clock gene by transgenic BAC rescue. Cell 89, 655667, 1997.
5. Bachmanov AA, Li X, Li S, Neira M, Beauchamp GK, Azen EA High-resolution
genetic mapping of the sucrose octaacetate taste aversion (Soa) locus on mouse
Chromosome 6. Mamm Genome 12:695699, 2001.
6. Bachmanov AA, Li X, Reed DR, Ohmen JD, Li S, Chen Z, Tordoff MG, de Jong
PJ, Wu C, West DB, Chatterjee A, Ross DA, Beauchamp GK. Positional cloning of
the mouse saccharin preference (Sac) locus. Chem Senses 26:925933, 2001.
7. Ballabio A. The rise and fall position cloning? Nat Genet 3:277281, 1993.
8. Bennett B. Congenic strains developed for alcohol- and drug-related phenotypes.
Pharmacol Biochem Behav 67:671681, 2000.
9. Bennett B, Beeson M, Gordon L, Carosone-Link P, Johnson TE. Genetic dissection
of quantitative trait loci specifying sedative/hypnotic sensitivity to ethanol: mapping
with interval-specific congenic recombinant lines. Alcohol Clin Exp Res
26:16151624, 2002.
10. Berrettini WH, Ferraro TN, Alexander RC, Buchberg AM, Vogel WH. Quantitative
trait loci mapping of three loci controlling morphine preference using inbred mouse
strains. Nat Genet 7:5458, 1994.
11. Burke DT, Carle GF, Olson MV. Cloning of large segments of exogenous DNA into
yeast by means of articicial chromosome vectors. Science 236:806812, 1987.
12. Caldarone B, Saavedra C, Tartaglia K, Wehner JM, Dudek BC, Flaherty L. Quanti-
tative trait loci analysis affecting contextual conditioning in mice. Nat Genet
17:335337, 1997.
13. Carr LG, Foroud T, Bice P, Gobbett T, Ivashina J, Edenberg H, Lumeng L, Li TK.
A quantitative trait locus for alcohol consumption in selectively bred rat lines. Alcohol
Clin Exp Res 22:884887, 1998.
14. Collins FS. Positional cloning moves from perditional to traditional. Nat Genet
9:347350, 1995.
15. Cowley AW, Jr., Roman RJ, Jacob HJ Application of chromosomal substitution
techniques in gene-function discovery. J Physiol 554:4655, 2004.
16. Cox DR, Burmeister M, Price ER, Kim S, Myers RM. Radiation hybrid mapping:
a somatic cell genetic method for constructing high-resolution maps of mammalian
chromosomes. Science 250:245250, 1990.
17. Darvasi A. Experimental strategies for the genetic dissection of complex traits in
animal models. Nat Genet 18:1924, 1998.
18. Demant P, Hart AA. Recombinant congenic strainsa new tool for analyzing
genetic traits determined by more than one gene. Immunogenetics 24:416422,
1986.
19. Demarest K, Koyner J, McCaughran J, Jr., Cipp L, Hitzemann R. Further character-
ization and high-resolution mapping of quantitative trait loci for ethanol-induced
locomotor activity. Behav Genet 31:7991, 2001.
1903_C005.fm Page 69 Wednesday, July 26, 2006 6:34 PM
20. Denny P, Lord CJ, Hill NJ, Goy JV, Levy ER, Podolin PL, Peterson LB, Wicker LS,
Todd JA, Lyons PA. Mapping of the IDDM locus Idd3 to a 0.35-cM interval containing
the interleukin-2 gene. Diabetes 46:695700, 1997.
21. Dukhanina OI, Dene HY, Deng AY, Choi CR, Hoebee B, Rapp JP. Linkage map and
congenic strains to localize blood pressure QTL on rat chromosome 10. Mamm
Genome 8:229235, 1997.
22. Ehringer MA, Thompson J, Conroy O, Yang F, Hink R, Bennett B, Johnson TE,
Sikela JM. Fine mapping of polymorphic alcohol-related quantitative trait loci can-
didate genes using interval-specific congenic recombinant mice. Alcohol Clin Exp
Res 26:16031608, 2002.
23. Fernandez-Teruel A, Escorihuela RM, Gray JA, Aguilar R, Gil L, Gimenez-Llort L,
Tobena A, Bhomra A, Nicod A, Mott R, Driscoll P, Dawson GR, Flint J. A quantitative
trait locus influencing anxiety in the laboratory rat. Genome Res 12:618626, 2002.
24. Flint J, Corley R, DeFries JC, Fulker DW, Gray JA, Miller S, Collins AC. A simple
genetic basis for a complex psychological trait in laboratory mice. Science
269:14321435, 1995.
25. Frankel WN, Johnson EW, Lutz CM. Congenic strains reveal effects of the epilepsy
quantitative trait locus, El2, separate from other El loci. Mamm Genome 6:839843,
1995.
26. Gershenfeld HK, Paul SM. Mapping quantitative trait loci for fear-like behaviors in
mice. Genomics 46:18, 1997.
27. Gish W, States DJ. Identification of protein coding regions by database similarity
search. Nat Genets 3:266272, 1993.
28. Glazier AM, Nadeau JH, Aitman TJ. Finding genes that underlie complex traits.
Science 298:23452349, 2002
29. Gould KA, Dietrich WF, Borenstein N, Lander ES, Dove WF. Mom1 is a semi-
dominant modifier of intestinal adenoma size and multiplicity in Min/+ mice. Genetics
144:17691776, 1996.
30. Hitzemann R, Malmanger B, Reed C, Lawler M, Hitzemann B, Coulombe S, Buck
K, Rademacher B, Walter N, Polyakov Y, Sikela J, Gensler B, Burgers S, Williams
RW, Manly K, Flint J, Talbot C. A strategy for the integration of QTL, gene expres-
sion, and sequence analyses. Mamm Genome 14:733747, 2003.
31. Ioannou PA, Amemiya CT, Garnes J, Kroisel PM, Shizuya H, Chen C, Batzer MA,
DeJong PJ. A new bacteriophage P1-derived vector for the propagation of large human
DNA fragments. Nat Genet 6:8489, 1994.
32. Iraqi F, Clapcott SJ, Kumari P, Haley CS, Kemp SJ, Teale AJ. Fine mapping of
trypanosomiasis resistance loci in murine advanced intercross lines. Mamm Genome
11:645648, 2000.
33. King DP, Zhao Y, Sangoram AM, Wilsbacher LD, Tanaka M, Antoch MP, Steeves
TDL, Vitaterna MH, Kornhauser JM, Lowrey PL, Turek FW, Takahashi JS. Positional
cloning of the mouse circadian clock gene. Cell 89:641653, 1997.
34. Korstanje R, Paigen B. From QTL to gene: the harvest begins. Nat Genet 31:235236,
2002.
35. Lander E, Kruglyak L. Genetic dissection of complex traits: guidelines for interpreting
and reporting linkage results. Nat Genet 11:241247, 1995.
36. Legare ME, Bartlett FS, Frankel WN. A major effect QTL determined by multiple
genes in epileptic EL mice. Genome Res 10:4248, 2000.
37. Levy-Lahad E, Wasco W, Poorkaj P, Romano DM, Oshima J, Pettingell CY, Jondro
PD, Schmidt SD, Wang K, Crowley AC, Fu YH, Guenette SY, Galas D, Nemens E,
1903_C005.fm Page 70 Wednesday, July 26, 2006 6:34 PM
70 Neurobehavioral Genetics
Wijsman EM, Bird TD, Schellenberg GD, Tanzi RE. Candidate gene for the chro-
mosome 1 familial Alzheimers disease locus. Science 269:973977, 1995.
38. Li X, Inoue M, Reed DR, Huque T, Puchalski RB, Tordoff MG, Ninomiya Y,
Beauchamp GK, Bachmanov AA. High-resolution genetic mapping of the saccharin
preference locus (Sac) and the putative sweet taste receptor (T1R1) gene (Gpr70) to
mouse distal Chromosome 4. Mamm Genome 12:1316, 2001.
39. Liang T, Spence J, Liu L, Strother WN, Chang HW, Ellison JA, Lumeng L, Li TK,
Foroud T, Carr LG. alpha-Synuclein maps to a quantitative trait locus for alcohol
preference and is differentially expressed in alcohol-preferring and -nonpreferring
rats. Proc Natl Acad Sci U S A 100:46904695, 2003.
40. Melo JA, Shendure J, Pociask K, Silver LM. Identification of sex-specific quantitative
trait loci controlling alcohol preference in C57BL/6 mice. Nat Genet 13:147153,
1996.
41. Moen CJ, Groot PC, Hart AA, Snoek M, Demant P. Fine mapping of colon tumor
susceptibility (Scc) genes in the mouse, different from the genes known to be somat-
ically mutated in colon. Proc Natl Acad Sci U S A 93:10821086, 1996.
42. Moisan MP, Courvoisier H, Bihoreau MT, Gauguier D, Hendley ED, Lathrop M,
James MR, Mormde P. A major quantitative trait locus influences hyperactivity in
the WKHA rat. Nat Genet 14:471473, 1996.
43. Moisan MP, Llamas B, Cook MN, Mormede P. Further dissection of a genomic locus
associated with behavioral activity in the Wistar-Kyoto hyperactive rat, an animal
model of hyperkinesis. Mol Psychiatry 8:348352, 2003.
44. Monaco AP. Isolation of genes from cloned DNA. Current opinion in Genetics and
Development 4:360365, 1994.
45. Nadeau JH, Singer JB, Matin A, Lander ES. Analysing complex genetic traits with
chromosome substitution strains. Nat Genet 24:221225, 2000.
46. Olson M, Hood L, Cantor C, Botstein D. A common language for physical mapping
of the human genome. Science 245:14341435, 1989.
47. Parimoo S, Patanjali SR, Kolluri R, Xu H, Wei H, Weissman SM. cDNA selection
and other approaches in positional cloning. Analyt Biochem 228:117, 1995.
48. Parrish JE, Nelson DL. Methods for finding genes a major rate-limiting step in
positional cloning. GATA 10:2941, 1993.
49. Peirce JL, Lu L, Gu J, Silver LM, Williams RW. A new set of BXD recombinant
inbred lines from advanced intercross populations in mice. BMC Genet 5:7, 2004.
50. Podolin PL, Denny P, Armitage N, Lord CJ, Hill NJ, Levy ER, Peterson LB, Todd
JA, Wicker LS, Lyons PA. Localization of two insulin-dependent diabetes (Idd) genes
to the Idd10 region on mouse chromosome 3. Mamma Genome 9:283286, 1998.
51. Podolin PL, Denny P, Lord CJ, Hill NJ, Todd JA, Peterson LB, Wicker LS, Lyons
PA. Congenic mapping of the insulin-dependent diabetes (Idd) gene, Idd10, localizes
two genes mediating the Idd10 effect and eliminates the candidate Fcgr1. J Immunol
159:18351843, 1997.
52. Ramos A, Moisan MP, Chaouloff F, Mormede P. Identification of female-specific
QTLs affecting an emotionality-related behavior in rats. Mol Psychiatry 4:453462,
1999.
53. Sherrington R, Rogaev EI, Liang Y, Rogaeva EA, Levesque G, Ikeda M, Chi H, Lin
C, Li G, Holman K, Tsuda T, Mar L, Foncin JF, Bruni AC, Montesi MP, Sorbi S,
Rainero I, Pinessi L, Nee L, Chumakov I, Pollen D, Brookes A, Sanseau P, Polinsky
RJ, Wasco W, DaSilva HAR, Haines JL, Pericak-Vance MA, Tanzi RE, Roses AD,
Fraser PE, Rommens JM, St George-Hyslop PH. Cloning of a gene bearing missense
mutations in early-onset familial Alzheimer's disease. Nature 375:754760, 1995.
1903_C005.fm Page 71 Wednesday, July 26, 2006 6:34 PM
54. Shimomura K, Low-Zeddies SS, King DP, Steeves TD, Whiteley A, Kushla J,
Zemenides PD, Lin A, Vitaterna MH, Churchill GA, Takahashi JS. Genome-wide
epistatic interaction analysis reveals complex genetic determinants of circadian behav-
ior in mice. Genome Res 11:959980, 2001.
55. Shizuya H, Birren B, Kim UJ, Mancino V, Slepak T, Tachiiri Y, Simon M. Cloning
and stable maintenance of 300-kilobase-pair fragments of human DNA in Escherichia
coli using an F-factor-based vector. Proc Nat Acad Sci USA 89:87948797, 1992.
56. Talbot CJ, Nicod A, Cherny SS, Fulker DW, Collins AC, Flint J. High-resolution
mapping of quantitative trait loci in outbred mice. Nat Genet 21:305308, 1999.
57. Terenina-Rigaldie E, Moisan MP, Colas A, Beauge F, Shah KV, Jones BC, Mormede
P. Genetics of behaviour: phenotypic and molecular study of rats derived from high-
and low-alcohol consuming lines. Pharmacogenetics 13:543554, 2003.
58. Threadgill DW, Hunter KW, Williams RW. Genetic dissection of complex and quan-
titative traits: from fantasy to reality via a community effort. Mamm Genome
13:175178, 2002.
59. Uberbacher EC, Mural RJ. Locating protein-coding regions in human DNA sequences
by a multiple sensor-neural network approach. Proc Natl Acad Sci USA
88:1126111265, 1991.
60. Wall JD, Pritchard JK. Haplotype blocks and linkage disequilibrium in the human
genome. Nat Rev Genet 4:587597, 2003.
61. Wehner JM, Radcliffe RA, Rossmann ST, Christensen SC, Rasmussen DL, Fulker
DW, Wiles M. Quantitative trait locus analysis of contextual fear conditioning in
mice. Nat Genet 17:331334, 1997.
62. Yoshikawa T, Watanabe A, Ishitsuka Y, Nakaya A, Nakatani N. Identification of
multiple genetic loci linked to the propensity for "behavioral despair" in mice.
Genome Res 12:357366, 2002.
63. Zhang QY, Dene HY, Deng AY, Garrett MR, Jacob HJ, Rapp JP. Interval mapping
and congenic strains for a blood pressure QTL on rat chromosome 13. Mamm Genome
8:636641, 1997.
1903_C005.fm Page 72 Wednesday, July 26, 2006 6:34 PM
1903_C006.fm Page 73 Wednesday, July 26, 2006 6:35 PM
6 Gene Expression
Richard A. Radcliffe
CONTENTS
Introduction.............................................................................................................. 73
6.1 The Gene ........................................................................................................ 74
6.2 Transcription: Basic Mechanism of mRNA Biogenesis ............................... 76
6.3 Regulation of Transcription ........................................................................... 80
6.4 Transcriptional Regulation through Signal Transduction ............................. 83
6.5 RNA Interference ........................................................................................... 85
6.6 Gene Silencing ............................................................................................... 86
6.7 Sources of Variation in Gene Expression...................................................... 87
6.8 Measuring Gene Expression .......................................................................... 88
References................................................................................................................ 92
INTRODUCTION
Gene expression, or transcription, is an important element of The Central Dogma,
the phrase used by Francis Crick1 to describe what was and still is thought to be the
basic and only flow of genetic information in a cell: DNA is transcribed to messenger
RNA (mRNA) which is then translated to protein. In this scheme, only nucleic acids
actually contain information that is encoded in the specific sequence of nucleotides
that make up the DNA or RNA. Cells can construct proteins from the information
contained in nucleic acids, but the transfer of information is one way; cells do not
have the capacity to reproduce nucleic acid sequence information from proteins
alone. While it is known that RNA performs more than just information functions,2
proteins are still thought to be the primary working molecules in a cell carrying out
diverse functions related to energy management, immune response, structure, trans-
port and storage, signal transduction, cell division, etc. Thus, implicit in the Central
Dogma are mechanisms for the tightly controlled regulation of mRNA and thus
protein amount, one of several mechanisms through which the cell maintains strict
control over the final activity of a protein. On a grander scale, the integrated
transcriptional activity of the entire genome (the entire DNA sequence in an organ-
ism) to a great extent determines the nature of a cell and its interactions with other
cells in an organism or with the environment.
Transcription is an essential mechanism throughout the development of a mul-
ticellular organism. The coordinate induction or repression of the thousands of genes
found in the organisms genome results in a wide array of cell phenotypes through
73
1903_C006.fm Page 74 Wednesday, July 26, 2006 6:35 PM
74 Neurobehavioral Genetics
the process of cell differentiation. Mature cells, including highly specialized post-
mitotic cells such as neurons, continue to make use of transcription throughout their
life. One reason for this is that active transcription is required to replace proteins
all of which have a finite duration, at least to the extent that is known, and therefore
must be replaced throughout the life of the cell. Cells also need to adapt to a dynamic
environment necessitating variable levels of some proteins. Environmental adapta-
tion through transcription is perhaps most highly developed in neurons of the central
nervous system (CNS) in which one of the most important and interesting functions
of the CNS takes place: learning and memory.3 The CNS is constantly being bom-
barded with sensory information, some of which is critical and must be dealt with
immediately, but most of which is extraneous and can be ignored. In either case,
processes of learning and memory aid the organism in distinguishing into which
category the incoming information belongs. These processes cannot take place in
the absence of a functioning transcription apparatus.
The fundamentals of transcription are essentially the same in all eukaryotic cells,
but the timing and nature of the regulatory events that determine which genes are
actively being transcribed vary from gene to gene. This chapter describes the fun-
damentals of transcription as well as the basic principles of transcriptional regulation.
Also included is a discussion of sources of variation in mRNA levels and a brief
description of some standard methods for measuring mRNA amounts from cell or
tissue samples.
Gene Expression 75
5GTGTATAAAATGATTGTT 5GTACCATTCCGAATGTAT
3CACATATTTTACTAACAA 3CATGGTAAGGCTTACATA
TATA Box Initiator
Coding Region
(100s -1000s bp)
1
H3 C
2
H3 C
5 5
Mature mRNA 3 HO
P-P-P- AAAAA
FIGURE 6.1 Basic structure and transcription of a hypothetical gene. Two binding elements
are shown, the TATA box and the initiator, along with their complementary DNA strand
(sequence on either side of the element is arbitrary). During transcription, the 5' end of the
nascent mRNA is capped with methyl-guanine and the 3' end is polyadenylated (step 1).
Introns are removed from the pre-mRNA in the final processing step toward the mature mRNA
molecule (step 2). In a very small number of cases, the mRNA is further processed with RNA
editing (not shown). Transcription takes place in the nucleus followed by transport of the
mature mRNA out of the nucleus to other cellular compartments where translational occurs.
coding region. The regulatory region can be further subdivided into the promoter
(the location on the DNA where the transcriptional machinery binds) and the regu-
lator binding sites (individual sites on the DNA to which proteins bind; also called
binding elements). Regulator binding sites that promote transcription are often clus-
tered in regions known as enhancers. Similarly, clusters of binding sites that suppress
transcription when bound by regulatory proteins are called silencers.6
The basis for the specificity of interactions between nucleic acids or between
nucleic acids and proteins is in the linear sequence of the four nucleotide bases that
are found in double-stranded DNA or single-stranded RNA: adenine, guanine,
cytosine, and thymine (uracil in RNA). This was the fundamental discovery made
by Watson and Crick some 50 years ago.7 It is the unidirectional order of these bases
(specified with relation to the 5' and 3' phosphodiester linkages between ribose or
deoxyribose sugar molecules in the RNA or DNA backbone) and their two- and
three-dimensional relationships to one another that contain specific information
about the structure of proteins and their temporal and spatial expression. Many
different types of regulatory and enzymatic proteins are able to read the nucleic
1903_C006.fm Page 76 Wednesday, July 26, 2006 6:35 PM
76 Neurobehavioral Genetics
Gene Expression 77
Preinitiation
complex
Nucleosome
modifiers
1
Activators
100s -1000s bp
2 3
PPP PPP
Pre-mRNA
5
5
FIGURE 6.2 The molecular machinery of transcription. General transcription factors facil-
itate the binding of Pol II onto the promoter; this is the preinitiation complex. In vivo, many
other components are needed to initiate transcription including the mediator complex, nucleo-
some modifiers, and activators (step 1). Phosphorylation of the Pol II tail by TFIIH causes it
to be released from the preinitiation complex and begin transcription (step 2). Sequence signals
in the transcript cause it to be released from Pol II and polyadenylated, thus terminating
transcription (step 3).
78 Neurobehavioral Genetics
TABLE 6.1
Common Promoter Elements
Binding Element
Consensus Transcription
Promoter Location Positiona Sequenceb Factorc
Initiator Core promoter element +1 PyPyAN(T/A)PyPy TBP
TATA box Core promoter element 35 to 20 TATAAA TBP
CAAT box Promoter-proximal element 200 to 70 CCAAT CBF, NF1,
C/EBP
GC box Promoter-proximal element 200 to 70 GGGCGG SP1
a Number of nucleotides from the start site with negative numbers indicating in the upstream direction;
i.e., toward the 5 end.
bThe consensus sequence is the ideal sequence for the most effective interaction with its factor; binding
elements can often tolerate slight differences from the consensus. Py = pyrimidine (C or T); N = any
(A, C, T, G).
c TBP: TATA binding protein; CBF: CAAT binding protein; NF1: nuclear factor 1; C/EBP:
that facilitate the binding of the preinitiation complex to the core promoter. These
factors, which include the mediator, a complex of more than 20 subunits, regulatory
proteins, and nucleosome-modifying proteins, are engaged in various activities
including covalent modification of histones and Pol II recruitment (see below).10,11
Once the preinitiation complex has formed at the core promoter, Pol II begins
transcription and attempts to separate itself from the complex. Several short tran-
scripts are synthesized, but the polymerase cannot enter into the elongation phase
until its C-terminal tail, which is composed of a long string of Tyr, Ser, Pro, and
Thr residues, is phosphorylated by the kinase activity of the factor TFIIH and other
kinases. Once the tail has been phosphorylated to a certain level, affinity of Pol II
for the general transcription factors decreases considerably allowing the polymerase
to move away from the core promoter and begin synthesizing a full-length transcript
during the elongation phase. Phosphorylation and dephosphorylation of the C-
terminal tail are important regulatory mechanisms of Pol II activity.1113
As Pol II enters into the elongation phase, the general transcription factors are
replaced with a new set of factors that bind favorably to the phosphorylated form
of the C-terminal tail. The new factors fulfill three basic functions: (1) promote
continued RNA synthesis by stimulating Pol II activity; (2) proofread and correct
the nascent RNA to ensure exact complementarity to the DNA template; and (3)
process the emerging RNA molecule with several types of covalent modifications
primarily for stabilization. Many of these factors bind to the C-terminal tail of
Pol II. As the RNA is synthesized, its 5 end emerges from the polymerase near
the C-terminal tail. Thus, the factors are in the ideal position for gaining access to
the newly synthesized RNA.1113
1903_C006.fm Page 79 Wednesday, July 26, 2006 6:35 PM
Gene Expression 79
P-TEFb is a central factor recruited to the C-terminal tail early in the elongation
phase. Along with other factors, it stimulates the elongation of the RNA molecule
by Pol II. P-TEFb has inherent kinase activity and it phosphorylates additional
residues on the Pol II tail. This activity aids in the recruitment of the elongation
factors hSPT5, TAT-SF1, and TFIIS, all of which promote the continued elongation
of the transcript. Pol II does not synthesize RNA at a constant rate; rather, the rate
is dependent on the specific sequence being transcribed and some sequences will
cause the polymerase to come to a near complete stop. One of the functions of TFIIS
is to smooth the progress of Pol II through these more involved sequences so that
it progresses at a faster overall rate. Combined with the nuclease activity of Pol II,
TFIIS also aids in the detection and replacement of mismatched nucleotides. This
proofreading capability is not as efficient as that which occurs during DNA synthesis.
Approximately 1 in 107 mismatched nucleotides escape detection during DNA rep-
lication, but transcription is somewhat less accurate to approximately 1 in 104 to 105
errors are made during transcription. Of course, the consequences of a DNA mutation
are potentially much more severe than the occasional single mutant protein translated
from a mismatched mRNA sequence.1113
When the new transcript reaches a length of 20 to 40 nucleotides, the 5 end is
modified with the addition of a methylated guanine, known as capping. Guanine is
added in an unusual 5-5 triphosphate linkage and then a methyl group is added to
the nitrogen in the 7 position of the guanine moiety (see Figure 6.2). The cap serves
as a signal for other processes including recruitment of the translation machinery.
As the transcript reaches full-length, a second processing step takes place, polyade-
nylation, that coincides with termination of transcription. CPSF and CstF are proteins
that have been recruited to the phosphorylated tail of Pol II to separate the nascent
mRNA from the RNA that is still being made and to facilitate polyadenylation.
Transcription of the poly-A signal sequence promotes release of CPSF and CstF
from the Pol II onto the new transcript. These factors cause the mRNA to be cleaved
from the still growing RNA strand and recruit poly-A polymerase which enzymati-
cally adds approximately 200 adenine molecules (from ATP) to the 3 end of the
transcript. Pol II continues to synthesize complementary RNA sometimes as long
as several hundred nucleotides, but eventually Pol II dissociates from the DNA and
releases the aberrant RNA molecule which is rapidly degraded. The signals that
cause dissociation of the polymerase may have to do with either the absence of a
5 guanine cap and/or with the transfer of CPSF and CstF from the tail of Pol II,
but the details of this have not been entirely worked out.12,13
For most eukaryotic genes, the primary transcript, or pre-mRNA, requires one
additional processing step, splicing, before it is mature and ready for the translation
of protein. The pre-mRNA usually consists of alternating segments known as exons
and introns that specify coding and noncoding portions of the gene, respectively.
All of the exons in a gene (more than 300 in some cases) must be joined in a
contiguous fashion in order for the transcript to be appropriately translated into the
protein product. Thus, introns are removed and exons are spliced together in a
reaction catalyzed by a large protein-RNA complex known as the spliceosome. The
spliceosome is directed by specific sequences at the exon/intron border which guide
the splicing reaction. Some primary transcripts have the capacity to be spliced in
1903_C006.fm Page 80 Wednesday, July 26, 2006 6:35 PM
80 Neurobehavioral Genetics
Gene Expression 81
Transcription factors can either facilitate or suppress transcription and are either
constitutive or inducible. Constitutive transcription factors are present in the cell at
all times and are typically activated through signal transduction pathways. Inducible
transcription factors are not usually present in the cell, but are transcribed and then
translated following some kind of intra- or extracellular signal. Positive transcription
factors are known as activators; negative factors are called repressors. Transcription
factors bind with high specificity to specific DNA sequences in the promoter or
elsewhere in the regulatory region of genes and interact with other proteins to
influence transcription. The proteinprotein interactions are of two general types:
recruitment, in which a protein facilitates the cooperative binding of one or more
other proteins to a specific genomic location; and allostery, in which one protein
triggers a conformational change in another eliciting some kind of functional mod-
ification in the second protein. Often both types of interactions take place. In many
cases, DNA-binding proteins bind to adjacent sites on the DNA, but this is not always
true. Interacting proteins may bind at very long distances from one another and are
able to make physical contact because of DNAs ability to loop back on itself. Indeed,
there are specific DNA-binding proteins that facilitate the bending of DNA so that
distant sites can be brought in close proximity to one another.19,20
Transcription factors typically have distinct DNA-binding domains and protein-
binding domains. The DNA-binding domains have high affinity for specific DNA
nucleotide sequences and often bind to DNA as dimers with an -helix functioning as
the DNA recognition site. Other parts of the DNA-binding domain interact with the
DNA backbone to align the DNA recognition helix in the proper position in the major
groove of DNA. Some of the more common DNA-binding domains found in transcrip-
tion factors include homeodomain proteins, zinc-containing DNA-binding domains
(includes zinc finger proteins), leucine zippers, and helix-loop-helix proteins.21
One of the important functions of activator proteins is to recruit Pol II to the
promoter. Typically this is not a direct interaction with Pol II, but rather recruitment
takes place indirectly through interactions with preformed polymerase-associated
protein complexes, such as mediator or TFIID, which are often already associated
with Pol II. Often the concerted action of many activators recruits the transcriptional
machinery. For a simple example, consider two distinct activators that contact medi-
ator at different sites. Either one alone provides insufficient binding energy, but the
combined binding energy of the two is able to effectively recruit the complex.
Transcriptional initiation often requires the presence of many activators that work
together in such an integrated manner. Each gene does not have its own specific
activator; rather, the combined activity of a group of activators present in the right
combination and at the right time is required for gene expression. This kind of
combinatorial control is a key element for the appropriate cell-, time-, and response-
dependent expression of many genes.8
The DNA in chromatin is generally inaccessible for transcription because pro-
moters are tightly bound by histone proteins. Some activators recruit specific proteins
that alter nucleosomes so that the transcriptional machinery can access the gene.
This occurs in one of two ways: nucleosome remodeling or through covalent mod-
ifications. Activators recruit nucleosome remodeling complexes that change the
nature of DNA-histone interactions. These complexes can free the promoter either
1903_C006.fm Page 82 Wednesday, July 26, 2006 6:35 PM
82 Neurobehavioral Genetics
by moving the position of the histone relative to the promoter or by causing the
promoter to become less tightly bound to the histone. Either situation promotes the
binding of Pol II and its associated proteins. The second type of alteration involves
reversible covalent modifications to histone proteins. Activators recruit histone acety-
lases that add acetyl groups to the N-terminal tails of histones. The tails are rich in
lysine residues whose positive charge normally facilitates tight binding of the histone
to the negatively charged phosphate portion of the DNA backbone. Acetylation
effectively neutralizes the tails positive charge causing the nucleosome structure to
relax facilitating the binding of the transcriptional machinery to the promoter. In
addition, certain proteins and protein complexes such as TFIID have specific domains
known as bromodomains that bind with high affinity to acetylated histones further
promoting the initiation of transcription. Remodeling and histone modification typ-
ically act in combination to promote transcription, sometimes the activity of one
enhancing that of the other.5,22
Histone tails can also be modified with the addition of methyl groups by histone
methylases. This can occur in an analogous manner as acetylation, but histone
methylation is usually associated with inhibition of transcription. Thus, chromo-
domain-containing proteins interact with methylated histones to prevent transcription
by inhibiting the binding of transcriptional activators. Methylation also recruits
proteins with methylase activity that propagate methylation locally to unmodified
histones (see below). The covalent modification of histone tails is a dynamic process
and it is believed that it is the overall pattern of histone modifications that is an
important factor in transcriptional regulation.5,22
As a result of the integrated activity of many activators and the proteins with
which they interact, only certain genes are transcribed at any given time. Transcrip-
tional repressors also contribute to the transcriptional status of a gene. In general,
repressors behave very similarly to activators, i.e., they bind with high affinity to
specific DNA sequences and also bind to other proteins, but with the end result
being the inhibition of initiation. A repressor can bind to a specific DNA site allowing
it to interact with the transcriptional machinery thereby preventing initiation through,
for example, allosteric modulation. Repressors can also inhibit transcription more
indirectly by preventing the binding of activators either to DNA or to the proteins
with which the activators interact. Repressors can also prevent transcription through
the recruitment of histone deacetylases. These enzymes remove acetyl groups from
the N-terminal tails of histones thereby restoring nucleosome structure.23
Messenger RNA is ultimately degraded through the sequential activity of specific
decay enzymes found localized to discrete cytoplasmic bodies. The initiating step
of decay is the deadenylation of the mRNA. From there, it proceeds through one of
two known pathways that end in its digestion by the activity of either 53 or
35 exonucleases. As might be expected, the decay of mRNA is highly regulated,
a consequence of which is that the mRNA for constitutively expressed genes tends
to have a longer half-life than for genes that are expressed in response to external
signals. The signal for degradation is generally, but not exclusively, found in the 3
untranslated region (3UTR) of the mature mRNA. There are specific elements in
the 3UTR whose bound regulatory proteins recruit the decay enzymes. Conversely,
certain regulatory proteins are able to stabilize an mRNA, prolonging its life in the
1903_C006.fm Page 83 Wednesday, July 26, 2006 6:35 PM
Gene Expression 83
84 Neurobehavioral Genetics
1
R AC
G G ATP Cell membrane
PKA
4
ADP ATP
P 5
FIGURE 6.3 Activation of CREB through stimulation of a typical signal transduction pathway.
The process is initiated by the binding of a small molecule ligand to a cell surface receptor
(R) embedded in the cell membrane (step 1). A subsequent conformational change in the
receptor causes guanosine-diphosphate (GDP) to be replaced with guanosine-triphosphate
(GTP) on the trimeric G-protein (G) and its GTP-containing subunit to become dissociated
(step 2). The /GTP complex activates the enzyme adenylyl cyclase (AC) which catalyzes the
formation of cyclic-AMP (cAMP) from ATP. cAMP triggers the dissociation of the catalytic
subunit of protein kinase A (PKA) which then diffuses into the nucleus and phosphorylates
CREB (steps 3 and 4). Phosphorylated CREB now binds to CRE on the DNA and promotes
transcription through its interactions with the preinitiation complex and CREB binding protein
(CBP) (step 5).
Gene Expression 85
86 Neurobehavioral Genetics
Gene Expression 87
88 Neurobehavioral Genetics
Gene Expression 89
90 Neurobehavioral Genetics
Gene Expression 91
92 Neurobehavioral Genetics
of tools and the scientists trained to use them that will start to really tap into the
power of this experimental strategy.
REFERENCES
1. Crick, F.H.C., On protein synthesis, in Symposium of the Society for Experimental
Biology XII, Academic Press, New York, 1958, 153.
2. Cech, T.R., The efficiency and versatility of catalytic RNA: implications for an RNA
world, Gene, 135, 33, 1993.
3. Kaczmarek, L., Molecular biology of vertebrate learning: is c-fos a new beginning?,
J. Neurosci. Res., 34, 377, 1993.
4. Belmont, A.S., Dietzel, S., Nye, A.C., Strukov, Y.G., and Tumbar, T., Large-scale
chromatin structure and function, Curr. Opin. Cell Biol., 11, 307, 1999.
5. Wolffe, A.P., Chromatin structure and the regulation of transcription, in Transcription
Factors, Locker, J., Ed., BIOS Scientific Publishers Ltd., Oxford, 2001, chap. 3.
6. Watson, J.D. et al., Molecular Biology of the Gene, 5th ed., Cold Spring Harbor
Laboratory Press, San Francisco, 2004.
7. Watson, J.D. and Crick, F.H.C., Genetical implications of the structure of deoxyri-
bonucleic acid, Nature, 171, 964, 1953.
8. Courey, A.J., Regulatory transcription factors and cis-regulatory regions, in Tran-
scription Factors, Locker, J., Ed., BIOS Scientific Publishers Ltd., Oxford, 2001,
chap. 2.
9. Garvie, C.W. and Wolberger, C., Recognition of specific DNA sequences, Mol. Cell,
8, 937, 2001.
10. Young, B.A., Gruber, T.M., and Gross, C.A., Views of transcription initiation, Cell,
109, 417, 2002.
11. Pugh, B.F., RNA polymerase II transcription machinery, in Transcription Factors,
Locker, J., Ed., BIOS Scientific Publishers Ltd., Oxford, 2001, chap. 1.
12. Maniatis, T. and Reed R., An extensive network of coupling among gene expression
machines, Nature, 416, 499, 2002.
13. Howe, K.J., RNA polymerase II conducts a symphony of pre-mRNA processing
activities, Biochim. Biophys. Acta, 1577, 308, 2002.
14. Schmauss, C. and Howe, J.R., RNA editing of neurotransmitter receptors in the
mammalian brain, Sci. STKE, 2002, pe26, 2002.
15. Sommer, B. et al., RNA editing in brain controls a determinant of ion flow in
glutamate-gated channels, Cell, 67, 11, 1991.
16. Brusa, R. et al., Early-onset epilepsy and postnatal lethality associated with an editing-
deficient GluR-B allele in mice, Science, 270, 1677, 1995.
17. Job, C. and Eberwine, J., Localization and translation of mRNA in dendrites and
axons, Nat. Rev. Neurosci., 2, 889, 2001.
18. Anderson, L. and Seilhammer. J., A comparsion of selected mRNA and protein
abundances in human liver, Electrophoresis, 18, 533, 1997.
19. Lefstin, J.A. and Yamamoto, K.R., Allosteric effects of DNA on transcriptional
regulators, Nature, 392, 885, 1998.
20. Ptashne, M. and Gann, A., Transcriptional activation by recruitment, Nature, 386,
569, 1997.
21. Fairall, L. and Schwabe, J.W.R., DNA binding by transcription factors, in Transcrip-
tion Factors, Locker, J., Ed., BIOS Scientific Publishers Ltd., Oxford, 2001, chap. 4.
1903_C006.fm Page 93 Wednesday, July 26, 2006 6:35 PM
Gene Expression 93
22. Fry, C.J. and Peterson, C.L., Chromatin remodeling enzymes: who's on first?, Curr.
Biol., 11, R185, 2001.
23. Maldonado, E., Hampsey, M., and Reinberg, D., Repression: targeting the heart of
the matter, Cell, 99, 455, 1999.
24. Wilusz, C.J. and Wilusz, J., Bringing the role of mRNA decay in the control of gene
expression into focus, Trends Genet., 20, 491, 2004.
25. Shore, P. and Sharrocks, A.D., Regulation of transcription by extracellular signals,
in Transcription Factors, Locker, J., Ed., BIOS Scientific Publishers Ltd., Oxford,
2001, chap. 6.
26. Mayr, B. and Montminy, M., Transcriptional regulation by the phosphorylation-
dependent factor CREB, Nat. Rev. Mol. Cell Biol., 2, 599, 2001.
27. Lonze, B.E. and Ginty, D.D., Function and regulation of CREB family transcription
factors in the nervous system, Neuron, 35, 605, 2002.
28. Waterhouse, P.M., Wang, M.B., and Lough, T., Gene silencing as an adaptive defence
against viruses, Nature, 411, 834, 2001.
29. Bartel, D.P., MicroRNAs: genomics, biogenesis, mechanism, and function, Cell, 116,
281, 2004.
30. Lim, L.P. et al., Vertebrate microRNA genes, Science, 299, 1540, 2003.
31. Ambros, V., The functions of animal microRNAs, Nature, 431, 350, 2004.
32. Nelson, P. et al., The microRNA world: small is mighty, Trends Biochem. Sci., 28,
534, 2003.
33. Meister, G. and Tuschl, T., Mechanisms of gene silencing by double-stranded RNA,
Nature, 431, 343, 2004.
34. Gartenberg, M.R., The Sir proteins of Saccharomyces cerevisiae: mediators of tran-
scriptional silencing and much more, Curr. Opin. Microbiol., 3, 132, 2000.
35. Gottschling, D.E., Gene silencing: two faces of SIR2, Curr. Biol., 10, R708, 2000.
36. Martienssen, R.A. and Colot, V., DNA methylation and epigenetic inheritance in
plants and filamentous fungi, Science, 293, 1070, 2001.
37. Richards, E.J. and Elgin, S.C., Epigenetic codes for heterochromatin formation and
silencing: rounding up the usual suspects, Cell, 108, 489, 2002.
38. Tilghman, S.M., The sins of the fathers and mothers: genomic imprinting in mam-
malian development, Cell, 96, 185, 1999.
39. Ryabinin, A.E., ITF mapping after drugs of abuse: pharmacological versus percep-
tional effects, Acta Neurobiol. Exp., 60, 547, 2000.
40. Graybiel, A.M., Moratalla, R., and Robertson H.A., Amphetamine and cocaine induce
drug-specific activation of the c-fos gene in striosome-matrix compartments and
limbic subdivisions of the striatum, P. Nat. Acad. Sci. U.S.A., 87, 6912, 1990.
41. Hitzemann, B. and Hitzemann, R., Genetics ethanol and the Fos response: a compar-
ison of the C57BL/6J and DBA/2J inbred mouse strains, Alcohol. Clin. Exp. Res.,
21, 1497, 1997.
42. Ryabinin, A.E. et al., Differential sensitivity of c-Fos expression in hippocampus and
other brain regions to moderate and low doses of alcohol, Mol. Psychiatr., 2, 32, 1997.
43. Jouvert, P. et al., Differential rat brain expression of EGR proteins and of the tran-
scriptional corepressor NAB in response to acute or chronic cocaine administration.
Neuromol. Med., 1, 137, 2002.
44. Nestler, E.J., Molecular basis of long-term plasticity underlying addiction, Nat. Rev.
Neurosci., 2, 119, 2001.
45. Monti, J. et al., Expression analysis using oligonucleotide microarrays in mice lacking
bradykinin type 2 receptors, Hypertension, 38, E1, 2001.
1903_C006.fm Page 94 Wednesday, July 26, 2006 6:35 PM
94 Neurobehavioral Genetics
7 Bioinformatics of Behavior
Elissa J. Chesler
CONTENTS
Summary .................................................................................................................. 96
7.1 What Is Bioinformatics? ................................................................................ 96
7.2 Database Design............................................................................................. 97
7.2.1 Data Storage beyond the Spreadsheet: Laboratory Information
Management Systems (LIMS) ........................................................... 97
7.2.2 Data Modeling.................................................................................... 98
7.2.2.1 The Flat Format .................................................................. 98
7.2.2.2 The Hierarchical Format..................................................... 98
7.2.2.3 The Relational Database Management System (RDBMS) 98
7.3 Natural Keys in Biological Databases ....................................................... 99
7.3.1 Genome Location as Reference....................................................... 100
7.3.2 Genes as References ........................................................................ 100
7.3.3 Genetic Reference Populations........................................................ 101
7.3.3.1 Genetic Correlations and Reference Populations ............ 102
7.3.3.2 GRPs for Mapping QTLs ................................................. 103
7.3.4 Gene Sets as References .................................................................. 104
7.3.4.1 Gene Ontology.................................................................. 104
7.3.4.2 The Interactome ................................................................ 104
7.4 Applications ................................................................................................. 104
7.4.1 Sequence Analysis Scoring Matrices and Phylogeny ..................... 105
7.4.1.1 Scoring Matrices............................................................... 106
7.4.1.2 Motif Search and Alignment............................................ 106
7.4.1.3 Structure Analysis............................................................. 106
7.4.2 QTL Candidate Gene Selection....................................................... 107
7.4.3 Microarray, Proteomic, and Other High-Throughput Gene
Set Analysis...................................................................................... 108
7.4.4 Text Mining...................................................................................... 109
7.4.5 Integrating the Genome and the Phenome for Systems-Level
Bioinformatics.................................................................................. 109
7.5 Toward a Bioinformatics of Behavior ......................................................... 110
Acknowledgments.................................................................................................. 111
References.............................................................................................................. 111
95
1903_C007.fm Page 96 Wednesday, July 26, 2006 6:35 PM
96 Neurobehavioral Genetics
SUMMARY
This chapter is intended to provide a brief introduction to biological databases for
two major purposesfirst, to familiarize readers with the structure and design of
databases for use in their own laboratories, and second, to illustrate examples of
public biological databases and approaches that have grown from early bioinformatic
methods.
Bioinformatics of Behavior 97
98 Neurobehavioral Genetics
Most readers are familiar with the flat data model, in which each database is a self-
contained table consisting of rows and columns. This is the typical entry format for
statistical analysis tools, in which each row is an experimental unit, subject, or record
and each column is an attribute or measured variable of that subject. It is critical to
most applications that the data in a single column be of a single typecharacter or
numeric. Users of common business spreadsheets have the luxury of entering data
of arbitrary types into these cells, rendering import into more sophisticated databases
and statistical packages a nightmare! A frequent example is the technician entered
comment, e.g., 3.45?, 52+, 6.39.6, or 43 gm. Using a rigorous data storage
technique, users can create a series of comment codes that can be later used as
filtering or analytic criteria. Other common violations of the flat file format, such
as entering data from multiple groups in separate areas of the spreadsheet rather
than creating a separate column to indicate the grouping variable, should be avoided.
The hierarchical format is a method for data storage in which each object is a subset
of a larger class of objects. Most users are familiar with this in the form of file
storage systems on Windows, Mac, and Unix operating systems, or even a simple
physical file cabinet. The limitation of the utility of such a tree-based system for
managing diverse, interrelated sets of data becomes readily obvious. Two folders
labeled QTL analysis, and microarray analysis do not provide a unique location
for a new paper on quantitative trait loci (QTL) analysis of microarray data. Most
individual flat files that comprise the various studies conducted in a single laboratory
each have different data structures and often get filed in such a fashion. As a result,
it is difficult to integrate them and analyze a novel, synthetic view of the data.
Relational databases are the most versatile system for storing large complex sets of
data. In a relational database, data are stored in multiple tables and operations are
performed on tables to result in new tables. The database can be assembled from a
set of flat-format tables through a design process called normalization. Each level
of normalization represents a dramatic reduction in the redundancy of data storage.
The purpose of this type of data model is to avoid a common problem in data storage
systems, that if there are two copies of a datum, one of them will be wrong. This
may seem paradoxical at first, but is a problem that most readers have experienced
1903_C007.fm Page 99 Wednesday, July 26, 2006 6:35 PM
Bioinformatics of Behavior 99
TABLE 7.1
Transitioning from Flat Tables to First Normal Form
Flat Format
MOUSE_ID Latency_Trial_1 Latency_Trial_2
A9483 3.42 2.59
A7842 5.43 5.52
First-Normal Form
MOUSE_ID Trial Latency
A9483 1 3.42
A9483 2 2.59
A7842 1 5.43
A7842 2 5.52
firsthand. Using a relational database eliminates this problem because changes made
to one datum will automatically be propagated throughout the entire database. In
the first normal form, each field contains different information. For an example, see
Table 7.1. If latency measures on a maze task were obtained twice for each individual
mouse, the record can be entered in columns with field names Mouse_ID, latency_1,
latency_2. This is a common data entry format for many statistical packages.
However, in the first normal form this would correspond to two records each with
column field names Mouse_ID, latency, score. There are four other levels of
normalization. The end result is a set of tables to which each possesses a primary
key (the unique value for each record in the table) and a set of foreign keys that
refer to records in other tables. Accession numbers are a common example of primary
keys. To create a purely normalized set of tables is a challenging effort that may
exceed the needs of most users, and ideally a balance is struck between efficiency
(obtained by a lack of redundant data) and complexity (obtained by a proliferation
of tables).
assumption that biological data could be brought together post hoc in such a fashion
probably emerged because of the fortuitous properties of sequence datait anneals
spontaneously. Numerous alignment algorithms have been developed to pull this
data together optimally, but in general, any sequence data can be combined with
other overlapping sequence data to identify and compare individuals, strains, and
species. This is not the case with typical experimental observations, for which
additional data often translate into additional data structure complexity. The unaided
human brain must be used to perform the only integration possible. Other natural
keys are present in much of the data collected by mouse geneticists, the gene and
the mouse strain. To the extent that these natural keys can be well defined and remain
stable, they act as references through which data can be aggregated indefinitely and
integrated analytically.
FIGURE 7.1 Genome sequence as reference. A screen image of Ensembl Contig View shows
the location of genes, cross-species synteny, microsatellite markers, microarray probes, and
polymorphisms in the region of a QTL for seizure susceptibility.
EFERENCE
CR
ACTGAT TI P
A
ACTATT
GE N
NE
L
000100111011111
001101110010101
Bioinformatics
000010111101101
011101110110111
Knowledge
FIGURE 7.2 Using genetic reference populations to meet the challenge of integrating data
across levels of biological scale. There is tremendous hope and expectation that compiling
biological data using bioinformatics methods will rapidly lead to knowledge. A strong exper-
imental design can greatly facilitate the process. Using a genetic reference population,
sequence polymorphisms, gene annotation, strain variation, molecular, behavioral, physiolog-
ical and neuroanatomical data can be integrated with great facility.
individuals that can be repeatedly sampled through time and space (Figure 7.2). In
this approach, the mouse strain becomes the natural key, allowing the analytic
integration of many diverse data types. A single matrix can be constructed of
attributes by mouse strain, and the data integration is performed using correlational
analyses, though various attributes, for example genotype at a QTL, can be ascribed
a causal role. Where causality cannot be obtained analytically, a new experiment
can be performed to assess the antecedence, necessity and sufficiency of the putative
cause. Two resources for the analysis of these populations are WebQTL
(www.genenetwork.org)1315 and the Mouse Phenome Database.16
A major application of the genetic reference population is that diverse trait data can
be aggregated and correlated indefinitely, allowing the identification of traits that
share pleiotropic regulation. The presence of a genetic correlation implies that the
traits share a common heritable source of variance, which is often the result of
multiple effects of a single polymorphism or mutation; however, numerous other
sources of apparent pleiotropy exist.17 While individual studies allow the determi-
nation of heritability of a phenotype and possible sources of genetic variation, i.e.,
QTLs, the aggregation of multiple studies allows for analysis of the stability of the
phenotype under different environmental conditions, and the interpretation of his-
torical data in the context of novel technologies. For example, in WebQTL, users
can examine relationships among the earliest behavioral phenotypes collected in the
BXD RI strains in the early 1980s to microarray based measures of gene expres-
sion,14,15 a technology that had not even been invented at the time of the initial data
collection! The tremendous advantage of reference populations is that results
obtained can be integrated analytically. WebQTL features a variety of analytic tools,
1903_C007.fm Page 103 Wednesday, July 26, 2006 6:35 PM
ranging from simple descriptive statistics, to complex and highly flexible multivariate
data analysis, with results referenced to external genomic resources.15
Some reference populations have the additional advantage that they can be used for
genetic mapping. These isogenic lines have some known genetic variation that can
be assembled into high precision genome-wide databases and associated with phe-
notypic data.
The recombinant inbred strains are the first and arguably still the best genetic
reference population for mapping. The largest set of these strains, the BXD RIs were
developed in the late 1970s and expanded in the 1990s by B. A. Taylor18 and recently
expanded again to a size of nearly 80 strains.19 The high density of recombinations
obtained results in a dense genotypic map that has been refined and incorporated
into WebQTL.13 These genotypes arise from just two progenitors that were allowed
to randomly segregate. This is an important contrast to the standard inbred strains,
which experienced numerous bottlenecks in their breeding history.20
The standard inbred strains are a tantalizing population for use as a genetic
mapping panel. Early approaches used limited single nucleotide polymorphism data
from a few strains to identify QTLs using these lines of mice,21 but were challenged
due to low statistical power, high error rates, and concerns about the genetic archi-
tecture of the population.22,23 As additional mouse SNP data became available3,4,24
this controversial approach has been revisited. However, the genetic architecture of
these strains is highly complex2 giving rise to concerns about the effects of population
admixture due to the breeding history of the strains that remain.25,26 This has resulted
in the need to restrict these analyses to a small fraction of the existing standard
inbred strainsfewer than that available in the larger RI sets. While the use of
mouse single nucleotide polymorphisms is an excellent approach for refinement of
the QTL interval and selection of additional informative strains for multiple cross
mapping,27,28 the use of these strains for genome wide scans is not yet sufficiently
accurate. The interest in this approach has led to the construction of a variety of
large SNP data sets. Gradually, this data is being incorporated into existing public
genome browsers, and will be a major boon to efforts at QTL candidate gene
identification. Construction of a large, powerful, randomly bred and genetically
diverse recombinant inbred strain panel is underway.29
The consomic lines30 are also a genetic reference population, but they have low
resolution and other limitations. They can be used to map sources of genetic variance
to a single chromosome, allowing for determination of the location of genes with
main effects on phenotypic variation. However, a significant problem with the
consomic lines is that most complex traits are under regulation of multiple genetic
loci, and phenotypes may be a consequence of epistasis, a situation that is not
modeled well in these lines. Occasional evidence of coadaptive alleles is observed,
such that having a pair of genes from common progenitors at two homozygous loci
results in a different phenotype than having either recombinant pair of alleles. In
consomic analysis, it would appear that these two loci have main effects on pheno-
type such that the introgressed chromosome has a main effect on the phenotype
1903_C007.fm Page 104 Wednesday, July 26, 2006 6:35 PM
relative to the background chromosome. The chief advantage of the consomic lines
is that they are closer to congenic status than other cross-progeny.
A major effort is being made to identify meaningful sets of genes that share a
common biological structure, function or compartment.31 Numerous tools exist for
the examination of representation of category members among a gene set identified
through other biological analysis. These tools also allow users to compare submitted
lists of genes with known biological pathways databases KEGG,32 PubMeds GRIF,
and other categorization schemes. Some of these are Web-based, including DAVID33
and the Gene Ontology Tree Machine34 http://genereg.ornl.gov/gotm/ (Figure 7.3),
whereas others are standalone executables including EASE.35 A major advantage of
the standalone tools is that they allow users to store and compare gene sets generated
locally. For gene sets associated with very subtle behavioral phenotypes, the custom
development of gene sets will be necessary because few categorization schemes
reflect in-depth characterizations of behavior at the present time.
Multiple databases have been developed to identify genes that physically interact.
This includes DIP,36,37 BIND,38,39 and MINT.40 While the categorization of the phys-
ical interactions between gene products is interesting, it is important to note that in
the case of behavioral analysis, the interactions of gene products are often indirect,
as in the interaction between neurotransmitter synthesis enzymes and receptors.
While these gene products may be highly correlated and coregulated, and may co-
occur in various tissues, the interaction between them is not physical and would not
be encapsulated in the interactome databases.
7.4 APPLICATIONS
A few applications and examples merit further elaboration, though it is impossible
for any text to remain current on this topic. As with any biological result, the user
should make use of documentation to know and understand the source of the data
and the analytic criteria by which results are presented. Regardless of the professional
appearance of the interface, the quality of the underlying data and the validity of
1903_C007.fm Page 105 Wednesday, July 26, 2006 6:35 PM
molecular function
biological process cellular component
binding
cell
cellular physiological
process process
nucleic translation intracellular
acid regulator
binding activity cell
growth cytoplasm
and/or
maintenance
translation cytoplasmic
factor vesicle
activity,
nucleic transport
acid coated
binding vesicle
intracellular protein
transport transport
clathrin-coated
translation vesicle
initiation intracellular
factor protein synaptic
activity transport vesicle
FIGURE 7.3 Directed acyclic graphs (DAGs) from WebGestalts GO analysis of genes that
are co-expressed in the brains of isogenic mice indicate covariance among translation initiation
factors, transport mechanisms and synaptic vesicle related genes.
the analytical approach should be evaluated critically. Biological databases are not
perfect animals, and are known to contain errors and inaccuracies. Often, the research
community is able to deposit data with little verification or review. The more
frequently used data are going to be more accurate. This does not diminish the value
of biological databases, but is a caveat to their use as one of many biological tools.
A major application for multiple sequence analysis methods is the identification and
search for common motifs. A popular system for this is MEME, which can be used
to compare several sequences for the identification of motifs, and MAST, which can
then be used to search for the conserved motif throughout the entire genome
(http://meme.sdsc.edu/meme/website/intro.html). These tools can be powerfully
applied to gene sets in which one suspects common functional motifs, and have been
applied to upstream DNA to identify novel transcription factor binding sites that
may be shared by a coregulated gene set. Another application of the search for motifs
is to examine biologically important and conserved sequences.
Protein structure data can be obtained for many gene products and visualized using
NCBIs Cn3D tool, obtained at http://www.ncbi.nlm.nih.gov/Struc-
ture/CN3D/cn3d.shtml. This user-friendly tool allows users to examine structure
and amino acid sequence relations from structural data retrieved from the Molecular
Modeling DataBase, which contains data from the Protein Data Bank
(http://www.rcsb.org/pdb/).4244 For example, Figure 7.4 shows the ovine serotonin
acetyltransferase, highlighted to reveal residues that are different in the mouse. Trait
relevant mis-sense polymorphisms are likely to influence protein structure or func-
tion. In this example, the ligand binding site is relatively conserved between the two
species, but other regions are much more variable. Using actual structure or results
from prediction, in concert with knowledge of sequence polymorphisms, one can
rapidly predict the molecular consequences of allelic variation and prioritize analysis
of polymorphic loci. The rapid development of structure prediction algorithms will
1903_C007.fm Page 107 Wednesday, July 26, 2006 6:35 PM
FIGURE 7.4 Structural analysis of Aanat, the serotonin n-acetyl transferase enzyme. con-
served amino acid sequence between mouse and sheep using Cn3D. Protein structure was
retrieved from Entrez Structure, and mouse amino acid sequence was retrieved via Ensembl.
Sequences were aligned and differences were highlighted using Cn3D.
greatly increase the number of structures available and will soon make this approach
readily applicable.45
LRS ? Frequency of the Peak LRS ? Red indicates the currently visible
Additive Effect ? SNP Density ? magnified region of chromosome 1.
Mtap2
50.00 Mb 80.00 Mb
Significant LRS = 17.02
Suggestive LRS = 10.22
17.1
15.0
12.5
10.0 +0.20
7.5 +0.15
5.0 +0.10
2.5 +0.05
0.00
50 52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 (-0.05)
(-0.10)
(-0.15)
(-0.20)
17.1
15.0
61.3
12.5
50.0
10.0 0.20
40.0
7.5 0.15
30.0
5.0 0.10
20.0
0.00 0%
-0.05
-0.10
-0.15
17.2 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 x
15.0
39.0
12.5
0.20 30.0
10.0
7.5 0.15
20.0
5.0 0.10
10.0
2.5 0.05
0.00 0%
-0.05
-0.10
-0.15
-0.20
FIGURE 7.5 From QTL to candidate gene. QTL analysis of motor protein Kif5a mRNA
abundance. Mtap2 lies in the significant regulatory locus on Chr. 1, Creb is near the peak
regulatory locus. The bottom panel is a whole genome scan from WebQTL. The middle panel
is a view of Chr. 1, and the top is WebQTL analyzer view, which shows the physical map
position of the QTL, the location of genes in the region, and the SNP density within genes
and across the region. The heavy curve indicates the likelihood ratio statistic across the
genome. The bars indicate the frequency of the QTL peak at a given location over 2000
bootstraps. Horizonal lines indicate suggestive and significant linkage thresholds based on
the permutation analysis. The light line indicates the additive effect of the DBA/2J allele. In
this example, the DBA/2J allele decreases expression of the kinesin 5a transcript.
(WebQTL) coregulation. Identifying genes that are hypothesized to have these specific
relationships in common is an important first step to defining a meaningful result.
Not all transcripts that are co-expressed are regulated by the same transcription regu-
latory pathways. New tools, including MOTIF http://motif.genome.jp/ and PAINT
http://www.dbi.tju.edu/dbi/tools/paint/index.php?op=FnetBuilder, can be used to
search for known transcription factor motifs among sets of coregulated transcripts.
Microarray analysts often then identify known pathways and systems in which these
genes reside. Numerous tools are available for this type of annotation, and several
incorporate the actual expression results. These include GenMAPP, 46,4 7
DAVID/EASE,33,35 and several commercial tools.
through some form of neural plasticity. As data are accumulated in the recombinant
or standard inbred strains, these databases will expand dramatically in their utility.
ACKNOWLEDGMENTS
Jeremy L. Peirce and Susan E. Bergeson provided helpful comments on this chapter.
REFERENCES
1. Waterston, R.H. et al., Initial sequencing and comparative analysis of the mouse
genome. Nature, 2002. 420(6915): p. 52062.
2. Yalcin, B. et al., Unexpected complexity in the haplotypes of commonly used inbred
strains of laboratory mice. Proc Natl Acad Sci U S A, 2004. 101(26): p. 97349.
3. Wiltshire, T. et al., Genome-wide single-nucleotide polymorphism analysis defines
haplotype patterns in mouse. Proc Natl Acad Sci U S A, 2003. 100(6): p. 33805.
4. Wade, C.M. et al., The mosaic structure of variation in the laboratory mouse genome.
Nature, 2002. 420(6915): p. 5748.
5. Lagos-Quintana, M. et al., Identification of novel genes coding for small expressed
RNAs. Science, 2001. 294(5543): p. 8538.
6. Ambros, V., microRNAs: tiny regulators with great potential. Cell, 2001. 107(7):
p. 8236.
7. Smaglik, P., Internet gateway planned for neuroinformatics data. Nature, 2000.
405(6787): p. 603.
8. McCullough, B.D. and B. Wilson, On the accuracy of statistical procedures in
Microsoft Excel 2000 and Excel XP. Computational Statistics & Data Analysis, 2002.
40: p. 713721.
9. Chesler, E.J. et al., Identification and ranking of genetic and laboratory environment
factors influencing a behavioral trait, thermal nociception, via computational analysis
of a large data archive. Neurosci Biobehav Rev, 2002. 26(8): p. 90723.
10. Platt, J.R., Strong inference. Certain systematic methods of scientific thinking may
produce more rapid progress than others. Science, 1964. 146(3642): p. 347353.
11. Austin, C.P. et al., The knockout mouse project. Nat Genet, 2004. 36(9): p. 9214.
12. Bult, C.J. et al., The Mouse Genome Database (MGD): integrating biology with the
genome. Nucleic Acids Res, 2004. 32(Database issue): p. D47681.
13. Wang, J., R.W. Williams, and K.F. Manly, WebQTL: web-based complex trait analysis.
Neuroinformatics, 2003. 1(4): p. 299308.
14. Chesler, E.J. et al., Genetic correlates of gene expression in recombinant inbred
strains: a relational model system to explore neurobehavioral phenotypes. Neuroin-
formatics, 2003. 1(4): p. 34357.
15. Chesler, E.J. et al., WebQTL: rapid exploratory analysis of gene expression and
genetic networks for brain and behavior. Nat Neurosci, 2004. 7(5): p. 4856.
16. Bogue, M.A. and S.C. Grubb, The Mouse Phenome Project. Genetica, 2004. 122(1):
p. 714.
17. Hodgkin, J., Seven types of pleiotropy. Int J Dev Biol, 1998. 42(3): p. 5015.
18. Taylor, B.A. et al., Genotyping new BXD recombinant inbred mouse strains and
comparison of BXD and consensus maps. Mamm Genome, 1999. 10(4): p. 33548.
19. Peirce, J.L. et al., Genetic architecture of the mouse hippocampus: identification of
gene loci with selective regional effects. Genes Brain Behav, 2003. 2(4): p. 23852.
20. Beck, J.A. et al., Genealogies of mouse inbred strains. Nat Genet, 2000. 24(1):
p. 235.
21. Grupe, A. et al., In silico mapping of complex disease-related traits in mice. Science,
2001. 292(5523): p. 19158.
1903_C007.fm Page 112 Wednesday, July 26, 2006 6:35 PM
22. Darvasi, A., In silico mapping of mouse quantitative trait loci. Science, 2001.
294(5551): p. 2423.
23. Chesler, E.J., S.L. Rodriguez-Zas, and J.S. Mogil, In silico mapping of mouse quan-
titative trait loci. Science, 2001. 294(5551): p. 2423.
24. Pletcher, M.T. et al., Use of a dense single nucleotide polymorphism map for in silico
mapping in the mouse. PLoS Biol, 2004. 2(12): p. e393.
25. Mhyre, T. et al., Heritability, correlations, and in silico mapping of locomotor behavior
and neurochemistry in inbred strains of mice. Genes, Brain, Behavior, 2005. 4:
20928.
26. Churchill, G.A., Researchers didnt scurry around fields looking for mice, says Jack-
son Lab scientist, BioMed Central, July 7, 2004.
27. Hitzemann, R. et al., Multiple cross mapping (MCM) markedly improves the local-
ization of a QTL for ethanol-induced activation. Genes Brain Behav, 2002. 1(4): p.
21422.
28. Shifman, S. and A. Darvasi, Mouse inbred strain sequence information and yin-yang
crosses for quantitative trait locus fine mapping. Genetics, 2005. 169(2): p. 84954.
29. Churchill, G.A. et al., The Collaborative Cross, a community resource for the genetic
analysis of complex traits. Nat Genet, 2004. 36(11): p. 11337.
30. Singer, J.B. et al., Genetic dissection of complex traits with chromosome substitution
strains of mice. Science, 2004. 304(5669): p. 4458.
31. Ashburner, M. et al., Gene ontology: tool for the unification of biology. Nat Genet,
2000. 25: p. 2529.
32. Ogata, H. et al., KEGG: Kyoto Encyclopedia of Genes and Genomes. Nucleic Acids
Res, 1999. 27(1): p. 2934.
33. Dennis, G., Jr. et al., DAVID: Database for Annotation, Visualization, and Integrated
Discovery. Genome Biol, 2003. 4(5): p. P3.
34. Zhang, B. et al., GOTree Machine (GOTM): a web-based platform for interpreting
sets of interesting genes using gene ontology hierarchies. BMC Bioinformatics, 2004.
5(1): p. 16.
35. Hosack, D.A. et al., Identifying biological themes within lists of genes with EASE.
Genome Biol, 2003. 4(10): p. R70.
36. Xenarios, I. et al., DIP, the Database of Interacting Proteins: a research tool for
studying cellular networks of protein interactions. Nucleic Acids Res, 2002. 30(1):
p. 3035.
37. Xenarios, I. et al., DIP: the database of interacting proteins. Nucleic Acids Res,
2000. 28(1): p. 28991.
38. Bader, G.D. et al., BINDThe Biomolecular Interaction Network Database. Nucleic
Acids Res, 2001. 29(1): p. 2425.
39. Alfarano, C. et al., The Biomolecular Interaction Network Database and related tools
2005 update. Nucleic Acids Res, 2005. 33 Database Issue: p. D41824.
40. Zanzoni, A. et al., MINT: a Molecular INTeraction database. FEBS Lett, 2002.
513(1): p. 13540.
41. Subramaniam, S., The Biology Workbencha seamless database and analysis envi-
ronment for the biologist. Proteins, 1998. 32(1): p. 12.
42. Berman, H.M. et al., The Protein Data Bank. Nucleic Acids Res, 2000. 28(1): p.
23542.
43. Berman, H.M. et al., The Protein Data Bank. Acta Crystallogr D Biol Crystallogr,
2002. 58(Pt 6 No 1): p. 899907.
44. Bhat, T.N. et al., The PDB data uniformity project. Nucleic Acids Res, 2001. 29(1):
p. 2148.
1903_C007.fm Page 113 Wednesday, July 26, 2006 6:35 PM
45. Berman, H.M. and J.D. Westbrook, The impact of structural genomics on the protein
data bank. Am J Pharmacogenomics, 2004. 4(4): p. 24752.
46. Dahlquist, K.D. et al., GenMAPP, a new tool for viewing and analyzing microarray
data on biological pathways. Nat Genet, 2002. 31(1): p. 1920.
47. Doniger, S.W. et al., MAPPFinder: using Gene Ontology and GenMAPP to create a
global gene-expression profile from microarray data. Genome Biol, 2003. 4(1): p. R7.
48. Chen, H. and B.M. Sharp, Content-rich biological network constructed by mining
PubMed abstracts. BMC Bioinformatics, 2004. 5(1): p. 147.
49. Homayouni, R. et al., Gene clustering by latent semantic indexing of MEDLINE
abstracts. Bioinformatics, 2005. 21(1): p. 10415.
50. Smalheiser, N.R. and D.R. Swanson, Using ARROWSMITH: a computer-assisted
approach to formulating and assessing scientific hypotheses. Comput Methods Pro-
grams Biomed, 1998. 57(3): p. 14953.
51. Grubb, S.C., G.A. Churchill, and M.A. Bogue, A collaborative database of inbred
mouse strain characteristics. Bioinformatics, 2004. 20(16): p. 28579.
1903_C007.fm Page 114 Wednesday, July 26, 2006 6:35 PM
1903_C008.fm Page 115 Wednesday, July 26, 2006 6:36 PM
CONTENTS
Introduction............................................................................................................ 115
8.1 Congenic Strains .......................................................................................... 116
8.2 Consomic Strains ......................................................................................... 119
8.3 Knockout/Congenic Strains ......................................................................... 119
8.4 Mapping QTLs with Knockout/Congenic Mice.......................................... 123
References.............................................................................................................. 126
INTRODUCTION
Genetically defined mouse strains have become invaluable in biomedical research
because of their ability to clarify genetic vs. environmental influences on a wide
variety of traits including behavior. They have provided the means to test multiple
genetically identical or nearly identical animals at one time, thus reducing the effects
of genetic variation to a minimum. Some of the most useful types include inbred,
co-isogenic, congenic, consomic, recombinant inbred, knockout, and transgenic
strains.
The most commonly used strain of mouse is the inbred strain. Since these strains
are made by brothersister mating mice for more than 20 generations, any two mice
within an inbred strain will be genetically identical at 99.99% of their genetic loci.
A co-isogenic strain is similar to an inbred strain except that it differs from its inbred
partner by only one locus.1 These co-isogenic strains are usually the result of
spontaneous mutations that occur within an inbred stock and have been very useful
in analyzing the effects of single genes on a particular behavior or trait. However,
the investigator must either wait for mutations to occur spontaneously or induce
them by use of mutagenesis techniques, both requiring much time and effort. Because
of the usefulness of these co-isogenic strains, in the 1940s, George Snell constructed
a breeding scheme to develop a new type of strain, called the congenic strain, which
approximates a co-isogenic strain. These congenic strains have become extremely
valuable resources in analyzing single gene effects. Below we describe how these
strains are made and some of their properties. We also describe a different use for
congenic strains in mapping genes that influence quantitative traits such as behavior.
115
1903_C008.fm Page 116 Wednesday, July 26, 2006 6:36 PM
Finally we describe a variation of the congenic strain, called the consomic strain,
which selects for genetic differences covering an entire chromosome instead of a
single genetic locus.
Strain G Strain B
Locus F
Selection for allele at Locus F
a b
Differential chromosome
a b
Cross I. X
Other 19 chromosome
b
(GXB)F1 (or N1)
a b
Backcross to Strain B
b X b
Cross 2-10. and at each generation
select for the a allele at
the differential locus
a
b
N2
b
N10
a a
a
B . G - a N10F1
FIGURE 8.1 Derivation of a congenic strain of mice. Selection for the a allele at locus F
of the G strain (donor) at every generation.
1903_C008.fm Page 118 Wednesday, July 26, 2006 6:36 PM
TABLE 8.1
Congenic and Consomic Strains of Mice and Amount of Contaminant Genetic
Material*
Average Amount of
Strain: Recipient Differential Genetic Unlinked Contaminant
Strain (donor) Region (average) (cM) Genetic Region (cM)
Consomics B6 (A) B6 (Spretus) 160 cM 3 cM
Congenics (>N10) B6 (various) 20 cM 3 cM
Speed congenics (N5) B6 (various) 40 cM 16 cM
Knockout/congenics
B6 (129) 20 cM 3 cM
(>N10)
Genome tagged mice
B6 (DBA/2, CAST) 4050 cM 48 cM
(N4-N5)
TABLE 8.2
Amount of Genomic Material That Is Retained as a Function of the
N Number of Backcross Generations
Congenic
Consomic
N (other 19 chrs) Unlinked Linked
2 1282 Mb* 1282 Mb 140.6 Mb
4 321 321 82.0
8 20 20 46.7
10 5.0 5.0 37.5
16 0.1 0.1 23.4
20 <0.01 <0.01 18.7
*Numbers assume a genome size of 2700 Mb, a recombinant map size of 1600 cM and
an average chromosome size of 135 Mb. References are 1, 2.
of it. Additional recombinational events in the interval can then be used to divide the
region for further analysis. Congenic strains where selection is based on flanking
markers are often called interval-specific congenic strains.7
Another breeding scheme that has become more popular in recent years, because
of the advent of high-throughput genotyping, is marker-assisted selection and results
in quasi-congenic strains, commonly called a speed congenic.8,9 Speed congenic
strains start the same way as conventional congenics. An F1 is made between
two inbred strains of mice and the donor allele is selected in each generation. However,
at every subsequent generation, founder mice are selected that (1) retain the differential
allele and (2) have a minimum of unlinked genetic material from the donor
1903_C008.fm Page 119 Wednesday, July 26, 2006 6:36 PM
strain. Here, the unlinked contaminant genetic material is detected by use of a genome
scan of DNA markers spaced at regular intervals on the 19 nonsyntenic chromosomes.
The optimal spacing of these markers should be at approximately 20 cM intervals.10,11
At least 20 mice from each generation must be typed in a genome scan to obtain mice
with a minimal number of unlinked genes. With this technique, one can make the
equivalent of an N10 congenic in 4 to 5 generations, thus, reducing the time for
completion by at least one half (Table 8.1). However, the creation of these speed
congenics also requires more technician time and genotyping costs because of the
number of mice and loci that must be typed. For a more thorough discussion of this
technique, see Markel et al.8
Strain G Strain B
Selection for markers spaced
abcdefghi evenly along chromosome, e.g., Chr 3
Differential chromosome
Cross I. X
Other 19 chromosomes
abcdefghi
abcdefghi abcdefghi
Backcross to Strain B
Cross 2-10.
X and at each generation select
for markers spaced along
chromosome
abcdefghi
N2
abcdefghi
N10
abcdefghi abcdefghi
abcdefghi
abcdefghi
B-3GN10F1
these 129 knockout strains to B6. This leads to a congenic strain with a section of
129 introgressed into the B6 background. Thus, when these knockout/congenic
strains are tested for a phenotype that differs from the B6 strain, this phenotype
1903_C008.fm Page 121 Wednesday, July 26, 2006 6:36 PM
Target locus
B6
Differential
From Chromosome
129 ES cell X
Other 19
chromosomes
Target locus
could either be due to the gene that is ablated or to the flanking 129 chromosomal
segment (Figure 8.3).17 This consideration is important for two reasons. First, one
should be cautioned when using knockout/congenic strains to understand that the
observed phenotype is not always due to the targeted locus. There have been several
reviews, covering the cautions attached to the use of knockout/congenics.1821 This
situation is particularly poignant when considering phenotypes that are quantitatively
different in B6 vs. 129 and when the phenotype is not predicted from the tissue
distribution of gene expression and/or the known function of the ablated gene.
Second, the knockout/congenic strains can be used, in a positive way, to scan the
genome for genes that influence quantitative traits such as behavior.17 Here, the
investigator can use these strains as a panel of congenic strains with different
segments of 129 from a wide variety of chromosomes introgressed into the B6
background. Thus, by testing these strains, the investigator will actually be testing
for the QTL location somewhere in the flanking regions surrounding the targeted
loci (see below). In this respect, there are more than 50 knockout/congenic strains
available through commercial sources. Since the average length of the differential
129 chromosomal segment in these strains is about 20 to 30 cM (see Table 8.3 for
a representative list), this assortment of knockout/congenic strains offers the
researcher a set of strains that will cover most of the mouse genome. Moreover,
often these strains provide a set of overlapping 129 chromosomal segments that will
map a particular trait. Here the region of overlap is often small and may define the
boundaries of the genetic region determining the trait. For example, in Figure 8.4,
the QTL could be narrowed to a short segment by the overlap between B and C.
1903_C008.fm Page 122 Wednesday, July 26, 2006 6:36 PM
TABLE 8.3
Length of Differential Chromosomal Segments in a Sampling of
Knockout/Congenic Strains*
Length of 129
Position in cM Chromosomal
Targeted Gene N Number Chr. of Target Gene Stretch**
Apoe 10 7 4 1622
Cd3z 8 1 87 910
Fcgr3 6 1 92 2535
Igl-5 6 16 10 1632
Il6 11 5 17 916
Il10 10 1 70 3341
Lck 12 4 59 814
Selp 10 1 86 2935
Tap1 10 17 19 35
*Information on position of targeted gene and number of backcross generations was taken from
www.jax.org.
Trait
A yes
B yes
C yes
D no
F no
FIGURE 8.4 Mapping QTLs with overlapping differential chromosomes. Striped areas indi-
cate the chromosomal stretch on the differential chromosome that is derived from 129. The
arrows indicate the region that maps the trait.
2500
2000
* * A
1500
Total distance traveled * *
1000
on Day 1 *
500
700 * B
500
-100
-300
129 Il10 B2m Cd8a Apoe Lipc Il7r Tap1 Fmr1
FIGURE 8.5 Exploratory activity and habituation of inbred and knockout/congenic mouse
strains. The knockout mouse strains tested included B6.129P2-Il10, B6.129P2-Fcgr3,
B6.129P2-B2m, B6.129S2-Il6, B6.129S2-Cd8a, B6.129S2-Cd4, B6.129P2-Apoe, B6.129P2-
Apoa1, B6.129P2-Lipc, B6.129P2-Tcrd, B6.129S7-Il7r, B6.129S2-Igl-5, B6.129S2-Tap1,
B6.129S2-Tnfsf5 and B6.129P2-Fmr1. Knockout mice are all on a B6 background and are
indicated by the name of the ablated target gene. Two inbred strains were also tested: B6/J
and 129P2 (129). At least 24 male mice of each strain were tested. Results are shown + s.e.m.
There was a significant effect of strain for first day activity (F16,491 = 12.206, p < .0001; Panel
A) and habituation (F16,491 = 8.885, p < .0001; Panel B). Tukeys posthoc tests were then used
to make pairwise comparisons. An * indicates that a knockout/congenic strain is significantly
different from B6.
Flaherty, unpublished results). Finally, and most likely, there could be B6 modifier
genes that are interacting with 129 flanking genes to produce a heightened effect.
This explanation seems the most reasonable based on the high frequency of genetic
background effects on behavioral traits.
For intersession habituation (Figure 8.5B), we have also found the B6-Il7r has an
abnormally low habituation score. Since the Il7r gene is located on Chr. 15, this
agrees with our previous data showing a strong gene affecting habituation on
this same chromosome.26 Moreover, we have subsequently confirmed the location
of a gene influencing habituation on Chr. 15 by testing a Chr. 15 consomic strain,
B6-15A.16 This strain was significantly less active in the open field (Figure 8.6A),
and habituated less (Figure 8.6B) than B6 mice.
Thus, both congenic and consomic strains have been successful tools for the
behavioral geneticist in investigating single gene effects as well as in mapping genes
1903_C008.fm Page 125 Wednesday, July 26, 2006 6:36 PM
1600
1400 A
1200 *
1000
Total Distance
800
Traveled (cm)
600
400
200
700
600 B
500
200
*
100
B6 B6-15A
FIGURE 8.6 Exploratory activity and habituation of B6 and B6-15A strains. Twenty-four
male mice of each strain were tested. Results are shown + s.e.m. B6 mice were significantly
more active on the first day of testing (T46 = 3.747, p = .0005; Panel A) and displayed more
habituation (T46 = 4.203, p = .0001; Panel B) than B6-15A mice. An * indicates a significant
difference between the two groups.
that influence behavioral traits. The advantage of using these strains for mapping is
that they can provide a convenient starting place for fine mapping studies. Here, the
creation of subcongenic strains that are derivatives of congenic strains and made by
further recombinational events inside the introgressed differential chromosomal
region can isolate smaller chromosomal segments that are small enough to be
convenient for positional cloning efforts. These new subcongenics can then be used
to make an overlapping map that isolates the behavioral QTL to a shorter segment
of chromosome (see Figure 8.3 for an example of mapping using overlapping
differential chromosomal regions). Since multiple and genetically identical subcon-
genic mice can be made, the testing of these subcongenic mice will help to resolve
weak effects that cannot be detected with more conventional intercross techniques.
1903_C008.fm Page 126 Wednesday, July 26, 2006 6:36 PM
REFERENCES
1. Silver, L.M. Mouse Genetics Concepts and Applications, Oxford University, New
York, 1995.
2. Flaherty, L. Congenic strains. in The Mouse in Biomedical Research, Vol. 1, eds.
Foster, H.L., Small, J.D. & Fox, J.G., 21522, Academic Press, New York, 1981.
3. Bennett, B. Congenic strains developed for alcohol- and drug-related phenotypes.
Pharmacol Biochem Behav, 67: 67181, 2000.
4. Radcliffe, R.A., Bohl, M.L., Lowe, M.V., Cycowski, C.S. & Wehner, J.M. Mapping
of quantitative trait loci for hypnotic sensitivity to ethanol in crosses derived from
the C57BL/6 and DBA/2 mouse strains. Alcohol Clin Exp Res, 24: 133542, 2000.
5. Morel, L., Yu, Y., Blenman, K.R., Caldwell, R.A. & Wakeland, E.K. Production of
congenic mouse strains carrying genomic intervals containing SLE-susceptibility
genes derived from the SLE-prone NZM2410 strain. Mamm Genome, 7: 3359, 1996.
6. Abiola, O. et al. The nature and identification of quantitative trait loci: a community's
view. Nat Rev Genet, 4: 9116, 2003.
7. Darvasi, A. Interval-specific congenic strains (ISCS): an experimental design for
mapping a QTL into a 1-centimorgan interval. Mamm Genome, 8: 1637, 1997.
8. Markel, P. et al. Theoretical and empirical issues for marker-assisted breeding of
congenic mouse strains. Nat Genet, 17: 2804, 1997.
9. Wakeland, E., Morel, L., Achey, K., Yui, M. & Longmate, J. Speed congenics: a
classic technique in the fast lane (relatively speaking). Immunol Today, 18: 4727,
1997.
10. Visscher, P.M. Speed congenics: accelerated genome recovery using genetic markers.
Genet Res, 74: 815, 1999.
11. Servin, B. & Hospital, F. Optimal positioning of markers to control genetic back-
ground in marker-assisted backcrossing. J Hered, 93: 2147, 2002.
12. Hudgins, C.C., Steinberg, R.T., Klinman, D.M., Reeves, M.J. & Steinberg, A.D.
Studies of consomic mice bearing the Y chromosome of the BXSB mouse. J Immunol,
134: 384954, 1985.
13. Santos, J. et al. A new locus for resistance to gamma-radiation-induced thymic
lymphoma identified using inter-specific consomic and inter-specific recombinant
congenic strains of mice. Oncogene, 21: 66803, 2002.
14. Nadeau, J.H., Singer, J.B., Matin, A. & Lander, E.S. Analysing complex genetic traits
with chromosome substitution strains. Nat Genet, 24: 2215, 2000.
15. Singer, J.B., Hill, A.E., Nadeau, J.H. & Lander, E.S. Mapping quantitative trait loci
for anxiety in chromosome substitution strains of mice. Genetics, 169: 85562, 2005.
16. Singer, J.B. et al. Genetic dissection of complex traits with chromosome substitution
strains of mice. Science, 304: 4458, 2004.
17. Bolivar, V.J., Cook, M.N. & Flaherty, L. Mapping of quantitative trait loci with
knockout/congenic strains. Genome Res, 11: 154952, 2001.
1903_C008.fm Page 127 Wednesday, July 26, 2006 6:36 PM
18. Crusio, W.E. Flanking gene and genetic background problems in genetically manip-
ulated mice. Biol Psychiatry, 56: 3815, 2004.
19. Wolfer, D.P., Crusio, W.E. & Lipp, H.P. Knockout mice: simple solutions to the
problems of genetic background and flanking genes. Trends Neurosci, 25: 33640,
2002.
20. Gerlai, R. Gene targeting: technical confounds and potential solutions in behavioral
brain research. Behav Brain Res, 125: 1321, 2001.
21. Gerlai, R. Afraid of complications in gene targeting? Anxiety plays a role! Trends
Neurosci, 25: 136, 2002.
22. Iakoubova, O.A. et al. Genome-tagged mice (GTM): two sets of genome-wide con-
genic strains. Genomics, 74: 89104, 2001.
23. Liu, D. et al. Mapping behavioral traits by use of genome-tagged mice. Am J Geriatr
Psychiatry, 12: 15865, 2004.
24. Liu, D. et al. Identifying loci for behavioral traits using genome-tagged mice. J
Neurosci Res, 74: 5629, 2003.
25. Bolivar, V.J., Caldarone, B.J., Reilly, A.A. & Flaherty, L. Habituation of activity in
an open field: A survey of inbred strains and F1 hybrids. Behav Genet, 30: 28593,
2000.
26. Bolivar, V. & Flaherty, L. A region on chromosome 15 controls intersession habitu-
ation in mice. J Neurosci, 23: 94358, 2003.
27. Cook, M.N., Bolivar, V.J., McFadyen, M.P. & Flaherty, L. Behavioral differences
among 129 substrains: implications for knockout and transgenic mice. Behav Neu-
rosci, 116: 60011, 2002.
28. Bothe, G.W.M., Bolivar, V.J., Vedder, M.J. & Geistfeld, J.G. Genetic and behavioral
differences among five inbred mouse strains commonly used in the production of
transgenic and knockout mice. Genes Brain Behav, 3: 14957, 2004.
29. Bolivar, V.J., Ganus, J.S. & Messer, A. The development of behavioral abnormalities
in the motor neuron degeneration (mnd) mouse. Brain Res, 937: 7482, 2002.
30. Bolivar, V.J., Manley, K. & Messer, A. Exploratory activity and fear conditioning
abnormalities develop early in R6/2 Huntingtons disease trangenic mice. Behav
Neurosci, 117: 123342, 2003.
31. Bolivar, V.J., Manley, K. & Messer, A. Early exploratory behavior abnormalities in
R6/1 Huntingtons disease transgenic mice. Brain Res, 1005: 2935, 2004.
1903_C008.fm Page 128 Wednesday, July 26, 2006 6:36 PM
1903_C009.fm Page 129 Thursday, July 27, 2006 8:34 PM
9 Animal Resources in
Behavioral Neurogenetics
Jean-Michel Lassalle
CONTENTS
129
1903_C009.fm Page 130 Thursday, July 27, 2006 8:34 PM
9.1 INTRODUCTION
Genetics is a differential approach of the intra-specific variation, either normal or
pathologic, based on several types of experimental methods. Most of the behavioral
and neural phenotypes are complex traits that require the development of specific
tools. Several animal species are currently used in the field of behavioral neuroge-
netics, from humans to rotifers, through dogs, rodents, birds, fishes and flies.
Whereas in animals it is possible to create strains according to some specific
criteria, and to use them as genuine tools to devise genetic experiments for obvious
ethical reasons, in humans experimental approaches are restricted to natural exper-
iments. These basically include family studies, twin studies, and adoption studies.
On the other hand, in animals, various true experimental tools were developed. They
are based on inbreeding, selection experiments, crossing experiments, spontaneous
mutations, experimental mutagenesis or, more recently, on genetic engineering.
Consequences of inbreeding
100
90
Portion of the genome
80
70
60
50
40
30 Individual homozygosity
Portion of the genome that is fixed
20
10
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Generations of inbreeding
FIGURE 9.1 Effect of inbreeding during 20 generations on the portion of the genome of an
animal that is homozygous at each generation (solid line) and the portion that is fixed
identically in two animals chosen as parents for the next generation. (Adapted from Silver.2)
1903_C009.fm Page 132 Thursday, July 27, 2006 8:34 PM
generations under brother sister mating. Most inbred strains of mice and rats have
greatly surpassed 20 generations of inbreeding, which ensures homozygosis at nearly
all loci. Nevertheless, not all species support inbreeding; birds generally become sterile
after two generations of inbreeding and it seems that Drosophila resists inbreeding.1
9.2.1.1 F1 Hybrids
Crossing two inbred strains produces F1 hybrids. For instance, B6D2 F1 results
from a cross between a C57BL/6 female (B6) and a DBA/2 male (D2), whereas
D2B6 F1 is the reciprocal F1 cross.
1903_C009.fm Page 133 Thursday, July 27, 2006 8:34 PM
The characteristic features of F1 hybrids are their genetic and phenotypic uni-
formity: although partly heterozygous and partly homozygous, depending on the
loci where the parents carry different or identical alleles, they are all isogenic. F1
hybrids present advantages that predispose them to specific applications. First of all,
F1 hybrids between two inbred strains often display hybrid vigor or heterosis that
results in improved development and longevity, better resistance to diseases and
stress, larger size and weight, increased prolificacy, etc.). They often also display
better behavioral responses to cognitive tests.4 Secondly, they are useful as hosts for
tissue transplants (tumors, skin, ovaries) from either parental strain and as recipients
for some deleterious mutations (transgenic mice). The comparison of reciprocal
crosses between the two F1s allows detection of maternal effects resulting from
cytoplasmic and/or pre- and postnatal maternal environmental influences. Last, they
can be repeatedly produced, as long as the inbred parental strains remain available
(samples of parental strains can be preserved for years as frozen embryos).
9.2.1.2 F2 Hybrids
Recipient Donor
strain strain
F1
BC1
Selection
BC2
Selection
BC 10
Heterozygous congenic
Intercross
Homozygous congenic
Conplastic strains are used to study cytoplasmic inheritance and to detect point
mtDNA mutations. Inherited mtDNA is 99.9% maternal and represents less than 1%
of the total cellular DNA in humans and 0.2% in the mouse.9
Originally designed to detect major gene effects, they are more and more used to
detect and locate quantitative trait loci (QTLs) on chromosomes, with increased
precision as the number of strains in a series increases. Thirty-six BXD RI strains
are now available from the Jackson laboratory and the number of strains in the BXD
series will soon be greatly increased. The University of Tennessee Health Science
1903_C009.fm Page 137 Thursday, July 27, 2006 8:34 PM
F1 (isogenic)
F2
(heterozygous)
20 G. brother x
sister mating
BXD RI set
(inbred,
isogenic)
These mutant mice are useful tools to elucidate basic biological processes and
to study relationships between gene mutations and disease phenotypes. Although
only recently used in mice, mutagenesis has been successfully employed in Droso-
phila to study learning and memory mutants early in the 1970s. A neuroscience
mutagenesis facility has been created for mice at the Jackson laboratory that is
available online.
They constitute useful animal models for studying the effects of chromosomal
aberrations in humans. Moreover, Robertsonian chromosomes in combination can
be used to produce whole chromosome trisomy for specific mouse chromosomes.
For example [Rb(6.16)24Lub Rb(16.17)7Bnr]F1 mice, produce trisomy for chro-
mosome 16.
DBA/1
Littles mice used in DBA
color experiments
DBA/2
C
CBA
CHI
X
CI2I
C3H/SI
C3H/Bi
C3H/An
C3H/He
C3H/He J
C3He B/Fe J
Dealers stock Bogg BALB/c ( OvaC57BL/6J
transferred to
)
in Ohio Albinos A/J
A/St
X
A/Bi
A/He
Cold Spring Harbor A/HeJ
Albinos C5B
C57BL /6
C57BL C57BL /10
Miss Lothrops
C57BR/cd
Stocks
C57BR C57BL /o
C57L
AKR
Furths A & R Stocks RF
SWR
European White Mice
SJL
Webster Swiss
(1909) 12 16 20 24 28 32 36 40 44 48 52 56 60
FIGURE 9.4 The origins and relationships of some of the inbred strains of mice.18
CBA/J, C3H/HeJ, C57BL/6J, DBA/2J, ), while others have the M.m. musculus
type (e.g., AKR/J, SJL/J, SWR/J, etc.).
Most laboratory mice have contributions from both Mus musculus musculus and
Mus musculus domesticus. There is evidence that smaller contributions also may
have come from Mus musculus molossinus and Mus musculus castaneus.3
C57BL
GEN. 0 20 40 60 80 100 120
10 Ch
10 Fo
10 Gn
10 Sf
10 Jax
10 Sc
10 Sn
10 Ph
10Y
<?> Ks
6By
6 lcr
6 He
6N
6 Bom
6 Jax
6Y
6Hb
Little 6Mi
How
Mcl
Fr
Rii
A
Rl
He
An
Hf
1Umc
Bcr
Ge
lcrf
Go
Fa
H
LacH
Gr
Ka
St
In July 2001, The Jackson Laboratory and Charles River Laboratories Interna-
tional, Inc., started to cooperate to supply JAX mice to researchers located in
European and Pacific Rim countries.
As detailed in the JAX Bulletin, (N11, Sept. 2001), their distribution agreement
stipulates that the Jackson laboratory ensures that JAX mice raised in Charles
Rivers European and Japanese facilities meet the highest standards for health and
genetic quality. The main points of this agreement are summarized below.
JAX animals bred in the United States, Europe and Japan originate from pedigreed
identified pairs shipped from The Jackson Laboratory. Authorized breeding colonies
are reinfused with new pedigreed stock from The Jackson Laboratory on a routine
basis to minimize spontaneously occurring genetic drift. Routine monitoring for
genetic quality is performed to ensure both genetic identity and purity.
JAX rodents bred in Europe and Japan are monitored for the same pathogenic and
opportunistic agents as are the JAX rodents raised at The Jackson Laboratory, on
the same basis (sampling plan, frequency of testing, diagnostic test methods).
Such genetic and health quality procedures ensure that all mice used in various
countries at different times match the highest standards for genetic uniformity and
makes particularly meaningful comparisons of results.
stage (5 days) and the pupa stage (4 days). The adult fly lives about 30 days in the
laboratory. Drosophila has a compact genome (The X/Y sex chromosomes and 3
pairs of chromosomes, the fourth of which being very tiny) that has been entirely
sequenced. The size of the genome is about 165 million bases and contains an
estimated 14,000 genes. Drosophila displays a rather large behavioral repertoire:
learning and memory, courtship, circadian rhythms, food search, olfaction, locomo-
tion, aggression and sleep-like processes can be studied in Drosophila, some of them
at both the larval and adult stages. Most methods used in forward and backward
genetics can be implemented in that species, including polygenic and single-gene
analysis, selective breeding, QTL analysis, mutational analysis (EMS and screening),
namely for memory and biological clocks, visible markers, transposons, genetic
engineering (inducible knockout, reporter genes such as -Galactosidase, GFP).
One limit of Drosophila comes from its resistance to inbreeding so that there
are no truly inbred strains in this species.
The Whole Mouse catalog (this Web site is for mice and rats, formerly
the Mouse and Rat Research Home Page located at CalTech):
http://www.rodentia.com/wmc/.
Mouse Genome Informatics (MGI): http://www.informatics.jax.org.
Neuroscience mutagenesis facility: http://nmf.jax.org.
Complex Trait Consortium (CTC): http://www.complextrait.org.
The Rat Genome Database (Ratmap group, Sweden)
Ratmap: http://ratmap.gen.gu.se/.
Rat Resources: http://ratmap.org/ratres.html.
International Committee on Standardized Genetic Nomenclature for Mice
Chairperson: Dr. Janan T. Eppig ([email protected]).
Rat Genome and Nomenclature Committee, Chairperson: Dr. Eberhard
Guenther ([email protected]).
Current nomenclature rules for naming genes
For mouse: http://www.informatics.jax.org/mgihome/nomen/gene.shtm1#genenom.
For rat: http://rgd.mcw.edu/nomen_rules.html.
Strain names registration
For mouse: http://www.informatics.jax.org/mgihome/submissions/
submissions_menu.shtml.
For rat: http://rgd.mcw.edu.
The C. elegans www server at UTSMC is the best on line reference for
C. elegans: (http://elegans.swmed.edu/).
elegans Net: (http://members.tripod.com/C.elegans/index.html links on C.
elegans and includes extensive introductory links (ACekit at
http://winw.nbr.wisc.edu/outreach/test/celegans.html ).
REFERENCES
1. Petit, C., Linfluence du mode de croisement sur la structure gntique des popula-
tions; la stabilit des populations exprimentales de faible effectif. Annales de gn-
tique, 1963, 6, 2935.
2. Silver, L.E., Mouse genetics, concepts and applications. 1995 Oxford University
Press. Available online at the Jackson laboratory Mouse Genome Informatics facility:
http://www.informatics.jax.org/silver/frame1.1.shtml and also the CTC Web site:
http://www.complextrait.org.
3. Festing, M.F.W., Inbred strains of mice and their characteristics In: Mouse Genome
Informatics (MGI) 1998 at http://www.informatics.jax.org.
4. Lassalle J.M., Le Pape G., and Mdioni J., A case of behavioral heterosis in mice:
quantitative and qualitative aspects of performance in a water-escape test. J. Comp.
Physiol. Psychol., 1976, 93, 116123.
5. Darvasi, A. and Soller, M., Advanced intercross lines, an experimental population
for fine genetic mapping. Genetics, 1995 141: 11991207.
6. Snell, G.D., Congenic resistant strains of mice. In: Origins of inbred mice, Morse,
H.C., Ed., Academic Press, New York, 1978 pp 131.
7. Flaherty, L., Congenic strains. In: The mouse in biomedical research, Vol. 1, Foster,
H.L., Small, J.D., Fox, J.G. Eds, Academic Press, New York, 1981 pp. 215222.
8. Nadeau, J.H., Singer, J.B., Martin, A., and Lander, E.S., Analysing complex genetic
traits with chromosome substitution strains. Nature Genetics, 2000, 24: 221225.
9. Lewin, B., 1987. Genes, Wiley, New York.
10. Mc Clearn, G.E. and Kakihana, R., Selective breeding for ethanol sensitivity: short
sleep and long sleep mice, in: Development of animal models as pharmacogenetic
tools. G.E. McClearn, R.A. Dietrich and V.G. Ervin, Eds., Research monograph No.
6, Bethesda, MD, US. 1978 Department of Health and Human Services, 147159.
11. Bailey, D.W., 1971. Recombinant inbred strains, an aid to finding identity, linkage,
and function of histocompatibility and other genes. Transplantation, 1971 11:
325327.
12. Taylor, B.A., Recombinant inbred strains: use in gene mapping. In: Origins of inbred
mice, Morse, H.C., Ed., Academic Press, New York, 1978 pp 423438.
13. Peirce, J.L., Lu, L. GU, J., Silver, L.M. and Williams, R., A new set of recombinant
inbred lines from advanced intercross populations in mice. BMC Genetics, 2004 5:7,
117. Article available online from: http://www.biomedcentral.com/14712156/5/7.
14. Williams, R.W. and Churchill, G.A., The collaborative cross: rationale, implemen-
tation and costs. National Institute of General Medical Sciences, Complex trait
workshop report. 1998 www.nigms.nih.gov/newx/reports/genetic_arch.html.
15. Demant, P. and Hart, A.A.M., Recombinant congenic strains a new tool for ana-
lyzing genetic traits determined by more than one gene. Immunogenetics 1986 24:
416422.
1903_C009.fm Page 147 Thursday, July 27, 2006 8:34 PM
16. Hsiao, K.A., Chapman, P., Nilsen, S., Eckman, C., Harigaya, Y., Youkin, S., Yang,
F., and Cole, G., 1996. Correlative memory deficits, An elevation, and amyloid
plaques in transgenic mice. Science, 1996, 274: 99102.
17. Greenspan, R.J., The induction, detection and isolation of mutations. In D. Goldowitz,
D. Wahlsten and R.E. Wimer Eds: Techniques for the Genetic Analysis of Brain and
Behavior. Elsevier Science Publishers BV, 1992, 93110.
18. Staats, J., Origins of the older inbred strains. In: Biology of the Laboratory Mouse.
Green, E.L., Ed., 1966, McGraw Hill.
19. Bailey, D.W., Definitions of inbred strains, substrains, sublines, and F1 Hybrids. In
Handbook of Genetically Standardized Jax Mice. R.R. Fox and B.A. Witham, Eds.
1997 Fifth Edition. The Jackson laboratory, Bar Harbor, Maine.
20. Green, E.L., Genetics and Probability in Animal Breeding Experiments. 1981 Oxford
University Press, New York.
1903_C009.fm Page 148 Thursday, July 27, 2006 8:34 PM
1903_C010.fm Page 149 Wednesday, July 26, 2006 6:37 PM
CONTENTS
Introduction............................................................................................................ 149
10.1 Effect size .................................................................................................... 150
10.2 Probability of a False Positive .................................................................... 152
10.3 Probability of Failing to Detect Something Real ....................................... 154
10.4 Estimating Effect Size from Published Data.............................................. 155
10.5 Methods for Estimating Sample Size ......................................................... 156
10.6 Power and Sample Size for Two Independent Groups .............................. 158
10.7 Comparison of Several Inbred Strains........................................................ 159
10.8 Comparing Specific Groups in a One-Way Design ................................... 161
10.9 Strain Treatment Experiments ................................................................. 162
10.10 Bridging the Gap: the 2 2 Design............................................................ 165
10.11 Sample Size for More Specialized Experiments ........................................ 167
Acknowledgments.................................................................................................. 167
References.............................................................................................................. 167
INTRODUCTION
Planning any experiment involves choice of an experimental design and sample size
for each of the groups in the study. The design of a study is dictated by the kind of
question one seeks to answer, and the choice of appropriate control groups is
determined by logic and established genetic principles. Approaches to design are
discussed in numerous articles and chapters, including many in the present volume.
Once the design is chosen, the researcher must decide on the sample size. This will
commonly be done when preparing a grant or thesis proposal. Increasingly, animal
ethics committees also request that the investigator justify the number of animals to
be used. Ethical and financial considerations, alike, demand that the minimum
effective number be employed. At the same time, if too few animals are studied, the
experiment may be unable to detect the very effects it was designed to investigate,
149
1903_C010.fm Page 150 Wednesday, July 26, 2006 6:37 PM
thereby wasting the animals as well as the researchers time and the funding agencys
money.
The proper sample size is determined by the experimenters choice of three
values. Two are probabilities of making either of two kinds of errors of statistical
inference, and the third is the size of an effect that one seeks to detect.17
1+
1
Small
2 = .5
Medium
2 = .75
Large
2 = 1.0
-2 -1 0 1 2 3 4
Standard deviations from 1
FIGURE 10.1 Overlap of scores in two normally distributed groups for three effect sizes
ranging from small to large. Note that most scores in a group are within two standard deviations
of its mean.
1903_C010.fm Page 151 Wednesday, July 26, 2006 6:37 PM
of the effects of the mutation can be assessed by comparing the difference between
group means to the standard deviation within a group. For brain weight, effect size
would be 20 mg/15 mg = 1.33, whereas for learning it would be 4 trials/5 trials
= 0.80. Relative to variation within a group, the mutation would appear to have a
relatively greater effect on brain weight than on maze learning.
The ratio () of group difference in mean (1 2) to standard deviation within
a group () is widely used as an index of effect size when two groups are being
compared, and this is the value that is commonly used for a sample size estimation.
Generally speaking, the smaller the effect that one would like to detect, the larger
must be the sample size.
Because sample size calculations are done before any data are collected, these
involve population values of parameters rather than estimates from sample data.
Symbols used for population parameters and sample statistics are given in Table 10.1.
For a study involving comparisons of two groups, the indicator of effect size is the
hypothesized true value of the population parameter = (1 2)/, whereas its
estimate from sample statistics is d = (M1 M2)/S. The value of d is slightly biased
and can be adjusted when precision is deemed important.14,26 In this presentation,
the true, population parameter is either a Greek letter or boldface, while the estimated
sample statistic is either italicized or the Greek letter with caret ^ to indicate an
estimate from data.
On the basis of my review of many studies of a wide variety of phenotypes,
guidelines are offered for values of that correspond to what are generally regarded
as very small, small, medium, large, and very large effects in neurobehavioral
genetics: = 0.25, 0.50, 0.75, 1.0, and 1.5, respectively (Table 10.2). These are
somewhat larger than values assigned to the same descriptors in psychological
research with humans by Cohen8 and Borenstein et al.5
When more than two groups are involved, as in a study of several inbred strains,
the difference between any two groups will be unique to that pair. Hence, an indicator
that is more broadly representative of the data is needed. Cohen8 has suggested a
convenient approach that compares the population standard deviation among group
means (M) to the standard deviation within groups (): f = M/. When an analysis
of variance (ANOVA) partitions total variance into components among and within
TABLE 10.1
Symbols for True Values of Parameters and Their Estimates From Sample
Statistics
Population Sample
Characteristic Parameter Statistic
Mean M
Standard deviation S
Effect size for 2 groups d
Effect size for J groups 2 est. 2
Cohens effect size for J groups f
Linear contrast of J groups est.
1903_C010.fm Page 152 Wednesday, July 26, 2006 6:37 PM
TABLE 10.2
Values Corresponding to Magnitudes of Effect Size for Animal Research in
Neurobehavioral Genetics
2 Groups J Groups
Size of Effect 2 Cohens f 2
Very small 0.25 1.5% 0.1 1%
Small 0.5 6% 0.2 4%
Medium 0.75 12% 0.35 11%
Large 1.0 20% 0.5 20%
Very large 1.5 36% 0.7 33%
groups, a good descriptor of effect size is 2, the proportion of total variance that
is attributable to differences among group means in a fixed effects analysis.13 The
f index has a simple relation to this ratio of variances: 2 = f2/(f2 + 1). Values of
f and 2 for effects of various sizes are compared in Table 10.2. When there are
only two groups in the study, 2 = 2/(2 + 4).
Cohen8 proposed that criteria for small, medium and large effects in a one-way
ANOVA design in psychological research be f = 0.1, 0.25 and 0.4, respectively, which
correspond to 2 = 0.01, 0.06 and 0.14. These values appear to be somewhat lower
than what is commonly observed in neurobehavioral genetics, for example in com-
parisons of several inbred strains. A multigroup difference that accounts for only 1%
of total variance would be seen as not merely small but almost trivially small. Many
inbred-strain studies have found effects exceeding 20% of total varianceand this
would be a reasonable criterion for a large effect in our field. Standards are proposed
for both f and 2 in Table 10.2 for a study with J groups. With these conventions,
the verbal labels of size of the effect correspond to similar values of 2 for a study
with two groups as well as for a large experiment with J groups.
such as a t test when comparing two groups or the F test in the ANOVA comparing
several groups, in order to assess whether the evidence is sufficient to justify rejection
of the null hypothesis. If the null that there was no real effect can be refuted, then
we may conclude, tentatively, that there was indeed an effect. This arcane logic,
clothed in a most unfortunate choice of words such as significance, often makes
it difficult to keep the characters in the drama of science in their proper roles.
The null may in fact be true or it may be false. If the null is true but we decide
to reject it on the basis of the results our study, then the outcome of our study is a
false positive result and, in statistical parlance, we commit a Type I error of inference.
By convention, we agree to set a criterion for rejecting the null hypothesis based on
the probability of committing a Type I error. Most commonly this criterion, sym-
bolized , is set at one chance in 20. If the data suggest that the probability of
committing a Type I error is less than 5%, the null may be rejected, but this does
not reveal the scientific significance of the result. Instead, it would be better thought
of as the false positive rate (FPR). If we are comparing two independent groups of
animals, perhaps one homozygous for a knockout (/) and the other the wild-type
littermate controls (+/+), and the t test indicates the FPR is .001 or one chance in
1,000, then it is a fairly safe bet that the knockout really did affect the phenotype
we measured.
In those situations where the null is indeed true, the probability of obtaining a
false positive result is not influenced in any way by the sample size. That is, even
an elegant study of large numbers of animals will lead us to commit a Type I error
on about 5% of the tests when the null is true, the same FPR as when only a few
animals are studied.
The choice of the Type I error criterion influences the calculated sample size.
Generally speaking, the smaller the value of , the larger must be the size of the
observed effect in order to warrant rejection of the null, and the larger the sample
size that must be employed. When the results of a simple study depend on just one
statistical test of the null hypothesis, then = .05 may be appropriate. If the study
involves K independent tests, however, or if the study is one of K similar studies
being done independently in different labs, then the Bonferroni corrected Type I
error probability /K is more appropriate. A more sophisticated approach to adjust-
ing for multiple tests has been proposed by Benjamini et al.4
If the researcher has a good idea of the direction of the likely effect of a mutation,
then it would be appropriate to conduct a one-tailed test of the null hypothesis that
two groups do not differ, whereas a two-tailed test is preferred if the direction of
the effect is uncertain. In an ANOVA with multiple groups, this distinction is not
pertinent.
The overbearing emphasis on null hypothesis testing in psychology has been the
subject of debate.7,12 One of the central issues is the credibility of the null in any
particular study. In a genetic linkage study, for example, the hypothesis that alleles
at a marker locus are not related to phenotypic variation is highly credible, and we
expect that most markers will not show evidence of linkage. In such a case, the
proper choice is critically important. Lander and Kruglyak16 proposed the use of
= .0001 for a genome-wide scan in linkage analysis, while Belknap et al.3 argued
for a lower level of as a foil against the demise of a study owing to insensitivity
1903_C010.fm Page 154 Wednesday, July 26, 2006 6:37 PM
to real linkage relationships. Given the surprising but common finding that a gene
knockout has little or no effect on many phenotypes, the null may also be credible
in this kind of research. For a multiple inbred strain study, on the other hand, it
would be astounding if no strain main effect were found. Instead, the more interesting
question in strain studies is usually the presence or absence of strain by treatment
interactions or strain correlations among phenotypes.
Whereas the null hypothesis is commonly taken to be no effect at all, methods
are available for evaluating the possibility that the observed effect exceeds some
nonzero effect size.17 This approach has not yet been widely applied in neurobe-
havioral genetics, but it is useful when researchers want to know whether a
mutation has, for example, something more than merely a small effect on a
phenotype.
TABLE 10.3
Four Estimates of Sample Size Needed to Detect a Large Effect Size of = 1.0
Sample Size in Each of Two Groups According to
Kraemer and Borenstein
Tails Power C, Thiemann14 Cohen8 Wahlsten26 et al.6
.05 1 80% 6.18 15 13 15 14
.05 1 90% 8.56 20 18 20 18
.05 1 95% 10.82 25 22 24 23
.05 1 99% 15.77 36 32 34 33
.05 2 80% 7.85 19 17 18 17
.05 2 90% 10.51 24 22 24 23
.05 2 95% 12.99 30 27 28 27
.05 2 99% 18.37 41 38 39 38
.01 1 80% 10.04 23 22 23 22
.01 1 90% 13.02 30 27 29 28
.01 1 95% 15.77 36 33 34 33
.01 1 99% 21.65 48 45 46 45
.01 2 80% 11.68 27 25 26 26
.01 2 90% 14.88 34 31 32 32
.01 2 95% 17.81 40 37 38 38
.01 2 99% 24.03 53 50 51 50
Mean of all 16 values: 31.3 28.7 30.1 29.2
Note: In the method of Wahlsten,25,26 the normal approximation to the noncentral t distribution involves
a constant C, that depends on the standard normal deviates for Type I error (z) and Type II error
(z), such that C, = (z + z)2. The sample size per group is then n = C,,/2 + 2. The resulting value
of n is rounded up to the nearest integer.
1903_C010.fm Page 158 Wednesday, July 26, 2006 6:37 PM
accuracy of any estimate of sample size will be undermined by small but inevitable
subject attrition during any large study, and the researcher is well advised to plan on
using a few more animals than the minimum indicated by a sample size calculation
in order to be sure that the final number of good data points is adequate.
80
Large Medium
Very = 1 = 0.75
Large n = 86
60 for 90%
= 1.5 Small Power
= 0.5
40
= .05
20 2-tailed
11 22 38
0
10 20 30 40 50
Number of Animals Per Group
FIGURE 10.2 Power vs. sample size for a two-tailed t test of the difference between two
independent groups when the null hypothesis is that the true group difference is = 0. The
sample size needed to yield a power of 90% is shown as an italicized numeral along the x axis.
1903_C010.fm Page 159 Wednesday, July 26, 2006 6:37 PM
0.8 = .05
Power
0.6
0.4
= .01
0.2
52 65 82 97
0.0
0 10 20 30 40 50 60 70 80 90 100
Number of Subjects Per Group
FIGURE 10.3 Power vs. sample size for a t test comparing two groups when the null
hypothesis is = 0 and the alternative is = 0.5. Number of tails (1 or 2) for the test is
indicated in the circle on each line. The sample size needed to yield a power of 80% is shown
as an italicized numeral along the x axis.
or a small effect, and it would seem reasonable to choose sample size in order to
have perhaps 90% power to detect a medium effect of = 0.75. Usually we would
expect a knockout to change a phenotype in a direction that makes the animal less
viable, vigorous or competent, and a one-tailed test would be appropriate. If only
one major phenotype is to be assessed, a choice of = .05 would make sense, but
= .01 would be a better choice if the study will involve measures of several
phenotypes, each to be assessed with a separate test of significance.
In the event that the investigator has only 10 or even fewer mutant animals for
the first study, it is still important to attend to the power curve. For example, if the
test will be done one-tailed with = .05, Figure 10.2 shows that power will be 90%
only for a very large effect size, whereas power will be about 35% if the real effect
size is = 0.75. Thus, by considering the question of power, the investigator will
be forewarned that failure to reject the null that = 0 in no way proves that the
knockout has no effect at all, because the chance of a Type II error occuring would
be quite high for a medium or even large effect. The purpose of doing an initial
study with fewer than 10 mutant animals and a similar number of normal sibs would
be to evaluate the possibility of a severely abnormal phenotype. If such a dramatic
deficiency were not found, it would be wise to use larger samples in future research
with that particular knockout.
0.8
0.6
= .05
= .01
0.4
0.2 Large f = .5
0.0
0 10 20 30 40 50 60 70 80
1.0
0.8
Power
0.6
0.4
0.8
0.6
0.4
0.2 Small f = .2
0.0
0 10 20 30 40 50 60 70 80
Sample Size per Group
FIGURE 10.4 Power vs. sample size for an ANOVA comparing means of 8 strains of animals
for three values of Cohens f index of effect size. Black dots indicate the sample size for
power of 90%.
1903_C010.fm Page 161 Wednesday, July 26, 2006 6:37 PM
in each case. For 4 strains, 31 mice per strain would suffice to yield power of 90%,
whereas for 8 strains the sample size per group would be considerably less. When
power is at issue, it is apparent that studying a smaller number of strains does not
reduce the total number of animals greatly. For 10, 8, 6, and 4 strains, the total
sample would need to be 170, 160, 144, and 124, respectively. For 2 strains, the
sample would be considerably smaller, but the generality of the results would be
severely restricted. It does not necessarily follow that a study of 10 strains will cost
about the same amount as with 6 strains, because we usually begin the process of
choosing strains with the most common ones that are also least expensive and then
move to less commonly studied strains that cost more per animal.
0.6 4
2
0.4
0.2 = .05
0.0 17 20 24 31 44
0 10 20 30 40 50
Sample Size per Group
FIGURE 10.5 Power vs. sample size for an ANOVA on different numbers of strains when
f = 0.35. The sample size needed to yield a power of 90% is shown as an italicized numeral
along the x axis.
1903_C010.fm Page 162 Wednesday, July 26, 2006 6:37 PM
TABLE 10.4
Contrasts among Males of Reciprocal Backcrosses to Strain A; Agonistic
Acts in the Home Cage Intruder Test
Contrast Coefficients to Detect Effect
Maternal
Mother Father Mean mtDNA Y Chromosome Environment
AB A 17 1 0 1
BA A 15 1 0 0
A AB 10 0 1 0
A B A 13 0 1 1
2 3 4
n per group 51 24 15
Note: Sample size is found using the method of Wahlsten25,26 with a one-tailed test at = .05 and
power of 95% when within = 3.0. As in the case of two groups (Table 10.3), C, = (z + z)2. For a
one degree of freedom contrast among J groups, n = Ccj2/(/)2 + 2.
1903_C010.fm Page 163 Wednesday, July 26, 2006 6:37 PM
In order to estimate sample size for a two-way design when each group is to
have the same sample size, one begins by positing cell means that would constitute
an effect of interest to the investigator. The values should be the kinds of means that
could realistically be observed for the strains, treatment, and phenotype in question,
as suggested by preliminary data or the published literature. If the investigator is
seriously interested in the possibility of a strain treatment interaction, then the
numerical model must express such an interaction. A computation is done to find
the sample size needed to detect the interaction as well as the main effects. In almost
all situations, the sample sizes needed to detect the three effects will be different,
but the study can actually be done with just one sample size, so the best option is
to choose the largest sample size needed to detect the smallest effect of interest.
This will ensure that power to detect the other two effects will be even larger than
the nominal level of power used in the calculation.
Consider the example shown in Table 10.5 where 4 strains reared in either
the usual lab environment or an enriched environment are trained to locate the sub-
merged platform in a water maze. Numbers represent mean latency to escape the
water after 4 days of training. It is expected that the standard deviation of latencies
within a strain will be about = 8.0 sec. In this case, there are major differences
among strains, a substantial effect of environment, and a noteworthy strain envi-
ronment interaction. For strain S1 the environmental treatment has no effect at all, it
improves latencies by 5 sec for strains S2 and S3, whereas it has a large effect of
10 sec on strain S4. This interaction appears to the researchers eye as something
important that ought to be detectable as a significant interaction effect by the ANOVA,
provided enough animals are tested to confer adequate power on the test.
TABLE 10.5
Calculation of Effect Sizes from Hypothetical Results of 4 Strain 2
Environment Design
A. Hypothetical Group Means with Interaction
Deviation
S1 S2 S3 S4 Row Mean
from G
E1 10 15 20 25 17.5 +2.5
E2 10 10 15 15 12.5 2.5
Column mean 10 12.5 17.5 20 G = 15
Dev. from G 5 2.5 +2.5 +5
B. Group Means under Null Hypothesis of Additivity
S1 S2 S3 S4 Row Mean
E1 12.5 15 20 22.5 17.5
E2 7.5 10 15 17.5 12.5
Column mean 10 12.5 17.5 20
C. Interaction Deviations of Cell Means in A from Additive Model in B
S1 S2 S3 S4
E1 2.5 0 0 +2.5
E2 +2.5 0 0 2.5
1903_C010.fm Page 164 Wednesday, July 26, 2006 6:37 PM
The next step in the planning process is to determine the three effect sizes. This
is relatively straightforward for the two main effects; find the mean of all values in
a row or column and then find the population standard deviation of all means along
one margin. For the strain effect, this will be the standard deviation of the four
marginal means 10, 12.5, 17.5 and 20, which is M = 3.95 or about 4. For the
environmental treatment we find M = 2.5. Thus, the effect sizes for strain and
treatment main effects are f = 4/8 = .5 and f = 2.5/8 = .31, respectively. Sample
sizes can be found from the tables in Cohen8 or using the SamplePower program by
Borenstein et al.6 The tabular approach requires two steps because the power tables
are set up for a one-way design. For = .05 and power of 90%, Cohens8 table 8.4.4
calls for 15 subjects per group when four groups have means with f = 0.5. The full
factorial study actually has 42 = 8 groups, however. Thus, the next step is to divide
the 15 subjects per strain between the two treatment conditions, resulting in a final
choice of 8 subjects per group. Similarly for the environmental factor, Table 8.4.4
calls for 57 subjects per group, but when these are distributed across the four strains,
the choice is 15 per cell. Full details of the method to find cell sample size from
tabled n are provided in Cohen.8
The SamplePower program uses the same shortcut for finding sample size for
the main effects, while concealing the internal workings of the calculations from the
user. One can enter means for the marginal values of each main effect, but it is not
possible to enter means for all 8 cells separately. Consequently, the program will
not compute the effect size for the interaction. Instead, the user must have a good
understanding of the statistical definition of interaction in order to compute its effect
size using either an electronic calculator, a spreadsheet such as Excel, or a flexible
data analysis program. The procedure is shown in Table 10.5.
The hypothesis of interaction is tested against the null hypothesis that strain and
environment effects are additive. Statistically, interaction is defined as departures
from additivity. For the example in panel A of Table 10.5, the grand mean G of all
eight group means is 15 sec, and the deviations of the two environment means from
this grand mean are +2.5 and 2.5 sec, these being the average effects of each
environment across the 4 strains. For the strains, deviations from 15 range from 5
to +5 sec. To find the deviation of each cell mean that is expected when the main
effects are additive, we simply add the separate deviations. For group S1E1, we add
its environmental deviation +2.5 sec to its strain deviation 5 sec to obtain 2.5 sec
below 15 or 12.5 sec. The results of these additions are shown in panel B of Table
10.5. Note that the marginal means in panel B are identical to those in panel A; the
only difference is that panel A expresses strain environment interaction, whereas
panel B does not. Next, we find the deviation of each cell in panel A from the
additive model in panel B and enter these interaction effects into panel C. Note that
for strains S2 and S3, the means expected under an additive relation (panel B) are
the same as proposed in panel A, while strains S1 and S4 deviate from the additive
model. Interaction effect size is then the standard deviation of these eight interaction
deviations (M = 1.77) divided by = 8 to yield effect size f = 0.22. This value can
then be used to find sample size from either Cohens table 8.4.4 or the SamplePower
program where 0.22 must be entered into the appropriate box, yielding an estimated
sample size of 38 animals in each of the eight groups.
1903_C010.fm Page 165 Wednesday, July 26, 2006 6:37 PM
The power curves for the three effects in the ANOVA are portrayed in Figure
10.6. As has been demonstrated for several plausible kinds of interactions,24 at a given
sample size the power to detect the interaction effect is substantially inferior to the
power to detect the main effects. Likewise, to achieve the same level of power as for
the main effects, the sample size to detect the interaction effect must be considerably
larger. The exact ratio of sample sizes to detect interaction and main effects depends
strongly on the specific kind of interaction. To date, no author has provided a satis-
factory rule for judging what constitutes a small, medium and large interaction in
neurobehavioral genetics. The only rule offered here is that the interaction should be
interesting or noteworthy when an experienced investigator examines the pattern of
cell means. Unfortunately, when power to detect an interaction is low, the pattern
may look interesting to the educated eye, yet fail to achieve significance at = .05.
Bausell and Li2 devote two chapters to power and sample size analysis for inter-
actions in several kinds of factorial designs. They note that power to detect interaction
tends to be lower than for main effects and that a formal power analysis is essential
when the investigator is interested in the possibility of interaction effects. Murphy and
Myors17 observe: In general, the study is most likely to have higher power for testing
main effects and lower power for testing complex interactions, and they warn that
very large samples may be needed for detecting complex interaction effects.
0.8
Strain
Strain x Env.
Power
0.6
0.4 Env.
0.2
0.0 8 15 38
0 10 20 30 40
Sample Size in Each of 8 Groups
FIGURE 10.6 Power vs. sample size for main and interaction effects in an ANOVA for an
experiment with 4 strains reared in 2 environments (Table 10.5). The sample size needed to
yield a power of 90% is shown as an italicized numeral along the x axis.
1903_C010.fm Page 166 Wednesday, July 26, 2006 6:37 PM
TABLE 10.6
Genotype (G1, G2) by Environment (E1, E2) Experiment That Expresses a
Noteworthy Interaction Where the Environmental Effect on Genotype 2 Is
Twice the Magnitude of the Effect on Genotype 1
Group G1E1 G1E2 G2E1 G2E2
Mean j 10 20 20 40 cj2 = cjj 2
G effect cj 0.5 0.5 +0.5 +0.5 1.0 15 225
100
nG
0
0.5 1.0 1.5 2.0
Main Effect Size (/ )
FIGURE 10.7 Sample size needed to yield power of 95% when = .05 for an experiment
in which 2 strains are reared in two environments and the environmental effect on one strain
is twice the size of the effect on the other strain, as shown in Table 10.6. Sample size needed
to detect the interaction is substantially greater than to detect the main effects over a wide
range of effect sizes.
ACKNOWLEDGMENTS
Preparation of this chapter was supported in part by grant 45825 from Natural Science
and Engineering Research Council of Canada and grant 2 R01 AA012714 from the
National Institutes of Health. The author is grateful to Elizabeth Munn for assistance.
REFERENCES
1. The Oxford English Reference Dictionary, Oxford, Oxford University Press, 1996.
2. Bausell, R.B. and Li, Y.-F., Power analysis for experimental research: a practical
guide for the biological, medical, and social sciences, Cambridge, UK, Cambridge
University Press, 2002.
3. Belknap, J.K., Mitchell, S.R., OToole, L.A., Helms, M.L., and Crabbe, J.C., Type I
and type II error rates for quantitative trait loci (QTL) mapping studies using recom-
binant inbred mouse strains, Behav. Genet., 26: 149160, 1996.
1903_C010.fm Page 168 Wednesday, July 26, 2006 6:37 PM
4. Benjamini, Y., Drai, D., Elmer, G., Kafkafi, N., and Golani, I., Controlling the false
discovery rate in behavior genetics research, Behav. Brain Res., 125: 279284, 2001.
5. Borenstein, M., Cohen, J., Rothstein, H., Schoenfeld, D., Berlin, J., and Lakatos, E.,
Power and Precision, Englewood, NJ, Biostat, 2001.
6. Borenstein, M., Rothstein, H., Cohen, J., Schoenfeld, D., and Berlin, J., SamplePower
2.0, Chicago, IL, SPSS, Inc., 2000.
7. Chow, S.L., Statistical Significance: Rationale, Validity, and Utility, London, Sage
Publications, 1996.
8. Cohen, J., Statistical Power Analysis for the Behavioral Sciences, Hillsdale, NJ,
Erlbaum, 1988.
9. Crabbe, J.C., Wahlsten, D., and Dudek, B.C., Genetics of mouse behavior: interactions
with laboratory environment, Science, 284: 16701672, 1999.
10. Crusio, W.E., A note on the effect of within-strain sample sizes on QTL mapping in
recombinant inbred strain studies, Genes Brain Behav., 3: 249251, 2004.
11. Egger, M., Smith, G.D., and Phillips, A.N., Meta-analysis. Principles and procedures,
British Medical Journal, 315: 15331537, 1997.
12. Harlow, L.L., Mulaik, S.A., and Steiger, A.H., Eds., What if There Were No Signifi-
cance Tests? Mahwah, NJ, Lawrence Erlbaum Associates, 1997.
13. Hays, W.L., Statistics, 4th edition, New York, Holt, Rinehart, Winston, 1988.
14. Hedges, L.V. and Olkin, I., Statistical Methods for Meta-Analysis, Orlando, Academic
Press, 1985.
15. Kraemer, H.C. and Thiemann, S., How Many Subjects? Statistical Power Analysis in
Research, Newbury Park, Sage Publications, 1987.
16. Lander, E. and Kruglyak, L., Genetic dissection of complex traits: Guidelines for
interpreting and reporting linkage results, Nat. Genet., 11: 241247, 1995.
17. Murphy, K.R. and Myors, B., Statistical Power Analysis: a Simple and General Model
for Traditional and Modern Hypothesis Tests, Mahwah, NJ, Lawrence Erlbaum Asso-
ciates, 2004.
18. Neumann, P.E., Three-locus linkage analysis using recombinant inbred strains and
Bayes theorem, Genetics, 128: 631638, 1991.
19. Paigen, K. and Eppig, J.T., A mouse phenome project, Mamm. Genome, 11: 715717,
2000.
20. Rosenberg, M.S., Adams, D.C., and Gurevitch, J., MetaWin Version 2.0, Sunderland,
MA, Sinauer Associates, 1997.
21. Sokolowski, M.B., Genetic analysis of behavior in the fruit fly, Drosophila melano-
gaster, in Techniques for the Genetic Analysis of Brain and Behavior. Focus on the
Mouse, D. Goldowitz, D. Wahlsten, and R.E. Wimer, Eds. Amsterdam, Elsevier,
497512, 1992.
22. Thomas, L. and Krebs, C.J., A review of statistical power analysis software, Bulletin
of the Ecological Society of America, 78: 126139, 1997.
23. Wahlsten, D., Behavioral genetics and animal learning, in Psychopharmacology of
Aversively Motivated Behaviors, H. Anisman and G. Bignami, Eds. New York, Ple-
num, 63118, 1978.
24. Wahlsten, D., Insensitivity of the analysis of variance to heredityenvironment inter-
action, Behav. Brain Sci., 13: 109120, 1990.
25. Wahlsten, D., Sample size to detect a planned contrast and a one degree-of-freedom
interaction effect, Psychol. Bull., 110: 587595, 1991.
26. Wahlsten, D., Experimental design and statistical inference, in Molecular-Genetic
Techniques for Behavioral Neuroscience, W.E. Crusio and R.T. Gerlai, Eds. Amster-
dam, Elsevier, 4057, 1999.
1903_C011.fm Page 169 Wednesday, July 26, 2006 6:54 PM
CONTENTS
Introduction............................................................................................................ 169
11.1 Association Studies: Definitions and Mechanisms ..................................... 170
11.1.1 Case Control Association Study ...................................................... 170
11.1.2 Family-Based Association Studies .................................................. 171
11.2 Advantage of Association Studies............................................................... 172
11.2.1 Power................................................................................................ 172
11.2.2 Linkage Disequilibrium ................................................................... 173
11.3 Limitations of Association Studies.............................................................. 173
11.3.1 False Positives: Linkage Disequilibrium
and Control Groups.......................................................................... 174
11.3.2 Statistical Power: Accept or Reject an Association........................ 175
11.3.3 Consequences in Neurobehavioral Genetics ................................... 175
11.4 Progress in Neurobehavioral Genetics ........................................................ 176
11.4.1 One Gene Many Phenotypes (the Case
of Phenotypical Heterogeneity) ....................................................... 177
11.4.2 One Phenotype Many Genes (the Case
of Genetic Heterogeneity)................................................................ 177
11.4.3 Interaction between the Environment and Genetic
Vulnerability..................................................................................... 178
11.5 Conclusions .................................................................................................. 180
References.............................................................................................................. 180
INTRODUCTION
In the beginning of the twenty-first century, the fiftieth anniversary of the discovery
of DNA was celebrated, and despite the rapid progress in both human genetics and
molecular biological technologies since, neurobehavioral genetics has provided us
with only its first promising successes. The human genome has a physical size of
3109 base pairs that encode and regulate approximately 32,000 genes. Some of
169
1903_C011.fm Page 170 Wednesday, July 26, 2006 6:54 PM
these genes play a role in the aetiology of psychiatric disorders. However, the model
of segregation for these disorders does not follow a simple Mendelian trait but
follows complex traits where both polygenetic and environmental factors influence
the susceptibility to neurobehavioral diseases. The use of association studies to
complement linkage analyses provides a powerful tool to identify the genetic com-
ponent in psychiatric disorders.
TABLE 11.1
Measures of Association
Phenotype Genetic Marker
Present Absent Total
Affected a b a+b
Controls c d c+d
Total a+c b+d N
Notes: 2 (qualitative approach) = (obscalc)/calc. Where a, b, c, d are the observed (obs) numbers,
and (a+b)*(a+c)/N is the calculated (calc) value for a, and so on for b, c, d.
instead of a difference, whereas, the OR value is the ratio of odds for each population.
The significance of the association can be computed from the OR.1 Finally, AR
provides the strength of the association and reflects the impact of the allele in
explaining the phenotype, since it is dependent on the OR and also the allele frequency
in the general population.
The case/control association study is a useful tool that provides a powerful
statistical test. However, one major limitation of association analysis can be caused
by population stratification, i.e., when cases and controls are not ethnically matched.
A/B C/D
A/C B/D
These tests include TRANSMIT program, which can compute association for
marker(s) or haplotype(s) to affected subjects or to one randomly selected from
families, either in the to absence of parental genotypes or to unknown haplotype
phase.10 These family-based association tests are now available in packages, e.g.,
FBAT.11
11.2.1 POWER
Whole genome scans followed by genetic linkage analyses are considered to be the
most powerful strategies to identify genes of heritable diseases. Linkage analyses
are successfully used for genetic disorders harboring Mendelian inheritance patterns,
but may not be applicable for common diseases that have a more complex pattern
of inheritance. Linkage analyses present several methodological limitations for deter-
mining polygenic disorders when the assumption of a single major gene is incorrect,
or genetic heterogeneity and sporadic cases are present, or penetrance of suscepti-
bility genotypes is reduced, or when the genetic aetiology is not well defined.12
These features all characterize the genetics of animal and human neurobehavior, and
therefore association studies may be more relevant.
In fact, linkage analyses are comparisons of the likelihood to observe a link
between markers and loci (transmitted together with greater frequency than 50%),
while association studies compare discrete traits and alleles (found together with
greater frequency than by chance alone). Linkage analyses attempt to detect physical
mapping on the chromosome while association studies look for a causal relation-
ship. Linkage analyses also need the screening of multigenerational families, while
association studies require only the studying of unrelated individuals within a
population, or triad families. According to these characteristics, association studies
should be the preferred initial choice rather than linkage studies. Association studies
are more appropriate for polygenic (many additive genes, each with a small effect),
multiloci (two or more genes involved), multifactorial (epigenetic factors, such as
environment, play a significant role in the individual vulnerability), heterogeneous
(phenotype contains different entities that are dependent on several factors) disor-
ders or traits.
Furthermore, the power of association studies vs. linkage studies to detect a
significant role of a gene involved in complex human disease has been compared
recently.13 When the genotypic RR is low (<2.0), then the number of families needed
to detect a significant linkage is so large that it becomes impractical to achieve
(12,000 to 300,000 for allele frequencies between 0.01 and 0.50). In comparison,
the number of triplets to be investigated with the TDT method is much lower (340
1903_C011.fm Page 173 Wednesday, July 26, 2006 6:54 PM
to 5800), for allele frequencies between 0.01 and 0.50. Simulations of the power
of association studies can be performed with the recent genetic power calculator
program (GPC; http://statgen.iop.kcl.ac.uk/gpc/).14 These results have shown that
association studies are relevant for genetic susceptibility or liability, but not for
genetic determinism in complex genetic diseases, such as neurobehavior disorders.
TABLE 11.2
Estimation of Each Sample Size Needed to Show a Significant Association
According to Allele Frequency in Controls and the Expected Difference
between Patients and Controls
Allele Frequency in Control (N)
phenotype) renders the choice of the control group even more critical. For example,
for research on a common disorder such as alcoholism, it may be more relevant to
recruit controls that drink alcohol regularly without developing alcohol abuse or
dependence, instead of abstainers, who may quickly develop alcoholism if they were
exposed to alcohol consumption.
There are not many psychiatric disorders for which some evidence exists about the
clear involvement of some vulnerability genes. Autism might be one of them. Jamain
1903_C011.fm Page 178 Wednesday, July 26, 2006 6:54 PM
environment (here aggression from the parents during childhood) are significantly
involved in the risk of aggression in adults.
A second interesting aspect is the greater chance to discover which environmen-
tal factors are interacting with genes, in order to propose a prevention strategy. For
example, the role of cannabis in schizophrenia has been the topic of numerous
discussions, the difficulty being to disentangle the role of cannabis as a marker (i.e.,
those who will ultimately develop schizophrenia are more prone to use cannabis)
or a risk factor (i.e., those who use cannabis increase their risk for schizophrenia).
One study, focusing on geneenvironment interactions, showed that cannabis is a
risk factor, but mainly for a subgroup of at-risk patients as having familial history
of psychosis.51
The adequate way to assess the existence of a geneenvironment interaction
depends on the selected sample. The golden standard would be based on the prospec-
tive collection of both environmental and lifestyle data and DNA, if (1) the disease,
(2) the vulnerability allele(s), and (3) major environmental risk factors are frequent
enough. Collecting DNA at the end of the cohort would expose to stratification bias
when DNA is not obtained from all patients and controls. For rare disorders, such as
autism, a case-control approach assessing geneenvironment interaction retrospec-
tively might be more appropriate, although more exposed to different risks. Such
biases could include unnatural selection (for more severe or treated patients), popu-
lation stratification, survivor bias (suicide being a major problem for bipolar disorder
for example) and imperfect recall (age at onset and number of lifetime major depres-
sive episodes are sometimes difficult to remember for old depressed patients).
Statistically, there are two ways to demonstrate a significant geneenvironment
interaction. Table 11.3 could be used, defining a, b, c and d as the frequency of affected
patients, taking into account presence vs. absence of the vulnerability factor and/or
the vulnerability allele. Testing the departure from the multiplicative model of inter-
action is the most frequently used way, comparing the relative risk of the population
b (exposed to both factors) from the relative risk of a (only environmental risk)
multiplied by the relative risk of d (only genetic risk). Relative risks are created with
the reference category c for which the relative risk is, by definition 1 (no exposition),
and using the 95% confidence interval of each relative risk to assess the significance
of the difference between these two relative risks. The alternative (the additive model)
is to use rate differences, the effect of genes (frequency d) plus the effect of environment
(frequency a) being smaller than the effect of both factors (b).
TABLE 11.3
Assessing GeneEnvironmental Interaction on the Basis of the Percentage
of Affected Patients Exposed (vs. Unexposed) to Each Factor
Environmental Exposure Genotype
Wild Type Variant
Exposed a b
Unexposed c d
1903_C011.fm Page 180 Wednesday, July 26, 2006 6:54 PM
11.5 CONCLUSIONS
Association studies have one of the highest risk of false-positive findings, are
exposed to numerous bias because of the selection of probands (frequently variable
in different studies, even when having the same diagnosis) and controls [from too
healthy (super-normal controls) to to close with the analysed phenotype (general
population)], are extremely sensitive to the problem of mismatching controls with
cases for ethnic background (stratification bias), and are particularly exposed to the
problem of the nonlinearity of linkage disequilibrium through the genome. Yet, the
association study approach is one of the most powerful, i.e., being able to discover
the role which has a small attributable risk, and has many advantages that are summed
up in this chapter. Thus, this approach should be used cautiously, in conjunction
with more specific techniques, and be more appropriate for complex traits, such as
polygenic (many additive genes each with a small effect), multiloci (two or more
genes involved), multifactorial (epigenetic factors such as environment play a sig-
nificant role in the individual vulnerability), heterogeneous (phenotype contains
different entities which are dependent on several factors) disorders or traits. As all
these characteristics may be particularly true for neuropsychiatric disorders, associ-
ation studies still have a role to play.
REFERENCES
1. Mantel, N. and Haenzel, W., Statistical aspects of the analysis of data from retro-
spective studies of disease, J. Natl. Cancer Inst., 22, 719, 1959.
2. Falk, C.T. and Rubinstein, P., Haplotype relative risk: an easy reliable way to construct
a proper control sample for risk calculation, Ann. Hum. Gen., 51, 227, 1987.
3. Ott, J., Statistical properties of the haplotype relative risk, Genet. Epidemiol., 6, 127,
1989.
4. Spielman, R.S., McInnis, R.E., and Ewens, W.J., Transmission Test for Linkage
Disequilibrium: The insulin gene region and insulin-dependent diabetus mellitus
(IDDM), Am. J. Hum. Genet., 52, 506, 1993.
5. Spielman, R. and Ewens, W., The TDT and other family-based tests for linkage
disequilibrium and association, Am. J. Hum. Genet., 59, 983, 1996.
6. Sham, P.C. and Curtis, D., An extended transmission/disequilibrium test (TDT) for
multi-allele marker loci, Ann. Hum. Genet., 59, 323, 1995.
7. Spielman, R.S. and Ewens, W.J., A sibship test for linkage in the presence of asso-
ciation: the sib transmission/disequilibrium test, Am. J. Hum. Genet., 62, 450, 1998.
8. Abecasis, G.R., Cardon, L.R., and Cookson, W.O., A general test of association for
quantitative traits in nuclear families, Am. J. Hum. Genet., 66, 279, 2000.
9. Martin, E.R. et al., A test for linkage and association in general pedigrees: the pedigree
disequilibrium test, Am. J. Hum. Genet., 67, 146, 2000.
10. Clayton, D., A generalization of the transmission/disequilibrium test for uncertain-
haplotype transmission, Am. J. Hum. Genet., 65, 1170, 1999.
11. Laird, N., Horvath, S., and Xu, X., Implementing an untried approach to family based
tests of association, Genet. Epidemiol., 19, S36, 2000.
12. Risch, S. et al., Ensuring the normalcy of normal volunteers, Am. J. Psychiatry,
147, 682, 1990.
1903_C011.fm Page 181 Wednesday, July 26, 2006 6:54 PM
13. Risch, N. and Merikangas, K., The future of genetic studies of complex human
diseases, Science, 273, 1516, 1996.
14. Purcell, S., Sham, P., and Daly, N.J., Parental phenotypes in family-based association
analysis, Am. J. Hum. Genet., 76, 249, 2005.
15. Patil, N. et al., Blocks of limited haplotype diversity revealed by high-resolution
scanning of human chromosome 21, Science, 294, 1719, 2001.
16. Chakravarti, A. and Nei, M., Utility and efficiency of linked marker genes for genetic
counseling. II. Identification of linkage phase by offspring phenotypes, Am. J. Hum.
Genet., 34, 531, 1982.
17. Jorde, L.B., Linkage disequilibrium as a gene-mapping tool, Am. J. Hum. Genet., 56,
11, 1995.
18. Owen, M. and McGuffin, P., Association and linkage: complementary strategies for
complex disorders, J. Med. Genet., 30, 638, 1993.
19. Peterson, A.C., et al., The distribution of linkage disequilibrium over anonymous
genome regions, Hum. Mol. Genet., 4, 887, 1995.
20. Nimgaonkar, V., In defense of association studies, Mol. Psychiatry, 2, 275, 1997.
21. Gibbons, R., Davis, J., and Hedeker, D., A comment on the selection of healthy
controls for psychiatric experiments, Arch. Gen. Psychiatry, 47, 785, 1990.
22. Shtasel, D. et al., Volunteers for biomedical research. Recruitment and screening of
normal controls, Arch. Gen. Psychiatry, 48, 1022, 1991.
23. Buckley, P. et al., Schizophrenia research: the problem of controls, Biol. Psychiatry,
32, 215, 1992.
24. Suarez, B.K. and Hampe, C.L., Linkage and association, Am. J. Hum. Genet., 54,
554, 1994.
25. Kendler, K., The super-normal control group in psychiatric genetics: possible evi-
dence for coaggregation, Psychiatric Genetics, 1, 45, 1990.
26. Greenberg, D.A., Linkage analysis of necessary disease loci versus susceptibility
loci, Am. J. Hum. Genet., 52, 135, 1993.
27. Cox, N.J. and Bell, G.I., Disease associations. Chance, artifact, or susceptibility
genes?, Diabetes, 38, 947, 1989.
28. Crowe, R.R., Candidate genes in psychiatry: an epidemiological perspective, Am. J.
Med. Genet., 48, 74, 1993.
29. Kidd, K.K., Association of disease with genetic markers: deja vu all over again, Am.
J. Med. Genet., 48, 71, 1993.
30. Paterson, D., Case control association studies in complex traits-the end of an era?,
Mol. Psychiatry, 2, 277, 1997.
31. Hodge, S.E., What association analysis can and cannot tell us about the genetics of
complex disease, Am. J. Med. Genet., 54, 318, 1994.
32. Baron, M., Association studies in psychiatry: a season of discontent, Mol. Psychiatry,
2, 278, 1997.
33. Inoue, K. and Lupski, J.R., Genetics and genomics of behavioural and psychiatric
disorders, Cur. Op. Gen Dev., 13, 303, 2003.
34. Gorwood, P., The generalized anxiety disorder and major depressive disorder comor-
bidity: an example of genetic pleitropy? Eur. Psychiatry, 19, 27, 2004.
35. Anguelova, M., Benkelfat, C., and Turecki, G., A systematic review of association
studies investigating genes coding for serotonin receptors and the serotonin trans-
porter: II. Suicidal behavior, Mol. Psychiatry, 8, 646, 2003.
36. Lin, P.Y., and Tsai, G., Association between serotonin transporter gene promoter
polymorphism and suicide: results of a meta-analysis. Biol. Psychiatry, 15, 1023,
2004.
1903_C011.fm Page 182 Wednesday, July 26, 2006 6:54 PM
37. Munafo, M.R., Clark, T., and Flint, J., Does measurement instrument moderate the
association between the serotonin transporter gene and anxiety-related personality
traits? A meta-analysis. Mol. Psychiatry, 10, 415, 2005.
38. You, J.S., Hu.,S., Chen, B., and Zhang, H.,G., Serotonin transporter and tryptophan
hydroxylase gene polymorphisms in Chinese patients with generalized anxiety dis-
order. Psychiatr. Genet., 15, 7, 2005.
39. Hoefgen, B., Schulze, T.G., Ohlraun, S., von Widdern, O., Hofels, S., Gross, M. et al.
The power of sample size and homogeneous sampling: association between the 5-
HTTLPR serotonin transporter polymorphism and major depressive disorder. Biol.
Psychiatry, 57, 247, 2005.
40. Cho, H.J., Meira-Lima, I., Cordeiro, Q., Michelon, L., Sham, P., Vallada, H., and
Collier, D.A., Population-based and family-based studies on the serotonin transporter
gene polymorphisms and bipolar disorder: a systematic review and meta-analysis.
Mol. Psychiatry. 2005 Apr. 12 (online).
41. Malhotra, A.K., Murphy, G.M., Jr, and Kennedy, J.L., Pharmacogenetics of psycho-
tropic drug response. Am. J. Psychiatry, 161, 780, 2004.
42. Leboyer, M., Bellivier, F., Nosten-Bertrand, M., Jouvent, R., Pauls, D., and Mallet,
J., Psychiatric genetics: search for phenotypes. Trends Neurosci., 21, 102, 1998.
43. Hariri, A.R., Mattay, V.S., Tessitore, A., Kolachana, B., Fera, F., Goldman, D., Egan,
M.F., and Weinberger, D.R., Serotonin transporter genetic variation and the response
of the human amygdala. Science, 297, 400, 2002.
44. Jamain, S. et al., Mutations of the X-linked genes encoding neuroligins NLGN3 and
NLGN4 are associated with autism. Nat. Genet., 34, 27, 2003.
45. Laumonnier, F. et al., X-linked mental retardation and autism are associated with a
mutation in the NLGN4 gene, a member of the neuroligin family. Am. J. Hum. Genet.,
74, 552, 2004.
46. Yan, J. et al., Analysis of the neuroligin 3 and 4 genes in autism and other neurop-
sychiatric patients. Mol. Psychiatry, 10, 329, 2005.
47. Gauthier, J., Bonnel, A., St-Onge, J., Karemera, L., Laurent, S., Mottron, L., Fom-
bonne, E., Joober, R., and Rouleau, G.A., NLGN3/NLGN4 gene mutations are not
responsible for autism in the Quebec population. Am. J. Med. Genet. B Neuropsychi-
atr. Genet., 132, 74, 2005.
48. Garrod, A., The incidence of alkaptonuria: a study in chemical individuality. Lancet,
2, 1616, 1902.
49. Hunter, D., Gene-environment interactions in human diseases. Nature reviews genet-
ics, 6, 87, 2005.
50. Caspi, A., McClay, J., Moffitt, T.E., Mill, J., Martin, J., Craig, I.W., Taylor, A., and
Poulton, R., Role of genotype in the cycle of violence in maltreated children. Science,
297, 851, 2002.
51. Henquet, C., Krabbendam, L., Spauwen, J., Kaplan, C., Lieb, R., Wittchen, H.U., and
van Os, J., Prospective cohort study of cannabis use, predisposition for psychosis,
and psychotic symptoms in young people. BMJ, 330, 11, 2005.
1903_C012.fm Page 183 Wednesday, July 26, 2006 6:55 PM
CONTENTS
12.1 INTRODUCTION
Traditionally, behavioral scientists and epidemiologists have tended to attribute
health and behavioral variation to environmental sources. The recent completion of
mapping the human genome opens a new era of discovering genetic influences on
behavioral and neuroscience phenotypes. Behavioral genetics offers a theoretical
and statistical approach for assessing both the genetic and environmental contribu-
tions to individual variation in health. This paper provides an introduction to the
concepts and statistical techniques used in behavioral genetic research.
183
1903_C012.fm Page 184 Wednesday, July 26, 2006 6:55 PM
not the only viable interpretations. These conditions may be due to adverse envi-
ronments in the womb.
One of the potential dangers in emphasizing the presence of genetic influences
and ignoring environmental factors is that society and policy makers will invest less
in changing environments. One must keep in mind that behavior genetics methods
operate under assumptions of the collective contribution of genes and environment,
not nature vs. nurture. So the interventions to improve health will be most effective
when combinations of environmental manipulations and gene therapies (when appro-
priate) are developed. In this article, a brief introduction to behavior genetic meth-
odology is presented, followed by a brief review of some recent research on health
using this approach from the past 15 years.
of their segregating genes and DZ twin pairs share on average 50%. Using data from
twin pairs and operating under this assumption, the heritability of a trait can be
analyzed using several techniques. Heritability is considered the proportion of phe-
notypic variance that is due to genetic variation. A popular simple method for cal-
culating heritability is to calculate twice the difference between the intraclass corre-
lations of MZ and DZ twin pairs (for the calculation of intraclass correlations see
Plomin et al.1). Shared environmental variance can be calculated by doubling the
intraclass correlation for the DZ twin pairs and subtracting the MZ intraclass corre-
lation. By subtracting the sum of the heritability and shared environmental estimates
from 1.00, an estimate of nonshared environmental variance is obtained.
While twin designs are a useful and interesting method for evaluating genetic
and environmental contributions, family studies are a popular alternative. Family
studies use the same basic analysis techniques as those in twin studies, but the
expected covariances for pairs of relatives are estimated. The following table high-
lights some of the common associations used in family studies.
1.00
1.00 MZ
0.50 DZ
E1 C1 A1 A2 C2 E2
e c h h c e
P1 P2
possibilities including new insights about diseases and how to treat and prevent
them. The trepidation about genetic approaches in the study of health by some
scientists may arise from the inherent power in identifying biological mechanisms
of disease and illness. Knowing the sequence of the genome, however, is only the
beginning. Equally important will be our knowledge of how the environment influ-
ences health, disease, and complex behaviors associated with health. This integrative
geneticenvironmental perspective places behavioral genetic approaches at the fore-
front of useful research designs and theory that will advance the search for answers
about health. Twin studies and other behavioral genetic designs have not been used
to their fullest potential to this end.
One of the areas of most interest in the search for answers about health and
disease is why disparities exist across economic and ethnic groups. It is a correct
assumption that basing investigations of health differentials across ethnic groups
solely on the basis of genetic differences will not yield accurate identification of the
mechanisms responsible for health disparities. Preconceived notions about geneti-
cally based racial inferiority have hindered advances in understanding and reducing
health disparities. Attempting to explain the differential health burden ethnic minor-
ities experience by genetic differences goes against probability given there are
considerably small genetic differences across racial groups and more variability
within each group. The role of genetic influences, however, cannot be completely
dismissed. The manner in which genes have the potential of playing a role in creating
health differentials requires further explanation. It is not genes defining individuals
from different ethnic groups that is key to the elucidation of health differentials per
se. Instead, describing health differentials as arising from insults to a complex system
represented by the interaction between genes and environments, which creates excess
burden of chronic illness and disease within some groups, is a more accurate
perspective.
In contrast to simply focusing on genetic explanations, there is ample information
that differences in environmental factors between ethnic groups account for dispar-
ities in health status. Previous research on the significant impact combinations of
socio-demographic and psychosocial factors in disease processes and complex
behaviors is perhaps our best indicator that science must avoid a reductionistic view.
Genetic reductionism assumes knowing and manipulating the genome will cure all
our ills. Rather, we must understand how genetic and environmental influences work
in concert to account for health conditions and the psychosocial variables that impact
health. Much of previous research has focused on the behaviors and social structures
that produce differences in health and disease across ethnic groups. One of the future
and formidable challenges to using the information ascertained from adding genetic
information to examinations of health differentials is to understand the underlying
effect genes have on health and aging within complex environments or contexts. We
may find that the polymorphisms that occur in genotypes are deleterious or protective
factors related to disease and health that are created, modified, or triggered by cultural
and contextual factors.
Complementary, interdisciplinary approaches are desperately needed to harness
the important findings that can from the Human Genome Project and continued
epidemiological research in the exploration of the underlying causes of health and
1903_C012.fm Page 188 Wednesday, July 26, 2006 6:55 PM
illness and the related psychosocial behaviors. Continued use of behavior genetic
designs (and modified designs such as those mentioned here) will significantly
advance our knowledge. Both conceptual and statistical advances in these methods
are still required. These methods have the potential to provide the backdrop for
exciting new revelations about how genes and environment work in concert to create
health and illness.
ACKNOWLEDGMENTS
Keith E. Whitfield is supported by the National Institute on Aging (5-RO1-AG13662-
04). Tracy L. Nelson is supported by USPHS Grant K01-DK-64647-01.
REFERENCES
1. Plomin R., DeFries, J. C. and McClearn, G. E. Behavioral Genetics: A Primer. W.
A. Freeman and Co. New York. 1990, P53 ff.
2. Barker, D. J. A new model for the origins of chronic disease. Medical Health Care
Philosophy, 4 (1), 3135, 2001.
3. Plomin, R. and Daniels, D. Why are children in the same family so different from
one another? Behavioral and Brain Sciences, 10 (1), 116, 1987.
4. Khoury, M. J., Beaty, T. H., and Cohen, B. H. Fundamentals of Genetic Epidemiology:
chapter 7, Oxford University Press, Chap 7, 1993.
5. Jinks, J. L. and Fulker, D. W. Comparison of the biomertical genetical, MAVA, and
classical approaches to the analysis of human behavior. Psychological Bulletin, 73,
311349, 1970.
6. Neale, M.C. and Cardon, L. R. (Eds.). Methodology for Genetic Studies of Twins and
Families. Dordrecht, the Netherlands: Kluwer Academic Press, 1992.
7. Loehlin, J. C. Latent Variable Models. Baltimore: Lawrence Erlbaum, 1987.
8. Jreskog, K. G. and Srbom, D. LISREL 7: A Guide to the Program and Applications
(2nd ed.). Chicago: SPSS, Inc., 1989.
9. Segal, N.L. The importance of twin studies for individual differences research. J.
Counsel. Dvlp., 68(6), 612622, 1990.
1903_C013.fm Page 189 Wednesday, July 26, 2006 7:04 PM
13 GeneEnvironment
Interactions
Byron C. Jones and Leslie C. Jones
CONTENTS
189
1903_C013.fm Page 190 Wednesday, July 26, 2006 7:04 PM
can include internal as well as external events. Thus, exposure to drugs, hormones,
toxins, or teratogens in utero, nutritional status, viral infections, early rearing con-
ditions, and social situations in adulthood may all be considered as environment.
Under this condition, individuals inherit the same relevant alleles and are exposed
to the same environment as other family members. For example, some forms of
antisocial behaviors show significant heritability. It is also nearly certain that those
at genetic risk are exposed to family members who exhibit many of the same
behavioral characteristics. The importance of this correlation has obvious implica-
tions for developmental and intervention concerns.2
Children who are verbally and socially gifted tend to seek out social situations that
match their tendencies. This is termed niche-picking. Scarr and McCartney3 have
exploited genotypeenvironmental correlation to help explain some of the phenom-
ena in child development, parental styles, treatment of behavior disorders, and
childfamily relations.
Phenotypic variability (or more properly, variance) can be partitioned into a number
of components. Thus Vg is the genetic part and consists of additive and dominant
genetic effects. Ve is the environmental source and V is measurement error. Vgxe
is the component left over when Vg, Ve, and V are accounted for (in the long run,
we assume that V sums to zero).
Some graphical representations are shown below.
G1, G2
Response
E1 E2
FIGURE 13.1 In this representation, we can see that the environment has an effect on the
phenotype, but both genotypes respond equally to the environmental change.
1903_C013.fm Page 192 Wednesday, July 26, 2006 7:04 PM
G1
Response
G2
E1 E2
FIGURE 13.2 In this figure, we observe that there is no effect of the environment, but an
effect of genotype.
G1
Response
G2
E1 E2
FIGURE 13.3 This figure shows, not only a significant effect of genotype, a possible effect
of environment, not only but also a significant genotypeenvironment interaction.
G1
Response
G2
E1 E2
FIGURE 13.4 In this figure, we see a significant interaction between genotype and environ-
ment but no main effect of either genotype or environment.
activity on the order of 1.4- to 1.7-fold and cells transfected with the l allele evince
increased serotonin uptake compared with cells transfected with the s allele. Lesch
then performed an association study between allelic configuration and affective
disorders. Despite the rather dramatic differences in 5HT biology, the allelic differ-
ences accounted for rather modest variance in anxiety-related traits.
1903_C013.fm Page 193 Wednesday, July 26, 2006 7:04 PM
respectively), and it was determined that the depression scores of high-risk children
doubled those of children who had the s/s genotype but positive social supports.10
The authors concluded that social supports further moderated the interaction between
the 5HTTLPR gene and childhood maltreatment, reducing the (increased) risk for
depression among children with the s allele.10 This study both supports the
geneenvironment interaction model and emphasizes further the relative importance
of environmental factors.
In summary of the findings of the four studies in agreement, it can be concluded
that the s allele of the 5HTTLPR significantly increases susceptibility to depression
only in the presence of SLEs, and this increased vulnerability can be at least partially
mitigated for by other environmental factors, including social supports.
Hariri and colleagues12 used an anxiety- and fear-provoking task, involving
photos of fearful and angry facial expressions, in conjunction with fMRI, to provide
evidence of (right) amygdala hyper-reactivity among subjects carrying the s allele
of 5HTTLPR. Indeed, the association between 5HTTLPR and amygdala hyper-
reactivity fit nicely with the current neurobiological hypothesis of depression.13
Of particular interest, however, was that the subjects tested in this study were healthy
volunteers with no history of psychiatric illness, including depression. It seems
likely, therefore, that the amygdala hyper-reactivity among subjects carrying the s
allele indicates that the 5HTTLPR may in fact mediate the intensity of an individuals
reaction to negative stimuli, and, hence, increased susceptibility to SLE-induced
depression.
Pezawas and colleagues14 used voxel-based morphometry (VBM) to test the
hypothesis that allelic configuration of the 5HTTLPR involves neural circuits that
function in emotional regulation. VBM is a highly sensitive method that measures
differences in gray matter volume between subjects. Using VBM, Pezawas and
colleagues detected reduced gray matter in carriers of the s allele in the right
amygdala and the perigenual anterior cingulate cortex (pACC). The volume of pACC
and amygdala gray matter were shown to be positively correlated and functionally
connected, as shown by fMRI. Two subregions of the pACC are functionally con-
nected to the amygdala, one showing positive and the other showing negative cor-
related activity. The authors proposed that the three brain regions may form a
feedback loop, modulating amygdala reactivity. In carriers of the s allele, the cou-
pling of the amygdala to these regions was impaired, especially the connection
between the amygdala and the positively associated subregion of the pACC.
Thus, a body of research findings is accumulating that suggests that the s allele of
the 5HTTLPR causes neurobiological differences evident even in normal subjects
with no history of psychiatric illness. These neurobiological differences do predict
personality traits related to depression in healthy s-carriers, but do not necessarily
cause depression. When life stressors are not reported, s-carriers do not have an
increased risk of developing depression; however, when a sufficient amount of life
stress is experienced, s-carriers exhibit an increased risk toward major episodes
of depression.
Because the subjects in both the fMRI and VBM studies were healthy individ-
uals, the authors also concluded that the phenotypic effects were developmental.11
This conclusion is in agreement with recent evidence that the impact of reduced
1903_C013.fm Page 195 Wednesday, July 26, 2006 7:04 PM
120
80
40
0
NC100 NC900 NC100 NC900 NC100 NC900
Non-Handled
Handled
FIGURE 13.5 Handling reduced serum corticosterone response to stress (right two panels) under
isolate and group housing conditions in the low aggressive animals (NC100). Moreover, group
housing in this line also reduced the response and both treatments had additive effects on the
corticosterone response. The high aggressive line presents a different story, as there was no effect
of handling under either housing condition. (Reprinted from Garipy, J. -L. et al., Pharmacology
Biochemistry and Behavior, 73(1), 717. Copyright 2002, with permission from Elsevier.)
16 16 Non-Handled
Non-Handled
Handled
14 Handled 14
fmol/mg protein-1
12 12
10 10
8 8
6 6
4 4
2 2
0 0
NC100 NC900 NC100 NC900
FIGURE 13.6 The effect of infantile handling on the density of dopamine D1 receptors. Note
here that handling dramatically increased the density of dopamine receptors in the nucleus
accumbens. There was no effect of differential housing on this measure and there was no
effect of either handling or housing on D1 receptor densities in the caudate-putamen.
(Reprinted from Garipy, J. -L. et al., Pharmacology Biochemistry and Behavior, 73(1), 717.
Copyright 2002, with permission from Elsevier.)
400 400
300 300
200 200
100 100
0 0
NC100 NC900 NC100 NC900
25 25
Non-Handled Non-Handled
Handled
Handled
20 20
15 15
Freeze
Attack
10 10
5 5
0 0
NC100 NC900 NC100 NC900
FIGURE 13.7 Handling had no effect on latency to attack in the low aggressive line but
significantly shortened this measure in the high aggressive mice. The same was true for a
number of attacks, although in the low aggressive line, there was a nonsignificant trend toward
an increase in attack frequency in the low aggressive mice. Latency to freeze was also affected
by handling, more so in the low line than in the high line, and the number of times that the
animal was seen motionless (freezing) was reduced in the low line and abolished in the high
line, although there was a clear floor effect in these animals.
for increased risk, there is the possibility that this knowledge can help equip us for
more effective means of prevention, prophylaxis, and treatment. In a recent, provoc-
ative article, Tienari and colleagues24 related childhood rearing environments to risk
for developing schizophrenia-spectrum disorder (SPD) in young adulthood. The
subjects consisted of adoptees whose mothers had been diagnosed with (SPD) and
a group of adoptees whose mothers had not been so diagnosed. The rearing envi-
ronment was evaluated in terms of factors indicating adversity. When the adoptees
reached young adulthood, all were assessed for SPD. Among those individuals whose
biological mothers were free of SPD, the environment, adverse or benign, had no
effect on the frequency diagnosed with SPD. In the group whose biological mothers
had been diagnosed with SPD, those who were reared in an adverse environment
had a significant increase in SPD diagnoses. The value of interventiongiven
biologically derived risk in this exampleis clear.
The research in animals gives us the prospect to understand what systems seem
to be more affected by geneenvironment and what kinds of environments are
important. The kinds of genetic material that are involved are also important. For
example, does geneenvironment involve specific structural genes or promoters, for
example 5HTTPLR? Are housekeeping genes involved? Do genes involved in
geneenvironment involve specific networks?
REFERENCES
1. Cardno, A. G. and Gottesman, I. I. Twin studies of schizophrenia: from bow-and-
arrow concordances to Star Wars Mx and functional genomics. Am J Med Genet 97:
1217, 2000.
2. Plomin, R., Loehlin, J. D. and DeFries, J. C. Genetic and environmental components
of environmental influences. Developmental Psychology 21: 391402, 1985.
3. Scarr, S. and McCartney, K. How people make their own environments: a theory of
genotypeenvironment effects. Child Development 54: 424435, 1983.
4. Cloninger, C. R. Neurogenetic adaptive mechanisms in alcoholism. Science 236:
410416, 1987.
5. Lesch, K. P. Geneenvironment interaction and the genetics of depression. J. Psychi-
atry Neurosci 29: 174184, 2004.
6. Anguelova, M., Benkelfat, C., Turecki, G. A systematic review of association studies
investigating genes coding for serotonin receptors and the serotonin transporter: I.
Affective disorders. Mol Psychiat 8: 574591, 2003.
7. Hoefgen, B., Schulze, T. G., Ohlraun, S., Widdern, O. V., Hofels, S., Gross, M.,
Heidmann, V., et al. The power of sample size and homogeneous sampling: associ-
ation between the 5-HTTLPR serotonin transporter polymorphism and major depres-
sive disorder. Biol Psychiat 57: 247251, 2005.
8. Caspi, A., Sugden, K., Moffitt, T. E., Taylor, A., Craig, I. W., Harrington, H. L.,
McClay, J. M., et al. Influence of Life Stress on Depression: Moderation by a
polymorphism in the 5-HTT gene. Science 301: 386389, 2003.
9. Kendler, K. S., Kuhn, J. W., Vittum, J., Prescott, C. A., and Riley, B. The interaction
of stressful life events and a serotonin transporter polymorphism in the prediction of
episodes of major depression. Arch Gen Psych 62: 529535, 2005.
1903_C013.fm Page 199 Wednesday, July 26, 2006 7:04 PM
10. Gillespie, N. A., Whitfield, J. B., Williams, B., Heath, A. C., and Martin, N. G. The
relationship between stressful life events, the serotonin transporter (5-HTTLPR) gen-
otype, and major depression. Psychol Med 35: 101111, 2005.
11. Kaufman, J., Yang, B-Z., Douglas-Palumberi, H., Houshyar, S., Lipschitz, D., Krystal,
J. H., and Gelernter, J. Social supports and serotonin transporter gene moderate
depression in maltreated children. PNAS 101(49):1731617321, 2004.
12. Hariri, A. R., Drabant, E. M., Munoz, K. E, Kolachana, B. S., Mattay, B. S., Egan,
M. F., and Weinberger, D. R. A susceptibility gene for affective disorders and the
response of the human amygdala. Arch Gen Psych 62: 146152, 2005.
13. Hamann, S. Blue Genes: wiring the brain for depression. Nature neurosc 8: 701703,
2005.
14. Pezawas, L., Meyer-Lindenberg, A., Drabant, E. M., Verchinski, B. A., Munoz,
K. E., Kolachana, B. S., Egan, M.F., Mattay, V.S., Hariri, A.R., and Weinberger, D.R.
5-HTTLPR polymorphism impacts human cingulate-amygdala interaction: a genetic
susceptibility mechanism for depression. Nature Neurosc 8(6): 828834, 2005.
15. Ansorge, M. S., Zhou, M., Lira, A., Hen, R., Gingrich, J. A. Early-life blockade of
the 5-HT transporter alters emotional behavior in adult mice. Science 306(5697):
879881, 2004.
16. Treiman, D. M. and Levine, S. Plasma corticosteroid response to stress in four species
of wild mice. Endocrinology 84: 676680, 1969.
17. Harlow, H. F., and Soumi, S. Social recovery by isolation-reared monkeys. Proc Natl
Acad Sci USA 68: 15341538, 1971.
18. Levine, S. Developmental determinants of sensitivity and resistance to stress. Psych-
oendocrinology 30: 939946, 2005.
19. Rosenzweig, M. R. and Bennett, E. L. Psychobiology of plasticity: effects of training
and experience on brain and behavior. Behavioural Brain Research 78: 5765, 1996.
20. Henderson, N. D. Genetic influences on the behavior of mice can be obscured by
laboratory rearing. J Comp Physiol Psychol 72: 505511, 1970.
21. Raymond, L. N., Tokuda, S., Reyes, E. and Jones, B.C. Differential immune response
in two handled inbred strains of mice. Physiology and Behavior, 37: 295297, 1986.
22. Garipy, J-L., Rodriguiz, R. M. and Jones, B. C. Handling and genetic effects on the
stress system, social behavior and dopamine function. Pharmacol Biochem Behav,
73(1): 717, 2002.
23. Cairns, R. B. Aggression from a developmental perspective: genes, environments and
interactions. Ciba Found Symp 194: 4556, 1996.
24. Tienari, P., Wynne, L. C., Sorri, A., Lahti, I., Laksy, K., Moring, J., Naarala, M.,
Nieminen, P. and Wahlberg, K.-E. Genotypeenvironment interaction in schizophre-
nia-spectrum disorder. Br J Psychiat 184: 216222, 2004.
1903_C013.fm Page 200 Wednesday, July 26, 2006 7:04 PM
1903_C014.fm Page 201 Wednesday, July 26, 2006 7:05 PM
CONTENTS
14.1 INTRODUCTION
Genetics was discovered in the nineteenth century by an Austrian monk named
Mendel, who spent years observing the reproduction of pea plants (in those
days there was no HBO). Mendel noticed that baby pea plants often inherited
certain characteristics of the mommy and daddy pea plants, such as height, eye
color, and personality. Mendel found that, by mating a certain pea plant with a
certain other pea plant, he could cause a third pea plant to go into a violent
jealous rage. What can we learn from these experiments? I have no idea.
Dave Barry
In the beginning, there was the OGOD model: one-gene one-disease.1 Even before
the age of molecular genetics, the study of patterns of inheritance made it clear that
this model was not applicable to the study of personality traits. When the first
replicated reports were published of positive associations between a candidate gene
and a personality traitdopamine D4 receptor gene (D4DR) and novelty seeking
201
1903_C014.fm Page 202 Wednesday, July 26, 2006 7:05 PM
14.2 EPISTASIS
Plomin1 provides an excellent explanation of genegene interactions, commonly
referred to as epistasis. He explains that while dominance is an intralocus interaction
with a given allele interacting in a nonadditive way with the allele at the same locus
on the homologous chromosomeepistasis refers to the interlocus interaction of
alleles at other loci. In Plomins words, [C]onsider two loci (A and B) that affect
a phenotypic character. Both the additive genetic values and the dominance devia-
tions are summed across the two loci. However, a particular combination of a certain
allele at locus A and another at locus B may influence the phenotype in ways not
explainable by the additive and dominance effects. Epistasis refers to this sort of
effect (p. 292). It should also be pointed out that epistatic effects are not limited
to the interaction of only two genes, but may include any number of players.
Increasingly sophisticated analyses are being proposed, which should help carry this
work forward. For a good discussion of sample size requirements for finding
genegene and geneenvironment interactions in various models of association
studies (case-control, affected sib, and case-parent/ case only designs), see Gauder-
man.23 Holmans24 discusses the conditions under which the use of a modification of
logistic regression should be used to examine possible genegene interactions in
linkage studies. An expanded method for model fitting which uses a conditional
logistic approach and allows for the analysis of genegene interactions at unlinked
loci is presented by Cordell et al.25
14.6 SUMMARY
What, as asked above, can we learn from these experiments? One important lesson
might be that when we begin to think about genetic engineering for human person-
ality, caveat emptorlet the buyer be very very beware. We already know that many
human genes are highly pleiotropic, and that changing a gene could have far-ranging
but unintended effects. Now things have become even more complex. Not only the
multiple effects of any given gene must be taken into account, but also all possible
interactions with any number of other genes. A second lesson, perhaps more imme-
diate than the first, is that meta-analysis may not be the most appropriate way to
understand the matrix of replications and non-replications of the effects of individual
candidate genes on personality or temperament. It seems increasingly likely that
summing studies across different populations will yield false negatives, as differing
allele frequencies for other genes interacting with the candidate gene will be impact-
ing upon the dependent variable. Systematic estimation of allele frequencies in
different populations, and wide-ranging examination of possible interactions
between known candidate genes will be required in order to elucidate the true nature
of genetic factors in personality, and no less, in the search for the genetic components
of major mental illnesses.
REFERENCES
1. Plomin, R., DeFries, J.C., McClearn, G.E., and Rutter, M. (1997). Behavioral Genet-
ics. NY: W.H. Freeman and Company.
2. Ebstein, R.P., Novick, O., Umansky, R., Priel, B., Osher, Y., Blaine, D., Bennett, E.R.,
Nemanov, L., Katz, M., and Belmaker, R.H., (1996). Dopamine D4 receptor (D4DR)
exon III polymorphism associated with the human personality trait of Novelty Seek-
ing. Nat Genet 12, 7880.
3. Benjamin, J., Li, L., Patterson, C., Greenberg, B.D., Murphy, D.L., and Hamer, D.H.
(1996). Population and familial association between the D4 dopamine receptor gene
and measures of Novelty Seeking. Nat Genet 12, 8184.
4. Cloninger, C.R., Przybeck, T., Svrakic, D., and Wetzel, R. (1994). The Temperament
and Character Inventory: A guide to its development and use. St. Louis: Center for
Psychobiology of Personality, Washington University.
1903_C014.fm Page 206 Wednesday, July 26, 2006 7:05 PM
5. Lesch, K.P., Bengel, D., Heils, A., Sabol, S.Z., Greenberg, B.D., Petri, S., Benjamin,
J., Muller, C.R., Hamer, D.H., and Murphy, D.L. (1996). Association of anxiety-
related traits with a polymorphism in the serotonin transporter gene regulatory region.
Science 274:15271531.
6. Osher, Y., Hamer, D., and Benjamin, J. (2000). Association and linkage of anxiety-
related traits with a functional polymorphism of the serotonin transporter gene reg-
ulatory region in Israeli sibling pairs. Mol Psychiat 5:216219.
7. Lesch, K.P. (2001). Molecular foundation of anxiety disorders. J Neural Transm
108:717746.
8. Lakatos, K., Nemoda, Z., Birkas, E., Ronai, Z., Kovacs, E., Ney, K., Toth, I., Sasvari-
Szekely, M., and Gervai, J. (2003). Association of dopamine receptor gene and
serotonin transporter promoter polymorphisms with infants response to novelty. Mol
Psychiatry 8:9097.
9. Szekely, A., Ronai, Z., Nemoda, Z., Kolmann, G., Gervai, J., and Sasvari-Szekely,
M. (2004). Human personality dimensions of persistence and harm avoidance asso-
ciated with D4DR and 5-HTTLPR polymorphisms. Am J Med Genet 126B:106110.
10. Ebstein, R.P., Segman, R., Benjamin, J., Osher, Y., Nemanov, L., and Belmaker, R.H.
(1997). 5-HT2C Serotonin receptor gene polymorphism associated with human per-
sonality trait of reward dependence: interactins with D4DR and D3DR polymor-
phisms. Am J Med Genet 74:6572.
11. Kuhn, K.U., Meyer, K., Nothen, M.M., Gansicke, M., Papassotiropoulos, A., and
Maier, W. Allelic variants of dopamine receptor D4 (DRD4) and serotonin receptor
5HT2c (HTR2c) and temperament factors: replication tests. Am J Med Genet
88(2):16872.
12. Boulton, A.A. and Eisenhofer, G. (1998). Catecholamoine metabolism: From molec-
ular understanding to clinical diagnosis and treatment. Overview. Adv Pharmacol
42:273292.
13. Mannisto, P.T. (1998). Catechol O-methyltransferase: characterization of the protein,
its gene, and the preclinical pharmacology of COMT inhibitors. Adv Pharmacol
42:324328.
14. Weinshilboum, R. and Dunnette, J. (1981). Thermal stability and the biochemical
genetics of erythrocyte catechol O-methyl-transferase and plasma dopamine-beta-
hydroxylase. Clin Genet 19:426437.
15. Lotta T., Vidgren J., Tilgmann C., Ulmanen I., Melen K., Julkunen I., and Taskinen
J. (1995). Kinetics of human soluble and membrane-bound COMT: a revised mech-
anism and description of the thermolabile variant of the enzyme. Biochemistry
34:42024210.
16. Lachman, H.M., Papolos, D.F., Saito, T., Yu, Y.M., Szumlanski, C.L., and Weinshil-
boum, R.M. (1996). Human COMT pharmacogenetics: description of a functional
polymorphism and its potential application to neuropsychiatric disorders. Pharma-
cogenetics 6:243250.
17. Benjamin, J., Osher, Y., Kotler, M., Gritsenko, I., Nemanov, L., Belmaker, R.H., and
Ebstein, R.P. (2000). Association between TPQ traits and three functional polymor-
phisms: D4DR, 5-HTTLPR, and COMT. Mol Psychiat 5:96100.
18. Strobel, A., Lesch, K.P., Jatzke, S., Paetzold, F., and Brocke, B. (2003). Further
evidence for a modulation of NS by DR exon III, 5-HTTLPR, and COMT val/met
variants. Mol Psychiatry 8:371372.
19. McClearn, G. E. (2004). Nature and nurture: Interaction and coaction. Am J Med Gen
124B: 124130.
1903_C014.fm Page 207 Wednesday, July 26, 2006 7:05 PM
20. Murphy, D.L., Uhl, G.R., Holmes, A., Ren-Patterson, R., Hall, F.S., Sora, I., Detera-
Wadleigh, S., and Lesch, K.P. (2003). Experimental gene interaction studies with
SERT mutant mice as models for human polygenetic and epistatic traits and disorders.
Genes Brain Behav 2:350364.
21. Moore, J.H. (2003). The ubiquitous nature of epistasis in determining susceptibility
to common human diseases. Hum Hered 56:7382.
22. Anttila, S., Illi, A., Kampman, O., Mattila, K.M., Lehtimaki, T., and Leinonen, E.
(2004). Interaction between NOTCH4 and COMT genotypes in schizophrenia patients
with poor response to typical neuroleptics. Pharmacogenetics 14: 303307.
23. Gauderman, W.J. (2002). Sample size requirements for association studies of
genegene interaction. Am J Epidemiol 155: 478484.
24. Holmans, P. (2002). Detecting genegene interactions using affected sib-pair analysis
with covariates. Hum Hered 53:92102.
25. Cordell, H.J., Barratt, B.J., and Clayton, D.G. (2004). Case/pseudocontrol analysis
in genetic association studies: A unified framework for detection of genotype and
haplotype associations, genegene and geneenvironment interactions, and parent-
of-origin effects. Genet Epidemiol 26:167185.
1903_C014.fm Page 208 Wednesday, July 26, 2006 7:05 PM
1903_C015.fm Page 209 Wednesday, July 26, 2006 7:05 PM
15 Schizophrenia: Study
of a Genetically Complex
Phenotype
Michael F. Pogue-Geile and Irving I. Gottesman
CONTENTS
ABSTRACT
Genetic research on the phenotype of schizophrenia is reviewed to illustrate the strat-
egies, problems, and findings encountered in the study of a genetically complex,
relatively common, dysfunctional behavioral phenotype. Schizophrenia provides an
excellent case history for these purposes because it has one of the longest histories
and most socially costly presence of any psychopathological phenotype. Family, twin,
and adoption studies are consistent in indicating that genetic effects on schizophrenia
are both important and complex. Genetic model-fitting to risk data for different
relative classes also suggests that: (1) environmental experiences that are uncorrelated
among siblings probably interact with the relevant genotypes to contribute to (or protect
from) schizophrenia, (2) action of abnormal alleles in multiple genes seems likely
209
1903_C015.fm Page 210 Wednesday, July 26, 2006 7:05 PM
for most schizophrenia cases, and (3) phenotypically normal (non-schizophrenic) indi-
viduals with unexpressed genetic liability exist, confusing molecular genetic studies
as false negatives. Genetic linkage and association studies attempting to locate and
identify individual genes of probably modest effect have to-date produced numerous
positive and negative findings with frequent failures to replicate. However, a second
generation of more sophisticated studies suggests increased reasons for cautious opti-
mism that the nature of the genetic effects on schizophrenia will be revealed in the
foreseeable future.
15.1 INTRODUCTION
Schizophrenia, which was first defined and initially named dementia praecox by
the German psychiatrist Emil Kraepelin in 1896,31 is a diagnosis based upon abnor-
malities of experience and behavior that include hallucinations, delusions, disorga-
nized speech, affective flattening, and bizarre behavior (for current criteria see DSM
IV1). Individuals with the diagnosis suffer severe personal and family distress as
well as frequent chronic impairments in social and occupational functioning. Not
only is the illness usually clinically severe, it is also relatively common and wide-
spread, with approximately a 1% lifetime risk in the general population, world-
wide.16,19 It is therefore a major public health problem warranting concerted effort
to understand its etiology as a guide to effective prevention and treatment.
The aim of the present chapter is a circumscribed one. For pedagogical purposes,
we will consider schizophrenia as an example of a genetically complex behavioral
phenotype and will selectively survey research on genetic effects in order to demon-
strate the major questions, strategies, problems, and findings that have been con-
fronted in its study and that may therefore be relevant to other behavioral phenotypes.
This is a daunting task, as schizophrenia has both one of the longest and most active
histories of genetic research of any psychopathological phenotype, beginning with
Rdin57 in 191612,22 and continuing until today. With these caveats, it should be clear
that the following is of necessity not a comprehensive and critical review of all genetic
findings on schizophrenia.8,10,21,29,43
Schizophrenia 211
genetic causes is a complex one, then it can be argued that the particular phenotype
may not be the most useful one for genetic analysis. In which case, it makes sense
to begin to define new phenotypes and endophenotypes using genetic validating
criteria that may reflect gene effects more simply.13,61 This boot strapping strategy
in which the results of genetic studies serve to change the definition of the phenotype
studied is an important approach in general and one that appears to have particular
relevance to schizophrenia. However, for reasons of space we will focus here primarily
on the clinically defined phenotype of schizophrenia itself and not on potentially
related phenotypes, such as schizoaffective disorder, schizotypal personality disorder,
or various cognitive and biological abnormalities in the long causal chain from gene
to endophenotype to a named clinical phenotype, despite their potential interest and
importance.13,52,53,66 For similar reasons, we will also not consider in detail any poten-
tial genetic differences among the various operational definitions of clinical schizo-
phrenia itself.
TABLE 15.1
Risk of Definite or Probable Schizophrenia among Relatives of
Schizophrenic Probands Aggregated across Studies
% Genes Shared
Relationship to Schizophrenic with Observed
Proband Schizophrenic Proband Morbid Risk (%)1
Monozygotic cotwins 100 48
Dizygotic cotwins 50 17
Siblings 50 9
Offspring (1 parent affected) 50 13
Parents 50 6
Second degree2 25 34
General population 0 1
1 Estimate of the probability of receiving a diagnosis of schizophrenia at some time during ones life
(all risks age-corrected, except twins).
2 Includes half-siblings, uncles/aunts, nephews/nieces, and grandchildren.
Source: After Gottesman I.I., Schizophrenia Genesis: The Origins of Madness. WH Freeman & Co.,
New York, 1991. With permission.
primarily from observing an increased age adjusted lifetime risk for schizophrenia
among the different relative classes (e.g., parents, offspring, siblings, etc.) compared
with the lifetime risk among control samples or the general population, which is
usually estimated to be about 1%.19 As can be seen in Table 15.1, risks for schizo-
phrenia are increased among all classes of relatives, indicating familiality for schizo-
phrenia, which by itself is consistent either with the effects of shared genes and/or
shared environmental experiences.
If, for the sake of argument at this stage, we were to presume that this observed
resemblance were due only to shared genes and not shared environment, then the data
suggest several tentative interpretations. First, there is a relative risk (risk among
family members/risk in the general population) among individuals sharing 50% of
genes with a schizophrenic proband (i.e., offspring and siblings) of approximately 10
times the general population rate. Although substantial, a risk of 13% among offspring
of a schizophrenic parent is considerably less than the 50% predicted from a rare,
single gene with dominance and complete penetrance (i.e., all individuals with a
disease genotype [cf. Huntingtons disease] become schizophrenic). The similarity
between risks in offspring and siblings argues against genetic recessive effects. Some-
what of an anomaly is the risk among parents, who also share 50% of their genes
with a schizophrenic proband, but only have about half the risk of offspring and
siblings. This discrepancy is usually interpreted as being a result of the common-sense
screening for mental health that occurs in the mating process and the resulting reduced
reproductive fitness (number of offspring) among schizophrenic patients. The risk
among second-degree relatives (e.g., uncles/aunts, nephews/nieces, and grandchil-
dren), who share 25% of their genes with the proband, is also less than half of the
risk among first-degree relatives, the prediction if only a single-gene were involved
1903_C015.fm Page 213 Wednesday, July 26, 2006 7:05 PM
Schizophrenia 213
(see below). Therefore, if these family data were due solely to genetic effects, their
pattern is inconsistent with a simple, single-gene transmission, but instead tentatively
suggests a complex mode of transmission involving perhaps multiple genes and/or
incomplete penetrance (i.e., not all individuals with a schizophrenia genotype develop
schizophrenia). Of course, based only on family study data, these findings could also
all be explained by experiences shared within families. Thus, although the family
study data are consistent with genetic influences on schizophrenia, they cannot prove
their existence (or rather, fail to rule them out).
Schizophrenia 215
effects and any observed phenotypic resemblance between adoptees and adoptive
relatives can be attributed to shared environmental effects. The two inferences, how-
ever, depend on the absence of selective placement of adoptees in adoptive homes
that induces a correlation between their genotypes and rearing environments that are
relevant to the phenotype under study. Selective placement is often indexed by an
observed resemblance for relevant phenotypes between biological parents (as an
estimate of the adoptees genotype) and adoptive parents (as an estimate of the
adoptees rearing environment). For example, it could be that adoptees were selec-
tively placed in adoptive homes based on matching the religion or language spoken
(e.g., Spanish or English) of the adoptive parents with that of the adoptees biological
parents as part of an adoption agency policy attempting to increase similarity between
adoptees and their adoptive families for ethnic/cultural characteristics. In such a
situation (selective placement for and environmental transmission of language spo-
ken) it would be obviously incorrect to attribute any observed resemblance for lan-
guage spoken between adoptees and biological relatives to genetic effects. Instead,
adoptees and biological relatives spuriously resemble each other for language spoken
due to the effects of selective placement and cultural transmission from adoptive
parents to adoptees. Similar examples (e.g., selective placement for hair color) could
be developed in which selective placement spuriously inflates estimates of environ-
mental effects based on resemblance between adoptees and their adoptive relatives.
Unfortunately, this important assumption of the adoption design is difficult to evaluate
for studies of schizophrenia because so little is known about which phenotypes are
relevant and should be correlated between biological parents and adoptive parents.
More optimistically, if researchers do not know what to measure to detect selective
placement, then it seems unlikely that adoption agencies do either.
Adoption studies of relatively rare, categorical phenotypes such as schizophrenia
rarely use unselected random samples of adoptees, but typically employ some
selective ascertainment strategy to increase statistical power and efficiency. Two
complimentary techniques have been used for estimating genetic effects based on
the resemblance between adoptees and their biological relatives. The adoptees
method ascertains index families through a schizophrenic biological parent who
has adopted away an offspring and control families through a demographically
matched non-schizophrenic biological parent who has adopted away an offspring.
The rate of schizophrenia in the adopted away offspring (adoptees) of the index
and control biological parents is then assessed and compared. If genetic influences
contribute to schizophrenia, then schizophrenia should be more frequent in the index
adoptees compared with the control adoptees. In contrast, the adoptees families
method ascertains index probands who are schizophrenic adoptees and control
probands who are demographically matched non-schizophrenic adoptees. The rate
of schizophrenia in the biological relatives (adoptees families) of the index
adoptees is then assessed and compared with that among the biological relatives of
the control adoptees.
The adoption strategy was pioneered in schizophrenia research by Heston in
Oregon in 196617 and Rosenthal56 and Kety26 from the NIMH who studied adoptees
and their families in Denmark. The results across these and other adoption studies68
have been consistent in indicating significant resemblance for schizophrenia between
1903_C015.fm Page 216 Wednesday, July 26, 2006 7:05 PM
biological relatives and adoptees using both adoptees and adoptees families methods,
thus suggesting the presence of genetic influences (see Gottesman16 for a detailed
review). For example, a final report from Kety and colleagues of their Danish adoptees
families study based on personal interviews found a morbid risk for schizophrenia
(using DSM-III criteria) among the first degree biological relatives of schizophrenic
adoptees of 8% compared to about 1% in the first degree biological relatives of control
adoptees.23,27 Inclusion of diagnoses of schizoaffective disorder, schizotypal, and
paranoid personality disorder as part of a schizophrenia spectrum increases the
prevalences to 24% vs. 5%, respectively. The convergence of evidence indicating
genetic influences from both adoption and twin studies is impressive. Although each
kind of study makes assumptions that could be argued separately, they are different
assumptions, and thus strengthen the case for an important role for genetic effects
among the distal causes of schizophrenia.
Schizophrenia 217
Schizophrenia 219
epistatic oligogenic model with 50% penetrance of the joint two locus genotype due
to nonshared experiences (to predict MZ twin risk satisfactorily). The probability
that a MZ twin of a schizophrenic proband shares alleles at two loci with the proband
is 100% (because MZ twins share all their genes) and given a penetrance of 50% the
model predicts a risk for MZ twins of 50% (100% 50% penetrance = 50%
predicted vs. about 50% observed). The probability that a first-degree relative shares
alleles at two loci with a proband is 25% (probability of sharing at one locus, 50%,
probability of sharing at second locus, 50%, = 25%) and thus the model predicts a
risk of 12.5% for first-degree relatives (25% 50% penetrance = 12.5% predicted
vs. about 12% observed). Similarly, the probability that second-degree relatives share
two loci with a proband is 6.25% (25% 25% = 6.25%) and their predicted risk
would be 3% (6.25% 50% penetrance = 3.125% predicted vs. 3% observed). This
rough example shows how well even a simple multigene model can fit the observed
data and MFT models fit similarly well with the heritability of liability estimated at
about 80%.39
Given such results, it is very likely that schizophrenias genetic contributions
involve two or more gene loci acting together. However, it is difficult to distinguish
among epistatic oligogenic models with a few necessary genes, MFT models with
many possible genes of quite small effect, or MFT models with some genes of
moderate effect along with many of small effect. Attempts at distinguishing some
of the multigene alternatives have investigated the properties of integrative mixed
models, in which there is both one gene of major effect and a polygene background
both contributing to schizophrenia risk.45 Although fitting mixed models to individual
data sets has usually resulted in a failure to detect a single gene of major effect,67
analyses of aggregated data sets suggest some interesting possibilities. Gottesman
and McGue14 found the following three general scenarios that were statistically
consistent with observed familial risks: (1) presence of a highly penetrant, but very
rare, single major locus and high polygene heritability, (2) presence of a low pene-
trance, but relatively common, single locus and high polygene heritability, and, (3)
absence of a single locus and high polygene heritability. Risch55 used the same data
and found that several oligogenic models also fit the observed data well. In summary,
it appears quite likely that most cases of schizophrenia are caused by two or more
gene loci acting along with nonshared environmental influences, although the further
details of the number and importance of the loci remain speculative at this time.9 It
is also important to note that the individual genes in such multigene systems need
not be specific to schizophrenia liability.
in the same journal issue.25 Since these initial reports, the history of linkage studies
in schizophrenia has been an active one, but until recently with few positive findings
and fewer replications. This brief history can be divided into two phases. Early linkage
studies often seemed to be predicated on a search for the gene for schizophrenia,
despite the evidence discussed above suggesting the involvement of multiple genes.
Designs often employed large multiplex families, with many affected members in an
attempt to identify families segregating a single major gene. Candidate chromosomal
regions were usually investigated based on suggestions from other observations.
Sample sizes were usually relatively small based on the optimistic assumption of
finding a gene of major effect. Methods of analysis usually emphasized parametric
linkage approaches that depend on specifying a model of transmission that includes
values for: dominance/recessivity, frequency of phenocopies, penetrance of geno-
types, and frequency of alleles. Incorrect specification of these parameters should
decrease power to detect linkage, however, testing of multiple models tends to
inflate alpha error rates. Phenotype definitions were multiple, including diagnoses
of schizophrenia, schizophrenia-spectrum, and often even more inclusive diagnostic
groupings. Failures to replicate positive findings during this early phase were often
optimistically interpreted in terms of genetic heterogeneity across studies.
The frustrations of these initial studies and methodological developments have
encouraged a more sophisticated and realistic wave of studies and reports.29,42,44,49
Increasingly the view is that the aim of linkage studies is to identify multiple genes,
probably each of small to modest effect. Studies now typically utilize complete
genome scan approaches and, given the multitude of statistical tests applied, failures
to replicate are more often interpreted from the perspective of statistical false positives,
leading to proposals for more stringent statistical criteria.33 Designs have been increas-
ingly based on affected sibling pairs and analyzed using nonparametric affected
pedigree member techniques.34,69 In contrast to the parametric methods, these analytic
approaches do not require specification of a transmission model and include only
schizophrenic relatives, thus avoiding the potential problems of incorrect model spec-
ification and diagnostic false negatives among non-schizophrenic relatives, although
at a cost of statistical power. Sample sizes have also become much larger. Meta-
analyses and pooling of data across studies have also become more common (some-
times mandated by grant funders) and have been very useful. Although still very much
of a moving target, results from some of these newer studies have generated consid-
erable enthusiasm recently about potential replicated linkages on chromosomes 1, 2,
6, 8, 13, and 22 and others that have been reported by several groups.43,49,59,63 However,
skeptics remain28 and it will be for future work to resolve these controversies.2,29,35,41
Schizophrenia 221
linkage disequilibrium with a locus that does. Linkage disequilibrium occurs when
alleles at different loci become correlated. Association studies have the advantage
of being much more powerful than linkage studies to detect loci of small effect;
however, they are much less sensitive than linkage studies to loci beyond a narrow
distance surrounding the marker. Association studies employ either case-control
designs or various sorts of within-family designs to avoid spurious findings due to
population stratification. Increasingly, association studies utilize multiple polymor-
phisms (i.e., single nucleotide polymorphisms, SNPs) within a gene, in order to
measure haplotypes, which increases power. A haplotype is the sequence of alleles
across loci on a particular chromosome. Given their precision, genome-wide asso-
ciation studies have been impractical to date, perhaps requiring well over 100,000
markers to screen the entire genome.56 New technology using microchips that allows
rapid genotyping of individuals for such dense maps of polymorphisms, however,
is becoming less expensive and will make this strategy increasingly useful. Tech-
niques, such as DNA pooling, in which genotyping is performed on pools of
groups of individuals (e.g., all cases vs. all controls), simultaneously, are also
reducing genotyping costs, although the statistical issues surrounding performing
thousands of tests remain complicating factors.
For these reasons, association studies to date have employed some sort of
candidate strategy to narrow the search to a smaller number of markers. Candidate
strategies have either been based on hypotheses drawn from models of pathology
or on positional information provided by linkage studies. For schizophrenia, there
have been numerous such association studies using candidate genes most frequently
chosen because they might be relevant to schizophrenias hypothesized pathophys-
iologies (e.g., genes coding for characteristics affecting dopamine or glutamate
neurotransmission). To date, most such association results have been negative and
positive findings have been rarely replicated. One particular variation of this strategy
that should hold special promise is exemplified by the history of the RGS4 (Regulator
of G protein signaling 4, chr 1) gene. RGS4 was identified as a candidate through
a gene expression study of post-mortem brain in schizophrenia, which found it to
be under-expressed.42 Based on this gene-expression finding, gene association stud-
ies using RGS4 as a candidate have been performed, with promising results to date.4
As knowledge of pathology increases and more genes are identified and sequenced,
these approaches should become more fruitful.48,49
In contrast, positional candidate strategies have generated considerable excite-
ment recently. Based on promising results from linkage studies mentioned above,
markers in several candidate chromosomal regions have been investigated using
association techniques. Although still controversial, several genes have been pro-
posed as potential contributors to schizophrenia liability based on this strategy. The
following genes have had at least some positive replications (along with some
negative): Dystrobrevin-binding protein 1 (DTNBP1 dysbindin, chr 6),64 Neuregulin
1 (NRG1, chr 8),62 and D-amino acid oxidase (DAO) and D-amino acid oxidase
activator (DAOA) (chr 13)5. Future study will show whether the current enthusiasm
for these particular genes is well placed; denser SNP-otyping may reveal
unexpected candidate genes very close to the genes named here.
1903_C015.fm Page 222 Wednesday, July 26, 2006 7:05 PM
15.7 CONCLUSIONS
The aim of this chapter is to provide a case history of the primary strategies used
to investigate genetic influences on the complex phenotype of schizophrenia, as well
as to provide a sketch of current findings. The Internet is your best friend for keeping
up with insertions and deletions in the corpus of received knowledge. From the long
history of research on schizophrenia, the following points seem certain enough to
conclude: (1) genes play a major causal role in schizophrenia, (2) effects of multiple
genes, combined with environmental and stressful experiential factors, epigenetic,
and stochastic factors are safe bets for explaining most schizophrenia cases, (3) overall
effects at any single gene locus are small with unhelpful positive predictive power,
(4) some sort of genetic heterogeneity across schizophrenia patients is very likely,
(5) environmental experiences that are uncorrelated among siblings probably interact
with the relevant genotypes to contribute to (or protect from) schizophrenia liability,
and (6) non-schizophrenic individuals with unexpressed genetic liability exist and
may become transmitters of liability. It is very likely that most other psychological
and psychopathological phenotypes share many of these features.
Although we have covered many aspects of genetic research on schizophrenia, we
have also omitted many important areas due to space constraints.18 Among those that
may prove most relevant to future genetic studies are work on environmental influences,
geneenvironment interaction, etiological heterogeneity, endophenotype and phenotype
definition, developmental neurobiology,65 and evolutionary issues. Despite these omis-
sions, we have tried to communicate at least a hint of the clinical importance, intellectual
challenge, and excitement to be found in the etiological study of schizophrenia.
REFERENCES
1. American Psychiatric Association: Diagnostic and Statistical Manual of Mental Dis-
orders (4th ed.), American Psychiatric Association Press, Washington, DC, 1994.
2. Badner, J.A. and Gershon, E.S., Meta-analysis of whole-genome linkage scans of
bipolar disorder and schizophrenia. Mol. Psychiatry, 2002, 7: 405411.
3. Cardno, A.G. and Gottesman, I.I., Twin studies of schizophrenia: From bow-and-
arrow concordances to star wars Mx and functional genomics. American Journal of
Medical Genetics, 2000, 97, 1217.
4. Chowdari, K.V., Mirnics K., Semwal, P. et al., Association and linkage analyses of
RGS4 polymorphisms in schizophrenia. Hum. Mol. Genet., 2002, 11: 13731380.
5. Chumakov, I., Blumenfeld M., Guerassimenkom, O., et al., Genetic and physiological
data implicating the new human gene for D-amino acid oxidase in schizophrenia.
Proc. Natl. Acad. Sci. USA, 2002, 99: 1367513680.
6. Falconer, D.S., The inheritance of liability to certain diseases estimated from inci-
dence among relatives. Annals of Human Genetics, 1965, 29: 5176.
7. Fischer, M., Psychoses in the offspring of schizophrenic monozygotic twins and their
normal co-twins. British Journal of Psychiatry, 1971, 115: 981990.
8. Gershon, E.S. and Cloninger, C.R., Genetic Approaches to Mental Disorders. Amer-
ican Psychiatric Association Press, Washington, DC, 1994.
9. Gottesman, I.I., Complications to the complex inheritance of schizophrenia. Clinical
Genetics, 1994, 46, 116123.
1903_C015.fm Page 223 Wednesday, July 26, 2006 7:05 PM
Schizophrenia 223
10. Gottesman, I.I., Schizophrenia Genesis: The Origins of Madness. WH Freeman &
Co., New York, 1991.
11. Gottesman, I.I. and Bertelsen A., Confirming unexpressed genotypes for schizophre-
nia: Risks in the offspring of Fischers Danish identical and fraternal discordant twins.
Archives of General Psychiatry, 1989, 46: 867872.
12. Gottesman, I.I. and Bertelsen, A., Editorial: Legacy of German psychiatric genetics:
Hindsight is always 20/20. American Journal of Medical Genetics, 1996, 67: 317322.
13. Gottesman, I.I. and Gould, T.D., The endophenotype concept in psychiatry:
Etymology and strategic intentions. Am. J. Psychiatry. 2003, 160: 636645.
14. Gottesman, I.I., and McGue, M., Mixed and mixed-up models for the transmission
of schizophrenia. In D. Cichetti, W. Grove (Eds.), Thinking Clearly about Psychology:
Essays in Honor of Paul E. Meehl. University of Minnesota Press, Minneapolis, 1991,
pp. 295312.
15. Gottesman, I.I. and Shields J., A polygenic theory of schizophrenia. Proceedings of
the National Academy of Sciences of the United States of America, 1967, 58: 199205.
16. Gottesman, I.I. and Shields, J., Schizophrenia: The Epigenetic Puzzle. Cambridge
University Press, New York, 1982.
17. Heston, L.L., Psychiatric disorders in foster home reared children of schizophrenic
mothers. British Journal of Psychiatry, 1966, 112: 819825.
18. Hirsch, S.R. and Weinberger, D.R., Schizophrenia. Blackwell, Oxford, UK, 1995.
19. Jablensky, A., Schizophrenia: Recent epidemiologic issues. Epidemiologic Reviews,
1995, 17, 1020.
20. James, J.W., Frequency in relatives for an all-or-none trait. Annals of Human Genetics,
1971, 35: 4749.
21. Kendler, K.S. and Diehl, S.R., The genetics of schizophrenia: A current, geneticepi-
demiologic perspective. Schizophrenia Bulletin, 1993, 19: 261285.
22. Kendler, K.S. and Zerbin-Rdin, E., Abstract and review of Studien ber Vererbung
und Intstehung geistiger Strungen. I. Zur Vererbung und Neuentstehung der Demen-
tia Praecox, American Journal of Medical Genetics, 1996, 67: 338342.
23. Kendler, K.S., Gruenberg, A.M., and Kinney, D.K., Independent diagnoses of adop-
tees and relatives as defined by DSM-III in the provincial and national samples of
the Danish adoption study of schizophrenia. Archives of General Psychiatry, 1994,
51: 456468.
24. Kendler, K.S., McGuire, M., Gruenberg, A.M., OHare, A., Spellman, M., and Walsh,
D., The Roscommon family study: I. Methods, diagnosis of probands and risk of
schizophrenia in relatives. Archives of General Psychiatry, 1993, 50: 527540.
25. Kennedy, J.L., Giuffra, L.A., Moises, H.W., Cavalli-Sforza, L.L., Pakstis, A.J., Kidd,
J.R., Castiglione, C.M., Sjogren, B., Wetterberg, L., Kidd, K.K., Evidence against
linkage of schizophrenia to markers on chromosome 5 in a northern Swedish pedigree.
Nature, 1988, 336: 167170.
26. Kety, S.S., Rosenthal, D., Wender, P.H., Schulsinger, F., The types and prevalence of
mental illness in the biological and adoptive families of adopted schizophrenics. In
D. Rosenthal, S.S. Kety (Eds.): The Transmission of Schizophrenia. Pergamon Press,
Oxford, 1968, pp. 345362.
27. Kety, S.S., Wender, P.H., Jacobsen, B., Ingraham, L.J., Jansson, L., Faber, B., Kinney,
D.K., Mental illness in the biological and adoptive relatives of schizophrenic adoptees:
Replication of the Copenhagen study in the rest of Denmark. Archives of General
Psychiatry, 1994, 51: 442455.
28. Kidd, K.K., Can we find genes for schizophrenia? American Journal of Medical
Genetics (Neuropsychiatric Genetics), 1997, 74: 104111.
1903_C015.fm Page 224 Wednesday, July 26, 2006 7:05 PM
29. Kirov G., ODonovan, M.C., and Owen, M.J., Finding schizophrenia genes. The
Journal of Clinical Investigation, 2005, 115: 14401448.
30. Klning, U., Schizophrenia in twins: Incidence and risk factors. Department of Psy-
chiatric Demography: University of Aarhus, Aarhus, Denmark. Unpublished doctoral
dissertation, 1996.
31. Kraepelin, E., Psychiatrie, Barth, Leipzig, 1896, pp. 426441.
32. Kringlen, E. and Cramer, G., Offspring of monozygotic twins discordant for schizo-
phrenia. Archives of General Psychiatry, 1989, 46: 873877.
33. Lander, E. and Kruglyak, L., Genetic dissection of complex traits: Guidelines for
interpreting and reporting linkage results. Nature Genetics, 1995, 11: 241247.
34. Lander, E. and Schork, N.J., Genetic dissection of complex traits. Science, 1994, 265,
20372048.
35. Lewis, C.M. et al. Genome scan meta-analysis of schizophrenia and bipolar disorder,
part II: Schizophrenia. Am. J. Hum. Genet. 2003, 73: 3448.
36. Luxenburger, H., Vorlufiger Bericht ber psychiatrische Serienuntersuchungen an
Zwillingen. Zeitscrift fr die gesamte Neurologie und Psychiatrie, 1928, 116:
297326.
37. Maier, W., Lichtermann, D., Minges, J., Hallmeyer, J., Heun, R., Benkert, O., and
Levinson, D.F., Continuity and discontinuity of affective disorders and schizophrenia.
Archives of General Psychiatry, 1993, 50, 871883.
38. McGue, M., When assessing twin concordance, use the probandwise not the pairwise
rate. Schizophrenia Bulletin, 1992, 18: 171176.
39. McGue, M. and Gottesman, I.I., Genetic linkage in schizophrenia: perspectives from
genetic epidemiology. Schizophrenia Bulletin, 1989, 15: 453464.
40. McGuffin, P. and Sturt, E., Genetic markers in schizophrenia. Human Heredity, 1986,
36: 6588.
41. McGuffin, P., Owen, M.J., and Gottesman, I.I., Psychiatric Genetics and Genomics.
London and New York: Oxford University Press, 2004.
42. Mirnics, K., Middleton, F.A., Stanwood, G.D., Lewis, D.A., and Levitt, P., Disease-
specific changes in regulator of G-protein signaling 4 (RGS4) expression in schizo-
phrenia. Mol. Psychiatry, 2001, 6: 293301.
43. Moises, H.W., Yang, L., Kristbjarnarson, H., Wiese, C. et al., An international two-
stage genome-wide search for schizophrenia susceptibility genes. Nature Genetics,
1995, 11: 321324.
44. Moldin, S.O. and Gottesman, I.I., Genes, experience, and chance in schizophrenia:
Positioning for the 21st century. Schizophrenia Bulletin, 1997, 23: 547561.
45. Morton, N.E. and MacLean, C.J., Analysis of family resemblance: III. Complex
segregation analysis of quantitative traits. American Journal of Human Genetics,
1974, 26: 489503.
46. Nance, M.A., Invited editorialHuntingtons disease: Another chapter rewritten.
American Journal of Human Genetics, 1996, 59, 16.
47. Onstad, S., Skre, I., Torgersen, S., and Kringlen, E., Twin concordance for DSM-III-
R schizophrenia. Acta Psychiatrica Scandinavica, 1991, 83: 395401.
48. Owen, M.J., ODonovan, M.C., and Harrison, P.J., Schizophrenia: Disorder of the
synapse? British Medical Journal, 330: 158159, 2005.
49. Owen, M.J., Williams, N.M., and ODonovan, M.C., The molecular genetics of
schizophrenia: New findings promise new insights. Mol. Psychiatry, 2004, 9: 1427.
50. Plomin, R., Owen, M.J., and McGuffin P., The genetic basis of complex human
behaviors. Science, 1994, 264: 17331739.
1903_C015.fm Page 225 Wednesday, July 26, 2006 7:05 PM
Schizophrenia 225
51. Pogue-Geile, M.F. and Rose R.J. Psychopathology: A behavior genetic perspective.
In T. Jacob (Ed.), Family Interaction and Psychopathology. Plenum Press, New York,
1987, pp. 629650.
52. Pogue-Geile, M.F. (2003). Schizophrenia spectrum disorders. In Encyclopedia of the
Human Genome, D. Cooper (Ed.), Vol. 5, pp. 185189. John Wiley & Sons, Ltd.,
Chichester, UK.
53. Prescott C.A. and Gottesman, I.I., Genetically mediated vulnerability to schizophre-
nia. Psychiatric Clinics of North America, 1993, 16, 245267.
54. Risch, N., Genetic linkage and complex diseases, with special reference to psychiatric
disorders. Genetic Epidemiology, 1990, 7: 316.
55. Risch, N., Linkage strategies for genetically complex traits: I Multilocus models.
American Journal of Human Genetics, 1990, 46: 222228.
56. Risch, N. and Merikangas, K., The future of genetic studies of complex human
diseases. Science, 1996, 273, 15161517.
57. Rosenthal, D., Wender, P.H., Kety, S.S., Schulsinger, F., Welner, J., stergaard, L.,
Schizophrenicsoffspring reared in adoptive homes. In D. Rosenthal, S.S. Kety (Eds.),
The Transmission of Schizophrenia. Pergamon Press, Oxford, 1968, pp. 377391.
58. Rdin, E., Zur Vererbung und Neuentstehung der Dementia Praecox. Springer, Berlin,
1916.
59. Schizophrenia Linkage Collaborative Group for chromosomes 3,6, and 8: Additional
support for schizophrenia linkage on chromosomes 6 and 8: A multicenter study.
American Journal of Medical Genetics (Neuropsychiatric Genetics), 1996, 67,
580594.
60. Sherrington, R., Brynjolfsson, J., Petursson, H., Potter, M., Dudleston, K., Barra-
clough, B., Wasmuth, J., Dobbs, M., and Gurling, H., Localization of a susceptibility
locus for schizophrenia on chromosome 5. Nature, 1988, 336, 164167.
61. Shields J. and Gottesman I.I., Cross-national diagnosis of schizophrenia in twins.
Archives of General Psychiatry, 1972, 27, 725730.
62. Stefansson, H. et al., Neuregulin 1 and susceptibility to schizophrenia. Am. J. Hum.
Genet. 71: 877892, 2002.
63. Straub, R.E., MacLean, C.J., ONeill, F.A., Burke, J., Murphy, B., Duke, F., Shinkwin,
R., Webb, B.T., Zhang, J., Walsh, D., Kendler, K.S., A potential vulnerability locus
for schizophrenia on chromosome 6p2422: Evidence for genetic heterogeneity.
Nature Genetics, 1995, 11, 287293.
64. Straub, R.E. et al. Genetic variation in the 6p22.3 gene DTNBP1, the human ortholog
of the mouse dysbindin gene, is associated with schizophrenia. Am. J. Hum. Genet,
2002, 71: 337348.
65. Thompson, J.L., Pogue-Geile, M.F., and Grace, A.A. Developmental pathology,
dopamine, and stress: A model for the age of onset of schizophrenia symptoms.
Schizophrenia Bulletin, 2004, 30, 875900.
66. Thompson, J.L., Watson, J.R., Steinhauer, S.R., Goldstein, G., and Pogue-Geile, M.F.
Indicators of genetic liability to schizophrenia: A sibling study of neuropsychological
performance. Schizophrenia Bulletin, 2005, 31, 8596.
67. Vogler, G., Gottesman, I.I., McGue, M.K., Rao, D.C., Mixed-model segregation
analysis of schizophrenia in the Lindelius Swedish pedigrees. Behavior Genetics,
1990, 20: 461472.
1903_C015.fm Page 226 Wednesday, July 26, 2006 7:05 PM
68. Wahlberg, K-E., Wynne, L.C., Oja, H., Keskitalo, P., Pyklinen, L., Lahti, I., Moring,
J., Naarala, M., Sorri, A., Seitamaa, M., Lksy, K., Kolassa, J., Tienari, P., Geneenvi-
ronment interaction in vulnerability to schizophrenia: Findings from the Finnish
adoptive family study of schizophrenia. American Journal of Psychiatry, 1997, 154:
355362.
69. Weeks, D.E. and Lange, K., The affected-pedigree-member method of linkage anal-
ysis. American Journal of Human Genetics, 1988, 42: 315326.
70. Wong, A.H.C., Gottesman, I.I., and Petronis, A., Phenotypic differences in genetically
identical organisms: the epigenetic perspective. Human Molecular Genetics, 2005,
14, 1118.
71. Woolf, C.M., Does the genotype for schizophrenia often remain unexpressed because
of canalization and stochastic events during development? Psychological Medicine,
1997, 27, 659668.
1903_C016.fm Page 227 Wednesday, July 26, 2006 8:36 PM
16 Genetics of Major
Affective Disorders
Wade Berrettini
CONTENTS
Introduction............................................................................................................ 227
16.1 Bipolar Disorders ......................................................................................... 228
16.1.1 Genetic Epidemiology: Family Studies........................................... 228
16.1.2 Genetic Epidemiology: Twin Studies .............................................. 228
16.1.3 Molecular Linkage Studies .............................................................. 230
16.1.4 Linkage Disequilibrium Studies ...................................................... 231
16.2 Recurrent Unipolar Disorders...................................................................... 236
16.2.1 Genetic Epidemiology: Family Studies........................................... 236
16.2.2 Genetic Epidemiology: Twin Studies .............................................. 237
16.2.3 Molecular Linkage Studies .............................................................. 237
Summary ................................................................................................................ 237
References.............................................................................................................. 238
INTRODUCTION
This chapter reviews some aspects of the genetic epidemiology and molecular genetic
research on bipolar disorders (BDs) and recurrent unipolar disorders (RUPs). Genetic
concepts of linkage and linkage disequilibrium (LD) are reviewed. Given that the
inherited susceptibilities for BD and RUP are explained by multiple genes of small
effect, simulations indicate that universal confirmation of vulnerability genes cannot
be expected due to power issues, sampling variation, and genetic heterogeneity. With
this background, several valid linkages of BD to genomic regions are reviewed,
including some that may be shared with schizophrenia. These results suggest that
nosology must be changed to reflect the genetic origins of the multiple disorders
that are collectively described by the term BD. The briefer history of RUP molecular
linkage and LD studies is also reviewed.
227
1903_C016.fm Page 228 Wednesday, July 26, 2006 8:36 PM
TABLE 16.1
Concordance Rates for Affective Illness in Monozygotic and
Dizygotic Twins*
Monozygotic Twins Dizygotic Twins
Concordant Pairs/Total Pairs (%) Concordant Pairs/Total Pairs (%)
Study
Luxenberger20 3/4 75.0 0/13 0.0
Rosanoff et al.21 16/23 69.6 11/67 16.4
Slater22 4/7 57.1 4/17 23.5
Kallman23 25/27 92.6 13/55 23.6
Harvald and Hauge24 10/15 66.7 2/40 5.0
Allen et al.18 5/15 33.3 0/34 0.0
Bertelsen et al.17 32/55 58.3 9/52 17.3
Totals 95/146 65.0 39/278 14.0
*Data not corrected for age. Diagnoses include both bipolar and unipolar illness.
N=30
60
50
40 1977
1977
30 2003
N=37
20 N=37
10
0
MZ DZ
FIGURE 16.1 Two BD twin studies, conducted 25 years apart, are compared for concor-
dance rates in monozygotic (MZ) and dizygotic (DZ) twins. These studies yield very similar
results, including estimates of heritability of ~80%. Twins were clinically ascertained
through a BPD proband. Co-twins were concordant if BPD or RUP was present. From
McGuffin P., et al. Arch Gen Psychiatry. 2003;60:497502, and Bertelsen A, et al. Br J
Psychiatry. 1977;130:330351.
has been achieved to detect the locus initially described. For example, if a locus
increases risk for BP illness by a factor of two, it may be necessary to study
approximately 200 affected sibling pairs in order to have adequate (90%) power to
detect such a locus.31
Unfortunately, few studies address this key issue. If 200 affected sibling pairs
are required to achieve adequate (90%) power to detect a previously described locus,
then a publication with less than 150 sibling pairs does not address the central issue
of confirmation. However, such power-limited publications may have an important
role in meta-analyses, in that they identify invaluable sources of additional data.
Comprehensive scans of the human genome have been completed with sufficient
numbers (e.g., >100) of BP individuals.3240 If a major locus (explaining >50% of
the risk in >50% of BPD persons) existed, it would have been detected in many of
these studies. Thus, no such major locus exists for BPD. There are several confirmed
reports of loci of smaller effect, which can be termed susceptibility loci. These loci
are neither necessary nor sufficient for disease, but increase risk for the disorder in
a non-Mendelian manner.
From these genome scans and from additional, smaller studies, a picture has
emerged in which there are approximately 10 confirmed DBD linkage regions across
the genome. It is highly probable that additional confirmed BD linkages will be
identified through future linkage studies. These BP linkage regions are confirmed
by virtue of at least one study with strong statistical significance (p < 0.0001) and
at least two confirmatory studies (p < 0.01). As noted elsewhere,41 in some cases,
these confirmed BP linkage regions overlap with schizophrenia linkage reports,
suggesting that the same loci may be involved in some aspects of both disorders. In
Figure 16.2, these are mapped onto an ideogram of the human genome.
The studies which support these findings are listed in Table 16.2, with the iden-
tified primary report cited as the first study with genome-wide statistical significance.
Two methodological approaches have been used to conduct meta-analyses of
BP linkage studies.73,74 Badner and Gershon73 analyzed linkage results using a
multiple scan probability approach, in which p values are combined across studies,
after adjusting for the size of the linkage region. These authors concluded that
two genomic regions, 13q32 and 22q,1113 were the most promising loci for DBD
disorder. Segurado et al.75 used the method of Levinson et al.74 which ranks the
p values across the genome of each study, then sums the rankings for each genomic
bin. In this approach, no genomic region reached genome-wide significance,
although the region that seemed most promising was the pericentromeric region
of 18.75
FIGURE 16.2 (SEE COLOR INSERT FOLLOWING PAGE 236) Confirmed linkage loci
for bipolar disorder (B), schizophrenia (S), and both disorders (*) on a diagram of the human
genome.
are two different nucleotides (from the possible four found among homo sapiens:
adenine [A], guanine [G], thymidine [T], and cytosine [C]). SNPs with a common
minor allele (frequency of approximately 20% or more) occur approximately every
1,000 base pairs of DNA.76 Analysis of closely spaced SNPs in outbred populations
suggests a complex pattern of inheritance in which recombination is inhibited in
a small region of DNA, such that blocks of DNA (containing multiple SNPs) tend
to be inherited intact over many generations.77 Thus, blocks of DNA are shared
among present-day individuals who may have had a common ancestor 10,000
generations ago. These blocks are variable in length and often contain multiple
SNPs, but among outbred human populations the block length rarely exceeds
approximately 100,000 base pairs. Alleles of SNPs within a block form a haplotype
(a set of alleles) that is usually inherited together across many generations. Such
SNPs are said to be in strong LD with one another. LD refers to the fact that
two (or more) alleles can be found together in unrelated individuals more often
than predicted by chance. The interested reader is referred to primary reports
concerning LD.77
LD is a useful tool to investigate the relatively small genomic regions that
have been implicated in the genetic origination of DBD through linkage studies
(Table 16.2). In this process, SNPs spaced across genes in the linkage region are
assessed in large groups (ideally at least several hundreds) of ethnically matched
1903_C016.fm Page 233 Wednesday, July 26, 2006 8:36 PM
TABLE 16.2
Confirmed Linkage Regions in Bipolar Disorder
18p11 Berrettini et al., Stine et al., 1995; Nothen et al., 1999; Paternal parent-of-
1994 and 1997 Turecki et al., 1999; Bennett et al., 2002 origin effect; see
Schwab et al., 1998;
Lin and Bale, 1997
21q22 Straub et al., 1994 Detera-Wadleigh et al., 1996; Smyth
et al., 1996; Kwok et al., 1999;
Morissette et al., 1999; Aita et al., 1999
22q11 Kelsoe et al., 2001 Detera-Wadleigh et al., 1999; Lachman Velocardiofacial
et al., 1997 syndrome region;
also a SZ locus: Gill
et al., 1996
18q22 Stine et al., 1995 McInnes et al., 1996; McMahon et al., See Freimer et al.,
1997; De Bruyn et al., 1996; McInnis 1996
et al., 2003
12q23 Morissette et al., Ewald et al., 2002; Maziade et al., 2002; Primary report in a
1999 Ekholm et al., 2003; Curtis et al., 2003; Canadian isolate;
Dawson et al., 1995 Abkevich et al., 2003
8q24 Cichon et al., 2001 McInnis et al., 2003; Dick et al.,
2003
13q32 Detera-Wadleigh Kelsoe et al., 2001; Potash et al., 2003; See Brzustowicz et al.,
et al., 1999 Liu et al., 2004; Badenohop et al., 2001 1999; Blouin et al.,
1998; Chumakov
et al., 2002
16p12 Ewald et al., 2002 Ekholm et al., 2003; Dick et al., 2003
4q32 Ekholm et al., 2003 Adams et al., 1998; McInnis et al., 2003;
Liu et al., 2004
4p15 Blackwood et al., Ewald et al., 2002; Cichon et al., 2001; See Ginns et al., 1998
1996 Morissette et al., 1999; Detera-
Wadleigh et al., 1999
cases and controls. Investigators compare allele and genotype frequencies among
groups of cases and controls.
There have been a multitude of LD studies in BD over the past decade. In a
typical report, allele and genotype frequencies in the BD cases and controls are
examined at a single candidate gene variant. If nominally significant differences in
allele or genotype frequencies are found between groups, the authors might conclude
that the variant influences the risk for BD disorder.
Most often, these studies have assembled a small group of BD patients and
unaffected controls from a population. These studies have typically employed one
1903_C016.fm Page 234 Wednesday, July 26, 2006 8:36 PM
with some studies having limited power to detect a small effect size. Thus, it is
understandable that conflicting studies are reported.
Another intensively studied candidate gene is the serotonin transporter (5HTT),
a functional candidate gene for which multiple BDLD studies have been published.
The 5HTT represents a logical candidate gene, as many antidepressants act through
binding to the 5HTT protein.99 There are two variants of the 5HTT that have been
studied in BD, and both have functional significance, based on in vitro analysis of
these noncoding polymorphisms. The first variant is an insertion/deletion polymor-
phism in the 5HTT promoter region. The shorter allele has much less transcriptional
activity than the longer allele.100 Moreover, the shorter allele has been associated
with anxiety-related personality traits in humans.101 The second variant is a variable
number of tandem repeats (VNTRs) polymorphism in intron 2. The two most
common alleles are the 10 and 12 repeats, which confer differential transcriptional
activity in an embryonic stem cell line.102 Collier et al.100 first reported that the 5HTT
intron 2 VNTR allele 12 was in LD with BD among patients from the U.K. Collier
et al. also reported that the short allele of the 5HTT promoter variant was more
common among 454 European BD and RUP patients, compared with 570 European
controls, although the statistical significance was marginal (p = 0.03), emphasizing
the small effect size involved. Analysis by genotype suggested that homozygosity
for the short allele was associated with BD (p < 0.05) and RUP ( p < 0.01).
Rees et al.104 confirmed the observation of Collier et al.100 that allele 12 of the
intron 2 VNTR was in LD with BD among 171 BD probands, compared with 121
controls (p = 0.031). Similarly, Rees et al. studied BD patients and controls, reporting
an excess of BD patients among individuals homozygous for the shorter promoter
allele, implying a recessive mode on inheritance. Note that the sample sizes for Rees
et al.104 and for Collier et al.103 were in the hundreds.
Vincent et al.105 studied an initial sample of approximately 100 BD probands
from Canada, confirming the observation that the promoter short allele was in LD
with BD, compared with approximately 100 controls; however, he then failed to
confirm this observation in a second set of approximately 100 BD probands. Sam-
pling variation and the small effect size, coupled with limited power of this sample
size, are probable explanations for these results.
Gutierrez et al.106 studied BD probands and controls from Spanish origin. They
reported no evidence for LD with 5HTT alleles. This may be secondary to the ethnic
background of patients or to small sample size. Bocchetta et al.107 studied approxi-
mately 55 Sardinian parentchild BD trios, finding no evidence for transmission
disequilibrium in the 5HTT gene, although sample size was a limiting factor in their
conclusions. Studying 123 BD parentchild trios of European origin, Mundo et al.108
reported no evidence for LD with the 5HTT promoter alleles. Mynett-Johnson
et al.109 studied approximately 100 Irish BD parentchild trios from multiplex fam-
ilies, reporting that a haplotype including the shorter promoter allele and a 3'UTR
SNP conferred risk for BD.
Kirov et al.110 studied 122 parentchild trios of British ethnic background, with
no nominally significant results at either polymorphism. In a study of 50 Indian BD
patients and controls using the VNTR variant, no evidence for LD was reported,111
this result being limited by the small sample size. From another ethnic perspective,
1903_C016.fm Page 236 Wednesday, July 26, 2006 8:36 PM
SUMMARY
Family, twin, and adoption studies of BD and RUP disorders are reviewed. They
are, in general, consistent with substantial heritable components to risk, with the
1903_C016.fm Page 238 Wednesday, July 26, 2006 8:36 PM
BD disorders having higher heritability than the RUP disorders. Multiple regions of
the genome, including 18p11, 18q22, 12q24, 21q21, 13q32, 4p15, 4q32, 16p12,
8q24, and 22q11, have been implicated by several independent groups in the genetic
origins of BD. It is likely that most of these regions will yield susceptibility genes
within the next 5 years through the application of LD mapping methods to large
sample sizes. LD approaches to candidate genes have yielded several promising
prospects, including G72 and BDNF for DBD. For each of these candidate genes,
there are several independent BD populations yielding data consistent with the
existence of one or more haplotypes as susceptibility sequences. Only several
genome scans for RUP disorders have been published, and confirmations are
required.
REFERENCES
1. Gershon, ES, Hamovit, J, Guroff, JJ, et al. A family study of schizoaffective, bipolar
I, bipolar II, unipolar, and normal control probands. Arch Gen Psychiatry
1982;39:115767.
2. Weissman, MM, Gershon, ES, Kidd, KK, et al. Psychiatric disorders in the relatives
of probands with affective disorder. Arch Gen Psychiatry 1984;41:1321.
3. Baron, M, Gruen, R, Anis, L, Kane, J. Schizoaffective illness, schizophrenia and
affective disorders: morbidity risk and genetic transmission. Acta Psychiatrica Scand
1983;65:25362.
4. Winokur, G, Tsuang, MT, Crowe, RR. The Iowa 500: Affective disorder in relatives
of manic and depressed patients. Am J Psychiatry 1982;139:20912.
5. Winokur G, Coryell, Keller, M, et al. A family study of manic-depressive (bipolar I)
disease. Is it a distinct illness separable from primary unipolar depression? Arch Gen
Psychiatry 1995;52:36773.
6. Helzer, JE and Winokur, G. A family interview study of male manic-depressives.
Arch Gen Psychiatry 1974;31,737.
7. James, NM and Chapman, CJ. A genetic study of bipolar affective disorder. Br J
Psychiatry 1975;126:44956.
8. Johnson, GFS and Leeman, MM. Analysis of familial factors in bipolar affective
illness. Arch Gen Psychiatry 1977;34:107483.
9. Angst, J, Frey, R, Lohmeyer, R, and Zerben-Rubin, E. Bipolar manic depressive
psychoses: results of a genetic investigation. Hum Genet 1980;55:23754.
10. Tsuang, MT, Winokur, G and Crowe, RR. Morbidity risks of schizophrenia and
affective disorders among first-degree relatives of patients with schizophrenia, mania,
depression, and surgical conditions. Br J Psychiatry 1980;137:497504.
11. Taylor, MA, Berenbaum, SA, Jampala, VC and Cloninger, CR. Are schizophrenia
and affective disorder related? Preliminary data from a family study. Am J Psychiatry
1993;150:27885.
12. Maier, W, Lichtermann, D, Minges, J, et al. Continuity and discontinuity of affective
disorders and schizophrenia. Results of a controlled family study. Arch Gen Psychiatry
1993;50:87183.
13. Gershon, ES, DeLisi, LE, Hamovit, J, et al. A controlled family study of chronic
psychoses. Arch Gen Psychiatry 1988;45:32836.
14. Kendler, KS, McGuire, M, Gruenberg, AM, et al. The Roscommon family study.
Arch Gen Psychiatry 1993;50:52740.
1903_C016.fm Page 239 Wednesday, July 26, 2006 8:36 PM
15. Potash, JB, Willour, VL, Chiu, YF, et al. The familial aggregation of psychotic
symptoms in bipolar disorder pedigrees. Am J Psychiatry 2001;158:125864.
16. Potash, JB, Chiu, Y-F, MacKinnon, DF, et al. Familial aggregation of psychosis in a
replication set of 69 bipolar pedigrees: Am J Med Genet 2003;116B:907.
17. Bertelsen, A, Harvald, B, and Hauge, M. A Danish twin study of manic-depressive
disorders. Br J Psychiatry 1977;130:33051.
18. Allen, MG, Cohen, S, Pollin, W, Greenspan, SI. Affective illness in veteran twins.
A diagnostic review. Am J Psychiatry 1974;131:12349.
19. McGuffin, P, Rijsdijk, S, Andrew, M, et al. The heritability of bipolar affective
disorder and the genetic relationship to unipolar depression. Arch Gen Psychiatry
2003;60: 497502.
20. Luxenberger, H. Psychiatrisch-neurologische Zwillings pathologie. Zentralblatt fur
diagesamte Neurologie and Psychiatrie 1930;14, 567,14580.
21. Rosanoff, AJ, Handy, L, and Plesset, IR. The etiology of manic-depressive syndromes
with special reference to their occurrence in twins. Am J Psychiatry 1935;91:72562.
22. Slater, E. The inheritance of manic-depressive insanity. Proceedings of the Royal
Society of Medicine 1953;29:98190.
23. Kallman, F. In: Hoch, PH and Zubin, J(eds). Depression. New York:Grune & Strat-
ton;1954:124.
24. Harvald, B, Hauge, M. Biopolar disorders, in, Genetics and the Epidemiology of
Chronic Diseases. (PHS Publ No 1163). Edited by Neal, JV, Shaw, MW, and Shull,
WJ. Washington, DC, U.S. Department of Health, Education, and Welfare, 1965 pp
6176.
25. Mendlewicz, J and Rainer, J.D. Adoption study supporting genetic transmission in
manic-depressive illness. Nature 1977;268:3279.
26. Wender, H, Kety, SS, Rosenthal, D, et al. Psychiatric disorders in the biological and
adoptive families of adopted individuals with affective disorders. Arch Gen Psychiatry
1986;43:9239.
27. Cadoret, RJ, Woolson, R, and Winokur, G. The relationship of age of onset in unipolar
affective illness to risk of alcoholism and depression in parents. J Psychiatr Res
1977;13:13742.
28. Von Knorring, AL, Cloninger, CR, Bohman, M, and Sigvardsson, A. An adoption study
of depressive disorders and substance abuse. Arch Gen Psychiatry 1983;40:94350.
29. Lander, E, and Kruglyak, L. Genetic dissection of complex traits: Guidelines for
interpreting and reporting linkage results. Nature Genet 1995;11:2417.
30. Suarez, B, Harpe, CL, and Van Eerdewegh, P. Problems of replicating linkage claims
in psychiatry. In: Gershon ES, Cloninger CR (eds). Genetic Approaches to Mental
Disorders. Proceedings of the 82nd Annual Meeting of the American Psychopatho-
logical Assn., Washington (DC): American Psychiatric Press;1994:2346.
31. Hauser, ER, Boehnke, M, Guo, SW, and Risch, N. Affected sib pair interval mapping
and exclusion for complex genetic traits. Genetic Epidemiology 1996;13:11737.
32. Detera-Wadleigh, SD, Badner JA, Berrettini, WH, et al. A highdensity genome scan
detects evidence for a bipolar-disorder susceptibility locus on 13q32 and other poten-
tial loci on 9q32 and 18p11.2. Proc Natl Acad Sci USA 1999;96:560456099.
33. Dick, DM, Foroud, T, Flury, L, et al. Genome-wide linkage analyses of bipolar
disorder: a new sample of 250 NIMH Genetics Initiative pedigrees. Am J Hum Genet
2003;73:10714.
34. McInnis, MG, Lan, T-H, Willour, VL, et al. Genome-wide scan of bipolar disorder
in 65 pedigrees: supportive evidence for linkage on 8q24, 18q22, 4q32, 2p12, and
13q12. Mol Psychiatry 2003;8:28898.
1903_C016.fm Page 240 Wednesday, July 26, 2006 8:36 PM
35. Bennett, P, Segurado, R, Jones, I, et al. The Wellcome trust UK-Irish bipolar affective
disorder sibling pair genome screen: First stage report. Mol Psychiatry 2002;7: 189200.
36. Rice, JP, Goate, A, Williams, JT, et al: Initial genome scan of the NIMH Genetics
Initiative bipolar pedigrees: Chromosomes 1, 6, 8, 10, and 12. Am J Med Genet
1997;74:24753.
37. Ekholm, JM, Kieseppa, T, Hiekkalinna, T, et al. Evidence of susceptibility loci on
4q32 and 16p12 for bipolar disorder. Hum Mol Genet 2003;12:190715.
38. Cichon, S, Schumacher, J, Muller, DJ, et al. A genome screen for genes predisposing
to bipolar affective disorder detects a new susceptibility locus on 8q. Hum Mol Genet
2001;10:293344.
39. Kelsoe, JR, Spence, MA, Loetscher, E, et al. A genome survey indicates a suscepti-
bility locus for bipolar disorder on chromosome 22. Proc Natl Acad Sci
2001;98:58590.
40. Liu, J, Juo, SH, Dewan, A, et al. Evidence for a putative bipolar disorder locus on
2p13-16 and other potential loci on 4q31, 7q34, 8q13, 9q31, 10q21 24, 13q32, 14q21
and 17q11-12. Molecular Psychiatry 2003;8:33342.
41. Berrettini, WH: Evidence for shared susceptibility in bipolar disorder and schizo-
phrenia. Am J Med Genet 2003;123C:5964.
42. Berrettini, W, Ferraro, T, Choi, H, et al. Linkage studies of bipolar illness. Arch Gen
Psych 1997;54:329.
43. Stine, OC, Xu, J, Koskela, R, et al. Evidence for linkage of bipolar disorder to
chromosome 18 with a parent-of-origin effect. Am J Hum Genet 1995;57:138494.
44. Nothen, MM, Cichon, S, Rohleder, H, et al. Evaluation of linkage of bipolar affective
disorder to chromosome 18 in a sample of 57 German families. Mol Psychiatry
1999;4:76.
45. Turecki, G, Grof, P, Cavazzoni, P, et al. Lithium responsive bipolar disorder, unilin-
eality and chromosome 18: A linkage study. Am J Med Genet 1999;88:41115.
46. Schwab, SG, Hallmayer, J, Lerer, B, et al. Support for a chromosome 18p locus
conferring susceptibility to functional psychoses in families with schizophrenia, by
association and linkage analysis. Am J Hum Genet 1998;63:1139.
47. Lin, JP, and Bale, SJ. Parental transmission and D18S37 allele sharing in bipolar
affective disorder. Genetic Epidemiology 1997;14:6658.
48. Straub, RE, Lehner, T, Luo, Y, et al. A possible vulnerability locus for bipolar affective
disorder on chromosome 21q22.3. Nature Genet 1994;8:2916.
49. Detera-Wadleigh, SD, Badner, JA, Goldin, LR, et al. Affected sib-pair analyses reveal
support of prior evidence for a susceptibility locus for bipolar disorder on 21q. Am
J Hum Genet 1996;58:127985.
50. Smyth, C, Kalsi, G, Brynjolfsson, J, et al. Further tests for linkage of bipolar affective
disorder to the tyrosine hydroxylase gene locus on chromosome 11p15 in a new series
of multiplex British affective disorder pedigrees. Am J Psychiatry 1996;153:2714.
51. Kwok, JB, Adams, LJ, Salmon, JA, et al. Non-parametric simulation-based statistical
analyses for bipolar affective disorder locus on chromosome 21q22.3. Am J Med
Genet 1996;88:99102.
52. Morissette, J, Villeneuve, A, Bordeleau, L, et al. Genome-wide search for linkage of
bipolar affective disorders in a very large pedigree derived from a homogeneous
population in Quebec points to a locus of major effect on chromosome 12q23-q24.
Am J Med Genet 1999;88:56787.
53. Aita, VM, Liu, J, Knowles, JA, et al. A comprehensive linkage analysis of chromo-
some 21q22 supports prior evidence for a putative bipolar affective disorder locus.
Am J Hum Genet 1999;64:21017.
1903_C016.fm Page 241 Wednesday, July 26, 2006 8:36 PM
54. Detera-Wadleigh, SD, Badner, JA, Berrettini, WH, et al. A high density genome scan
detects evidence for a bipolar-disorder susceptibility locus on 13q32 and other poten-
tial loci on 1q32 and 18p11.2. Proc Natl Acad Sci USA 1999;96:56049.
55. Lachman, HM, Kelsoe, JR, Remick, RA, et al. Linkage studies support a possible
locus for bipolar disorder in the velocardiofacial syndrome region on chromosome
22. Am J Med Genet 1997;74:1218.
56. Gill, M, Vallada, H, Collier, D, et al. A combined analysis of D22S278 marker alleles
in affected sib-pairs: support for a susceptibility locus for schizophrenia at chromo-
some 22q12. Am J Med Genet 1996;67:405.
57. McInnes, LA, Escamilla, MA, Service, SK, et al. A complete genome screen for
genes predisposing to severe bipolar disorder in two Costa Rican pedigrees. PNAS
1996;93:130605.
58. McMahon, FJ, Hopkins, PJ, Xu, J, et al. Linkage of bipolar affective disorder to
chromosome 18 markers in a new pedigree series. Am J of Hum Genet 1997;61:
1397404.
59. De Bruyn, A, Souery, D, Mendelbaum, K, et al. Linkage analysis of families with
bipolar illness and chromosome 18 markers. Biol Psychiat 1996;39:67988.
60. Freimer, NB, Reus, VI, Escamilla, MA, et al. Genetic mapping using haplotype,
association and linkage methods suggests a locus for severe bipolar disorder (BPI)
at 18q22-q23. Nature Genet 1996;12:43641.
61. Ewald, H, Flint, T, Kruse, TA, et al. A genome-wide scan shows significant linkage
between bipolar disorder and chromosome 12q24.3 and suggestive linkage to chro-
mosomes 1p22-21, 4p16, 6q14-22, 10q26 and 16p13.3. Mol Psychiatry 2002;7:
73444.
62. Maziade, M, Roy, MA, Rouillard, E, et al. A search for specific and common
susceptibility loci for schizophrenia and bipolar disorder: a linkage study in 13 target
chromosomes. Mol Psychiatry 2002;6:68493.
63. Curtis, D, Kalsi, G, Brynjolfsson, J, et al. Genome scan of pedigrees multiply affected
with bipolar disorder provides further support for the presence of a susceptibility
locus on chromosome 12q23-q24, and suggests the presence of additional loci on 1p
and 1q. Psychiatr Genet 2003;13:7784.
64. Dawson, E, Parfitt, E, Roberts, Q, et al. Linkage studies of bipolar disorder in the
region of the DarierAEs disease gene on chromosome 12q23-24.1. Am J Med Genet
1995;60:94102.
65. Abkevich, V, Camp, NJ, Hensel, CH, et al. Predisposition locus for major depression
at chromosome 12q22-12q23.2. Am J Hum Genet 2003;73:127181.
66. Badenohop, RF, Moses, MJ, Scimone, A, et al. A genome screen of a large bipolar
affective disorder pedigree supports evidence for a susceptibility locus on chromo-
some 13q. Mol Psychiatry 2001;6:396.
67. Brzustowicz, LM, Honer, WG, Chow, EWC, et al. Linkage of familial schizophrenia
to chromosome 13q32. Am J Hum Genet 1999;65:1096103.
68. Blouin, JL, Dombroski, B, Nath, SK, et al. Schizophrenia susceptibility loci on
chromosomes 13q32 and 8p21. Nat Genet 1998;20:70.
69. Chumakov, I, Blumenfeld, M, Guerassimenko, O, et al. Genetic and physiological
data implicating the new human gene G72 and the gene for D-amino acid oxidase in
schizophrenia. PNAS 2002;99:133657.
70. Adams, LJ, Mitchell, PB, Fielder, SL, et al. A susceptibility locus for bipolar affective
disorder on chromosome 4q35. Am J Hum Genet 1998;62:108491.
71. Blackwood, DHR, He, L, Morris, SW, et al. A locus for bipolar affective disorder on
chromosome 4p. Nature Genet 1996;12:42730.
1903_C016.fm Page 242 Wednesday, July 26, 2006 8:36 PM
72. Ginns, EI, St Jean, P, Philibert, RA, et al. A genome search for chromosomal loci
linked to mental health wellness in relatives at high risk for bipolar affective disorder
among the Old Order Amish. PNAS 1998;95:15531.
73. Badner, JA, and Gershon, ES. Meta-analysis of whole-genome linkage scans of
bipolar disorder and schizophrenia. Mol Psychiatry 2002;7:40511.
74. Levinson, DF, Levinson, MD, Sequardo, R, and Lewis, CM. Genome scan meta-
analysis of schizophrenia and bipolar disorder, part I: Methods and power analysis.
Am J Hum Genet 2003;73:1733.
75. Segurado, R, Detera-Wadleigh, SD, Levinson, DF, et al. Genome scan meta-analysis
of schizophrenia and bipolar disorder part III: bipolar disorder. Am J Hum Genet
2003;3:4962.
76. Venter, JC, Adams, MD, Myers, EW, et al. The sequence of the human genome.
Science 2001;291: 130451.
77. Gabriel, SB, Schaffner, SF, Nguyen, H, et al. The structure of haplotype blocks in
the human genome. Science 2002;296:22257.
78. Johnson, JW, and Ascher, P. Glycine potentiates the NMDA response in cultured
mouse brain neurons. Nature 2003;325:52931.
79. Schumacher, J, Jamra, RA, Freudenberg, J, et al. Examination of G72 and D-amino
acid oxidase as genetic risk factors for schizophrenia and bipolar affective disorder.
Mol Psychiatry 2004;9:2037.
80. Hattori, E, Liu, C, Badner, JA, et al. Polymorphisms at the G72/G30 gene locus on
13q33 are associated with bipolar disorder in two independent pedigree series. Am J
Hum Genet 2003;72:113140.
81. Chen, YS, Akula, N, Detera-Wadleigh, SD, et al. Findings in an independent sample
support an association between bipolar affective disorder and the G72/G30 locus on
chromosome 13q33. Mol Psychiatry 2004;9:8792.
82. Sklar, P, Gabriel, SP, McInnis, MG, et al. Family-based association study of 76
candidate genes in bipolar disorder: BDNF is a susceptibility locus. Mol Psychiatry
2002;7:57993.
83. Neves-Pereira, M, Mundo, E, Muglia, P, et al. The brain-derived neurotrophic factor
gene confers susceptibility to bipolar disorder: evidence from a family-based associ-
ation study. Am J Hum Genet 2002;71:6515.
84. Egan, MF, Goldberg, TE, Kolachana, BS, et al. Effect of COMT Val108/158 Met
genotype on frontal lobe function and risk for schizophrenia. PNAS 2001;98: 691722.
85. Nakata, K, Ujike, H, Sakai, A, et al. Association study of the BDNF gene with bipolar
disorder. Neurosci Lett 2003;337:1720.
86. Hong, CJ, Huo, SJ, Yen, FC, et al. Association study of a brain-derived neurotrophic
factor genetic polymorphism and mood disorders, age of onset and suicidal behavior.
Neuropsychobiology 2003;48:1869.
87. Harada, S, Agarwal, DP, and Goedde, HW. Aldehyde dehydrogenase deficiency as
cause of the flushing reaction to alcohol in Japanese. Lancet ii 1982;8253:982.
88. Thomasson, HR, Edenberg, HJ, Crabb, DW, et al. Alcohol and aldehyde dehydroge-
nase genotypes and alcoholism in Chinese men. Am J Hum Genet 1991;48:67781.
89. Craddock, N, Daniels, J, Roberts, E, et al. No evidence for allelic association between
bipolar disorder and MAOA gene polymorphisms. Am J Med Genet 1995;60:3224.
90. Lim, LC, Powell, J, Sham, P, et al. Evidence for a genetic association between alleles
of MAOA gene and bipolar affective disorder. Am J Med Genet 1995;60:32531.
91. Nothen, MM, Eggermann, K, Albus, M, et al. Association analyses of the MAOA
gene in bipolar affective disorder by using family based controls. Am J Human Genet
1995;57:9757.
1903_C016.fm Page 243 Wednesday, July 26, 2006 8:36 PM
92. Parsian, A, and Todd, RD. Genetic association between monoamine oxidase and
manic-depressive illness: comparison of relative risk and haplotype relative risk data.
Am J Med Genet 1997;74:4759.
93. Furlong, RA, Ho, L, Rubinszstein, JS, et al. Analysis of the MAOA gene in bipolar
affective disorder by association studies, meta-analyses and sequencing of the pro-
moter. Am J Med Genet 1999;88:398406.
94. Preisig, M, Bellivier, F, Fenton, BT, et al. Association between bipolar disorder and
MAOA gene polymorphisms: results of a multicenter study. Am J Psychiatry
2000;157:94859.
95. Kawada, Y, Hattori, M, and Dai, XY. Possible association between MAOA gene and
bipolar affective disorder. Am J Hum Genet 1995;56:3356.
96. Muramatsu, T, Matsushita, S, Kanba, S, et al. MAO genes polymorphisms and mood
disorder. Am J Med Genet 1997;19:4946.
97. Turecki, G, Grof, P, Cavazzoni, P, et al. MAOA: Association and linkage studies with
lithium responsive bipolar disorder. Psychiatric Genet 1999;9:136.
98. Kunugi, H, Ishida, S, Kato, T, et al. A functional polymorphism in the promoter
region of MAOA gene and mood disorders. Mol Psychiatry 1999;4:3935.
99. Ramammorthy, S, Bauman, AL, Moore, KR, et al. Antidepressant and cocaine sen-
sitive human transporter: Molecular cloning, expression and chromosomal localiza-
tion. Proc Natl Acad Sci 1993;90:25426.
100. Collier, DA, Arranz, MJ, Sham, P, et al. The serotonin transporter is a potential
susceptibility factor for bipolar affective disorder. Neuroreport 1996;7:16759.
101. Lesch, KP, Bengel, D, Heils, A, et al. Association of anxiety related traits with a poly-
morphism in the serotonin transporter gene regulatory region. Science 1996; 274: 152731.
102. Fickerstrand, CE, Lovejoy, EA, and Quinn, JP. An intronic polymorphic domain often
associated with susceptibility to affective disorders has allele dependent differential
enhancer activity in embryonic stem cells. FEBS Lett 1999;458:1714.
103. Collier, DA, Stober, G, Li, T, et al. A novel functional polymorphism within the
promoter of the serotonin transporter gene: possible role in susceptibility to affective
disorders. Mol Psychiatry 1996a;1:45360.
104. Rees, M, Norton, N, Jones, T, et al. Association studies of bipolar disorder at the
human serotonin transporter gene(hSERT;5HTT). Mol Psychiatry 1997;2:398402.
105. Vincent, JB, Masellis, M, Lawrence, J, et al. Genetic association analysis of serotonin
system genes in bipolar affective disorder. Am J Psychiatry 1999;156:1368.
106. Gutierrez, B, Arranz, MJ, Collier, DA, et al. Serotonin transporter gene and risk for
bipolar affective disorder: An association study in Spanish population. Biol Psychiatry
1998;43:8437.
107. Bocchetta, A, Piccardi, MP, Palmas, MA, et al. Family-based association study
between bipolar disorder and DRD2, DRD4, DAT, and SERT in Sardinia. Am J Med
Genet 1999;88:5226.
108. Mundo, E, Walker, M, Tims, H, et al. Lack of linkage disequilibrium between
serotonin transporter protein gene (SLC6A4) and bipolar disorder. Am J Med Genet
2000;96:37983.
109. Mynett-Johnson, L, Kealey, C, Claffey, E, et al. Multimarker haplotypes within the
serotonin transporter gene suggest evidence of an association with bipolar disorder.
Am J Med Genet 2000;96:8459.
110. Kirov, G, Rees, M, Jones, I, et al. Bipolar disorder and the serotonin transporter gene:
A family-based association study. Psychol Med 1999;29:124954.
111. Saleem, Q, Ganesh, S, Vijaykumar, M, et al. Association analysis of 5HT transporter
gene in bipolar disorder in the Indian population. Am J Med Genet 2000;96:1702.
1903_C016.fm Page 244 Wednesday, July 26, 2006 8:36 PM
112. Mendes de Oliveira, JR, Otto, PA, Vallada, H, et al. Analysis of a novel functional
polymorphism within the promoter region of the serotonin transporter gene (5-HTT)
in Brazilian patients affected by bipolar disorder and schizophrenia. Am J Med Genet
1998;81:2257.
113. Kunugi, H, Hattori, M, Kato, T, et al. Serotonin transporter gene polymorphisms:
ethnic difference and possible association with bipolar affective disorder. Mol Psy-
chiatry 1997;2:45762.
114. Furlong, RA, Ho, L, Walsh, C, et al. Analysis and meta-analysis of two serotonin
transporter gene polymorphisms in bipolar and unipolar affective disorders. Am J
Med Genet 1998;81:5863.
115. Coyle, N, Jones, I, Robertson, E, et al. Variation at the serotonin transporter gene
influences susceptibility to bipolar affective puerperal psychosis. Lancet
2000;356:14901.
116. Ospina-Duque, J, Duque, C, Carvajal-Carmona, L, et al. An association study of bipolar
mood disorder (type I) with the 5-HTTLPR serotonin transporter polymorphism in a
human population isolate from Colombia. Neurosci Lett 2000;292:199202.
117. Mendlewicz, J, Massat, I, Souery, D, et al. Serotonin transporter 5HTTLPR polymor-
phism and affective disorders: No evidence of association in a large European mul-
ticenter study. Eur J Hum Genet 2004;12:37782.
118. Weissman, MM, Wickramaratne, P, Adams, PB, et al: The relationship between panic
disorder and major depression: a new family study. Arch Gen Psychiatry 1993;
50:76780.
119. Mendlewicz, J and Baron, M. Morbidity risks in subtypes of unipolar depressive
illness: Differences between early and late onset forms. Br J Psychiatry
1981;139:4636.
120. Weissman, MM, Merikangas, KR, Wickramaratne, P, et al. Understanding the clinical
heterogeneity of major depression using family data. Arch Gen Psychiatry 1986;
43:4304.
121. Bland, RC, Newman, SC, and Orn, H. Recurrent and nonrecurrent depression. Arch
Gen Psychiatry 1986;43:10859.
122. Stancer, HC, Persad, E, Wagener, DK, and Jorna, T. Evidence for homogeneity of
major depression and bipolar affective disorder. J Psychiatr Res 1987;21:3753.
123. Ekholm, JM, Kieseppa, T, Hiekkalinna, T, et al. Evidence of susceptibility loci on
4q32 and 16p12 for bipolar disorder. Hum Mol Genet 2000;12:190715.
124. Reich, T, Van Eerdewegh, P, Rice, J, et al. The familial transmission of primary major
depressive disorder. J Psychiatr Res 1987;21:61324.
125. Kendler, KS, Neale, MC, Kessler, RC, et al. The clinical characteristics of major
depression as indices of the familial risk to illness. Br J Psychiatry 1994;165:6672.
126. Kendler, KS, Gardner, CO, and Prescott, CA. Clinical characteristics of major depres-
sion that predict risk of depression in relatives. Arch Gen Psychiatry 1999;56:322.
127. Merikangas, KR, Risch, NJ, and Weissman, MM. Co-morbidity and co-transmission
of alcoholism, anxiety and depression. Psychol Med 1994;24:6980.
128. Winokur, G, Cadoret, R, Dorsab, J, et al. Depressive disease: a genetic study. Arch
Gen Psychiatry 1971;24:13544.
129. Nurnberger, JI, Foroud, T, Flury, L, et al. Evidence for a locus on chromosome 1 that
influences vulnerability to alcoholism and affective disorder. Am J Psychiatry
2001;158:71824.
130. Sullivan, P, and Kendler, K. Genetic case-control studies in neuropsychiatry. Arch
Gen Psychiatry 2001;58:101524.
1903_C016.fm Page 245 Wednesday, July 26, 2006 8:36 PM
131. Bierut, LJ, Heath, AC, Bucholz, KK, et al. Major depressive disorder in a community-
based twin sample: Are there different genetic and environmental contributions for
men and women? Arch Gen Psychiatry 1999;56:55763.
132. Kendler, KS, Pedersen, N, Johnson, L, et al. A pilot Swedish twin study of affective
illness, including hospital and population-ascertained subsamples. Behav Genet
1995;25:21732.
133. Lyons, MJ, Eisen, SA, Goldberg, J, et al. A registry-based twin study of depression
in men. Arch Gen Psychiatry 1998;55:46872.
134. Kendler, KS, and Prescott, CA. A population-based twin study of lifetime major
depression in men and women. Arch Gen Psychiatry 1999;56:3944.
135. McGuffin, P, Katz, R, Watkins, S, and Rutherford, J: A hospital-based twin register of
the heritability of DSM-IV unipolar depression. Arch Gen Psychiatry 1996;53:12936.
136. Holmans, P, Zubenko, GS, Crowe, RR, et al. Genome-wide significant linkage to
recurrent early-onset major depressive disorder on chromosome 15q. Am J Hum Genet
2004;74:115467.
137. Zubenko, GS, Maher, B, Hughes, HB, et al. Genome-wide linkage survey for genetic
loci that influence the development of depressive disorders in families with recurrent
early-onset major depression. Am J Med Genet 2003;123B:118.
138. Zubenko, GS, Hughes, HG, Stiffler, JS, et al. Sequence variations in CREB1 co-
segregate with depressive disorders in women. Mol Psychiatry 2003;8:61118.
139. McInnis, MG, Lan, TH, Willour, VL, et al. Genome-wide scan of bipolar disorder
in 65 pedigrees: supportive evidence for linkage at 8q24, 18q22, 4q32, 2p12, and
13q12. Mol Psychiatry 2003;8:28898.
140. Cichon, S, Schumacher, J, Muller, DJ, et al. A genome screen for genes predisposing
to bipolar affective disorder detects a new susceptibility locus on 8q. Hum Mol Genet
2001;10:293344.
141. Liu J, Juo SH, Dewan A, et al. Evidence for a putative bipolar disorder locus on
2p13-16 and other potential loci on 4q31, 7q34, 8q13, 9q31, 10q21-24, 13q32, 14q21
and 17q11-12. Mol Psychiatry 2004;8:33342.
142. Kelsoe, JR, Spence, MA, Loetscher, E, et al. A genome survey indicates a possible
susceptibility locus for bipolar disorder on chromosome 22. Proc Natl Acad Sci
USA 2001;98:58590.
143. Dick, DM, Foroud, T, Flury, L, et al. Genome-wide linkage analyses of bipolar
disorder: a new sample of 250 NIMH Genetics Initiative pedigrees. Am J Hum Genet
2003;73:10714.
144. Potash, JB, Chiu, Y-F, MacKinnon, DF, et al. Familial aggregation of psychosis in a
replication set of 69 bipolar pedigrees: Am J Med Genet 2003;116B:907.
145. Berrettini, W, Ferraro, T, Goldin, L, et al. Chromosome 18 DNA markers and manic
depressive illness: evidence for a susceptibility gene. Proc Natl Acad Sci USA
1994;91:591821.
1903_C016.fm Page 246 Wednesday, July 26, 2006 8:36 PM
1903_C017.fm Page 247 Wednesday, July 26, 2006 7:06 PM
CONTENTS
247
1903_C017.fm Page 248 Wednesday, July 26, 2006 7:06 PM
17.1 OVERVIEW
In this chapter we describe our research on the behavior genetics of personality and
subjective well-being (SWB) in our chimpanzee cousins, a research program that
we plan to extend to the other great ape speciesorangutans, gorillas, and bonobos.
Although we had been familiar with classic behavior genetics designs before initi-
ating this research, we quickly realized that these approaches were not suitable for
studying great apes in naturalistic and semi-naturalistic settings. Our subjects could
not be selectively bred, they could not be conveniently grouped into sets of monozy-
gotic and dizygotic twins, and their pedigrees did not resemble trees but, instead,
kudzu. Fortunately, we stumbled upon two types of pedigree analysis that allowed
us to hack through the familial wilderness and estimate the relative contributions of
genetic and environmental effects on individual differences in chimpanzee person-
ality and SWB. The first part of this chapter will describe reasons why the behav-
iorgenetic study of chimpanzees is interesting and compelling. The second part
will describe how multiple regression analysis can be used to estimate relevant
variance components and will highlight the advantages and disadvantages of this
method. The third part will describe the use of mixed-model equations and a
restricted maximum likelihood analysis to estimate variance and covariance com-
ponents. Finally, we will offer concluding thoughts and suggest other potential uses
for these analyses.
1903_C017.fm Page 249 Wednesday, July 26, 2006 7:06 PM
Our primary obstacle was the low number of discrete family units in the sample. The
relative absence of discrete family units dramatically reduced the number of geneti-
cally independent pairs, thereby precluding use of most conventional techniques for
genetic analysis. Fortunately, approaches that consider the interrelationships between
all possible pairs of individuals in a sample enabled us to surmount this obstacle.
* The relationship between dependability and subjective well-being in chimpanzees may be accounted
for by the fact that the chimpanzee dependability factor is comprised of several items defining human
low-neuroticism.
1903_C017.fm Page 251 Wednesday, July 26, 2006 7:06 PM
studbook was the essential component to unraveling the behavior genetics of chim-
panzee personality and SWB.20 Since our studies, a new edition has been released.21
where i < j and Yi and Yj are the phenotype scores for individuals i and j.
Variance and covariance estimates can be estimated with a simple latent variable model
(see Figure 17.1). In this model, phenotype scores are the result of additive genetic
(A), zoo (Z), and nonshared environment effects plus error (E). Additive genetic effects
will be correlated between individuals to the extent that two individuals are genetically
related (Rij, Wrights coefficient of relatedness). One thing to note is that, while
computing Rij is a relatively simple task in samples containing mostly nuclear family
units, it is considerably more difficult to do with complex pedigrees, especially those
where inbreeding has occurred. Fortunately, free software packages, e.g., Animal
Breeders Toolkit,25 can compute the degree of relatedness between all pairs of indi-
viduals in a sample and provide inbreeding coefficients by which the degree of
relatedness must be adjusted (see Appendix A in this chapter for more details).
The correlation of zoo effects is denoted rzij and equal to 1 if the pair lives in the
same zoo enclosure and 0 if they do not. Finally, the nonshared environment effects
plus error are assumed to be uncorrelated.22 If we let a2, z2, and e2 be the variances
attributable to additive genetic, zoo, and error, respectively, it then follows that
var(Y) = a2 + z2 + e2
Rij
Ai Aj
a a
Ei e Yi Yj e Ej
z z
Zi Zj
rzij
FIGURE 17.1 The correlation in trait Y for any given pair of animals, i and j, is described
as a function of six latent variables with variances equal to 1. Latent variables include additive
genetic (A), shared zoo (Z ), and nonshared environment plus error (E) effects. Paths a, z and
e are the effects of each variable on the trait of interest. Rij is Wrights coefficient of relationship
and rzij is a dummy-coded variable equal to 1 if a pair lives in the same zoo and 0 if it does
not. (From Behavior Genetics, Weiss, A., King, J.E. and Figueredo, A.J. The heritability of
personality factors in chimpanzees (Pan troglodytes), 2000, vol. 30, 213221. With permission
from Kluwer Academic Publishers.)
These values can then be substituted in the first equation22 to yield a formula that
can be solved as the linear regression equation
The squared differences among all possible pairs are regressed onto (1 Rij)
and (1 rzij). The unstandardized beta weight for the variable reflecting (1 Rij) is
equal to the variance due to genetic effects (a2) and the unstandardized beta weight
for the variable reflecting (1 rzij) is equal to the variance due to shared zoo effects
(z2). Finally, the intercept term is equal to the variance due to the nonshared envi-
ronment plus error (e2).22
h2 = a2 / (a2 + z2 + e2)
The major advantage of SDS is that, if the genetic relatedness is known for all possible
pairs of individuals, there is no need for a specialized software package. We conducted
our analysis using the PROC REG command in SAS,29 but it could just as easily
have been computed using the regression modules of any other statistics package.
The use of all possible pairs of subjects in the SDS procedure means that each
subject will be present in n 1 pairs. This lack of independence means that standard
procedures for assessing the statistical significance of these estimates will be
incorrect. The statistical significance of these estimates can, however, be addressed
by Monte Carlo simulation, in which the variable representing degree of relatedness
is randomly assigned to subject pairs. Repeated iterations of the analysis can then
be used to estimate the sampling distribution of these estimates. We used 1,000
iterations to estimate these sampling distributions.30 The proportion of heritability
estimates from the simulation that were greater than or equal to the estimates
derived when the degrees of relatedness were properly assigned was used as an
estimate of p.
yi = i + ai + mi + ci + ei.
A more general form of this equation can be used to describe phenotypic deviations
of each individual from a sample or population mean. This general equation is based
** This approach can be used to estimate variances and covariance of and among multiple traits.29,30
1903_C017.fm Page 255 Wednesday, July 26, 2006 7:06 PM
on a series of vectors and matrices. The first of these vectors, i, is a vector of deviations
from a grand mean attributable to fixed effects (zoos in our case). Other vectors include
a, m, and c, which indicate the degree to which subjects deviations from the zoo mean
are the result of additive genetic, heritable maternal, and nonheritable maternal effects,
respectively. The matrices include X and Za, which indicate each individuals zoo
membership and identifying number, respectively. The final matrices are Zm, and Zc,
which indicate the identifying number of each chimpanzees motherused to assess
heritable maternal and nonheritable maternal effects, respectively.***
The products of these vectors and matrices and the residual term, e, which
represents deviations resulting from nonshared environmental effects plus error, can
be used to predict, y, a vector of phenotypes, in a sample or population.
*** If the source of these effects is the same, as when two types of maternal effects are investigated, the
matrices will be identical.
1903_C017.fm Page 256 Wednesday, July 26, 2006 7:06 PM
personality trait that was heritable,26 so we focused exclusively on the genetic and
environmental sources of variances and covariance for dominance and SWB.33
Along with the parameter estimates, MTDFREML provides a statistic (2log) indi-
cating how well the model replicated the actual data. The lower this statistic, the better
the model fit. We compared seven different models using two different approaches.
The first approach was based on Akaikes information criterion (AIC).36 AIC is based
on the notion that model fit must be balanced with model parsimony and, hence,
penalizes the degree of model fit by the number of parameters estimated (k).
AIC = 2log + 2k
The model with the lowest AIC, then, is the one that best balances model fit
and parsimony. In our analysis this model was one in which 66% of the dominance
and 40% of the SWB variance were heritable.33 The heritability estimate for domi-
nance was nearly identical to what we obtained in univariate analyses with SDS26
and the heritability estimate for SWB was nearly identical to what we obtained using
SDS and another estimation procedure, Method R.37,38 The covariance between
dominance and SWB was almost entirely due to shared genes. Furthermore, heritable
maternal effects accounted for approximately 22% of the SWB, but none of the
variance in dominance.33 Finally, largely independent nonshared environmental
effects plus error accounted for 27% of the remaining dominance and 35% of the
remaining SWB variance.33
The second approach, nested models comparisons, is based on the fact that the 2log
statistic is distributed as a chi-square. Hence, one can test whether parameters are
significant by comparing the 2log of models with and without those parameters.39
If the 2log difference between models is not significant, this indicates that the
additional parameters were not contributing to the model beyond what would be
expected by chance and, hence, the model with fewer parameters should be adopted.
In our analysis, comparing models using nested models comparison indicated
that there was no significant difference between the model that we chose based on
the comparison of the AIC statistics and a model that was identical except for not
including heritable maternal effects. We therefore accepted this more parsimonious
model that did not include heritable maternal effects as a plausible alternative.33
17.3.3 CONSIDERATIONS
As with SDS, using MTDFREML to solve for variance and covariance components
has advantages and disadvantages. These should be considered carefully when plan-
ning a behavior genetic study in samples with complex pedigrees.
1903_C017.fm Page 257 Wednesday, July 26, 2006 7:06 PM
Because breeding values reflect the genetic contribution to a given phenotype, they
could be used to improve the power of genetic association studies. Genetic association
1903_C017.fm Page 258 Wednesday, July 26, 2006 7:06 PM
studies typically compare individuals who are high in some phenotype with those
who are low or average in the phenotype. Because the phenotype is a blend of additive
and nonadditive genetic variance as well as environmental effects, the phenotype
score is only an approximation of an individuals genetic predisposition for that trait.
The use of breeding values in these studies would increase the likelihood of detecting
associations.
Similarly, pedigree analyses can also output the degree to which individuals differ
from their expected mean based on measured environmental effects. This potentially
useful measure is the obverse of breeding values and would show individual differ-
ences in susceptibility to environmental modification of the trait being studied. An
interesting question is whether this environmental susceptibility is itself heritable.
Those studying the environmental determinants of behavior could attempt to associate
different environmental events and conditions to these environmental susceptibility
values and reduce the likelihood that genetic effects confound their findings.
AUTHORS NOTE
We wish to thank Virginia Landau and the staff of the ChimpanZoo program and
all the raters at the different zoos for making our research possible. We also wish
to thank our past collaborators A. J. Figueredo and R. Mark Enns. Finally, we would
like to thank R. R. McCrae and Antonio Terracciano for their comments on an earlier
manuscript.
1903_C017.fm Page 259 Wednesday, July 26, 2006 7:06 PM
REFERENCES
1. Purvis, A. (1995). A composite estimate of primate phylogeny. Philosophical Trans-
action of the Royal Society of London, Series B: Biological Sciences 348, 405421.
2. Sarich, V.M. and Wilson, A.C. (1967). Immunological time scale for hominid evo-
lution. Science 158, 12001203.
3. Wilson, A.C. and Sarich, V.M. (1969). A molecular time scale for human evolution.
Proceedings of the National Academy of Sciences U.S.A. 63.
4. Rumbaugh, D.M. and Savage-Rumbaugh, E.S. (1994). Language in comparative
perspective. In Animal Learning and Cognition, N.J. Mackintosh, ed. (San Diego,
CA: Academic Press), pp. 307333.
5. Chomsky, A.N. (1972). Language and the Mind (New York: Harcourt Brace
Jovanovich).
6. Hare, B., Call, J., and Tomasello, M. (2001). Do chimpanzees know what conspecifics
know? Animal Behaviour 61, 139151.
7. Whiten, A. (1997). The Machiavellian mindreader. In Machiavellian Intelligence II:
Extensions and Evaluations, A. Whiten and R.W. Byrne, eds. (New York: Cambridge
University Press), pp. 144173.
8. de Waal, F.B.M. (1996). Good Natured: The Origins of Right and Wrong in Humans
and Other Animals. (Cambridge, MA: Harvard University Press).
9. de Waal, F.B.M. (2001). The Ape and the Sushi Master Cultural Reflections by a
Primatologist (New York: Basic Books).
10. Whiten, A., Goodall, J., McGrew, W.C., Nishida, T., Reynolds, V., Sugiyama, Y.,
Tutin, C.E.G., Wrangham, R.W., and Boesch, C. (1999). Cultures in chimpanzees.
Nature 399, 682685.
11. McCrae, R.R. and Costa, P.T., Jr. (2003). Personality in Adulthood: A Five-Factor
Theory Perspective (New York: Guilford).
12. Bouchard, T.J., Jr. and Loehlin, J.C. (2001). Genes, evolution, and personality. Behav-
ior Genetics 31, 243271.
13. McCrae, R.R. (2001). Trait psychology and culture: Exploring intercultural compar-
isons. Journal of Personality 69, 819846.
14. Lykken, D.T. and Tellegen, A. (1996). Happiness is a stochastic phenomenon. Psy-
chological Science 7, 186189.
15. Eid, M. and Diener, E. (2004). Global judgements of subjective well-being: Situa-
tional variability and long-term stability. Social Indicators Research 65, 245277.
16. Diener, E. (1996). Traits can be powerful, but are not enough: Lessons from subjective
well-being. Journal of Research in Personality 30, 389399.
17. King, J.E. and Figueredo, A.J. (1997). The Five-Factor Model plus Dominance in
chimpanzee personality. Journal of Research in Personality 31, 257271.
18. King, J.E. and Landau, V.I. (2003). Can chimpanzee (Pan troglodytes) happiness be
estimated by human raters? Journal of Research in Personality 37, 115.
19. King, J.E., Weiss, A., and Farmer, K.H. (2005). A chimpanzee (Pan troglodytes)
analogue of cross-national generalization of personality structure: zoological Parks
and an African Sanctuary. Journal of Personality 73, 389410.
20. Fulk, R. (1999). Chimpanzee Species Survival Plan: Master Plan 19981999, North
Carolina Zoological Park: Asheboro, NC.
21. Ross, S. (2003). North American Regional Studbook for the Chimpanzee (Pan trogl-
odytes), Lincoln Park Zoo: Chicago.
1903_C017.fm Page 260 Wednesday, July 26, 2006 7:06 PM
22. Grimes, L.W. and Harvey, W.R. (1980). Estimation of genetic variances and covari-
ances using Symmetric Differences Squared. Journal of Animal Science 50, 632644.
23. Bruckner, C.M. and Slanger, W.D. (1986). Symmetrical differences squared and
analysis of variance procedures for estimating genetic and environmental variances
and covariances for beef-cattle weaning weight: I. Comparison via simulation. Journal
of Animal Science 63, 17791793.
24. Bruckner, C.M. and Slanger, W.D. (1986). Symmetrical differences squared and
analysis of variance procedures for estimating genetic and environmental variances
and covariances for beef-cattle weaning weight: II. Estimates from a data set. Journal
of Animal Science 63, 17941803.
25. Golden, B.L., Snelling, W.M., and Mallinckrodt, C.H. (1992). Animal Breeders
Toolkit: Users Guide and Reference Manual, Colorado State University Agricultural
Experiment Station: Fort Collins.
26. Weiss, A., King, J.E., and Figueredo, A.J. (2000). The heritability of personality
factors in chimpanzees (Pan troglodytes). Behavior Genetics 30, 213221.
27. Bouchard, T.J., Jr. (2004). Genetic influence on human personality traits: A survey.
Current Directions in Psychological Science 13, 148151.
28. Rowe, D.C. (1994). The Limits of Family Influence (New York: Guilford).
29. SAS Institute, I. (1999). SAS/STAT Users Guide, Version 8 (Cary, NC: SAS Institute).
30. Cheverud, J.M., Falk, D., Vannier, M., Konigsberg, L., Helmkamp, R.C., and Hilde-
bolt, C. (1990). Heritability of brain size and surface features in rhesus monkeys
(Macaca mulatta). Journal of Heredity 81, 5157.
31. Boldman, K.G., Kriese, L.A., Van Vleck, L.D., Van Tassel, C.P., and Kachman, S.D.
(1995). A Manual for Use of MTDFREML [draft], Animal Research Station, United
States Department of Agriculture: Clay Center, NE.
32. Henderson, C.R. (1975). Best linear unbiased estimation and prediction under a
selection model. Biometrics 32, 423447.
33. Weiss, A., King, J.E., and Enns, R.M. (2002). Subjective well-being is heritable and
genetically correlated with Dominance in chimpanzees (Pan troglodytes). Journal of
Personality and Social Psychology 83, 11411149.
34. Meyer, K. (1989). Restricted maximum likelihood to estimate variance components
for animal models with several random effects using a derivativefree algorithm.
Genetics, Selection, Evolution 21, 317340.
35. Meyer, K. (1991). Estimating variances and covariances for multivariate animal
models by restricted maximum likelihood. Genetics, Selection, Evolution 23.
36. Akaike, H. (1987). Factor analysis and AIC. Psychometrika 52, 317332.
37. Weiss, A., King, J.E., and Enns, R.M. (1999). Can we breed a happier ape? The
heritability of subjective well-being in zoo chimpanzees (Pan troglodytes). Behavior
Genetics 29, 374.
38. Reverter, A., Golden, B.L., Bourdon, R.M., and Brinks, J.S. (1994). Method R
variance-components procedure: Application on the simple breeding value model.
Journal of Animal Science 72, 22472253.
39. Loehlin, J.C. (1998). Latent Variable Models: An Introduction to Factor, Path, and
Structural Analysis (Mahwah, NJ: Lawrence Erlbaum Associates).
40. Meyer, K. (1998). DFREML, 3.1 ed., Self Published: Armidale, NSW, Austra-
lia.Meyer, K. (1998). DFREML, 3.1 ed., Self Published: Armidale, NSW, Australia.
41. King, J.E., Weiss, A., and Sisco, M.M. Age differences in chimpanzee personality,
under review.
1903_C017.fm Page 261 Wednesday, July 26, 2006 7:06 PM
86 16 21
87 37 24
88 8 40
89 8 28
90 37 44
91 . .
92 53 36
93 . 41
94 16 23
95 37 45
96 8 11
.
.
.
The first step is to reorder and check the pedigree file:
chkord ped.old > ped.chk 2 > ped.bad
This command reorders the pedigree so that parents precede offspring and checks
whether there are any problems, e.g., individuals that appear as both fathers and
mothers. This procedure is required for computing the degrees of relatedness and it
is generally a good idea to apply this to any pedigree file including those that will
be used in other pedigree analyses such as MTDFREML. The ped.old file is the
1903_C017.fm Page 262 Wednesday, July 26, 2006 7:06 PM
original space-delimited pedigree file. The ped.chk file is the properly ordered
version of the pedigree file, i.e., where parents precede their offspring, which is
required for the next steps. The ped.bad file will contain a listing of any errors,
e.g., individuals that appear as mothers and fathers in the file.
The second step is to create a file containing the inverse numerator relationship
matrix (A1) and a file containing inbreeding coefficients (f) for the subjects in the
analysis from the reordered pedigree file that we created in the previous step.
ainv -i ped.chk -o ped.inbreed -v ped.nrm
where ped.inbreed and ped.nrm are the files containing the inbreeding coef-
ficients for all subjects and the inverse numerator relationship matrix for all possible
pairs of subjects, respectively.
The third step is to invert the inverse numerator relationship matrix:
invert -i ped.nrm -o ped.output
Here, ped.nrm is the file derived from the previous command and ped.out-
put contains a matrix, A, indicating the degree of relatedness between all pairs of
individuals.
Unfortunately, the format of the matrix might be difficult to use in a conventional
statistics packages. Hence, our next two steps use ABTK to change the format of
the ped.output file so that each row includes the stud numbers of the individuals
within a pair and their degree of relatedness. The commands to do this are as follows:
d2s ped.output ped.newoutput
s2t ped.newoutput
Here, ped.output is the file we created in the prior step and ped.newout-
put is the tree version of this file where each row represents a single pair of
individuals.
If there is no inbreeding, i.e., the ped.inbreed file is empty; the previous
step was the last. However, if there was some inbreeding, then the degree of relat-
edness for every pair of individuals in which an inbred animal was present needs to
be adjusted by dividing the degree of relatedness by the following:
( fi + 1)( f j + 1)
where fi is equal to the inbreeding coefficient of one individual in the pair and fj is
equal to the inbreeding coefficient of the other individual of the pair. If an individual
is not inbred, the inbreeding coefficient is equal to 0.
These file names are only examples; any file name can be used.
1903_C018.fm Page 263 Wednesday, July 26, 2006 7:06 PM
CONTENTS
18.1 Personality.................................................................................................... 263
18.1.1 What Is Personality and How Is It Measured? ............................... 263
18.2 Heritability ................................................................................................... 266
18.3 Molecular Genetics ...................................................................................... 266
18.4 Related Phenotypes ...................................................................................... 267
18.4.1 Fibromyalgia .................................................................................... 267
18.4.2 Altruism............................................................................................ 268
18.5 The Social Personality ................................................................................. 270
18.6 Linkage Studies............................................................................................ 271
18.7 Importance.................................................................................................... 273
Acknowledgments.................................................................................................. 273
References.............................................................................................................. 273
18.1 PERSONALITY
263
1903_C018.fm Page 264 Wednesday, July 26, 2006 7:06 PM
universal aspects of people, not how they differ from each other. Traditional theories
of personality, of which Freuds remains the outstanding example, purport to explain
what is man, what makes him tick, to suggest a physiology of personality. If
genomics aims to elucidate the entire genome of a species, and biochemistry and
physiology aim to account for the functions of that genome and its protein products,
and if genetics is limited to the study of the consequences of genetic differences
between members of the species, then by analogy we can say that this first sense of
personality is a genomic approach. Its subjects include the unconscious, moti-
vation, instincts, defense mechanisms, the life cycle, interpersonal relations, condi-
tioning and self-actualization. While this branch of psychology has had a major
influence on twentieth-century culture, much of it is psychoanalytic, metaphoric,
and untestable, it has hardly been studied by geneticists.
A second sense of the term personality, and the one we will use throughout
the remainder of this chapter, is the individual psychological aspects of people that
make them recognizable, which is to say different from one another. In this sense
we say that Jones is miserly, Smith is gregarious, and, even, that Brown is a
character. The personalities of fictitious characters like Falstaff, Othello, and Scarlet
OHara seem familiar to millions. However, this intuitive grasping of personality is
based on one or two outstanding traits of each individual and there are no obvious
or systematic relationships among the various examples of traits mentioned above.
A general, empirical theory of personality differences must include a battery of
dimensional variables, and the number of variables must be manageable. The
attributes chosen should account for most of the variance in personality, and hope-
fully also make clinical sense to the psychologist, that is, somehow relate to what
seem to be fundamental and universal aspects of personality in the first sense of the
term as described above. It is moreover a part of the definition of a personality trait
or a personality type that the description is relatively enduring and stable for the
individual concerned, not a temporary condition.
Virtually all approaches to personality assessment used in genetic research have
used either direct laboratory evaluation of dimensional behaviors or pencil-and-paper
questionnaires that inventory dimensional personality traits. Laboratory assessments
are objective but labor-intensive both for collecting data and for analyzing them.
Furthermore, they sample a single moment in time. Self- or other report question-
naires may be more subjective, but have the advantage that they typically sample
subjects characteristics over extended periods.
The most widely used general pencil-and-paper personality inventories assume
that rating an individual accurately on the important dimensions of personality gives
a reasonable and economical description of that individuals personality. The number
and makeup of dimensions chosen are typically derived by one of the following
three approaches.
Approaches based on psychopathology derive the personality traits from psychi-
atric illnesses. Psychiatric illnesses are assumed to be extreme variants, and therefore
the clearest expressions, of normal personality. Thus the Minnesota Multiphasic
Personality Inventory2 grades people on the degree to which they resemble patients
diagnosed with hysteria, obsessivecompulsive disorder, schizophrenia, and so on.
This approach is often reflected in popular discourse; we say that someone is
1903_C018.fm Page 265 Wednesday, July 26, 2006 7:06 PM
compulsive, hysterical, and so on, but the approach has not been widely used in
genetic research on human personality.
Theoretical approaches derive from a theory of personality in the first sense
described above. Cloninger developed the Tridimensional Personality Questionnaire
(TPQ)3 based on a biological model employing the monoamines dopamine, norad-
renaline, and serotonin, which function as neurotransmitters in subcortical brain
systems. Animal and human findings from lesioning, imaging and pharmacological
studies involving these pathways implicate serotonin (5-HT) pathways in behaviors
devoted to avoiding harm or escaping punishment, and to the associated emotion of
anxiety (called harm avoidance (HA) in the TPQ). Human subjects with high HA
would be called neurotic on many other instruments; those with low HA are
stable or healthy or well adjusted. Similar studies implicate noradrenergic
(NE) pathways in approach behaviors, which the TPQ calls reward dependence
(RD). Human beings high on RD are sentimental and affectionate; those with low
RD are tough and pragmatic. Finally dopaminergic (DA) pathways are implicated
in drug use, sensation seeking and explorative behaviors, and emotions like curiosity
and recklessness, which the TPQ calls novelty seeking (NS). Individuals with low
NS are deliberate and frugal. The eight possible combinations of the two extremes
of the distributions of these three dimensions yield personality constellations held
to reflect clinically recognized personality disorders such as antisocial psychopathy
and obsessivecompulsive personality disorder. A later version of the TPQ is the
Temperament and Character Inventory (TCI).
The third approach derives from everyday speech. Natural language is considered
to have proved itself during evolution to be a reasonably accurate and adaptive way of
describing people. Personality inventories have been constructed by listing thousands of
personality descriptors culled from dictionaries of everyday language and reducing them
to sizable numbers by factor analysis. The most popular such inventory in clinical and
academic use is the neopersonality inventory (NEO-PI)4 which assesses five main per-
sonality factors. Neuroticism resembles TPQ HA. Extraversion is the degree of interest
in other people and in social dominance. Openness to new ideas and experiences is the
opposite of traditionalism. Agreeableness is a measure of how likeable an individual is.
Conscientiousness assesses conscience, thoughtfulness, planning and order, and their
opposites: spontaneity and lack of prudence. Extraversion and conscientiousness (with
opposite signs) correlate with TPQ NS. While details of dimensions differ between
instruments, a number of recent factor analyses concur that five main factors adequately
describe human personality differences, and these have been termed the big five.
Given the uncertainty inherent in measuring personality traits, it seems advisable,
where possible, to study putative personality traits simultaneously with more than
one instrument and/or laboratory assessment. Good correlations between apparently
similar traits in two or more assessments, and converging findings of associations
between a given polymorphism and those traits in two or more assessments, enhance
our confidence in those findings. This is all the more necessary, given the difficulty
in replicating molecular genetic findings in personality genetics, similar to the
problems encountered in other complex phenotypes.
Additionally, personality geneticists study traits of particular interest, which may
not be included in general personality inventories (see below).
1903_C018.fm Page 266 Wednesday, July 26, 2006 7:06 PM
18.2 HERITABILITY
Personality dimensions measured by self-report questionnaires such as the NEO and
TPQ are moderately heritable (4060%).58
18.4.2 ALTRUISM
We have also examined an intriguing dimension of human personality, altruism, in
a large nonclinical population for association with several common polymorphisms
germane to a broader view of human personality than that examined with the usual
inventories (TPQ or NEO).40
The paradox of human altruism, helping others and thereby reducing ones own
fitness, has confounded evolutionary biologists since the days of Darwin. Not only
is altruism a puzzle for evolutionary biologists but this trait has also perplexed
psychologists who have questioned whether there is such a thing as an altruistic
personality.41,42 However, altruistic behavior is commonplace, and a unique feature
of human altruism is that it extends beyond Hamiltons concept of inclusive fit-
ness,43 which explains altruistic acts by including helping genetically related indi-
viduals, and even beyond reciprocal altruism and reputation-based altruism.44 Out
of all the animals, only humans practice wholesale mutual aid among genetically
unrelated individuals.
Although theoretical understanding of the evolutionary45 and psychological
mechanisms46 that underlie altruism has greatly increased in the past several decades,
almost nothing is known regarding specific genes contributing to this behavior.
However, several twin studies4752 reported significant heritability of prosocial atti-
tudes. As expected of a trait that is partially influenced by genes, prosocial or
altruistic dispositions show individual differences with origins in early childhood
and stability over developmental time.53,54
The mechanisms facilitating prosocial behavior and human altruism are of inter-
est not only to evolutionary biology, but also to students of psychopathology. Psy-
chiatric research has for obvious reasons focused on negative factors that adversely
affect mental well-being. However, there is growing interest in positive or protective
factors that play a role in promoting normal development. Prosocial or altruistic
attitudes, such as cooperativeness, helpfulness, sharing, and being empathetic, are
important elements that facilitate social networks thought to promote mental well-
being. For example, prosocial behavior correlates with childrens academic excel-
lence, allows them to resist social pressures for antisocial activities, and to engage
themselves with empathy in others emotional experiences.55 Additionally, positive
social interactions undoubtedly eased by altruistic behaviors have been shown to
exert powerful beneficial effects on health outcomes and longevity.56 Importantly,
self-report measures of altruism correlate with peers and teachers evaluation.53,57
Any account of prosocial behavior and altruism in humans should include the
identification of particular genetic polymorphisms partially contributing to this trait.
Toward the goal of identifying specific genes associated with altruism and prosocial
attitudes, 354 nonclinical families with multiple siblings were inventoried for scores
on the selflessness scale.58 This questionnaire measures the propensity to ignore
ones own needs and serve the needs of others, or in other words, altruism. Subjects
were also inventoried on Cloningers TPQ59 since the reward subscale of this ques-
tionnaire taps into elements of human altruism, such as empathy. We examined two
dopaminergic genes in these subjects that we hypothesized might contribute to
prosocial or altruistic traits, based on the role a single variant of these genes plays
1903_C018.fm Page 269 Wednesday, July 26, 2006 7:06 PM
(RSMS) and the concern for appropriateness scale (CFA), which together comprise
the social psychological construct of self-presentation. The two have generally been
found to be orthogonal96,97 and more psychometrically sound than Snyders scale.98
The RSMS measures acquisitive self-presentation, or getting ahead, an active and
flexible social approach aimed at gaining power and status. The CFA measures a
defensive and fearful social approach associated with conformity, aimed at gaining
acceptance and approval and avoiding social threats. Whereas both self-presenta-
tional styles reflect social orientations with a high degree of concern for social cues
and social approval, the RSMS correlates positively with measures of social adap-
tation such as self-esteem and extraversion,97 whereas the CFA correlates negatively
with such measures.
As discussed above, AVPR1A molds social behavior across the vertebrates, and
it makes sense that self-presentation, a construct that characterizes the style (dom-
inating, fitting in) an individual adopts in participating in group activities should
also be linked to this receptor. We further hypothesized that the SRQ construct of
conflict, and possibly that of power, would tap into the aggressive style of behav-
iors that are also molded by vasopressin.
We also thought it of some interest to examine sibling relationships because
such relationships are often the first dyadic social relationship that individuals
experience, and the style people adopt with their siblings may have ramifications
for their subsequent relationships with nonrelated persons. Indeed, we observe in
this cohort a correlation between the CFA and two of the SRQ scales (conflict and
warmth).
Suggestive linkage was observed between both microsatellites (RS1 and RS3)
and the SRQ Conflict Scale (RS1: chi-square = 13.65, LOD = 2.96, p = 0.0001;
RS3: chi-square = 14.54, LOD = 3.16, p = 0.00007) and the CFA self-presentational
style (RS1: chi-square = 8.25, LOD = 1.79 p = 0.002; RS3: chi-square = 8.81,
LOD = 1.91, p = 0.002).
The provisional linkage observed between microsatellites in the AVPR1A pro-
moter region and scores on the CFA and the SRQ conflict scale constitutes the first
substantial evidence that vasopressin, which mediates species-specific social behav-
ior across the vertebrates, may play a similar role in humans. Although social
behavior and the acquisition of social skills in humans constitute the cornerstone of
society and culture, few if any genetic studies have attempted to relate individual
differences in social skills to specific genes. Moreover, social skills are relevant to
a variety of psychiatric disorders including autism, schizophrenia, and externalizing
behavior problems in children. From an evolutionary perspective, it is not surprising
that a hormone that almost universally affects a spectrum of social and affiliative
behaviors in lower animals85 has a parallel role in the human, a social animal highly
dependent on group interactions for individual and species survival.
identical self-report questionnaires, the TPQ, whereas the Fullerton et al.100 report
used Eysencks psychoticism scale and found linkage to neuroticism.103 Subjects who
score high on harm avoidance can be described as worrying and pessimistic, fearful
and doubtful, and shy and fatigable. High scorers on Eysencks neuroticism dimension
can be similarly described. The Cloninger et al.99 genome scan used subjects recruited
from families with an alcoholic proband as part of the COGA study.104 The highest
LOD score (3.2) using 291 microsatellite markers for a site to harm avoidance was
near marker D8S1106 (8p 21-23) at 27 cM. The Fullerton British study genotyped
561 extremely discordant and concordant sibling pairs selected from a population of
34,580 families. The observed linkage to neuroticism near marker D8S277 (8 cM)
gave a LOD score of 2.9.
In our own study102 we have now genotyped altogether 24 microsatellite markers
in 377 families. Using three methods (maximum likelihood binomial or MLB, MER-
LIN and an associated one parameter model) we observed significant results (p-values
from 0.002 to 0.0004) for linkage to harm avoidance in this region. A peak multipoint
LOD score of 2.76 (p-value 0.0002) was obtained with the MLB method. The region-
wide empirical p-value was 0.002 [0.001 0.0046]. Although the peak position varied
somewhat according to the method (D8S1048 for MLB, D8S1463 for the two other
methods), for three methods D8S1810 (~60 cM) is within 1 to 2 cM of the peak for
harm avoidance. This marker is of particular interest since it is proximate (<0.5 cM)
to the core haplotype that in several recent studies shows significant association with
schizophrenia near neuroregulin 1.105,106 Although association studies with microsat-
ellite markers need to be interpreted cautiously, using the Haplotype Trend Regression
test one marker, D8S499 (~60 cM), showed an empirical p-value of 2 105 for allele
3, which confers a decreased harm avoidance score.
Altogether, our current linkage and association results suggest the possibility
that the same locus near the neuroregulin 1 gene on chromosome 8p confers risk
for both an anxiety-related personality trait as well as schizophrenia. We hypothesize
that this common genetic factor may contribute to emotional lability during early
development which constitutes a predisposing factor for major psychosis.
Curtis et al.108 recently reanalyzed the original dataset first reported by Cloninger et
99
al. using a new method of linkage analysis for quantitative traits that deals with extended
pedigrees. As well as supporting linkage of HA to D8S549 as originally reported,99 this
method also produces an MALOD of 2.4 (p = 0.002) near DBH and several positive
LOD scores for novelty seeking, the largest being MALODs of 3.1 (p = 0.0004) near
D12S391 and 3.4 (p = 0.0003) near D17S1299. There is no support for linkage of novelty
seeking or HA to the regions around DRD4 and 5-HTTLPR, respectively.
Neale et al.107 analyzed a genome scan for neuroticism on a sample of 129 sib-
pair families (113 with a single sibling pair, 18 with multiple sibling pairs) containing
a total of 201 possible sibling pairs, ascertained for concordance on nicotine depen-
dence. Using Merlin-Regress, and replicated peaks for neuroticism described by
prior studies, on chromosomes 1 and 11 with LOD scores of 2.52 and 1.97 (p-value
of 0.003 and 0.0108), respectively, and have some evidence for a novel finding on
chromosome 12 with a LOD of 2.85 (p-value of 0.0014).
1903_C018.fm Page 273 Wednesday, July 26, 2006 7:06 PM
18.7 IMPORTANCE
Advances in the molecular genetics of human personality could have major
implications for both psychology and psychiatry. In the field of psychology,
alleles or alleleallele interactions affecting personality traits with large effect
sizes would invite renewed explorations of the neurochemical or neuroanatomical
effects involved. Discoveries of such effects could potentially revolutionize under-
standing of the physical bases of the complex and hitherto intractable phenotypes
of personality traits. Such discoveries might also raise the specter of a pill for,
or against, every emotion or trait, a sort of plastic surgery of the psyche. History
provides ample evidence of our readiness to use and abuse psychoactive sub-
stances. Thus far, however, findings have been limited to modest effect sizes on
monoamine systems that were already predicted by knowledge outside the field
of genetics.
In psychiatry, chromosomal regions or interactions between regions, or alleles
or alleleallele interactions, affecting personality traits with large effect sizes would
invite new approaches to genetics for mental illnesses. We have already alluded to
the idea that mental illnesses may be extreme manifestations of normal personality
traits, or of maladaptive combinations of these traits. In analogous fashion, idio-
pathic hypertension and adult-onset diabetes may represent harmful expressions of
genes, or harmful interactions between genes, that originally evolved to code for
proteins that enable normal blood pressure and carbohydrate metabolism. Modern
environments may differ from those in which those genes were adaptive. Genes
affecting personality traits may also have been selected when both the external
environment and intrapersonal and interpersonal existence differed from modern
experience. Discoveries of chromosomal regions or alleles affecting, for example,
harm avoidance, invite investigations of those same regions or alleles in clinical
anxiety disorders. In practice, findings so far have been so few that any promising
leads probably will be followed up with initial investigations in all or almost all
psychiatric disorders.
The region on chromosome 8 implicated in harm avoidance will undoubtedly
be further scrutinized for linkage to psychiatric disorders. To the extent that future
genetic findings in personality are unexpected, they may suggest novel approaches
to psychiatric illnesses.
ACKNOWLEDGMENTS
This research was partially supported by the Israel Science Foundation founded by
the Israel Academy of Sciences and Humanities (RPE).
1903_C018.fm Page 274 Wednesday, July 26, 2006 7:06 PM
REFERENCES
1. Benjamin, J., Ebstein, R.P. & Belmaker, R., H. Molecular Genetics and the Human
Personality, 355 (American Psychiatric Publishing, Inc., Washington, D.C., 2002).
2. Butcher, J.N. MMPI-2: Minnesota Multiphasic Personality Inventory-2: manual for
administration, scoring, and interpretation, ix, 212 p. (University of Minnesota Press,
Minneapolis, 2001).
3. Cloninger, C.R. A unified biosocial theory of personality and its role in the develop-
ment of anxiety states. Psychiatr Dev 4, 167226 (1986).
4. Costa, P.T., Jr. & McCrae, R.R. Stability and change in personality assessment: the
revised NEO Personality Inventory in the year 2000. J Pers Assess 68, 8694 (1997).
5. Jang, K.L., Livesley, W.J. & Vernon, P.A. Heritability of the big five personality
dimensions and their facets: a twin study. J Pers 64, 57791 (1996).
6. Heiman, N., Stallings, M.C., Hofer, S.M. & Hewitt, J.K. Investigating age differences
in the genetic and environmental structure of the tridimensional personality question-
naire in later adulthood. Behav Genet 33, 17180 (2003).
7. Heath, A.C., Cloninger, C.R. & Martin, N.G. Testing a model for the genetic structure
of personality: a comparison of the personality systems of Cloninger and Eysenck.
J Pers Soc Psychol 66, 76275 (1994).
8. Bergeman, C.S. et al. Genetic and environmental effects on openness to experience,
agreeableness, and conscientiousness: an adoption/twin study. J Pers 61, 15979
(1993).
9. Ebstein, R.P. et al. Dopamine D4 receptor (D4DR) exon III polymorphism associated
with the human personality trait of Novelty Seeking. Nat Genet 12, 7880 (1996).
10. Benjamin, J. et al. Population and familial association between the D4 dopamine
receptor gene and measures of Novelty Seeking. Nat Genet 12, 814 (1996).
11. Lesch, K.P. et al. Association of anxiety-related traits with a polymorphism in the
serotonin transporter gene regulatory region [see comments]. Science 274, 152731
(1996).
12. Lusher, J.M., Chandler, C. & Ball, D. Dopamine D4 receptor gene (DRD4) is asso-
ciated with Novelty Seeking (NS) and substance abuse: the saga continues. Mol
Psychiatry 6, 4979 (2001).
13. Comings, D.E. et al. Studies of the 48 bp repeat polymorphism of the DRD4 gene
in impulsive, compulsive, addictive behaviors: Tourette syndrome, ADHD, patholog-
ical gambling, and substance abuse. Am J Med Genet 88, 35868 (1999).
14. Mel, H. et al. Additional evidence for an association between the dopamine D4
receptor (D4DR) exon III seven-repeat allele and substance abuse in opioid dependent
subjects: Relationship of treatment retention to genotype and personality. Addiction
Biology 3, 473481 (1998).
15. Kotler, M. et al. Excess dopamine D4 receptor (D4DR) exon III seven repeat allele
in opioid-dependent subjects. Mol Psychiatry 2, 2514 (1997).
16. Li, T. et al. Association analysis of the dopamine D4 gene exon III VNTR and heroin
abuse in Chinese subjects. Mol Psychiatry 2, 4136 (1997).
17. Muramatsu, T., Higuchi, S., Murayama, M., Matsushita, S. & Hayashida, M. Asso-
ciation between alcoholism and the dopamine D4 receptor gene. J Med Genet 33,
1135 (1996).
18. Lerman, C. et al. Depression and self-medication with nicotine: the modifying influ-
ence of the dopamine D4 receptor gene [see comments]. Health Psychol 17, 5662
(1998).
1903_C018.fm Page 275 Wednesday, July 26, 2006 7:06 PM
19. Shields, P.G. et al. Dopamine D4 receptors and the risk of cigarette smoking in
African-Americans and Caucasians. Cancer Epidemiol Biomarkers Prev 7, 4538
(1998).
20. Lerman, C. et al. Interacting effects of the serotonin transporter gene and neuroticism
in smoking practices and nicotine dependence. Mol Psychiatry 5, 18992 (2000).
21. Hu, S., Brody, C.L., Fisher, C. et al. Interaction between the serotonin transporter
gene and neuroticism in cigarette smoking behavior. Mol Psychiatry 5, 1818 (2000).
22. Faraone, S.V., Doyle, A.E., Mick, E. & Biederman, J. Meta-analysis of the association
between the 7-repeat allele of the dopamine D(4) receptor gene and attention deficit
hyperactivity disorder. Am J Psychiatry 158, 10527 (2001).
23. Cadoret, R.J. et al. Associations of the serotonin transporter promoter polymorphism
with aggressivity, attention deficit, and conduct disorder in an adoptee population.
Compr Psychiatry 44, 88101 (2003).
24. Kent, L. et al. Evidence that variation at the serotonin transporter gene influences
susceptibility to attention deficit hyperactivity disorder (ADHD): analysis and pooled
analysis. Mol Psychiatry 7, 90812 (2002).
25. Retz, W., Thome, J., Blocher, D., Baader, M. and Rosler, M. Association of attention
deficit hyperactivity disorder-related psychopathology and personality traits with the
serotonin transporter promoter region polymorphism. Neurosci Lett 319, 1336 (2002).
26. Manor, I. et al. Family-based association study of the serotonin transporter promoter
region polymorphism (5-HTTLPR) in attention deficit hyperactivity disorder. Am J
Med Genet 105, 915. (2001).
27. Arbelle, S. et al. Relation of shyness in grade school children to the genotype for the
long form of the serotonin transporter promoter region polymorphism. Am J Psychi-
atry 160, 6716 (2003).
28. Hariri, A.R. et al. Serotonin transporter genetic variation and the response of the
human amygdala. Science 297, 4003 (2002).
29. Caspi, A. et al. Influence of life stress on depression: moderation by a polymorphism
in the 5-HTT gene. Science 301, 3869 (2003).
30. Kluger, A.N., Siegfried, Z. & Ebstein, R.P. A meta-analysis of the association between
DRD4 polymorphism and novelty seeking. Mol Psychiatry 7, 7127 (2002).
31. Schinka, J.A., Letsch, E.A. & Crawford, F.C. DRD4 and novelty seeking: results of
meta-analyses. Am J Med Genet 114, 6438 (2002).
32. Sen, S., Burmeister, M. & Ghosh, D. Meta-analysis of the association between a
serotonin transporter promoter polymorphism (5-HTTLPR) and anxiety-related per-
sonality traits. Am J Med Genet 127B, 859 (2004).
33. Schinka, J.A., Busch, R.M. & Robichaux-Keene, N. A meta-analysis of the associa-
tion between the serotonin transporter gene polymorphism (5-HTTLPR) and trait
anxiety. Mol Psychiatry 9, 197202 (2004).
34. Okuyama, Y. et al. Identification of a polymorphism in the promoter region of DRD4
associated with the human novelty seeking personality trait. Mol Psychiatry 5, 649
(2000).
35. Buskila, D. Fibromyalgia, chronic fatigue syndrome, and myofascial pain syndrome.
Curr Opin Rheumatol 13, 11727 (2001).
36. Offenbaecher, M. et al. Possible association of fibromyalgia with a polymorphism in
the serotonin transporter gene regulatory region. Arthritis Rheum 42, 24828 (1999).
37. Cohen, H., Buskila, D., Neumann, L. & Ebstein, R.P. Confirmation of an association
between fibromyalgia and serotonin transporter promoter region (5- HTTLPR) poly-
morphism, and relationship to anxiety-related personality traits. Arthritis Rheum 46,
8457 (2002).
1903_C018.fm Page 276 Wednesday, July 26, 2006 7:06 PM
38. Ebstein, R.P. et al. No association between the serotonin transporter gene regulatory
region polymorphism and the Tridimensional Personality Questionnaire (TPQ) tem-
perament of harm avoidance. Mol Psychiatry 2, 2246 (1997).
39. Buskila, D., Cohen, h., Neuman, L. & Ebstein, R.P. An association between fibro-
myalgia and the dopamine D4 receptor exon III repeat polymorphism and relationship
to novelty seeking personality traits. Mol Psychiatry 9, 7301 (2004).
40. Bachner-Melman, R. et al. Dopaminergic polymorphisms associated with self-report
measures of human altruism: A fresh phenotype for the dopamine D4 receptor. Mol
Psychiatry 10, 3335 (2005).
41. Gergen, K.-J., Gergen, M.-M. & Meter, K. Individual orientations to prosocial behav-
ior. Journal of Social Issues 28, 105130 (1972).
42. Batson, C.D. The Altruism Question: Toward a Social Psychological Answer, ix, 257
p. (L. Erlbaum, Associates, Hillsdale, N.J., 1991).
43. Hamilton, W.D. The genetical evolution of social behaviour. I. J Theor Biol 7, 116
(1964).
44. Sigmund, K. & Hauert, C. Primer: altruism. Curr Biol 12, R2702 (2002).
45. Fehr, E. & Fischbacher, U. The nature of human altruism. Nature 425, 78591 (2003).
46. Eisenberg, N. Prosocial behavior, empathy, and sympathy. In Well Being: Positive
Development across the Life Course. Crosscurrents in Contemporary Psychology
(eds. Bornstein, M.H. & Davidson, L.) 253265 (Lawrence Erlbaum Associates,
Publishers, Mahwah, NJ, 2003).
47. Emde, R.N. et al. Temperament, emotion, and cognition at fourteen months: the
MacArthur Longitudinal Twin Study. Child Dev 63, 143755 (1992).
48. Loat, C.S., Asbury, K., Galsworthy, M.J., Plomin, R. & Craig, I.W. X inactivation as
a source of behavioural differences in monozygotic female twins. Twin Res 7, 5461
(2004).
49. Rushton, J.P., Fulker, D.W., Neale, M.C., Nias, D.K. & Eysenck, H.J. Altruism and
aggression: the heritability of individual differences. J Pers Soc Psychol 50, 11928
(1986).
50. Loehlin, J.C. & Nichols, R.C. Heredity, Environment, & Personality: A Study of 850
Sets of Twins, xii, 202 p. (University of Texas Press, Austin, 1976).
51. Matthews, K.-A., Batson, C.D., Horn, J. & Rosenman, R.-H. Principles in his nature
which interest him in the fortune of others: The heritability of empathic concern for
others. Journal of Personality 49, 237247 (1981).
52. Zahn-Waxler, C., Robinson, J.-L. & Emde, R.-N. The development of empathy in
twins. Developmental Psychology 28, 10381047 (1992).
53. Eisenberg, N. et al. Prosocial development in early adulthood: a longitudinal study.
J Pers Soc Psychol 82, 9931006 (2002).
54. Eisenberg, N. et al. Consistency and development of prosocial dispositions: a longi-
tudinal study. Child Dev 70, 136072 (1999).
55. Bandura, A., Caprara, G.V., Barbaranelli, C., Gerbino, M. & Pastorelli, C. Role of
affective self-regulatory efficacy in diverse spheres of psychosocial functioning. Child
Dev 74, 76982 (2003).
56. Seeman, T.E. Health promoting effects of friends and family on health outcomes in
older adults. Am J Health Promot 14, 36270 (2000).
57. Caprara, G.V., Barbaranelli, C., Pastorelli, C., Bandura, A. & Zimbardo, P.G. Proso-
cial foundations of children's academic achievement. Psychol Sci 11, 3026 (2000).
58. Bachar, E. et al. Rejection of life in anorexic and bulimic patients. International
Journal of Eating Disorders 31, 4348 (2001).
1903_C018.fm Page 277 Wednesday, July 26, 2006 7:06 PM
59. Cloninger, C.R. A systematic method for clinical description and classification of
personality variants. A proposal. Arch Gen Psychiatry 44, 57388 (1987).
60. Lowe, N. et al. Joint Analysis of the DRD5 Marker Concludes Association with
Attention-Deficit/Hyperactivity Disorder Confined to the Predominantly Inattentive
and Combined Subtypes. Am J Hum Genet 74, 34856 (2004).
61. Gaunt, T.R., Cooper, J.A., Miller, G.J., Day, I.N. & O'Dell, S.D. Positive associations
between single nucleotide polymorphisms in the IGF2 gene region and body mass
index in adult males. Hum Mol Genet 10, 1491501 (2001).
62. Quesada, A. & Micevych, P.E. Estrogen interacts with the IGF-1 system to protect
nigrostriatal dopamine and maintain motoric behavior after 6-hydroxdopamine
lesions. J Neurosci Res 75, 10716 (2004).
63. Shavali, S., Ren, J. & Ebadi, M. Insulin-like growth factor-1 protects human dopam-
inergic SH-SY5Y cells from salsolinol-induced toxicity. Neurosci Lett 340, 7982
(2003).
64. Stefansson, H. et al. Neuregulin 1 and susceptibility to schizophrenia. Am J Hum
Genet 71, 87792 (2002).
65. Bach, L.A. & Leeding, K.S. Insulin-like growth factors decrease catecholamine
content in PC12 rat pheochromocytoma cells. Horm Metab Res 34, 48791 (2002).
66. Offen, D. et al. Protective effect of insulin-like-growth-factor-1 against dopamine-
induced neurotoxicity in human and rodent neuronal cultures: possible implications
for Parkinson's disease. Neurosci Lett 316, 12932 (2001).
67. Clarkson, E.D. et al. IGF-I and bFGF improve dopamine neuron survival and behav-
ioral outcome in parkinsonian rats receiving cultured human fetal tissue strands. Exp
Neurol 168, 18391 (2001).
68. Guan, J. et al. N-terminal tripeptide of IGF-1 (GPE) prevents the loss of TH positive
neurons after 6-OHDA induced nigral lesion in rats. Brain Res 859, 28692 (2000).
69. Zawada, W.M. et al. Growth factors improve immediate survival of embryonic dopam-
ine neurons after transplantation into rats. Brain Res 786, 96103 (1998).
70. Hemberger, M. et al. H19 and Igf2 are expressed and differentially imprinted in
neuroectoderm-derived cells in the mouse brain. Dev Genes Evol 208, 393402.
71. Pham, N.V., Nguyen, M.T., Hu, J.F., Vu, T.H. & Hoffman, A.R. Dissociation of IGF2
and H19 imprinting in human brain. Brain Res 810, 18 (1998).
72. Mori, H. et al. Expression of mouse igf2 mRNA-binding protein 3 and its implications
for the developing central nervous system. J Neurosci Res 64, 13243 (2001).
73. Albrecht, S. et al. Variable imprinting of H19 and IGF2 in fetal cerebellum and
medulloblastoma. J Neuropathol Exp Neurol 55, 12706 (1996).
74. Pera, E.M., Wessely, O., Li, S.Y. & De Robertis, E.M. Neural and head induction by
insulin-like growth factor signals. Dev Cell 1, 65565 (2001).
75. Pera, E.M., Ikeda, A., Eivers, E. & De Robertis, E.M. Integration of IGF, FGF, and
anti-BMP signals via Smad1 phosphorylation in neural induction. Genes Dev 17,
30238 (2003).
76. Ding, Y.C. et al. Evidence of positive selection acting at the human dopamine receptor
D4 gene locus. Proc Natl Acad Sci U S A 99, 309314 (2002).
77. Wang, E. et al. The Genetic Architecture of Selection at the Human Dopamine
Receptor D4 (DRD4) Gene Locus. Am J Hum Genet 74, 931944 (2004).
78. Bouchard, T.J., Jr. & Loehlin, J.C. Genes, evolution, and personality. Behav Genet
31, 24373 (2001).
79. Ebstein, R.P., Benjamin, J. & Belmaker, R.H. Personality and polymorphisms of
genes involved in aminergic neurotransmission. Eur J Pharmacol 410, 205214
(2000).
1903_C018.fm Page 278 Wednesday, July 26, 2006 7:06 PM
80. Giraud, T., Pedersen, J.S. & Keller, L. Evolution of supercolonies: the Argentine ants
of southern Europe. Proc Natl Acad Sci U S A 99, 60759 (2002).
81. Krieger, M.J. & Ross, K.G. Identification of a major gene regulating complex social
behavior. Science 295, 32832 (2002).
82. Keller, L. & Parker, J.D. Behavioral genetics: a gene for supersociality. Curr Biol 12,
R1801 (2002).
83. Young, L.J., Lim, M.M., Gingrich, B. & Insel, T.R. Cellular mechanisms of social
attachment. Horm Behav 40, 1338 (2001).
84. Taylor, S.E. et al. Biobehavioral responses to stress in females: tend-and-befriend,
not fight-or-flight. Psychol Rev 107, 41129 (2000).
85. Ferguson, J.N., Young, L.J. & Insel, T.R. The neuroendocrine basis of social recog-
nition. Front Neuroendocrinol 23, 20024 (2002).
86. Winslow, J. & Insel, T.R. Vasopressin modulates male squirrel monkeys' behavior
during social separation. Eur J Pharmacol 200, 95101 (1991).
87. Thompson, R., Gupta, S., Miller, K., Mills, S. & Orr, S. The effects of vasopressin
on human facial responses related to social communication. Psychoneuroendocrinol-
ogy 29, 3548 (2004).
88. Kim, S.J. et al. Transmission disequilibrium testing of arginine vasopressin receptor
1A (AVPR1A) polymorphisms in autism. Mol Psychiatry 7, 5037 (2002).
89. Bachner-Melman, R. et al. Linkage between vasopressin receptor AVPR1A promoter
region microsatellites and measures of social behavior in humans. Journal of Indi-
vidual Differences 36, 45160 (2005).
90. Hammock, E.A. & Young, L.J. Variation in the vasopressin V1a receptor promoter
and expression: implications for inter- and intraspecific variation in social behaviour.
Eur J Neurosci 16, 399402 (2002).
91. Lennox, R.-D. & Wolfe, R.-N. Revision of the Self-Monitoring Scale. Journal of
Personality and Social Psychology 46, 13491364 (1984).
92. Buhrmester, D. & Furman, W. Perceptions of sibling relationships during middle
childhood and adolescence. Child Dev 61, 138798 (1990).
93. Bachner-Melman, R. et al. Association between a vasopressin receptor AVPR1A
promoter region microsatellite and eating behavior measured by a self-report ques-
tionnaire (Eating Attitudes Test) in a family-based study of a non-clinical population.
Int J Eat Disord 36, 45160 (2004).
94. Furman, W. & Buhrmester, D. Children's perceptions of the qualities of sibling
relationships. Child Dev 56, 44861 (1985).
95. Snyder, M. Self-monitoring of expressive behavior. Journal of Personality and Social
Psychology 30, 526537 (1974).
96. Wolfe, R.N., Lennox, R.D. & Cutler, B.L. "Getting along" and "getting ahead":
Empirical support for a theory of protective and acquisitive self-presentation. Journal
of Personality and Social Psychology 50, 356361 (1986).
97. Miller, M.L., Omens, R.S. & Delvadia, R. Dimensions of social competence: Per-
sonality and coping style correlates. Personality and Individual Differences 12,
955964 (1991).
98. Day, D., Schleicher, D.J., Unckless, A.L. & Hiller, N.J. Self-Monitoring personality
at work: A meta-analytic investigation of construct validity. Journal of Applied Psy-
chology 87, 390401 (2002).
99. Cloninger, C.R. et al. Anxiety proneness linked to epistatic loci in genome scan of
human personality traits. Am J Med Genet 81, 3137 (1998).
1903_C018.fm Page 279 Wednesday, July 26, 2006 7:06 PM
100. Fullerton, J. et al. Linkage analysis of extremely discordant and concordant sibling
pairs identifies quantitative-trait Loci that influence variation in the human personality
trait neuroticism. Am J Hum Genet 72, 87990 (2003).
101. Zohar, A.H. et al. Tridimensional personality questionnaire trait of harm avoidance
(anxiety proneness) is linked to a locus on chromosome 8p21. Am J Med Genet 117B,
669 (2003).
102. Dina, C. et al. Fine mapping of a region on chromosome 8p gives evidence for a
QTL contributing to individual differences in an anxiety-related personality trait, TPQ
harm avoidance. Am J Med Genet B 132, 1048 (2004).
103. Eysenck, H.J. The Scientific Study of Personality, (Routledge & Kegan Paul, London,
1952).
104. Schuckit, M.A. et al. A genome-wide search for genes that relate to a low level of
response to alcohol. Alcohol Clin Exp Res 25, 3239 (2001).
105. Stefansson, H. et al. Neuregulin 1 and susceptibility to schizophrenia. Am J Hum
Genet 71, 87792 (2002).
106. Stefansson, H. et al. Association of neuregulin 1 with schizophrenia confirmed in a
Scottish population. Am J Hum Genet 72, 837 (2003).
107. Neale, B.M., Sullivan, P.F. & Kendler, K.S. A genome scan of neuroticism in nicotine
dependent smokers. Am J Med Genet B 132, 659 (2005).
108. Curtis, D. Re-analysis of collaborative study on the genetics of alcoholism pedigrees
suggests the presence of loci influencing noveltyseeking near D12S391 and
D17S1299. Psych Gen 14, 1515 (2004).
1903_C018.fm Page 280 Wednesday, July 26, 2006 7:06 PM
1903_C019.fm Page 281 Wednesday, July 26, 2006 7:06 PM
CONTENTS
Summary ................................................................................................................ 281
19.1 Introduction .................................................................................................. 282
19.2 Mice.............................................................................................................. 282
19.3 Humans......................................................................................................... 286
19.4 Mice and Humans ........................................................................................ 287
References.............................................................................................................. 288
SUMMARY
Aggression is a complex, social behavior involving at least two individuals. Some
conceptual and methodological issues relevant to genetic analyses of aggression in
mice and humans are considered.
In animals, there are several types of aggression. In male and female mice, these
are offense, defense, infanticide, and predation. In females, there are also types of
aggression associated with pregnancy and lactation. There are strain differences for
each type, and genes have been identified for offense and defense. At present, 38
genes have been reported to affect some aspect of offense. There are also environ-
mental effects that include prenatal maternal environment, postnatal maternal envi-
ronment, postweaning environment, and test situation. All of these must be considered
in critically evaluating experimental data and experimental design for the genetics of
a type of aggression. The behaviors observed and measured are also a factor. Some
of these issues are discussed in detail for offense in males.
Although there may be different types of aggression in humans, it is difficult to
identify these. Consequently, aggression in humans is often treated as a single,
unitary trait. This lumping together of different types of aggression may render
difficult the interpretation and evaluation of genetic analyses for human aggression.
Regardless, there is some evidence that individual differences in personality traits
associated with aggressiveness are due to genetic variants. Also, three genes may
281
1903_C019.fm Page 282 Wednesday, July 26, 2006 7:06 PM
have been shown to have effects on impulsive aggression, and there is a genotype
interaction for one of these. However, there is no proof at present that criminal
behaviors associated with physical assault are substantially heritable, other than as
interactions of specific genes with specific environments.
19.1 INTRODUCTION
There are many definitions of aggression and of aggressive behavior. One of these
is overt behavior involving intent to inflict noxious stimulation to or behave destruc-
tively toward another organism.28 Another of these is any form of behavior directed
toward the goal of harming or injuring another living being who is motivated to
avoid such treatment.3 These definitions specify broadly the behavioral domain
considered in this chapter.
It has long been recognized that, at least in animals, there are many types of
aggression or aggressive behavior. These differ in mechanism, eliciting stimuli,
development, function, and phylogeny. Consequently, they may also differ in their
genetics. That is to say, the same genes may not be causes of differences or act in
development of each type of aggression.
Regardless, aggression has long been of interest as a behavioral trait for genetic
analyses.16,17 Although most of the research on genetics of aggression has been done
with mice, much of the scientific and public interest in the genetics of aggression is
sparked by concerns about the role of nature and nurture in human aggression. This
includes the causes of not only adaptive types of aggression, but also aggression in
psychopathologies, crime, and war.5,28 Here, some conceptual and methodological
issues involved in research on the genetics of aggression are considered, first for mice
and then for humans.
19.2 MICE
In male and female mice, the four types of aggression are offense, defense, infan-
ticide, and predation.16 There are also female specific types of aggression which
occur during pregnancy or lactation.4 Between strain differences exist for each type
of aggression,10 and genes have been identified for offense and defense.16,19,21,23
Offense and defense by males and females (not pregnant or lactating) are ago-
nistic behaviors. Infanticide and predation by males and females are not agonistic
behaviors. The agonistic behaviors have the common function of adaptation to
situations involving physical conflict between members of the same species. Each
type of aggressive behavior can be characterized in mice by function and bite targets.
These are listed in Table 19.1. For each type, there are also differences in mechanism
and development. Thus, it may be expected that although some genes may affect all
types of aggression, others will affect only one or a few kinds. Also, conceptual and
methodological issues in identifying genes for each type of aggression may be
different. Because most of the current research on the genetics of aggression in mice
focuses on offense in males, this part of the chapter will consider this behavior in
mice. Information on the other types of aggressive behavior in mice can be found
in references 2, 6, 14, 16.
1903_C019.fm Page 283 Wednesday, July 26, 2006 7:06 PM
Aggression 283
TABLE 19.1
Types of Mouse Aggression
Type Bite Target Function
Offense Back, rump, tail Obtaining and restraining resources
Defense Face, shoulders Self-protection from injury by
others
Pregnancy Anywhere Prevent reproductive termination
Lactation Flanks, neck Prevent reproductive termination
Infanticide Anywhere Reproductive termination
Predation (insects) Head, thorax Food
Elsewhere, we have reviewed in detail the many aspects of the environment that
can have an effect on offensive aggression in mice17,18,21 and that must be considered
in critically evaluating research results and in designing experimental studies on the
genetics of offense. These include prenatal maternal environment, postnatal parental
environment, postweaning environment, and test conditions. Also, the behavioral
measures used are a factor. Effects of test conditions and behavioral measure will
be considered in more detail.
Offense is obviously a social behavior involving at least two mice, and most
genetic analyses of aggression in mice use dyadic tests. For genetic analyses, there
are three types of dyadic tests. These are homogeneous set test, the panel of testers,
and the standard opponent test. In the homogeneous set test, all encounters are
between mice of the same genotype (strain or F1 hybrids). In the panel of testers,
encounters are between mice of the same or different genotype (strains or F1
hybrids), and each experimental group of genotypes (strains or F1 hybrids) is tested
against a panel of genotypes (strains or F1 hybrids). The panel of testees (experi-
mental group) and panel of testers (opponent group) may be of the same or different
array of genotypes (strains or F1 hybrids). Because in both the homogeneous set
test and panel of testers, the genotypes of both individuals in the dyadic encounter
must be known, only isogenic populations may be used in the two paradigms.
Isogenic populations include inbred strains, recombinant inbred strains, consomic
strains, recombinant congenic strains, congenic strains, coisogenic strains, and F1
hybrids of two such strains. With these two paradigms, only recombinant inbred
strains, recombinant congenic, congenic (such as knockout mutants), and coisogenic
strains (such as transgenes and insertional mutants) can be used to identify genes
with effects on offense. In the standard opponent test, all encounters are between
mice of one or another genotype and mice of a single, standard genotype (stock,
strain, or F1 hybrid). Since only the genotype of the standard opponent must be
known, the standard opponent test can be used with mice from isogenic or hetero-
genic populations. The latter include F2, F3, and other Fn hybrids, backcrosses,
heterogeneous stocks, and selected lines. These can be used to map quantitative trait
loci (QTLs) with effects on offense.
Strain or genetic differences in male offense can depend on the genotype not
only of the testee (experimental group), but also of the tester (opponent group). For
1903_C019.fm Page 284 Wednesday, July 26, 2006 7:06 PM
Aggression 285
inbred strains. There are several reports of phenotypic correlations among latency
to attack, number of attacks, and accumulated attack time. Also, it has been reported
that in tests with A/J males as standard opponents, the percentage of males that
attack is correlated with many other measures of offense across several strains.
However, once the QTL is identified and congenic strains for it are established, there
should be complete analyses of this genes effects on offense. This can and should
always be done for congenic or coisogenic strains with knockout mutants, other
mutants, or transgenes.
Just as drugs do not always affect all aspects of offense, genes may not always
do so. For example, it has been reported that the proportion of mice attacking has
a mode of inheritance across several strains and their hybrids different from that for
number of attacks and accumulated attack time. Also, effects of genes on the Y
chromosome on offense are dependent on the behavioral measure. In a standard
opponent test, the CBA/H and NZB Y chromosomes differ in effect on proportion
of males that attack at least once but not in effect on latency to attack. Similarly,
the Nos1 mutant affects the frequency but not latency of attacks in a standard
opponent test. Hence, in both types of research, complete ethological or behavioral
analysis is essential.
These and other methodological issues must be considered in evaluating reports
of gene effects on offense and in designing research to test gene effects on offense.
To date, there is some evidence for effects of 38 genes on male offense.17,21 These
are listed in Table 19.2. Several Slc6a327 and Gnai225 genes have been studied
intensively, leading to proposals for a role of genes expressed in the hippocampus,11
TABLE 19.2
Genes and Male Offense
Gene Name Chromosome
Adora1a Adenosine1A receptor 1
Adora2a Adenosine2A receptor 10
Adra2C Adrenergic alpha 2C receptor 5
Ar Androgen receptor X
Avpr1b Argenine vasopressin 1B receptor 1
B2m Beta2-microglobulin 12
Bcr Breakpoint cluster region 10
CamK2a -Calcium/calmodulin kinase II 18
Ckb Brain creatinine kinase 12
Comt Catecho-o-methyl transferase 16
Cyp19 Aromatase 9
Esr1 -Estrogen receptor 6
Esr2 -Estrogen receptor 12
Fyn Fyn tyrosine kinase 10
Gad1 Glutamnic acid decarboxylase 65 9
Gdi-1 Guanosine diphosphate dissociation inhibitor X
Girgeo-22 Gene trap ROSA-b-Geo 22 10
(Continued)
1903_C019.fm Page 286 Wednesday, July 26, 2006 7:06 PM
TABLE 19.2
Genes and Male Offense (Continued)
Gene Name Chromosome
Gnai2 Guanine nucleotide binding protein, Alpha 9
inhibiting 2
Hrh1 Histamine 1 receptor 6
Htr1b 5-HT1B receptor 9
Il-6 Interleukin-6 5
Maoa Monoamine oxidase A X
Mm2 Membrane metalo enopeptidase 3
Ncam Neural cell adhesion molecule 9
Nos1 Nitric oxide synthase 1 5
Nos3 Nitric oxide synthase 3 5
Nr3e1 Nuclear receptor family2 group E member 10
Oxt Oxytocin 2
Penk Enkephalin 4
Pet-1 Pet-1 ETS factor 1
Reg-2 Regulator of G-protein signaling 1
Slc6a3 Dopamine transporter 13
Slc6a4 Serotonin transporter 11
Sts Steroid sulfatase X/Y pseudoautosomal
Tacr1 Tachnykinine-1 receptor 6
Tgfa Transforming growth factor 6
Trpc2 TRP ion channel 7
19.3 HUMANS
Most researchers believe that, as with animals, there are many types of human
aggression. Although some have suggested that offense, defense, and predation are
distinct types of aggression in humans, it is widely recognized that it is difficult to
define the different types of aggression in humans. In part, this is because different
motor patterns of offense, defense, and predation in humans may be masked by the
use of extrinsic weapons. Other distinctions among types of human aggression
include hostile vs. instrumental aggression3 and, within the hostile category, between
impulsive and nonimpulsive aggression.19 Also, there are further distinctions within
any of these categories. These are physical vs. verbal, active vs. passive, and direct
vs. indirect. These different parsing of the aggressive phenotype could have an effect
on the results of genetic analyses. But distinctions are rarely made in research on
the genetics of human aggression. Rather, aggression often is treated as a single
unitary trait. This lumping together of different types of aggression may render
difficult genetic analyses. However, it has recently been suggested but not proven
that all aggression in humans is of the defensive type.1
1903_C019.fm Page 287 Wednesday, July 26, 2006 7:06 PM
Aggression 287
The methods for investigating human aggression are described by Baron and
Richardson3 and by Volavka.28 The actual behavior may be directly observed or
inferred from self- or archival reports. Observed behavior is frequently studied in
controlled experiments in the laboratory. More recently, it has also been observed
in normal social and cultural settings. Human aggression is not only investigated as
behavior (for example, the Child Behavior Check List) but also as a personality trait.
Some tests used are the Minnesota Multiphasic Personality Invertory (MMPI) and
Buss-Durkee Hostility Inventory. The implication from results of these tests is that
in the same interpersonal situation, some personality types are more prone than
others to be aggressive.
Many genetic studies of human aggression have been for traits inferred from
personality tests. These have often focused on impulsive or antisocial personality traits
that are associated with aggression.7,26 There is indication that some of the individual
differences in such traits are due, at least in part, to genetic variation. Also, variants
of the genes for monoamine oxidase A (MAOA), tryptophan hydroxylase, and the
5HT1B receptor may be associated with some aspects of impulsivity, including aggres-
sion. Recently, the effects of MAOA variants on measures of male aggression have
been shown to be dependent on maltreatment in childhood8 or heroin addiction.12
Other genetic studies have been concerned with antisocial or criminal behavior
based on archival data. However, the data do not prove that variations in murder,
assault, rape, or other offenses associated with physical attacks are substantially
heritable. The failure to detect a nonzero heritability for criminal behavior involving
physical attacks may be a function of the low incidence of criminal violence. Large
sample sizes are required to detect nonzero heritability for behaviors of low inci-
dence. These have not been available to date for criminal behaviors involving
physical attacks. However, this may also be due to interactions of genotype and
environment.8,12
All of this research should be critically evaluated in the context of the chapters
in this volume on twin studies and family studies. Regardless, it is clear that in
humans as in mice only part of the individual differences in any type of aggression
is due to genetic variants.
REFERENCES
1. ALBERT DJ, WALSH ML, JONIK RH: Aggression in humans: What is its biological
foundation? Neuroscience and Biobehavioral Reviews, 1993, 17, 405425.
2. ALLEVA E: Assessment of aggressive behavior in rodents. In: PM Conn (Ed):
Paradigms for the study of behavior. Methods in Neurosciences Vol. 14. Academic
Press, New York, 1993, pp 111137.
3. BARON RA, RICHARDSON DR: Human Aggression 2nd ed. Plenum Press, New
York, 1994.
4. BJORKQVIST K, NIEMELA P: Of Mice and Women: Aspects of Female Aggression.
Academic Press, New York, 1992.
5. BOCK GR, GOODE JA: Genetics of Criminal and Antisocial Behaviour. John Wiley
& Sons, New York, 1996.
6. BRAIN PF, HAUG M, KAMIS A: Hormones and different tests for aggression with
particular reference to the effects of testosterone metabolites. In: J Balthazart, E Prove,
R. Gilles (Eds): Hormones and Behaviour in Higher Vertebrates. Springer-Verlag,
Berlin, 1983, pp 290304.
7. CAREY G: Genetics and violence. In: AJ Reiss Jr., KA Miczek, JA Roth (Eds)
Understanding and Preventing Violence Vol. 2: Biobehavioral Influence. National
Academy Press, Washington, 1994, pp 2158.
8. CASPI A, MCCLAY J, MOFFITT TE, MILL J, MARTIN J, CRAIG I W, TAYLOR
A, POULTON R: Role of genotype in the cycle of violence in maltreated children.
Science, 2002, 297, 851854.
9. CHIAVEGATTO S, NELSON RJ: Interaction of nitric acid and serotonin in aggres-
sive behavior. Hormones and Behavior, 2003, 44, 233241
10. CRAWLEY JN, BELKNAP JK, COLLINS A, CRABBE JC, FRANKEL W, HEND-
ERSON N, HITZEMANN RJ, MAXSON SC, MINER LL, SILVA AJ, WEHNER
JM, WYNSHAW-BORIS A, PAYLOR R: Behavioral phenotypes of inbred mouse
strains. Psychopharmacology, 1998, 132, 107124.
11. Feldker DE, Datson N.A., Veenema AH, Proutski V, LATHOUWERS D , DE KLOET
E R, VREUGDENHIL E: (2003b). Gene chip analysis of hippocampal gene expres-
sion profiles of short- and long-attack-latency mice: technical and biological impli-
cations. Journal of Neuroscience Research, 2003, 74, 701716.
12. GERRA G, GAROFANO L, BOSARI S, PELLEGRINI C, ZAIMOVIC A, MOI G,
BUSSANDRI M, MOI A, BRAMBILLA F, MAMELI A, PIZZAMIGLIO M, DON-
NINI C: Analysis of monoamine oxidase A (MAO-A) promoter polymorphism in
male heroin-dependent subjects: behavioural and personality correlates. J Neural
Transm. 2004, 111, 61121.
13. HAHN ME, SCHANZ N: Issues in the genetics of social behavior: revisited. Behavior
Genetics, 1996, 26, 463470.
14. JONES SE, BRAIN PF: Performance of inbred and outbred laboratory mice in
putative tests of aggression. Behavior Genetics, 1987, 17, 8796.
15. LE ROY I, MORTAUD S, TORDJMAN S, DONSEZ-DARCEL E, CARLIER M,
DEGRELLE H, ROUBERTOUX PL: Genetic correlation between steroid sulfatase
concentration and initiation of attack behavior in mice. Behavior Genetics, 1999, 29,
131136.
16. MAXSON SC: Potential genetic models of aggression and violence in males. In: P
Driscoll (Ed) Genetically Defined Animal Models of Neurobehavioral Dysfunctions.
Birkhauser, Boston, 1992, pp 174188.
1903_C019.fm Page 289 Wednesday, July 26, 2006 7:06 PM
Aggression 289
20 Genetic Analysis of
Emotional Behaviors Using
Animal Models
Andr Ramos and Pierre Mormde
CONTENTS
Summary ................................................................................................................ 291
20.1 Introduction................................................................................................. 291
20.2 Definitions .................................................................................................. 292
20.3 Measurements ............................................................................................. 292
20.3.1 Unconditioned Stress Responses, Forced Exposure...................... 293
20.3.2 Unconditioned Stress Responses, Choice Tests............................. 295
20.3.3 Conditioning Experiments.............................................................. 295
20.3.4 Interpreting Emotionality through a Multivariate Approach......... 296
20.4 Genetic Analysis......................................................................................... 298
20.4.1 Selection Experiments.................................................................... 298
20.4.2 Quantitative Genetic Analysis........................................................ 299
20.4.3 Genetic Links between Different Traits......................................... 299
20.4.4 The QTL Chase .............................................................................. 300
20.4.5 Gene Targeting and Transgenesis .................................................. 303
References.............................................................................................................. 303
SUMMARY
The aim of this chapter is to introduce the reader to the universe of experimental
approaches available for the genetic study of emotionality, the focus being on the
use of animal models. First, some preliminary conceptual issues will be discussed
and then an overview of the main methods of measuring and interpreting the so-
called emotional behaviors will be presented. Finally we present the classical and
modern strategies to determine the genetic basis of variance in emotionality-related
traits.
20.1 INTRODUCTION
Studying the genetic aspects of emotions, in either humans or animal models, is a
particularly challenging task for three main reasons: emotions are difficult to define;
emotions are difficult to measure; and emotions are difficult to genetically analyze.
291
1903_C020.fm Page 292 Thursday, July 27, 2006 8:41 PM
20.2 DEFINITIONS
The interpretation of terms derived from the word emotion may vary among authors,
but most researchers agree that emotions are subjective experiences, that they appear
under nonordinary circumstances and that they involve behavioral and physiological
changes.46 In the field of experimental psychobiology, the concept of emotionality
has been used widely for laboratory animals, as defined by Hall.28
The term emotionality is defined as the state of being emotional. This state consists of
a group of organic, experiential and expressive reactions and denotes a general upset or
excited condition of the animal. Emotionality can be thought of as a trait The reader
is warned against interpreting emotionality as a thing or faculty. It is merely a convenient
concept for describing a complex of factors. the term emotion is conventionally used
to designate the experience which results from emotional stimulation.
20.3 MEASUREMENTS
A number of paradigms have been used to measure emotional behaviors of exper-
imental animals. Basically, these experimental paradigms consist of exposing
1903_C020.fm Page 293 Thursday, July 27, 2006 8:41 PM
animals to aversive stimuli while observing and analyzing their behavior. The aver-
sive stimuli may vary in nature from being essentially physical (electrical footshocks,
food deprivation, submersion in water, etc.) to those considered to be mainly psy-
chological (novel environments, strongly illuminated areas, open spaces, exposure
to the smell of predators, heights, social instability, etc.). Another important variable
is the ability of the animal to avoid or escape the aversive stimulus. Most experi-
mental situations use unconditioned responses to a number of novel environments,
or learning paradigms.
100
Outer locomotion
75
50
25
0
SHR WKY BN WF LEW FIS
15
Inner locomotion
10
0
SHR WKY BN WF LEW FIS
5
Fecal boli
0
SHR WKY BN WF LEW FIS
FIGURE 20.1 Open field. Locomotion in the peripheral (outer) and central (inner) part and
defecation score measured in 12 males each from six rat strains (SHR = Spontaneously
Hypertensive Rat; WKY = Wistar Kyoto; BN = Brown Norway; WF = Wistar Furth; LEW =
Lewis; FIS = Fischer 344). (Reprinted from Behav. Brain Res., 85, Ramos, A. et al., A multiple
test study of anxiety-related behaviours in six inbred rat strains, 5769, Copyright 1997, with
permission from Elsevier.)
1903_C020.fm Page 295 Thursday, July 27, 2006 8:41 PM
approach food and have a lower food consumption.42,62 These tests have been devel-
oped mainly for the screening of anxiolytic drugs.
Among other tests of reactivity, we can cite the acoustic startle reflex26,62 and
ultrasonic pup vocalizations, elicited during the first 2 weeks of life by isolating
the pups from their mother,52 but this list is far from complete.
20
10
0
SHR WKY BN WF LEW FIS
15
Total arm entries
10
0
SHR WKY BN WF LEW FIS
FIGURE 20.2 Elevated plus maze. Time spent in the open arms and number of total arm
entries measured in 12 males each from six rat strains. (Reprinted from Behav. Brain Res.,
85, Ramos, A. et al., A multiple test study of anxiety-related behaviours in six inbred rat
strains, 5769, Copyright 1997, with permission from Elsevier.)
tests can be applied to the same group of animals. In both cases, results normally
suggest that different behavioral reactions and different experimental contexts may
be associated to different dimensions of emotionality.
Single test situations. The multivariate or factorial analysis has been successfully
applied to distinguish several dimensions of emotional reactivity in experimental
animals in single experimental designs. The elevated plus maze paradigm is taken as
an example because it is probably the best documented. Cruz and collaborators15
measured 13 behavioral variables in the elevated plus maze from 30 male Wistar rats.
The first and most important two factors are seen as representing variables of anxiety
and locomotion. Percentage of open arm entries, time in open arms, and time in the
closed arms (negatively related) are the variables with the highest loadings on the first
factor. These measures were changed to one direction by administration of anxiolytic
drugs and to the opposite direction by anxiogenic drugs and were therefore thought
to measure anxiety. The second factor reflects mainly the number of entries in the
closed arms (locomotion). The total number of arm entries, which is normally used
as the main index of locomotion in the plus maze, is an ambiguous measure since
this variable loads simultaneously on both anxiety and locomotion factors, appearing
to be more contaminated by emotional responses than the variable number of closed
arm entries. Finally, time spent in the center of the plus maze loaded on a third factor.
Multiple test situations. This kind of analysis is most powerful to investigate the
data obtained from the same animals in different test situations. For instance, Trullas
and Skolnick62 studied the behavior of 16 inbred mouse strains in the open field and
the plus maze. The variables that saturated factor 1 (44.5% of total variance) were
all related to time spent or number of crosses into the open arms of the elevated plus
maze. Since the time spent in the open arms correlated with both the acoustic startle
reflex and hyponeophagia in a novel environment, factor 1 was interpreted as reflect-
ing the level of anxiety of the animal. Variables related to ambulation, both in the
open field and plus maze, loaded on factor 2 (27.5% of total variance), which appears
to measure general activity. Here again, a third factor (10.3% of total variance) was
defined almost exclusively as time spent in the central platform of the plus maze. A
similar analysis was conducted in our laboratory on the data obtained for 19 variables
from four behavioral tests: open field, plus maze, black and white box, and social
interaction.45 The subjects were male and female rats from six inbred strains. The
first factor (36.6% of total variance) was also defined by variables classically asso-
ciated with anxiety (locomotion in the central part of the open field, entries and time
spent in the open arms of the plus maze, behavior in the black and white box). The
second factor (30.3% of total variance) was defined by measures of locomotion (total
and outer locomotion in the open field, locomotion in the closed arms of the plus
maze). Interestingly enough, time of social interaction in a novel environment loaded
on a third factor (18.3% of total variance), together with defecation rates in the open
field and black and white box. Such an independence between behaviors measured
in different anxiety tests, such as the plus maze and social interaction, has already
been described by File.21 These data show that the three measures of open field
behaviorouter locomotion, inner locomotion and defecationload on different
factors and are therefore independent for the most part.
1903_C020.fm Page 298 Thursday, July 27, 2006 8:41 PM
Those strains/lines have been used extensively for quantitative genetic analysis
and are now being used in molecular genetic studies.
Flint et al.24 were the first to report, in 1995, QTLs associated with emotional
behaviors in mice. In rats, the first QTLs for emotionality were mapped in 1996 by
Moisan et al.37 (locomotor activity in a novel environment, WKY WKHA F2
intercross) and in 1999 by Ramos et al., using an F2 intercross from Lewis and SHR
inbred strains.48 These initial studies very often identified loci that were rather
specific for one or a few behavioral responses and that were not replicated across
different strains. Both these aspects brought some doubts about the psychological
significance of these traits and the biological interest of these loci. More recent
studies showed that some QTLs are likely to harbor genes with more general
emotional effects, whereas others affect specific behaviors that may correspond to
particular forms of fear and anxiety or even reflect other unrelated traits. For exam-
ple, an elegant work by Turri et al.64 showed that a QTL on mouse chromosome 15
has a broad range of behavioral effects that are consistent with a presumed influence
on fear or anxiety, but not on general locomotion. On the other hand, a QTL on
chromosome 1, which was thought to be anxiety-related after having been replicated
in different studies,23 was shown in fact to have an overall influence on exploratory
behaviors in both aversive and nonaversive situations. Finally, one locus on chro-
mosome 6 showed a very specific effect on a single variable from one particular
test, the elevated plus maze. In rats, a multivariate approach applied to QTL mapping
pointed to similar conclusions by identifying two loci (on chromosomes 1 and 19)
for very specific behaviors and one locus (on chromosome 5) affecting several
anxiety-related measures, such as avoidance responses, fear conditioning, and behav-
ior in the plus maze and open-field tests.20 QTLs with pleiotropic effects on numerous
fear-related measures are seen as good candidates for harboring genes expected to
be relevant for human anxiety.23
Some QTLs for emotionality may have pleiotropic effects on other types of
behaviors, potentially related to pathologies that can be comorbid with anxiety
disorders, such as alcoholism. In 1998, the group of Lucinda Carr, using the selected
rat lines P (alcohol-preferring) and NP (alcohol-nonpreferring), identified a powerful
QTL for alcohol consumption on the mid-portion of chromosome 4.1,6 A few months
later, a QTL located near this chromosome region was shown to have a strong
influence on the locomotion in the inner part of the open field, an index of emo-
tionality, in Lewis/SHR F2 rats.48 In spite of having their peaks more than 15 cM
apart, the confidence intervals of these two QTLs overlapped. The existence of a
QTL on chromosome 4 for inner locomotion in the open field (named Ofil1) was
corroborated by a second study involving two selected rat lines derived from the
Lewis and SHR strains.39 More recently, another whole-genome QTL analysis based
on the F2 intercross of two other rat strains that contrast for alcohol intake, namely
HEP (high-ethanol preferring) and WKY (Wistar-Kyoto) led to the identification of
a significant locus on chromosome 4, approximately midway between the two above-
mentioned QTLs (Figure 20.3), that strongly affects ethanol preference.59
Considering that anxiety is one of the predisposing factors involved in alcohol
abuse,41 the combination of findings described above led to the hypothesis that these
different QTLs might correspond in fact to one single locus with pleiotropic effects
on alcohol drinking and emotionality. This hypothesis has been corroborated so far.
The latest data from two rat lines derived from HEP and WKY, which were selected
1903_C020.fm Page 302 Thursday, July 27, 2006 8:41 PM
6
5
4
3
2
1
0
0 20 40 60 80 100 120 140 160 180 200
Mb
FIGURE 20.3 QTL map. Approximate positions of three QTLs independently mapped on
rat chromosome 4 for: inner locomotion in the open field (emotionality) in Lewis and SHR
rats;48 preference for 10% ethanol in P and NP rats;1,6 preference for 5% ethanol in HEP and
WKY rats.59 Physical map positions were estimated by comparing linkage data from the
original publications and marker coordinates (in base pairs) from the Rat Genome Database
(http://rgd.mcw.edu/).
of this strategy was shown, for example, by the replication of a QTL for open-field
behavior located on mouse chromosome 1.58 Two other alternative strategies, high-
resolution mapping using commercially available outbred animals and quantitative
complementation testing, were recently used to dissect this specific mouse QTL into
smaller, more precise components as well as to identify one of its underlying genes.65
REFERENCES
1. BICE P, FOROUD T, BO R, CASTELLUCCIO P, LUMENG L, LI TK, CARR LG:
Genomic screen for QTLs underlying alcohol consumption in the P and NP rat lines.
Mamm Genome, 1998, 9:949955.
2. BIGNAMI G: Selection for high rates and low rates of avoidance conditioning in the
rat. Anim Behav, 1965, 13:220221.
3. BROADHURST PL: The Maudsley reactive and nonreactive strains of rats : A survey.
Behav Genet, 1975, 5:299319.
4. BRUNELLI SA, HOFER MA, WELLER A: Selective breeding for infant vocal
response: a role for postnatal maternal effects? Dev Psychobiol, 2001, 38:221228.
5. BRUSH FR: Genetic determinants of individual differences in avoidance learning:
behavioral and endocrine characteristics. Experientia, 1991, 47:10391050.
6. CARR LG, FOROUD T, BICE P, GOBBETT T, IVASHINA J, EDENBERG H,
LUMENG L, LI T-K. A quantitative trait locus for alcohol consumption in selectively
bred rat lines. Alcohol Clin Exp Res, 1998, 22:884887.
7. CASTANON N, HENDLEY ED, FAN X-M, MORMDE P: Psychoneuroendocrine
profile associated with hypertension or hyperactivity in spontaneously hypertensive
rats. Am J Physiol, 1993, 265:R1304R1310.
8. CASTANON N, PEREZ-DIAZ F, MORMEDE P: Genetic analysis of the relation-
ships between behavioral and neuroendocrine traits in roman high and low avoidance
rat lines. Behav Genet, 1995, 25:371384.
9. CLMENT Y, CALATAYUD F, BELZUNG C: Genetic basis of anxiety-like behav-
iour: A critical review. Brain Res Bull, 2002, 57:5771.
10. COURVOISIER H, MOISAN MP, SARRIEAU A, HENDLEY ED, MORMEDE P:
Behavioral and neuroendocrine profile of stress reactivity in the WKHA/WKY inbred
rat strains : a multifactorial and genetic analysis. Brain Res, 1996, 743:7785.
1903_C020.fm Page 304 Thursday, July 27, 2006 8:41 PM
11. CRABBE JC: Animal Models in Neurobehavioral Genetics: Methods for Estimating
Genetic Correlation. In BC JONES, P MORMDE (Eds): Neurobehavioral Genetics:
Methods and Applications. CRC Press, Boca Raton, 1999, pp 121138.
12. CRAWLEY JN: Neuropharmacologic specificity of a simple animal model for the
behavioral actions of benzodiazepines. Pharmacol Biochem Behav, 1981, 15:695699.
13. CRUSIO WE, SCHWEGLER H, VAN ABELEEN JHF: Behavioral responses to
novelty and structural variation of the hippocampus in mice. I. Quantitativegenetic
analysis of behavior in the open-field. Behav Brain Res, 1989, 32:7580.
14. CRUSIO WE, VAN ABELEEN JHF: The genetic architecture of behavioural
responses to novelty in mice: a quantitative-genetic analysis. Heredity, 1986, 56:5563.
15. CRUZ APM, FREI F, GRAEFF FG: Ethopharmacological analysis of rat behavior
on the elevated plus-maze. Pharmacol Biochem Behav, 1994, 49:171176.
16. DA SILVA GE, RAMOS A, TAKAHASHI RN: Comparison of voluntary ethanol
intake by two pairs of rat lines used as genetic models of anxiety. Braz J Med Biol
Res, 2004, 37:15111517.
17. DEFRIES JC, GERVAIS MC, THOMAS EA: Response to 30 generations of selection
for open-field activity in laboratory mice. Behav Genet, 1978, 8:313.
18. DSAUTS C, BIDANEL JP, MORMEDE P: Genetic study of behavioral and
pituitary-adrenocortical reactivity in response to an environmental challenge in pigs.
Physiol Behav, 1997, 62:337345.
19. EYSENCK HJ: Psychopathology and Personality: Extraversion, Neuroticism and
Psychoticism. In A GALE, JA EDWARDS (Eds): Individual Differences and Psy-
chopathology. Academic Press, London, 1997, pp 1330.
20. FERNNDEZ-TERUEL A, ESCORIHUELA RM, GRAY JA, AGUILAR R, GIL L,
GIMNEZ-LLORT L, TOBEA A, BHOMRA A, NICOD A, MOTT R, DRISCOLL
P, DAWSON GR, FLINT J: A quantitative trait locus influencing anxiety in the
laboratory rat. Genome Res, 2002, 12:618626.
21. FILE SE: The biological basis of anxiety. In HY MELTZER, D NEROZZI (Eds):
Current Practices and Future Developments in the Pharmacotherapy of Mental Dis-
orders. Elsevier Science Publishers B.V. Amsterdam, 1997, pp 159165.
22. FINN DA, RUTLEDGE-GORMAN MT, CRABBE JC: Genetic animal models of
anxiety. Neurogenetics, 2003, 4:109135.
23. FLINT J: Analysis of quantitative trait loci that influence animal behavior. J Neuro-
biol, 2003, 54:4677.
24. FLINT J, CORLEY R, DEFRIES JC, FULKER DW, GRAY JA, MILLER S, COL-
LINS AC: A simple genetic basis for a complex psychological trait in laboratory
mice. Science, 1995, 269:14321435.
25. FUJITA O, ANNEN Y, KITAOKA A: Tsukuba High- and Low-emotional strains of
rats (rattus norvegicus) : an overview. Behav Genet, 1994, 24:389415.
26. GLOWA JR, HANSEN CT: Differences in response to an acoustic startle stimulus
among forty-six rat strains. Behav Genet, 1994, 24:7984.
27. HAHN ME, HEWITT JK, ADAMS M, TULLY T: Genetic influences on ultrasonic
vocalizations in young mice. Behav Genet, 1987, 17:155166.
28. HALL CS: Emotional behavior in the rat I. Defecation and urination as measures of
individual differences in emotionalitly. J Comp Psychol, 1934, 18:385403.
29. HANDLEY SL, MCBLANE JW: An assessment of the elevated X-Maze for studying
anxiety and anxiety-modulating drugs. J Pharmacol Toxicol Method, 1993, 29:129138.
30. HENDLEY ED, ATWATER DG, MYERS MM, WHITEHORN D: Dissociation of
genetic hyperactivity and hypertension in SHR. Hypertension, 1983, 5:211217.
1903_C020.fm Page 305 Thursday, July 27, 2006 8:41 PM
31. HENDLEY ED, OHLSSON WG: Two new inbred rat strains derived from SHR :
WKHA, hyperactive, and WKHT, hypertensive, rats. Amer J Physiol, 1991,
261:H583H589.
32. HENDLEY ED, WESSEL DJ, VAN HOUTEN J: Inbreeding of Wistar-Kioto rat
strain with hyperactivity but without hypertension. Behav Neural Biol, 1986, 45:116.
33. LANDGRAF R, WIGGER A: High vs. low anxiety-related behavior rats: an animal
model of extremes in trait anxiety. Behav Genet, 2002, 32:301314.
34. LESCH K-P: Molecular foundation of anxiety disorders. J Neural Transm, 2001,
108:717746.
35. LIEBSCH G, MONTKOWSKI A, HOLSBOER F, LANDGRAF R: Behavioural
profiles of two Wistar rat lines selectively bred for high or low anxiety-related
behaviour. Behav Brain Res, 1998, 94:301310.
36. LISTER RG: Ethologically-based animal models of anxiety disorders. Pharmacol
Ther, 1990, 46:321340.
37. MOISAN MP, COURVOISIER H, BIHOREAU MT, GAUGUIER D, HENDLEY
ED, LATHROP M, JAMES MR, MORMEDE P: A major quantitative trait locus
influences hyperactivity in the WKHA rat. Nat Genet, 1996, 14:471473.
38. MORMEDE P, DANTZER R, BLUTHE R, CARITEZ J: Differences in adaptive
abilities of three breeds of chinese pigs. Genet Sel Evol, 1984, 16:85102.
39. MORMDE P, MONEVA E, BRUNEVAL C, CHAOULOFF F, MOISAN M-P:
Marker-assisted selection of a neuro-behavioural trait related to behavioural inhibition
in the SHR strain, an animal model of ADHD. Genes Brain Behav, 2002, 1:111116.
40. OHTA R, MATSUMOTO A, HASHIMOTO Y, NAGAO T, MIZUTANI M: Behav-
ioral characteristics of rats selectively bred for high and low avoidance shuttlebox
responses. Cong Anom, 1995, 35:223229.
41. PANDEY SC: Anxiety and alcohol abuse disorders: a common role for CREB and
its target, the neuropeptide Y gene. Trends Pharmacol Sci, 2003, 24:456460.
42. PARE WP: Hyponeophagia in Wistar Kyoto (WKY) rats. Physiol Behav, 1994,
55:975978.
43. PORSOLT RD, BERTIN A, JALFRE M: Behavioural despair in rats and mice:
strain differences and the effects of imipramine. Eur J Pharmacol, 1978, 51:291294.
44. POTENZA MN, BRODKIN ES, JOE B, LUO X, REMMERS EF, WILDER RL,
NESTLER EJ, GELERNTER J: Genomic regions controlling corticosterone levels
in rats. Biol Psychiatry, 2004, 55:634641.
45. RAMOS A, BERTON O, MORMEDE P, CHAOULOFF F: A multiple test study of
anxiety-related behaviours in six inbred rat strains. Behav Brain Res, 1997, 85:5769.
46. RAMOS A, MORMEDE P: Stress and emotionality : a multidimensional and genetic
approach. Neurosci Biobehav Rev, 22:3357, 1998.
47. RAMOS A, MELLERIN Y, MORMDE P, CHAOULOFF F: A genetic and multi-
factorial analysis of anxiety-related behaviours in Lewis and SHR intercrosses. Behav
Brain Res, 1998, 96:195205.
48. RAMOS A, MOISAN M-P, CHAOULOFF F, MORMDE C, MORMDE P: Iden-
tification of female-specific QTLs affecting an emotionality-related behavior in rats.
Mol Psychiatry, 1999, 4:453462.
49. RAMOS A, KANGERSKI AL, BASSO PF, SANTOS JES, ASSREUY J, VEN-
DRUSCOLO LF, TAKAHASHI, RN: Evaluation of Lewis and SHR rat strains as a
genetic model for the study of anxiety and pain. Behav Brain Res, 2002, 129:113123.
50. RAMOS A, CORREIA EC, IZDIO GS, BRSKE GR: Genetic selection of two new
rats lines displaying different levels of anxiety-related behaviors. Behav Genet, 2003,
33:657668.
1903_C020.fm Page 306 Thursday, July 27, 2006 8:41 PM
CONTENTS
Introduction............................................................................................................ 307
21.1 The Fruit Fly D. melanogaster as a Model Organism................................ 308
21.2 Polygenic and Single-Gene Analysis .......................................................... 308
21.2.1 Polygenic Analysis ........................................................................... 309
21.2.2 Single-Gene Analysis ....................................................................... 309
21.2.3 Polygenic vs. Single-Gene Analysis of Behavior ............................ 310
21.3 Food Search Behavior.................................................................................. 311
21.4 Role of the Foraging Gene in Other Organisms......................................... 312
Acknowledgments.................................................................................................. 314
References.............................................................................................................. 314
INTRODUCTION
Food search behavior is a complex trait influenced by many internal and external
factors. It can involve sensory modalities such as smell, taste, and vision; motor
functions such as locomotion; internal environmental cues such as the degree of
starvation or satiation; and external environmental cues such as the quality, quantity,
and distribution of the food supply. All of this information is likely integrated in
the brain and an output (e.g., feed or search more) is then generated. In the
Sokolowski laboratory, we use the fruit fly Drosophila melanogaster as our model
organism to genetically dissect the components of food search behavior. In partic-
ular, we are interested in understanding the biological basis for naturally occurring
variation in larval and adult food search behaviors as well as the biochemical
pathways that contribute to these variations. In this chapter, we discuss the fruit
fly D. melanogaster as a model organism for behaviorgenetic analysis. We com-
pare and contrast the polygenic analysis of naturally occurring behavioral variation
307
1903_C021.fm Page 308 Wednesday, July 26, 2006 7:07 PM
and the single-gene mutant analysis approaches and discuss food search behavior
in the larval and adult fruit fly D. melanogaster. We conclude by discussing how
this research can be extended to additional organisms.
normal wild-type phenotype. Mutant screens can also generate conditional muta-
tions that are dependent on factors such as temperature, light, or diet. Conditional
mutations are useful for studies of the developmental and/or acute functions of a
gene. In subsequent analyses, mutants can also be mutated (mutation can be gen-
erated in a mutant genetic background) to find new mutations that suppress or
enhance the original phenotype.35
Numerous single-gene mutations that affect behavior (e.g., learning and memory,
courtship, circadian rhythms, olfaction, phototaxis) have been identified and char-
acterized in D. melanogaster.22,32,45,48 In the past, it was unusual for single-gene
behavioral mutants to be generated on defined genetic backgrounds. Once a mutant
is generated, it is important to know (1) how the genetic background affects the
expression of the behavior, and (2) whether the mutant phenotype is selected against
in the laboratory.11 If so, careful control of genetic background, whether through
making strains homozygous or continuous backcrossing to a well-defined control
strain, should be carried out.
One area of research that has been virtually ignored is the study of whether a
gene originally identified by mutant analyses is variable in natural populations (an
exception is the work of Costa and colleagues6). This can be accomplished by
analyzing natural populations to determine if identified SNPs show associations with
natural variation in behavior.
likely to vary in nature? How, at the molecular level, does genetic background
influence the expression of the behavioral phenotype?
conservation of most of the for splice sites.15,31 We are currently investigating whether
a genetic polymorphism found in the 3UTR of the PRKG1 mRNA is associated
with obesity in a sample of morbidly obese and normal-weight humans. In the future,
it will be of interest to investigate the association of other recently identified PRKG1
SNPs in humans with alterations in food-related behaviors.
Carrying on with the hypothesis that PKGs role in foraging is conserved across
the animal kingdom, Ben-Shahar et al.4 investigated the role of for-PKG in the
honeybee, Apis mellifera. Following the cloning of Amfor, an ortholog of the Droso-
phila foraging gene, they showed that the age-related transition by honeybees from
in-hive work (nursing) to out-of-hive work (foraging) is associated with an increase
in the expression of the honeybee foraging gene, Amfor. Furthermore, they demon-
strated that this effect is not a byproduct of the nurses being 2 to 3 weeks younger
than foragers. They generated same-age nurses and foragers by colony manipulation
and found that precocious foragers still had higher expression of Amfor RNA and
PKG activity. Additionally, cGMP treatment of in-hive workers elevated PKG activ-
ity and caused precocious foraging behavior. Thus, the same gene is involved in
food-related behaviors in the honeybee and the fly; however, the regulation of the
gene differs. Drosophila for is involved in allelic variation in behavior in the fly,
whereas the honeybee Amfor gene is involved in behavioral plasticity during the
lifetime of the individual. While other genes are involved in the switch from nurse
to forager in the honeybee,52 it was sufficient to manipulate the levels of for-PKG
to cause a change in the percentage of forager honeybees.
Scheiner et al.37,38 used lessons learned from the honey bee to ask function-
related questions in the fruit fly. Interestingly, sensory responsiveness in honey bees
has been correlated with different aspects of foraging behavior and different types
of learning. Scheiner et al.39 hypothesized that the fly for gene might also influence
sensory responsiveness and nonassociative learning in Drosophila. They used tech-
niques originally developed for honeybee studies to investigate the role of for-PKG
in adult fruit fly sensory responsiveness. A fruit fly extends its proboscis when its
foreleg is stimulated with sucrose. This is called a proboscis extension response
(PER). Scheiner and colleagues measured sucrose responsiveness of food, but not
water-deprived flies. They elicited PER following stimulation of sucrose receptors
found on the terminal segment of the foreleg known as the tarsus. Habituation of
the PER occurred when, after repeated stimulations, the flies no longer gave a PER
in response to sucrose stimulation. The results showed that rovers are more respon-
sive to sucrose than sitters and sitter mutants, and that this was true for flies exposed
to short or long periods of food deprivation. rovers also showed less habituation
than sitters and sitter mutants; rovers and sitters with similar sucrose responsiveness
were used for these habituation studies. Thus for-PKG affects sensory responsiveness
to sucrose and habituation in D. melanogaster. Like rover flies, forager bees show
higher levels of sensory responsiveness than do nurse bees and sitter flies.
Genetic manipulation of PKG in the nematode C. elegans, reveals two modes
of locomotion behavior in the presence of food: roaming and dwelling.17 Roamers
have high-speed locomotion with infrequent turns, while dwellers show low-speed
locomotion with frequent turns. The worm ortholog of the foraging gene, egl-4,
affects this feeding behavior. Interestingly, contrary to the case in Drosophila and
1903_C021.fm Page 314 Wednesday, July 26, 2006 7:07 PM
the honey bee, lower levels of PKG in C. elegans is associated with increased
roaming behavior. It is of interest to know whether PKG plays a role in feeding
behavior of other organisms. Recently, Fitzpatrick and Sokolowski15 constructed
protein phylogenies using 32 PKG sequences that include 19 species and proposed
five different evolutionary histories that can explain the link between PKG and
feeding behaviors in fruit flies, honeybees and worms. Studies such as these start
to bridge the gap between genetic model organisms and organisms not traditionally
used for genetic analyses.16
Sokolowski and collaborators have implicated a cGMP signal transduction path-
way in the regulation of food search behavior in flies and honeybees. The next step
is to identify additional components of this food search pathway. This can be
accomplished by using: (1) the single-gene mutant approach to generate mutations
in genes that interact with for, and (2) the QTL analysis approach to localize naturally
occurring variants that act as minor modifiers of foraging behavior in Drosophila.
We are currently mapping genes identified from these types of studies.
The rover/sitter model system is being developed to understand both the mech-
anistic and evolutionary significance of behavioral variation. With this in mind, we
are on the road to developing an understanding of how food search behavior is
affected at the level of the gene, molecule, nervous system, organism, population,
and speciesin both the laboratory and in nature.
ACKNOWLEDGMENTS
This chapter is dedicated to the memory of Dr. Lionel Peypelut, a dear friend and
colleague. The research described here from the Sokolowski laboratory was funded
by NSERC, CIHR and the Canada Research Chairs Program to M.B.S.
REFERENCES
1. Anholt, R.R. and Mackay, T.F.C., Quantitative genetic analysis of complex behaviors
in Drosophila, Nat. Rev. Genet., 5:838849, 2004.
2. Asburner, M., Hawley, S., and Golic, K., Drosophila: A Laboratory Handbook, Cold
Spring Harbor Laboratory, Cold Spring Harbour, New York, 2004.
3. Bell, W.J., Searching Behavior: The Behavioral Ecology of Finding Resources, Chap-
man & Hall, New York, 1991.
4. Ben-Shahar, Y., Ribichon, A., Sokolowski, M.B., and Robinson, G.E., Influence of
gene action across different time scales on behavior, Science, 296:741744, 2002.
5. Carlson, J., Olfaction in Drosophila: from odor to behavior, Trends. Genet.,
12:175180, 1996.
6. Costa, R., Peixoto, A.A., Barbujani, G., and Kyriacou, C.P., A latitudinal cline in
Drosophila clock gene, Proc. R. Soc. Biol. Sci. 250:4349, 1992.
7. de Belle, J.S. and Sokolowski, M.B., Heredity of rover/sitter: alternative foraging
strategies of Drosophila melanogaster, Heredity, 59:7383, 1987.
8. de Belle, J.S. and Sokolowski, M.B., Heredity of rover/sitter foraging behavior in
Drosophila melanogaster, genetic localization to chromosome-2L using compound
autosomes, J. Insect Behav., 2:291299, 1989.
1903_C021.fm Page 315 Wednesday, July 26, 2006 7:07 PM
9. de Belle, J.S., Hilliker, A.J., and Sokolowski, M.B., Genetic localization of foraging
(for): a major gene for larval behavior in Drosophila melanogaster, Genetics,
123:157164, 1989.
10. de Belle, J.S., Sokolowski, M.B., and Hilliker, A.J., Genetic analysis of the foraging
microregion of Drosophila melanogaster, Genome, 36:94101, 1993.
11. de Belle, J.S. and Heisenberg, M., Expression of Drosophila mushroom body muta-
tions in alternative genetic backgrounds: a case study of the mushroom body miniature
gene (mbm), Proc. Natl. Acad. Sci. U.S.A., 93(18):98759880, 1996.
12. De Luca, M., Roshina, N.V., Geiger-Thornsberry, G.L., Lyman, R.F., Pasyukova,
E.G., and Mackay, T.F.C., Dopa-decarboxylase affects variation in Drosophila lon-
gevity, Nat. Genet. 34:429433, 2003.
13. Falconer, D.S. and MacKay, T.F.C., Introduction to Quantitative Genetics, 4th ed.,
Longman, New York, 1996.
14. Fanara, J.J., Robinson, K.O., Rollmann, S.M., Anholt, R.R., and Mackay, T.F.C,
Vanaso is a candidate quantitative trait gene for Drosophila olfactory behavior, Genet-
ics, 162(3):13211328, 2002.
15. Fitzpatrick, M.J. and Sokolowski, M.B., In search of food: exploring the evolutionary
link between cGMP-dependent protein kinase (PKG) and behavior, Integr. Comp.
Biol., 44:2836, 2004.
16. Fitzpatrick, M.J., Ben-Shahar, Y., Smid, H.M., Vet, L.E.M., Robinson, G.E., and
Sokolowski, M.B., Candidate genes for behavioral ecology, Trends Ecol Evol,
20:96104, 2005.
17. Fujiwara, M., Sengupta, P., and McIntire, S.L., Regulation of body size and behavioral
state of C. elegans by sensory perception and the EGL-4 cGMP-dependent protein
kinase, Neuron, 36:10911102, 2002.
18. Gleason, J.M., Nuzhdin, S.V., and Ritchie, M.G., Quantitative trait loci affecting a
courtship signal in Drosophila melanogaster, Heredity, 89:16, 2002.
19. Graf, S.A. and Sokolowski, M.B., The rover/sitter Drosophila foraging polymorphism
as a function of larval development, food patch quality and starvation, J. Insect Behav.,
2:301313, 1989.
20. Greenspan, R.J., A kinder, gentler genetic analysis of behavior: dissection gives way
to modulation, Curr. Opin. Neurobiol, 7:805811, 1997.
21. Greenspan, R.J., Fly Pushing, The Theory and Practice of Drosophila Genetics, Cold
Spring Harbor Laboratory Press, Cold Spring Harbour, New York, 2004.
22. Greenspan, R.J., E Pluribus Unum, Ex Uno Plura: Quantitative- and single-gene
perspectives on the study of behavior, Annu. Rev. Neurosci., 27:79105, 2004.
23. Hall, J.C., The mating of a fly, Science, 264: 17021714, 1994.
24. Harmer, S.L., Panda, S., and Kay, S.A., Molecular bases of circadian rhythms, Annu.
Rev. Cell Dev. Biol., 17:215253, 2001.
25. Kalderon, D and Rubin, G.M., cGMP-dependent protein kinase genes in Drosophila,
J. Biol. Chem., 264: 1073810748, 1989.
26. Mackay, T.F.C., Quantitative trait loci in Drosophila, Nat. Rev. Genet. 2:1120, 2001.
27. Mather, K. and Jinks, J.L., Biometrical Genetics, 3rd ed., Chapman & Hall, New York,
1982.
28. Nuzhdin, S.V., Pasyukova, E.G., Dilda, C.L., Zeng, Z.-B., and Mackay, T.F.C., Sex-
specific quantitative trait loci affecting longevity in Drosophila melanogaster, Proc.
Natl. Acad. Sci. U.S.A., 94:97349739, 1997.
29. Nuzhdin, S.V., Dilda, C.L. and Mackay, T.F.C., The genetic architecture of selection
response: inference from fine-scale mapping of bristle number quantitative trait loci
in Drosophila melanogaster, Genetics, 153:13171331, 1999.
1903_C021.fm Page 316 Wednesday, July 26, 2006 7:07 PM
30. Osborne, K., Robichon, A., Burgess, E., Butland, S., Shaw, R.A., Coulthard, A.,
Pereira, H.S., Greenspan, R.J., and Sokolowski, M.B., Natural behavior polymor-
phism due to a cGMP-dependent protein kinase of Drosophila, Science, 277:834836,
1997.
31. Ostravik, S., Natarajan, V., Tasken, K., Jahnsen, T., and Sandberg, M., Characteriza-
tion of the human gene encoding the type I alpha and type I beta cGMP-dependent
protein kinase (PRKGI) 1997, Genomics, 42:311318, 1997.
32. Pak, W.L. and Leung, H.T., Genetic approaches to visual transduction in Drosophila
melanogaster, Receptors Channels, 9(3):149167, 2003.
33. Pasyukova, E.G., Roshina, N.V., and Mackay, T.F.C., Shuttle craft: a candidate quan-
titative trait gene for Drosophila melanogaster lifespan, Aging Cell, 3:297307, 2004.
34. Pereira, H.S. and Sokolowski, M.B., Mutations in the larval foraging gene affect adult
locomotory behavior after feeding in Drosophila melanogaster, Proc, Natl. Acad. Sci.
U.S.A., 90:50445046, 1993.
35. Pereira, H.S., MacDonald, D.E., Hilliker, A.J., and Sokolowski, M.B., Chaser (Csr),
a new gene affecting larval foraging behavior in Drosophila melanogaster, Genetics,
140:263270, 1995.
36. Sawin, E.P., Harris, L.R., Campos, A.R., and Sokolowski, M.B., Sensorimotor trans-
formation from light reception to phototactic behavior in Drosophila larvae, J. Insect
Behav., 7:553567, 1994.
37. Scheiner, R. Erber, J., and Page, R.E., Tactile learning and the individual evaluation
of the reward in honey bees (Apis mellifera L.), J. Comp. Physiol., 185:110, 1999.
38. Scheiner, R. Erber, J., and Page, R.E., The effects of genotype, foraging role and
sucrose responsiveness on the tactile learning performance of honey bees (Apis
mellifera L.), Neurobiol. Learn. Mem., 76:138150, 2001.
39. Scheiner, R., Sokolowski, M.B. and Erber, J., Activity of cGMP-dependent protein
kinase (PKG) affects sucrose responsiveness and habituation in Drosophila melano-
gaster, Learn. Mem., 11:303311, 2004.
40. Shaver, S.A., Varnam, C., Hilliker, A.J., and Sokolowski, M.B., The foraging gene
affects adult but not larval olfactory related behavior in Drosophila, Behav. Brain
Res., 95:2329, 1998.
41. Sokolowski, M.B., Foraging strategies of Drosophila melanogaster: a chromosomal
analysis, Behav. Genet., 10:291302, 1980.
42. Sokolowski, M.B., Genetic analysis of behavior in the fruit fly, Drosophila melan-
gaster. In Techniques for the Genetic Analysis of Brain and Behavior, D. Goldwitz,
D. Wahlsten, and R.E. Winer, Eds., Elsevier Science Publishers, 1992.
43. Sokolowski, M.B. and Hansell, K.P., The foraging locus: behavioral tests for normal
muscle movement in rover and sitter Drosophila melanogaster larvae, Genetica,
85:205209, 1992.
44. Sokolowski, M.B., Pereira, H.S., and Hughes, K., Evolution of foraging behavior in
Drosophila by density dependent selection, Proc. Natl. Adac. Sci. U.S.A.,
94:73737377, 1997.
45. Sokolowski, M.B., Drosophila: genetics meets behavior, Nat. Rev. Genet., 2:879890,
2001.
46. Sokolowski, M.B. and Wahlsten, D., Gene-Environment interaction and complex
behavior. In Methods in Genomic Neuroscience, H. Chin, and S. Moldin, Eds., CRC
Press, New York, 2002.
47. Stocker, R.F., The organization of the chemosensory system in Drosophila melano-
gaster: a review, Cell Tissue Res., 273:326, 1994.
1903_C021.fm Page 317 Wednesday, July 26, 2006 7:07 PM
CONTENTS
22.1 Introduction ................................................................................................... 319
22.2 Mating Behavior of Drosophila melanogaster,
Visual and Acoustic Stimuli ........................................................................ 320
22.3 Pheromones and Biosynthetic Genes............................................................ 321
22.4 Gustatory and Olfactory Receptors............................................................... 324
22.5 Sex Determination Cascade and the Courtship
Behavior Master Gene Fruitless .................................................................. 326
References.............................................................................................................. 328
22.1 INTRODUCTION
Drosophila melanogaster is a marvelous material for behavioural genetic anal-
ysis. First, it has been shown to be highly amenable for genetic analysis. It has
a reduced developmental time (about 9 days at 25C) and a compact genome
(four pairs of chromosomes including three pairs of autosomes and one pair of
sex chromosomes). The classic genetic tools such as visible markers and balancer
chromosomes can be efficiently used together with new technology to elucidate
molecular mechanisms. The complete genome sequence has been known for 5
years1 and can be compared with those of other model organisms that appear in
the literature. Large parts of those of other Drosophila species are also available2
at http://flybase.bio.indiana.edu.
As far as behavior is concerned, Drosophila behaves in two states, which follow
each other in postembryonic development, as a crawling larva first and then as a
flying insect. The imaginal behaviors are quite sophisticated, involving instinctive
and plastic facets. Mating behavior is the most complex.
319
1903_C022.fm Page 320 Wednesday, July 26, 2006 7:08 PM
in the motor system.14 Other P insertions created mutations affecting the sine song
frequency including Beethoven in a chromosome II gene encoding a protein with
ATPase activity. Several sex determination genes, when mutated, also showed song
abnormalities. Altogether, 17 single genes were involved.15 Another approach with
quantitative genetics took advantage of the natural variation within and between
species.16,17 Several studies of D. melanogaster interstrain variations showed additive
effects of chromosomes II and III to control IPI, with a larger effect of the third
chromosome. But only one candidate gene, tipE, another channel gene on 3L, was
clearly included in a quantitative trait locus (QTL)actually, in the major QTL of
chromosome III. A recent QTL study of IPI differences in backcross hybrids between
D. simulans and D. sechellia, another species of the melanogaster subgroup, has
shown similar contributions of autosomal chromosomes to IPI and localized four
QTL on 2R and two on 3R. In this study, three candidate genes, maleless, croaker,
and fruitless are related to QTL.15
The courtship behavior may be subject to associative modification. Immature
virgin females and recently fertilized females stimulate males as much as mature
virgin females, but both reject courters. While virgins usually escape from the
courting male by decamping, the copulated females extrude their ovipositor which
blocks genital contact. The rejection behavior in fertilized females is triggered by a
peptide which is produced in the male accessory gland and transferred to females
during copulation.18 Males that have previously courted fertilized females show a
reduced level of courtship toward a virgin female.19
with some sex differences.49,50 Its heterologous expression in yeast could complement
the mutant ole1, deficient in desaturase activity, and established the D9 specificity
of the Drosophila enzyme as well as its preference for palmitic acid as substrate.51
Insertions of various P transposons in the promoter region of desat1 showed more
or less decrease in the level of all 7 unsaturated HC in both sexes including 7T and
7,11HD. As their excision often reverted the wild-type HC phenotype, the desat1
enzyme is clearly involved in a common step of biosynthesis, the desaturation which
transforms palmitic acid into palmitoleic acid (C16:1,D9/w7).25
Although the 7,11 HD/5,9 HD polymorphism was first hypothetized to be linked
to an allelic variation in the desat1 gene with each allele producing a different
desaturase isoform, the reality was more complex. Although the Tai desat1 gene
coded a polypeptide chain with 3 amino acid substitutions compared to that of
CantonS, the yeast transformation results with either coding sequence were similar,
infirming the simple hypothesis. Later a mini-dissection of the locus around the
desat1 gene revealed the presence of two more desaturase genes, desat2 and desat3.
Desat2 was characterized in the Tai strain. Its uncorrupted open reading frames
(ORF) encodes a polypeptide with 65% identity with that of desat1 but a smaller
amino domain. The heterologous expression of desat2 in yeast also complemented
the deficient mutant ole1, suggesting again a D9 specificity, but a different substrate;
myristoleic acid (C14:1,D9/w5) was mainly produced instead of palmitoleic acid.51
This position w5 for HC is abundant mainly in the African strains and there only
in females. Actually desat2 can be expressed only in African females that make
5,9HD. These facts support a critical role for both desat1 and desat2 in the first
desaturation of linear saturated fatty acids. Another study has shown that the knock-
down of the desat2 gene in ancestral populations during their migrations in Africa
was linked to the deletion of 16 bp in the promoter region.52
To finally make dienes, monoenic fatty acids have to go through a second
desaturation, taking place in females of the D. melanogaster species and probably
not in D. simulans. As the enzyme specificity is usually restricted to a certain type
and position of bond along the chain and a substrate either saturated or monoenic,
other desaturase genes were searched for. Indeed the published sequence of
D. melanogaster genome suggested the existence of more desaturase genes, one on
chromosome II supposed to be more specific for males and four more on chromo-
some III.1,49 Among these latter genes, one in the cytological region 68 had been
shown to be possibly important to make female dienes.53 Chertemps et al.54 then
completely characterized a new desaturase gene, called desatF, which is expressed
in D. melanogaster females but not in conspecific males. Using the interference
RNA technique, they were able to knock down this gene almost completely. A
dramatic result was observed in the cuticle of transformed females: the level of 7,11
HD was decreased by more than 80% while that of all 7-monoenes was much
increased. This clearly shows that desatF adds a second double bond to monoenic
fatty acids to make dienic ones. The description of another group of biosynthetic
genes has been initiated, those coding for fatty acid elongases.55
Another approach to try to localize genes controlling HC production was quan-
titative genetics QTL analyses of HC were performed in flies of recombinant lines
produced between D. simulans and D. sechellia, another member of the melanogaster
1903_C022.fm Page 324 Wednesday, July 26, 2006 7:08 PM
in the very neurons where their expression is restricted and the inactivation of the
same neurons by tetanus toxin have both led to a marked decrease of male courting
activity; moreover, both manipulations reduced occurrences of specific courtship
steps such as wing vibrations and attempted copulations. This strongly suggested
that Gr68a was indeed a good candidate receptor to attach a female contact phero-
mone such as 7,11 HD; direct binding experiments still have to be done. The fact
that these transformed males are still able to perform part of the courtship suggests
the involvement of other types of Gr, possibly involved in binding of other contact
pheromones.66 Other elements of the pheromone transduction cascade have also been
described, two chemosensory proteins CheA29a and CheB42a which share some
characteristics with oderant pheromone binding protein (OBP/PBP) and are expressed
only in the front legs of males as Gr68a67 and one ion channel subunit, ppk25. This
gene is expressed not only in legs, but also in antennae and several of its mutations
strongly affect the courtship behaviour of males.68
In the olfactory system, it has been clearly shown that most of the 62 Or genes
were expressed alone in olfactory sensory neurons, while the situation looked more
complex for Gr in gustatory neurons.6971 Another difference is the projection
pattern of the sensory neurons. While most olfactory sensory neurons expressing
a given Or converge to a single glomerulus, one functional subunit of the brain
antennal lobes, a similar structure to the vertebrate olfactory bulb, gustatory sensory
neuron primary projections are not in the brain but in different parts of the nervous
system and thus much less well characterized. For example all gustatory neurons
of both forelegs converge to the anterior part of the thoracic ganglion. In a pioneer
study, Possidente and Murphey72 have described the sexual dimorphism of these
projections.
Only 2 out of about 50 glomeruli7374 have been found to display a dimorphism
between males and females VA1v and mainly DA1, which is 62% larger in males
than in females.75 These glomeruli are anatomically homologous to the more dra-
matically dimorphic glomeruli of moths. In males of these insects there is a macro-
glomerulus whose function is to integrate the complexity of the female pheromone
bouquet. Interestingly these two large structures are targets for the fruitless tran-
scription factor which also target only one or two other glomeruli.69 The olfactory
sensory neurons which converge toward the Drosophila dimorphic glomeruli start
in trichoid sensillae like those sensitive to pheromones in moths;2 they express
olfactory receptors which have been identified as respectively Or47b and Or67d.
Could cVA not be a possible ligand for one of these olfactory receptors? cVA
has been shown to be detected by a subset of trichoid sensillae of the antennal third
segment58,61 and one piece of its transduction machinery has been identified. Flies
lacking the gene lush, which codes for its OBP, display much less aggregation
ability.67 This gene has been originally identified because some of its mutations lead
to specific attraction toward high concentrations of ethanol and other short-chain
alcohols which are avoided by wild-type flies. Actually, lush is expressed in about
150 trichoid sensillae on the ventral-lateral surface, in both males and females, whose
sensory neurons project into the VA1 and DA1 glomeruli.76 It is possible that another
ligand might be the volatile female pheromone hypothetized by Shorey and Bartell.20
1903_C022.fm Page 326 Wednesday, July 26, 2006 7:08 PM
Flies with a female caryotype but which are mutant for tra develop and behave
as males. In such flies, a female phenotype could be restored by a transgene that
carried the female-specific cDNA of tra under the control of a heat-shock promoter.
This transgene could also transform caryotypic male animals into sterile females.
Raised at 25C, both such transformed XX and XY displayed typical male courtship
while at the same time, rich in dienes, they were attractive to males. When the
expression of tra was forced by heat shock, applied during a limited period around
puparium formation, male behavior was abolished and replaced by female behavior.
Thus sexual behavior is irreversibly programmed during a critical period as a result
of the activity switch of tra.85
Another strategy used a collection of mutations which affected all male courtship
behavior and mapped in a small region of the third chromosome, at the fruitless(fru)
locus. The first fru mutant (fru 1) was induced by x-rays and described behaviorally
first as a male-sterile mutant, then as a male courter of both sexes.86 Actually, it
corresponded to a small deletion together with a chromosomal inversion in the
interval 90C91B.87 Later four additional alleles (fru 2, fru 3, fru 4 and fru-satori)
have been obtained by P element insertion mutagenesis and showed different behav-
ioral characteristics. For example fru-satori males do not court females but can court
males; fru2-4 males court both males and females; fru 4 males do not perform love
song whichever the courted sex.4,8,88 Although a recognizable insertion like P made
this gene cloning possible, molecular studies were laborious as fru had several
transcripts and several promoters were identified. Homologies with members of the
BTB family of transcriptional regulators that also contains zinc-finger motifs were
found. In the regulatory region, the presence of three putative binding sites for the
transformer gene product, similar to those found in doublesex clearly showed that
fru was a neural sex-determination switch downstream of transformer, occupying
in the sex determination cascade a position parallel to doublesex.89,90
While transcripts from three promoters are common for both sexes, transcripts
from promoter P1 are controlled by the sex determination cascade and produce in
males a specific form fruM that is longer. A series of elegant experiments have
allowed Demir and Dickson91 to demonstrate that male splicing specifies the main
part of male courtship behavior. They also showed that this mode of splicing is
sufficient to have an otherwise normal female court as a male.91 But how is the gene
organizing the male neurocircuit which performs the male courtship program?
A transgene, in which the GAL4 coding sequence had been introduced into the
fruM coding sequence, was constructed by two groups and used to direct a detailed
search of fru target cells.92,93 Most of the peripheral sensory systems were concerned,
that is about 20 to 30 gustatory neurons of both the prothoracic legs and the proboscis
but also more than 100 to 200 neurons of the olfactory and acoustic parts of the
antennae. The olfactory neurons projected primarily to 34 glomeruli of the antennal
lobes as reported earlier.
In the central part of the nervous system, fruM is expressed in a few scattered
groups of cells. One group is a bilateral cluster of about 60 neurons in the suboe-
sophageal ganglion where fruM expression could be blocked with the RNA inter-
ference technique.94 Surprisingly, such transformed males started to court wild-type
females ten times more quickly than wild types, apparently bypassing the sensory
1903_C022.fm Page 328 Wednesday, July 26, 2006 7:08 PM
steps of early courtship such as orientation and tapping usually occurring in wild
type males; but they quickly displayed all later steps, often all at the same time,94
suggesting that these suboesophageal neurons might inhibit the courtship start until
sufficient and adequate sensory stimuli allow it.
The question of sexual dimorphism in the number of and positions of brain cell
clusters involved in courtship behavior has also been investigated. Until recently,
clear differences had been found only in the abdominal ganglion.72,95 In 2005,
Stockinger et al.,93 using a fluorescent reporter gene directed by a specific P [GAL4]
element inserted in the fru gene, have only observed larger clusters in the male
superior protocerebrum. But, with a slightly different technique, Kimura et al.96
could identify two more groups of sexually dimorphic neuron clusters: in the medulla
of the optic lobes there were a few more interneurons only in males; marked
differences in a brain region dorsal to the antennal lobes were also characterized.
There, on each side, a cluster of about 30 neurons was characterized in males while
there were not more than 5 in females. Moreover, these neurons also displayed clear
sex differences in their projection patterns. These interneurons seem to start in the
suboesophageal ganglion where gustatory informationspossibly contact phero-
mone inputscollected by tarsal and labellar gustatory neurons might be integrated
and finish in the superior lateral protocerebrum.
Transformer and Fruitless are thus remarkable examples of switch genes able
to transform a complex behavioral program such as courtship. The behavioral trans-
formation is linked to structural modifications in the nervous system. In both cases,
they are not fully understood but seem limited to a few clusters of neurons. The
action of such master genes can be studied in more depth thanks to all the sophis-
ticated tools that Drosophila provides for cellular and molecular analyses.
REFERENCES
1. Myers, E.W., Sutton, G.G., Delcher, A.L. et al., A whole genome assembly of
Drosophila, Science, 287, 2196, 2000.
2. Ashburner, M., Golic, K. and Hawley, S., Drosophila, A Laboratory Handbook, 2nd
ed., Cold Spring Harbor Laboratory Press, Cold Spring Harbor, 2005.
3. Bastock, M. and Manning, A., The courtship of Drosophila melanogaster, Behaviour,
8, 85,1955.
4. Hall, J.C., The mating of a fly, Science, 264, 1702, 1994.
5. Markow, T.A., Behavioral and sensory basis of courtship success in Drosophila
melanogaster, Proc. Natl. Acad. Sci. USA, 84, 6200, 1987.
6. Jallon, J.M. and Hotta, Y., Nonchemical messages of the female Drosophila melan-
ogaster, in Taniguchi Symposia in Biophysics No7, Hotta, Y., Tokyo University Press,
Tokyo, 1983, 136.
7. Oguma, Y. et al., Courtship behavior and sexual isolation between Drosophila auraria
and D. triauraria in darkness and in light, J. Evol. Biol., 9, 803, 1996.
8. Yamamoto, D., Jallon, J.M. and Komatsu, A., Genetic dissection of sexual behavior
in Drosophila melanogaster, Annu. Rev. Entomol., 42, 551, 1997.
9. Venard, R. and Jallon, J.M., Evidence for an aphrodisiac pheromone of female
Drosophila, Experientia, 36, 211, 1980.
1903_C022.fm Page 329 Wednesday, July 26, 2006 7:08 PM
10. Ewing, A., Functional aspects of Drosophila courtship, Biol. Rev., 58, 275, 1983.
11. Rybak, F., et al., Acoustic communication in Drosophila melanogaster courtship: are
pulse- and sine-frequencies important for courtship success? Can. J. Zool., 80, 987,
2002.
12. Schilcher, F.V., A mutation which changes courtship song in Drosophila melano-
gaster, Behav. Genet., 7, 251, 1977.
13. Kulkarni, S.J., Steinlanf, A.F. and Hall, J.C., The dissonance mutant of courtship
song in Drosophila melanogaster, Genetics, 118, 267, 1988.
14. Yokokura, T., Ueda, R. and Yamamoto, D., Phenotypic and molecular characterization
of croaker, a new mating behavior mutant of Drosophila melanogaster, Japan J.
Genet., 70, 103, 1995.
15. Gleason, J. and Ritchie, M., Do quantitative trait loci for a courtship song difference
between Drosophila simulans and D. sechellia overlap with candidate genes and
intraspecific QTLs?, Genetics, 166, 1303, 2004.
16. Mackay, T.F. et al., Genetics and genomics of Drosophila mating behavior, Proc.
Natl. Acad. Sci. USA, 102, 6622, 2005.
17. Gleason, J. et al., A quantitative trait loci for cuticular hydrocarbons associated with
sexual isolation between Drosophila simulans and D. sechellia, Genetics, 171, 1, 2005.
18. Chen, P.S., Stumm-Zollinger, E., Aigaki, T., Balmer, J., Bienz, M. and Bohlen, P., A
male accessory gland peptide that regulates reproductive behavior of female Droso-
phila melanogaster, Cell, 54, 291, 1988.
19. Gailey, D.A., Jackson, F.R. and Siegel, R.W., Conditioning mutations in Drosophila
melanogaster affect an experience-dependent behavioral modification in courting
males, Behav. Genet., 106, 613, 1984.
20. Shorey, H.H. and Bartell, R., Role of volatile sex pheromone in stimulating male
courtship behavior in Drosophila melanogaster, Anim. Behav., 18, 159, 1970.
21. Carlson, D.A. et al., Sex attractant pheromones of the house fly: isolation, identifi-
cation and synthesis, Science, 174, 76, 1971.
22. Venard, R., Attractants in the courtship behavior of Drosophila melanogaster, in
Development and Neurobiology of Drosophila, Siddiqi, O., Babu, P., Hall, L.M. and
Hall, J.C., Plenum Press, New York, 1980, 457.
23. Antony, C. and Jallon, J.M., The chemical basis for sex recognition in Drosophila
melanogaster, J. Insect. Physiol., 28, 873, 1982.
24. Arienti, M., Analyse de la variabilit de quelques mcanismes impliqus dans le
comportement sexuel de populations differentes de Drosophila melanogaster, Ph.D.
thesis, Universit de Paris XI, Orsay, 1993.
25. Ueyama, M. et al., Mutations in the desat1 gene reduces the production of courtship-
stimulatory pheromones through a marked effect on fatty acids in Drosophila mela-
nogaster, Insect Biochem. Mol. Biol., 35, 911, 2005.
26. Pechine, J.M., Antony, C. and Jallon, J.M., Precise characterization of cuticular
compounds in young Drosophila, J. Chem. Ecol., 14, 1071, 1988.
27. Antony, C. et al., Compared behavioural responses of male Drosophila melanogaster
(Canton S) to natural and synthetic aphrodisiacs, J. Chem. Ecol., 11, 1617, 1985.
28. Klun, J.A. and Maini, S., Genetics basis of an insect communication system in the
European corn borer, Environ. Entomol., 11, 1084, 1982.
29. Jallon, J.M. and Pechine, J.M., Une autre race chimique de Drosophila melanogaster
en Afrique, CR. Acad. Sci., Paris, 309, 1551, 1989.
30. Savarit, F. et al., Genetic elimination of known pheromones reveals the fundamental
chemical basis of mating and isolation in Drosophila, Proc. Natl. Acad. Sci. USA,
96, 9015, 1999.
1903_C022.fm Page 330 Wednesday, July 26, 2006 7:08 PM
31. Butterworth, F.M., Lipids in Drosophila, a newly detected lipid in the male, Science,
163, 1256, 1969.
32. Bartelt, R.J. et al., Cis-vaccenyl acetate acts as an aggregation pheromone in Droso-
phila melanogaster, J. Chem. Ecol., 11, 1747,1986.
33. Jallon, J.M., Antony, C. and Benamar, O., Un anti-aphrodisiaque produit par les males
de Drosophila melanogaster, C. R. Acad. Sci., Paris, 292, 1147, 1981.
34. Jallon, J.M., A few chemical words exchanged by Drosophila during courtship, Behav.
Genet., 14, 441, 1984.
35. Scott, D. and Jackson, L.L., Interstrain comparison of male predominant aphrodisiacs
in Drosophila melanogaster, J. Insect Physiol., 34, 863, 1988.
36. Scott, D., Genetic variation for female mate discrimination in Drosophila melano-
gaster, Evolution, 48, 112, 1994.
37. Ferveur, J.F. and Sureau, G., Simultaneous influence on male courtship of stimulatory
and inhibitory pheromones produced by live sex-mosaics Drosophila melanogaster,
Proc. R. Soc. London B, 263, 967, 1996.
38. Grillet, M., Dartevelle, L. and Ferveur, J.F., A Drosophila male pheromone affects
female sexual receptivity, Proc. R. Soc. London B, 273, 315323, 2006.
39. Cobb, M. and Jallon, J.M., Pheromones, mate recognition and courtship stimulation
in the Drosophila melanogaster species sub-group, Anim. Behav., 39, 1058, 1990.
40. Coyne, J.A., Crittenden, A.P. and Mah, K., Genetics of a pheromonal difference
contributing to reproductive isolation in Drosophila, Science, 265, 1461, 1994.
41. Rouault, J., Capy, P. and Jallon, J.M., Variations of male cuticular hydrocarbons with
geoclimate variables: an adaptative mechanism in Drosophila melanogaster?, Genet-
ica, 110, 117, 2001.
42. Haerty, W. et al., Reproductive isolation in natural populations of Drosophila mela-
nogaster, Genetica, 116, 215, 2002.
43. Sureau, G. and Ferveur, J.F., Co-adaptation of pheromone production and behavioural
responses in Drosophila melanogaster males, Genet. Res., 74, 129, 1999.
44. Marcillac, F., Grosjean, Y. and Ferveur, J.F., A single mutation alters production and
discrimination of Drosophila pheromones, Proc. R. Soc. London B, 272, 303, 2004.
45. Ferveur, J.F. et al., World-wide variation in Drosophila melanogaster sex pheromones:
behavioural effects, genetic bases and potentially evolutionary consequences, Genet-
ica, 97, 73, 1996.
46. Coyne, J.A.,Wicker-Thomas,C. and Jallon, J.M., A gene responsible for a cuticular
hydrocarbon polymorphism in Drosophila melanogaster, Genet. Res., 73, 189, 1999.
47. Wicker-Thomas, C., Henriet, C. and Dallerac, R., Partial characterization of a fatty
acid desaturase gene in Drosophila melanogaster, Insect Biochem. Mol. Biol., 27,
963, 1997.
48. Pennanech, M. et al., Incorporation of fatty acids into cuticular hydrocarbons of male
and female Drosophila melanogaster, J. Insect Physiol., 43, 1111, 1997.
49. Jallon, J.M. and Wicker-Thomas, C., Genetic studies on pheromone production in
Drosophila, in Insect Pheromone Biochemistry And Molecular Biology, Blomquist,
G.J. and Vogt, R.G, Elsevier, Amsterdam, 2003, 253.
50. Marcillac, F. et al., A mutation with major effects on Drosophila melanogaster sex
pheromones, Genetics, 171, 1617, 2005.
51. Dallerac, R.C. et al., Delta 9 desaturase gene with a different substrate specificity is
responsible for the cuticular diene hydrocarbon polymorphism in Drosophila mela-
nogaster, Proc. Natl. Acad. Sci. USA, 97, 9449, 2000.
1903_C022.fm Page 331 Wednesday, July 26, 2006 7:08 PM
52. Takahashi, A.S. et al., The nucleotide changes governing cuticular hydrocarbon vari-
ation and their evolution in Drosophila melanogaster, Proc. Natl. Acad. Sci. USA,
98, 3920, 2001.
53. Wicker-Thomas, C. and Jallon, J.M., Role of Enhancer of zeste on the production of
Drosophila melanogaster pheromonal hydrocarbons, Naturwissenschaften, 87, 76,
2000.
54. Chertemps, T. et al., Female-specific desaturase gene responsible for diene hydrocar-
bon biosynthesis and courtship behaviour in D. melanogaster, Insect Molec. Biol., in
press, 2006.
55. Chertemps, T. et al., A new elongase selectively expressed in Drosophila male
reproductive system, Biochem. Biophys. Res. Commun., 333, 1066, 2005.
56. Begg, M. and Hogben, L., Chemoreceptivity of Drosophila melanogaster, Proc. R.
Soc. London B, 133, 1, 1946.
57. Stocker, R.F. and Gendre, N., Courtship behavior of Drosophila, genetically and
surgically deprived of basichonic sensilla, Behav. Genet., 19, 371, 1988.
58. Venard, R., Antony, C. and Jallon, J.M., Drosophila Chemoreceptors, in Neurobiology
Of Sensory Systems, Singh, N. and Strausfeld, N., Eds., Plenum Press, New York,
1989, 377.
59. Nayak, S.V. and Singh, R.N., Sensilla on the tarsal segments and the mouthparts of
adult Drosophila melanogaster, Int. J. Insect Morph. Embryol., 12, 273, 1983.
60. Buck, L. and Axel, R., A novel multigene family may encode odorant receptors: a
molecular basis for odor recognition, Cell, 65, 175, 1991.
61. Clyne, P.J., Warr, C.G. and Carlson, J.R., Candidate taste receptors in Drosophila,
Science, 287, 1830, 2000.
62. Scott, K. et al., A chemosensory gene family encoding candidate gustatory and
olfactory receptors in Drosophila, Cell, 104, 661, 2001.
63. Dunipace, L., Meister, S., McNealey, C. and Amrein, H., Spatially restricted taste
receptors in the Drosophila gustatory system, Curr. Biol., 11, 822, 2001.
64. Brand, A.H. and Perimon, N., Targeted gene expression as a means of altering cell
fates and generating dominant phenotypes, Development, 117, 401, 1993.
65. Bray, S. and Amrein, H., A putative Drosophila pheromone receptor expressed in
male-specific yaste neurons is required for efficient courtship, Neuron, 39, 1019, 2003.
66. Thorne, N. et al., Taste perception and coding in Drosophila, Cur. Biol., 14, 1065,
2004.
67. Xu, P.X. et al., Drosophila OBP LUSH is required for activity of pheromone-sensitive
neurons, Neuron, 45, 193, 2005.
68. Lin, H. et al., A Drosophila DEG/EnaC channel subunit is required for male response
to female pheromones, Proc. Natl. Acad. USA, 102, 12831, 2005.
69. Fishilevitch, E. and Vosshall, L.B., Genetic and functional subdivision of the Droso-
phila antennal lobe, Cur. Biol., 15, 1548, 2005.
70. Couto, A., Alenius, M., and Dickson, B.J., Molecular, anatomical and functional
organization of the Drosophila olfactory system, Cur. Biol., 15, 1535, 2005.
71. Amrein, H. and Thorne, N., Gustatory perception and behavior in Drosphila melan-
ogaster, Cur. Biol., 15, 673, 2005.
72. Possidente, D.R. and Murphey, R.K., Genetic control of sexually dimorphic axon
morphology on Drosophila sensory neurons, Dev. Biol., 132, 448, 1989.
73. Stocker, R., The organization of the chemosensory system in Drosophila melano-
gaster, Cell Tissue Res., 275, 3, 1994.
74. Laissue, P.P. et al., Three-dimensional reconstruction of the antennal lobe in Droso-
phila melanogaster, J. Comp. Neurology, 405, 543, 1999.
1903_C022.fm Page 332 Wednesday, July 26, 2006 7:08 PM
75. Kondoh, Y. et al., Evolution of sexual dimorphism in the olfactory brain of Hawaiian
Drosophila, Proc. R. Soc. Lond., 270, 1005, 2003.
76. Kim, M.S., Repp, A. and Smith, D.P., LUSH Odorant-Binding Protein mediates
chemosensory responses to alcohols in Drosophila melanogaster, Genetics, 150, 711,
1998.
77. Nothiger, R. and Steinman-Zwicky, M., Sex determination in Drosophila, Trends in
Genetics, 1, 209, 1985.
78. Baker, B., Sex in flies: the splice of life, Nature, 340, 521, 1989.
79. Christiansen, A.E. et al., Sex comes in from the cold: the integration of sex and
pattern, Trends in Genetics, 18, 510, 2002.
80. Hotta, Y. and Benzer, S., Courtship in Drosophila mosaics: sex-specific foci for
sequential action patterns, Proc. Natl. Acad. Sci. USA, 73, 4154, 1976.
81. Hall, J.C., Control of male reproductive behavior by the central nervous system of
Drosophila: dissection of a courtship pathway by genetic mosaics, Genetics, 92, 437,
1979.
82. Jallon, J.M. and Hotta, Y., Genetic and behavioural studies of female sex appeal in
Drosophila, Behav. Genet., 9, 257, 1979.
83. McKeown, M., Belote, J.M. and Boggs, R.T., Ectopic expression of the female
transformer gene leads to female transformation of chromosomally male Drosophila,
Cell, 53, 887, 1988.
84. Ferveur, J.F. et al., Genetic feminization of brain structures and changed sexual
orientation in male Drosophila, Science, 267, 902, 1995.
85. Arthur, B.I. et al., Sexual behavior in Drosophila is irreversibly programmed during
a critical period, Cur. Biol., 8, 1187, 1998.
86. Hall, J.C., Courtship among males due to a male-sterile mutation in Drosophila
melanogaster, Behav. Genet., 8, 125, 1978.
87. Gailey, D. and Hall, J.C., Behavior and cytogenetics of fruitless in Drosophila mel-
anogaster: different courtship defects caused by separate closely linked lesions,
Genetics, 121, 773, 1989.
88. Villella, A., Extended reproductive roles of the fruitless gene in Drosophila melan-
ogaster by behavioural analysis of new fruitless mutants, Genetics, 147, 1011, 1997.
89. Ito, H. et al., Sexual orientation in Drosophila is altered by the satori mutation in the
sex-determination cascade gene fruitless that encodes a zinc finger protein with a
BTB domain, Proc. Natl. Acad. Sci. USA, 93, 9687, 1996.
90. Ryner, L.C. et al., Control of male sexual behavior and sexual orientation in Droso-
phila by the fruitless gene, Cell, 87, 1079, 1996.
91. Demir, S. and Dickson, B.J., fruitless splicing specifies male courtship behaviour in
Drosophila, Cell, 121, 785,2005.
92. Manoli, D.S. et al., Male-specific fruitless specifies the neural substrates of Droso-
phila courtship behaviour, Nature, 436, 395, 2005.
93. Stockinger, P. et al., Neural circuitry that governs Drosophila male courtship behavior,
Cell, 121, 795, 2005.
94. Manoli, D.S. and Baker, B., Median bundle neurons coordinate behaviours during
Drosophila male courtship, Nature, 430, 564, 2004.
95. Lee, G. and Hall, J.C., Abnormalities of male-specific FRU protein and serotonin
expression in the CNS of fruitless mutants in Drosophila, J. Neurosc., 21, 513, 2001.
96. Kimura, K.I. et al., Fruitless specifies sexually dimorphic neural circuitry in the
Drosophila brain, Nature, 438, 229, 2005.
1903_C023.fm Page 333 Wednesday, July 26, 2006 8:38 PM
CONTENTS
Introduction............................................................................................................ 333
23.1 Visualizing Neuronal Circuits Thanks to Genetics ..................................... 338
23.2 Imaging Brain Activity ................................................................................ 339
23.3 Learning and Memory Mutants ................................................................... 340
23.4 Temporal Control of Gene Expression........................................................ 341
23.5 A Tool to Build Anatomo-Functional Maps ............................................... 343
23.6 Localization of Olfactory Memory in Drosophila MBs ............................. 343
23.7 Dynamic of Olfactory Memory Phases in Drosophila ............................... 345
References.............................................................................................................. 348
INTRODUCTION
Drosophila melanogaster is one of the most intensively studied organisms in biology,
and it serves as a model system for the investigation of many cellular, developmental
and behavioral processes common to other species, including humans. This holds
true in particular for brain studies, as Drosophila central nervous system is made of
neurons and glia that operate on the same fundamental principles as their mammalian
counterparts. Thus, most neurotransmitters are identical in flies and humans, and
despite the fact that the Drosophila brain has only 100,000 cells,1 it produces complex
behaviors and sustains various forms of learning and memory. Besides studies of
fundamental brain properties, fly models are being developed for a variety of neu-
rodegenerative disorders, and the field is beginning to harness the power of Droso-
phila genetics to dissect pathways of disease pathogenesis and identify potential
targets for therapeutic intervention. This approach is possible because about 50%
of human genes have a Drosophila ortholog.2 In addition, transgenic strategies that
allow to introduce human genes into Drosophila continue to expand the list of
modeled diseases, which now includes Parkinsons disease, Alzheimers disease,
Huntingtons disease and several spinocerebellar ataxias.3,4
333
1903_C023.fm Page 334 Wednesday, July 26, 2006 8:38 PM
During the past few decades, the pool of molecular genetics techniques that
apply to the fruit fly has increased enormously. It has been a long time since the
selection of randomly and chemically produced mutants was the most common
technique to analyze genes function. Nowadays, the extensive use of transposable
element-induced mutations allows a much faster identification and study of genes
involved in a given developmental or physiological pathway. Once combined with
the systematic generation of high-resolution deletions, this global approach should
eventually provide researchers with a complete mutant collection (see Table 23.1).5,6
Moreover, new techniques have arisen that permit one to directly disturb the
expression of a defined gene, without the need of performing a mutagenesis. Rong79
and colleagues have developed the technique of homologous recombination in
Drosophila. This technique, however, remains heavy. The development of the inter-
ference RNA (RNAi) combined with the P[GAL4] enhancer-trap system allows one
to inhibit the expression of any gene of interest in almost any group of cellsfor
example, in a particular brain circuit.10
Drosophila genome is one of the first multicellular eukaryotic genomes to be
completely sequenced, together with Caenorhabditis elegans.11,12 Genes prediction has
been carried out by two different consortiums (Berkeley Drosophila Genome Project,
BDGP, and Heidelberg FlyArray) with some interesting small differences in the total
gene number and limits.13,14 The results of these international collaborations have been
made public in different Web sites (Table 23.1). Flybase offers most of the information
available for each of the 14,000 to 16,000 predicted Drosophila genes (Humans are
predicted to have about 20,000 to 25,000 protein-coding genes).15,16 One can find in
Flybase not only the sequence of the different transcripts and proteins, but also the
different mutant alleles and phenotypes, predicted protein domains, molecular function,
similarities within other species, publications referring to this gene, and more.
The whole-genome sequencing strategy has opened a new era of molecular strate-
gies such as the studies of the transcriptome and protein interactome.17,18 Several papers
report the use of microarrays to study physiological pathways involved in circadian
rhythms, immunity, sex differences, or memory.1924 Recently, White and collaborators
used probe sequences tiled throughout the genome.14 More than 170,000 oligonucle-
otides were designed against each exon, intron, and intergenic region. Their purpose
was not only to detect the differential expression of all genes and their alternative
mRNAs throughout development, but also to identify new, transcribed regions.
Giot and colleagues25 published a two-hybridbased, proteinprotein interaction
map of the fruit fly proteome. They produced a draft map of 7,048 proteins and
20,405 interactions, each with an associated rate of confidence. The complete
description of the proteome (expression pattern of all proteins and their interactions)
is thought to provide the mechanistic basis for much of the physiology.
The exponential growth in the volume of accessible biological information has
generated a plurality of voices surrounding the annotation of genes and their
products. The Gene Ontology project (see Table 23.1) seeks to provide a unified
set of structured vocabularies to describe gene products in any organism. This
work includes building three extensive ontologies to describe molecular function,
biological process, and cellular component, and providing a community database
resource that supports the use of these ontologies. The annotated genome sequence
1903_C023.fm Page 335 Wednesday, July 26, 2006 8:38 PM
TABLE 23.1
Drosophilas Web Resources
TABLE 23.1
Drosophilas Web Resources (Continued)
Bloomington The main Drosophila stock center: with a Mutant Drosophila
http://www.flystocks.bio.indi collection of P-element strains, EP strains, strains.
ana.edu EMS, natural produced mutants, etc.
KC KC
Ca Ca
LH
LH
AGT AGT
AL AL
FIGURE 23.1 (SEE COLOR INSERT FOLLOWING PAGE 236) Drosophila olfactory
memory system. Olfactory sensory neurons project to the antennal lobes (ALs). From there,
projection neurons project through the antenno-glomerular tract (AGT) and connect mushroom
body (MB) dendrites localized in the calyx (Ca), as well as the lateral horn (LH). Each MB
is composed of about 2,500 neurons, the Kenyon cells (KC). Three types of KC project in
five lobes: /, / and .
1903_C023.fm Page 339 Wednesday, July 26, 2006 8:38 PM
AAAAA
GAL4-mRNA
GFP expressed
within MB
AAAAA
GFP-mRNA
GAL4
+
UAS
UAS
UAS
UAS
UAS
GFP
5IR 3IR
FIGURE 23.2 (SEE COLOR INSERT) The GAL4/UAS system. When an enhancer-trap
transposable element is inserted near transcription enhancers that control expression in a given
structure, the GAL4 gene that is contained in the P-element is expressed in the same structure.
If this fly contains also a reporter gene downstream of the UAS sequences, GAL4 will
transcribe the reporter.
Another learning and/or memory mutant is radish (rsh).57 The rsh gene was
localized within a 180-kb interval in the 11D-E region of the X chromosome, and
several candidate genes were identified.57 Recently, Chiang and colleagues58 reported
that the responsible gene for rsh phenotype was a phospholipase A2. However a
second team has reported that rsh encodes a novel protein with possible nuclear
localization motifs.46,59 This discrepancy illustrates the difficulty linked to working
with EMS-induced behavioral mutants.
P-elementbased behavioral screens for learning and memory mutants have also
been performed. Various mutants were issued from these mutagenesis, including
nalyot (a myb-related Adf1 transcription factor),60,61 leonardo (a zeta isoform of the
14-3-3 protein),62 and volado (two splice variants of an -integrin).63
Dubnau and colleagues24 recently performed a behavioral screen for long-term
memory mutants, in parallel to microarray experiments aimed to select genes with
altered expression after long-term memory training. This work led to the identifica-
tion of proteins involved in mRNA processing and translation.24
We recently described crammer (cer), a gene involved specifically in the setting
up of long-term memory.64 The P[GAL4] cer strain has a reduced long-term memory
but a normal short-term and middle-term memory. Interestingly, in the wild-type
strain, cer is transiently underexpressed three hours after long-term memory training.
As the Cer peptide is an inhibitor of cysteine proteinases, the decrease in its expres-
sion shortly after intensive training must lead to a transient activation of its cysteine
proteinase(s) target(s).64 Altogether these works demonstrate the power of the for-
ward-genetic approach: learning or memory mutants are first characterized without
knowing in advance the molecular function of the gene involved.
Last, some genes have been shown to be involved in Drosophila learning or
memory due to function similarity to well-described genes in other species. For
example, NF1, a GTPase-activating protein for Ras, is linked to the human disease
neurofibromatosis that sometimes leads to learning disabilities.65 Drosophila NF1
mutants show a defect in olfactory learning.66 Notch, a critical component of an
evolutionarily conserved signaling mechanism that regulates neural development, is
involved in hippocampal synaptic plasticity and in learning and memory in mice.67,68
It was recently shown that Notch signaling is required for long-term memory for-
mation in adult Drosophila.69,70
UAS
UAS
UAS
UAS
UAS
RNAi RNAi
RNAi RNAi
GAL 80ts
B Low temperature High temperature
GAL4, UAS-RNAi
ts
GAL 4 GAL 80 GAL 4 RNAi
RNAi
(RNAi activity during
+ adult life)
UAS
UAS
UAS
UAS
UAS
UAS
UAS
UAS
UAS
UAS
RNAi RNAi
RNAi RNAi
GAL4 DNA-binding domain. This tool allows also a temporally controlled repres-
sion of genes located near a P[UAS] insertion.75
STMs were wild-type. Flies with all lobes present or without / lobes had a normal
STM and LTM. Thus, MBs are necessary to perform LTM, and more particularly
the / vertical lobes.92 By expressing Shits1 in / lobes, it was further shown that
lobes outputs are required during LTM retrieval.93
If MBs play a pivotal role in Drosophila olfactory learning and memory, a recent
study suggests that brain structures located outside the MBs also participate in
olfactory memory.94 We have recently shown that the Drosophila brain is asymmet-
ric, as a small structure expressing Fasciclin IIthe asymmetric bodyis present
only in the right hemisphere. Interestingly, about 7% of Canton-Special wild-type
flies present a bilateral structure. Those symmetric flies showed no LTM at 4 days,
suggesting that asymmetry may be required for generating, maintaining, or retrieving
LTM.94 The asymmetric body is located near the central complex, a structure that
connects brain hemispheres. However, the functional links between the asymmetric
body and other brain circuits remain to be sorted out.
STM/MTM, and the other with ARM (Figure 23.4A). Indeed, dnc and rut retain a
significant level of early memory,102 suggesting that an adenyl cyclase-Rut indepen-
dent learning might exist. Moreover, ARM levels in rut and amn are near normal,57,93
while their labile memories are strongly affected. Thus ARM does not seem to
depend on STM/MTM. Instead a second learning process could give rise to STM2
phases and later to ARM (Figure 23.4).
What are the relationships between ARM and LTM? To answer that question,
the ala mutant was trained with the long protocol, and the memory of flies lacking
vertical / lobes was measured at 30 minutes and 5 hours after the training. Thirty-
minute memory was normal, but, surprisingly, 5-hour memory was close to zero.
Memory performance was normal at 5 hours when flies without vertical lobes were
trained with the short protocol93 (Figure 23.5). Why does a longer training give rise
to a weaker memory? ala flies display no LTM because they lack the vertical lobes,
the center for LTM. These flies show a normal ARM at 5 hours after the short
protocol, but they show no ARM after the long protocol. This result suggests that
ARM is erased after LTM conditioning. Thus the consolidated memory phases
generated by olfactory conditioning are exclusive (Figure 23.6).93 Why is ARM
erased after LTM conditioning? We propose that ARM could act as a gating mech-
anism for LTM formation, avoiding a heavy cascade of gene expression in absence
of intensive, spaced conditioning. Despite the relative simplicity of the Drosophila
dunce
A rutabaga amnesiac
DCO
LRN 1 STM 1 MTM
c A MP-dependent way
? ? radish
LRN 2 STM 2 ARM
B
STM 1
STM 2
MTM
ARM
Memory level
1h 2h 3h 4h 5h 6h
FIGURE 23.4 (A) Model of associative memory phases and (B) temporal dynamics of
memory phases generated by a single cycle of conditioning (short protocol). LRN: learning;
STM: short-term memory; MTM: middle-term memory; ARM: anesthesia-resistant memory.
1903_C023.fm Page 347 Wednesday, July 26, 2006 8:38 PM
60
50
1X
40
PI
30
20 5X
10
0
30 min 3h 5h
FIGURE 23.5 In flies without MB alpha lobes that normally sustain long-term memory, the
long protocol decreases memory performance at 5 hours in comparison with the short protocol.
Grey line: short protocol; black line: long protocol.
A dunce
rutabaga
amnesiac
DCO
LRN 1 STM 1 MTM LTM
cAMP-dependent way
? ? radish
LRN 2 STM 2 ARM
B
B
STM
MTM
LTM
level
Memory level
Memory
1h 2h 3h 4h 5h 6h
FIGURE 23.6 (A) Model of associative memory phases and (B) temporal dynamics of
memory phases generated by five spaced cycles of conditioning (long protocol). LRN: learn-
ing; STM: short-term memory; MTM: middle-term memory; ARM: anesthesia-resistant
memory; LTM: long-term memory.
brain, this model suggests a cognitive complexity more frequently associated with
mammalian models. It supports the idea that Drosophila is a valid model in which
to study some of the molecular and cellular mechanisms involved in normal or
pathological human memory.3
1903_C023.fm Page 348 Wednesday, July 26, 2006 8:38 PM
REFERENCES
1. Shimada, T. et al. Analysis of the distribution of the brain cells of the fruit fly by an
automatic cell counting algorithm, Physica A 350, 144, 2005.
2. Rubin, G.M. et al. Comparative genomics of the eukaryotes, Science 287 (5461),
2204, 2000.
3. Shulman, J.M. et al. From fruit fly to bedside: translating lessons from Drosophila
models of neurodegenerative disease, Curr Opin Neurol 16 (4), 443, 2003.
4. Bier, E. Drosophila, the golden bug, emerges as a tool for human genetics, Nat Rev
Genet 6 (1), 9, 2005.
5. Parks, A.L. et al. Systematic generation of high-resolution deletion coverage of the
Drosophila melanogaster genome, Nat Genet 36 (3), 288, 2004.
6. Thibault, S.T. et al. A complementary transposon tool kit for Drosophila melanogaster
using P and piggyBac, Nat Genet 36 (3), 283, 2004.
7. Rong, Y.S. and Golic, K.G. Gene targeting by homologous recombination in Droso-
phila, Science 288 (5473), 2013, 2000.
8. Rong, Y.S. et al. Targeted mutagenesis by homologous recombination in D. melan-
ogaster, Genes Dev 16 (12), 1568, 2002.
9. Rong, Y.S. and Golic, K.G. A targeted gene knockout in Drosophila, Genetics 157
(3), 1307, 2001.
10. Piccin, A. et al. Efficient and heritable functional knock-out of an adult phenotype
in Drosophila using a GAL4-driven hairpin RNA incorporating a heterologous spacer,
Nucleic Acids Res 29 (12), E55, 2001.
11. Adams, M.D. et al. The genome sequence of Drosophila melanogaster, Science 287
(5461), 2185, 2000.
12. Misra, S. et al. Annotation of the Drosophila melanogaster euchromatic genome: a
systematic review, Genome Biol 3 (12), R1, 2002.
13. Hild, M. et al. An integrated gene annotation and transcriptional profiling approach
towards the full gene content of the Drosophila genome, Genome Biol 5 (1), R3, 2003.
14. Stolc, V. et al. A gene expression map for the euchromatic genome of Drosophila
melanogaster, Science 306 (5696), 655, 2004.
15. Venter, J.C. et al. The sequence of the human genome, Science 291 (5507), 1304,
2001.
16. Consortium, I.H.G.S. Finishing the euchromatic sequence of the human genome,
Nature 431 (7011), 931, 2004.
17. Callahan, C.A. et al. Control of neuronal pathway selection by a Drosophila receptor,
Nature 376 (6536), 171, 1995.
18. Hoskins, R.A. et al. Heterochromatic sequences in a Drosophila whole-genome
shotgun assembly, Genome Biol 3 (12), R1, 2002.
19. Claridge-Chang, A. et al. Circadian regulation of gene expression systems in the
Drosophila head, Neuron 32 (4), 657, 2001.
20. McDonald, M.J. and Rosbash, M. Microarray analysis and organization of circadian
gene expression in Drosophila, Cell 107 (5), 567, 2001.
21. De Gregorio, E. et al. Genome-wide analysis of the Drosophila immune response by
using oligonucleotide microarrays, Proc Natl Acad Sci USA 98 (22), 12590, 2001.
22. Girardot, F., Monnier, V. and Tricoire, H. Genome wide analysis of common and
specific stress responses in adult Drosophila melanogaster, BMC Genomics 5 (1),
74, 2004.
23. Jin, W. et al. The contributions of sex, genotype and age to transcriptional variance
in Drosophila melanogaster, Nat Genet 29 (4), 389, 2001.
1903_C023.fm Page 349 Wednesday, July 26, 2006 8:38 PM
45. Heisenberg, M. Mushroom body memoir: from maps to models, Nat Rev Neurosci 4
(4), 266, 2003.
46. Davis, R.L. Olfactory learning, Neuron 44 (1), 31, 2004.
47. Quinn, W.G., Harris, W.A. and Benzer, S. Conditioned behavior in Drosophila mel-
anogaster, Proc Natl Acad Sci USA 71 (3), 708, 1974.
48. Quinn, W.G., Sziber, P.P. and Booker, R. The Drosophila memory mutant amnesiac,
Nature 277 (5693), 212, 1979.
49. Livingstone, M.S., Sziber, P.P. and Quinn, W.G. Loss of calcium/calmodulin respon-
siveness in adenylate cyclase of rutabaga, a Drosophila learning mutant, Cell 37 (1),
205, 1984.
50. Levin, L.R. et al. The Drosophila learning and memory gene rutabaga encodes a
Ca2+/Calmodulin-responsive adenylyl cyclase, Cell 68 (3), 479, 1992.
51. Han, P.L. et al. Preferential expression of the Drosophila rutabaga gene in mushroom
bodies, neural centers for learning in insects, Neuron 9 (4), 619, 1992.
52. Mohler, J.D. Developmental genetics of the Drosophila egg. I. Identification of 59
sex-linked cistrons with maternal effects on embryonic development, Genetics 85 (2),
259, 1977.
53. Davis, R.L. and Davidson, N. Isolation of the Drosophila melanogaster dunce chro-
mosomal region and recombinational mapping of dunce sequences with restriction
site polymorphisms as genetic markers, Mol Cell Biol 4 (2), 358, 1984.
54. Feany, M.B. and Quinn, W.G. A neuropeptide gene defined by the Drosophila mem-
ory mutant amnesiac, Science 268 (5212), 869, 1995.
55. Moore, M.S. et al. Ethanol intoxication in Drosophila: Genetic and pharmacological
evidence for regulation by the cAMP signaling pathway, Cell 93 (6), 997, 1998.
56. Toba, G. et al. The gene search system. A method for efficient detection and rapid
molecular identification of genes in Drosophila melanogaster, Genetics 151 (2), 725,
1999.
57. Folkers, E., Drain, P. and Quinn, W.G. Radish, a Drosophila mutant deficient in
consolidated memory, Proc Natl Acad Sci USA 90 (17), 8123, 1993.
58. Chiang, A.S. et al. radish encodes a phospholipase-A2 and defines a neural circuit
involved in anesthesia-resistant memory, Curr Biol 14 (4), 263, 2004.
59. Folkers, E., Waddell, S. and Quinn, W.G. Personal communication, 2004.
60. Boynton, S. and Tully, T. latheo, a new gene involved in associative learning and
memory in Drosophila melanogaster, identified from P element mutagenesis, Genet-
ics 131 (3), 655, 1992.
61. DeZazzo, J. et al. Nalyot, a mutation of the Drosophila myb-related Adf1 transcription
factor, disrupts synapse formation and olfactory memory, Neuron 27 (1), 145, 2000.
62. Skoulakis, E.M. and Davis, R.L. Olfactory learning deficits in mutants for leonardo,
a Drosophila gene encoding a 14-3-3 protein, Neuron 17 (5), 931, 1996.
63. Grotewiel, M.S. et al. Integrin-mediated short-term memory in Drosophila, Nature
391 (6666), 455, 1998.
64. Comas, D., Petit, F. and Preat, T. Drosophila long-term memory formation involves
regulation of cathepsin activity, Nature 430 (6998), 460, 2004.
65. Shen, M.H., Harper, P.S. and Upadhyaya, M. Molecular genetics of neurofibromatosis
type 1 (NF1), J Med Genet 33 (1), 2, 1996.
66. Guo, H.F. et al. A neurofibromatosis-1-regulated pathway is required for learning in
Drosophila, Nature 403 (6772), 895, 2000.
67. Costa, R.M., Honjo, T. and Silva, A.J. Learning and memory deficits in Notch mutant
mice, Curr Biol 13 (15), 1348, 2003.
1903_C023.fm Page 351 Wednesday, July 26, 2006 8:38 PM
90. McGuire, S.E., Le, P.T. and Davis, R.L. The role of Drosophila mushroom body
signaling in olfactory memory, Science 293 (5533), 1330, 2001.
91. Tully, T. et al. Genetic dissection of consolidated memory in Drosophila, Cell 79 (1),
35, 1994.
92. Pascual, A. and Preat, T. Localization of long-term memory within the Drosophila
mushroom body, Science 294 (5544), 1115, 2001.
93. Isabel, G., Pascual, A. and Preat, T. Exclusive consolidated memory phases in Droso-
phila, Science 304 (5673), 1024, 2004.
94. Pascual, A. et al. Neuroanatomy: brain asymmetry and long-term memory, Nature
427 (6975), 605, 2004.
95. Quinn, W.G. and Dudai, Y. Memory phases in Drosophila, Nature 262 (5569), 576,
1976.
96. Isabel, G. and Preat, T. Unpublished results, 2005.
97. Drier, E.A. et al. Memory enhancement and formation by atypical PKM activity in
Drosophila melanogaster, Nat Neurosci 5 (4), 316, 2002.
98. Yin, J.C. et al. Induction of a dominant negative CREB transgene specifically blocks
long-term memory in Drosophila, Cell 79 (1), 49, 1994.
99. Perazzona, B. et al. The role of cAMP response element-binding protein in Drosophila
long-term memory, J Neurosci 24 (40), 8823, 2004.
100. Yin, J.C. et al. A Drosophila CREB/CREM homolog encodes multiple isoforms,
including a cyclic AMP-dependent protein kinase-responsive transcriptional activator
and antagonist, Mol Cell Biol 15 (9), 5123, 1995.
101. Yin, J.C. et al. CREB as a memory modulator: induced expression of a dCREB2
activator isoform enhances long-term memory in Drosophila, Cell 81 (1), 107, 1995.
102. Tully, T. and Quinn, W.G. Classical conditioning and retention in normal and mutant
Drosophila melanogaster, J Comp Physiol [A] 157 (2), 263, 1985.
1903_C024.fm Page 353 Thursday, July 27, 2006 8:38 PM
CONTENTS
24.1 Summary........................................................................................................ 353
24.2 C. elegans as a Model Organism .................................................................. 353
24.2.1 C. elegans Has a Very Simple Nervous System .............................. 353
24.2.2 C. elegans Is Well Suited for Genetic Studies................................. 355
24.3 Behaviors of C. elegans and Mutations Affecting Them............................. 355
24.3.1 Chemosensory Behavior and Its Plasticity....................................... 356
24.3.2 Egg Laying........................................................................................ 359
24.3.3 Regulation of Locomotion in Response to Food ............................. 361
24.3.4 Ethanol Resistance............................................................................ 362
24.3.5 Social Feeding Behavior................................................................... 363
24.3.6 Memory............................................................................................. 364
24.4 Conclusion ..................................................................................................... 367
References.............................................................................................................. 367
Authors Note ........................................................................................................ 369
24.1 SUMMARY
This chapter provides an overview of the behavioral studies that can be undertaken
on an extremely simple organism, the nematode Caenorhabditis elegans. First, we
will briefly describe this animal model, which has become increasingly popular for
molecular and cellular biology studies, and then we will present a few examples of
behavioral studies that are conducted on this organism. We aim to show that this
model provides paradigms for questions central to many behaviors, and that they
can be addressed at a single cell/single gene resolution.
353
1903_C024.fm Page 354 Thursday, July 27, 2006 8:38 PM
can be grown easily on petri dishes seeded with Escherichia coli (Figure 24.1). Its
length is about 1 mm as an adult and it has 959 somatic cells. Among those, 302
of them are neurons, which can be divided into 116 classes on the basis of their
morphology (Figure 24.2). With a laser beam, it is possible to destroy a given neuron
in an anaesthetized animal to assess the involvement of this neuron in a behavior or
a function. Moreover, thealmost invariantconnectivity of all neurons has been
established by 3D reconstruction of serially sectioned animals.
At the molecular level, the nervous system of C. elegans shares a common
organization with the far-distant vertebrates: its main excitatory transmitter is
FIGURE 24.1 C. elegans on a culture plate. Worms of various stages can be seen, including
eggs, larvae and adults. Adults are approximately 1 mm long.
FIGURE 24.2 The nervous system of C. elegans. Most or all neurons of C. elegans are
visualized by expressing a gpc-2::GFP construct. Weakly, staining of muscle cells can be
seen. Arrows indicate the ventral nerve cord, a triangle indicates staining in the anterior
ganglia, an asterisk indicates the posterior ganglia.
1903_C024.fm Page 355 Thursday, July 27, 2006 8:38 PM
glutamate, its main inhibitor is GABA, the neuromuscular synapse is mainly cho-
linergic, and neuronal activity is regulated in part by two major modulators that
display a wide spectrum of effects in the CNS of vertebrates: dopamine and sero-
tonin. A variety of peptidesalbeit less numerous than in mammalsalso act as
neuromodulators.
FIGURE 24.3 (SEE COLOR INSERT FOLLOWING PAGE 236) DIC photograph of the
head of a worm. The outline of the pharynx (p) can be seen and the beginning of the gut (g).
The approximate positions of the amphid sensory neuron cell bodies are indicated with colored
circles. The dendrite (d) and an axon (a) of one of the cells are drawn. Below, the names of
the corresponding cells are given, and the compounds perceived by these cells.
1903_C024.fm Page 357 Thursday, July 27, 2006 8:38 PM
The exact molecular mechanisms that make this possible are not completely clear,
but part of the specificity arises from a functional asymmetry between the left and
the right AWC cell (Wes and Bargmann, 2001). In addition, odorant detection seems
modulated by a complex G protein signaling network, involving two to three redun-
dant stimulatory G subunits and an inhibitory G in each neuron (Lans
et al., 2006).
As in mammals, binding of an odorant to a GPCR and activation of heterotri-
meric G proteins is thought to activate downstream effector molecules. In the AWA
cells, G proteins probably activate TRPV channels resulting in a Ca2+ influx. In the
AWC cells, G proteins induce an increase in intracellular cGMP, mediated by
guanylyl cyclases, leading to opening of a cyclic nucleotide-gated channel. These
studies have shown that C. elegans uses quite similar sensory signaling pathways
as mammals.
Prolonged exposure of nematodes to an odorant reduces chemotaxis to that
odorant (Mori and Oshima, 1997). This behavior is odorant specific, modulated by
food and stress and depends on the odorant concentrations used during pre-exposure
and chemotaxis. This behavior provides a good basis to study olfactory adaptation,
desensitization or perhaps even olfactory learning. However, although several genes
have been identified that play a role in olfactory plasticity, it has been quite difficult
to identify additional genes. The main difficulty is that this behavior is dependent
on many factors and hence relatively variable. However, the careful dissection of
the behavioral variables of olfactory plasticity over the past decade, the current
knowledge of olfactory signaling, and the development of novel techniques to study
gene function and identify cells involved should bring novel insights into the molec-
ular mechanisms of olfactory plasticity.
C. elegans shows strong chemo-attraction to salts. In nature, the detection of
salts is essential for salt homeostasis and salts are probably cues for food. In the
laboratory, salt detection is mainly tested using two assays. In one assay, animals
are placed on a salt gradient and tested for chemotaxis to the highest concentration
(Ward, 1973). This single worm assay has enabled the identification of the sensory
neurons involved in salt detection using laser ablation techniques. The detection of
NaCl is mainly mediated by the ASE chemosensory neurons, but a residual response
is mediated by the ADF, ASG, and ASI neurons. A more recent study confirmed
the importance of the ASE neurons for chemotaxis to NaCl, and revealed a functional
asymmetry between the left and the right cell: ASE left is primarily sensitive to
sodium, whereas ASE right is primarily sensitive to chloride and potassium (Pierce-
Shimomura et al., 2001). In an alternative assay, the quadrant assay, animals are
given a choice between two salt concentrations, by putting them on a plate divided
into four quadrants that can each be filled independently with salt containing agar
(Wicks et al., 2000). This assay provides a robust method to screen for additional
chemotaxis mutants.
In rodents and flies, the main mechanism of salt detection involves ion influx
through degenerin/epithelial Na+ channels (DEG/ENaC), leading to membrane
depolarization and neurotransmitter release. These ENaC channels can be blocked
specifically with the drug amiloride. In rodents, 75% of the response to salt can be
blocked with amiloride; in humans the amiloride sensitivity is less pronounced,
1903_C024.fm Page 358 Thursday, July 27, 2006 8:38 PM
indicating that also other salt detection mechanisms exist. In C. elegans, chemotaxis
to salts is not blocked by amiloride. Perhaps the worm can help to elucidate the
molecular mechanisms of salt detection.
C. elegans is attracted to salt concentrations ranging from 0.1 to 100 mM. High
salt concentrations result in avoidance behavior. This latter response is probably due
to a general avoidance of high osmotic strength, mediated by the nociceptive neurons
ASH and ADL. Forward and reverse genetic approaches have identified several
genes involved in salt detection in C. elegans. These include a cGMP-gated channel
and several Ca2+ activated proteins, indicating that cGMP and Ca2+ are important
second messengers in salt detection. Surprisingly, our recent analysis of these
mutants in the quadrant assay showed that these animals still respond to 25100
mM salt, indicating that another independent high salt detection pathway exists
(Hukema et al., 2006). Thus far, no molecules involved in this pathway have been
identified.
Recently, we have adapted the quadrant assay to study the plasticity of the
response to salt. In this assay, animals are pre-exposed to relatively high, but attrac-
tive concentrations of NaCl (100 mM), and subsequently tested for chemotaxis to
NaCl. Animals pre-exposed to NaCl are no longer attracted to NaCl but even avoid
it. This behavior, which we call gustatory plasticity, is time and concentration
dependent, reversible and partially salt specific (Jansen et al., 2002). The finding
that prolonged exposure to NaCl results in avoidance behavior suggests that gustatory
plasticity is more than adaptation or desensitization. We used two methods to identify
sensory neurons involved in this process. We tried to rescue the plasticity defect of
a G protein subunit mutant, gpc-1, by expressing the gpc-1 gene in specific cells.
In addition, we interfered with cell function by expressing a dominant mutant version
of an ion channel in specific sensory neurons. Using these techniques, we have
shown that gustatory plasticity requires a balance of antagonistic sensory inputs,
involving attractive signals via the ASE salt detection neurons, and aversive signals
via the ASI chemosensory neurons and the nociceptive neurons ASH and ADL
(Hukema et al., 2006).
In a candidate-gene approach, we analyzed gustatory plasticity of 66 mutant
strains (Hukema et al., submitted). In this study, we identified several signaling
cascades and neurotransmitters involved in gustatory behavior. Taking into account
the cellular circuitry and the expression patterns of the various signaling molecules,
we propose a model for the molecular and cellular mechanisms that regulate gusta-
tory plasticity. We are currently performing additional cell-specific rescue experi-
ments to test whether the different signaling molecules indeed function in the cells
as proposed in our model. First, the gustatory response requires salt detection
transmitted by a cGMP and Ca2+ pathway and another unknown high salt detection
pathway. These salt-attraction signals are probably mediated by the ASE neurons.
This signal is balanced by an unknown signal detected by the ASI neurons, trans-
mitted by G proteins, probably activating a thus far unknown guanylyl cyclase.
cGMP subsequently activates a cGMP-gated channel resulting in a Ca2+ influx and
activation of several proteins, ultimately leading to a salt-aversion signal. Another
aversion signal is provided by the nociceptive neurons. At least two G alpha subunits
1903_C024.fm Page 359 Thursday, July 27, 2006 8:38 PM
Odors Odors
Touch Touch
Temperature Temperature
Food Food
FIGURE 24.4 Egg laying as a model of integration. The egg-laying behavior is the result of
the integration of various sensory input. Negative stimuli (such as bad odors, vibrations, cold,
etc.) block the egg-laying process (right side of the figure). When all the sensory inputs are
appropriate (left side) egg-laying occurs. The question marks indicate stimuli that are thought
to influence egg laying although they have not been experimentally demonstrated. The advan-
tage of the C. elegans model is that it is amenable to the dissection of the stimuli at a one
cell/ one gene resolution.
Many causes can lead to an egl-d phenotype, including the inability of the animal
to lay eggs (for instance, morphological defects of the vulva). To distinguish which
egl-d mutants are truly unable to lay eggs, from those who could lay eggs but suffer
from misregulation, these mutants are tested for their response to serotonin, a strong
stimulator of egg laying, which acts directly on the vulva muscles. The egl-d mutants
who are responsive to serotonin must possess a functional egg-laying apparatus.
Therefore, their egl-d phenotype is supposed to be a misregulation mechanism.
Almost a dozen genes produce such mutants; unfortunately, few have been identified
yet. Interestingly, several egl-c mutants have abnormally short HSN processes, which
prevents the formation of synapses between HSNs and head neurons. This suggests
that the HSN neurons are indeed in a repressed (hyperpolarized) state under
unfavorable conditions and are derepressed in favorable conditions. The molecular
mechanisms involved are still unknown. The use of cameleon proteins, which are
GFP-based indicators of intracellular calcium, will certainly be helpful tools to
further understand this process.
1903_C024.fm Page 361 Thursday, July 27, 2006 8:38 PM
which promote social feeding, while signals from unidentified neurons repress social
feeding.
Second, Coates and de Bono (2002) follow up on the finding that NPR-1
regulates social feeding. By carefully analyzing the expression pattern of the npr-1
gene, using a npr-1::GFP fusion construct, and subsequent cell specific expression
of the NPR-1 215V gene in candidate cells in npr-1 animals they implicated the
AQR, PQR and URX neurons in the regulation of social feeding. This finding was
confirmed by genetically blocking the electrical activity of these neurons. In addition
to inhibition of social feeding via NPR-1 in these cells, social feeding is stimulated
by signaling through a cGMP-gated channel. The AQR, PQR and URX neurons are
directly exposed to the body fluid, which seems to function analogously to blood.
Together, these findings suggest a model in which the body cavity neurons, AQR,
PQR and URX, regulate the choice between solitary and social feeding by integrating
NPR-1 mediated signals, which inhibit aggregation, and cGMP mediated signals,
which stimulate aggregation (Coates and de Bono, 2002). At present, it is unclear
where these signals originate.
24.3.6 MEMORY
Although it may sound incredible to many mammalian neurobiologists, worms can
display some forms of associative memory. The most popular assay, described in
this section, is based on the storage of temperature information. Thirty years ago,
when C. elegans genetics was still in its infancy, Hedgecock and Russel (1975)
showed that when worms are placed in a thermal gradient in absence of food, they
move to the temperature where they last were in presence of food (isothermal
tracking). It is thought that, when in the wild, recollection of environmental param-
eters such as temperature and hygrometry can help them to find their food. On the
laboratory bench, it is just a paradigm showing that the animals are able to associate
the presence of food with the sensation of temperature and to store this information.
In absence of food, the isothermal tracking behavior is kept for several hours (Figure
24.5); then the worms will cross isotherms randomly to seek food at other temper-
atures. If food is encountered, a new acquisition period begins.
By laser-ablation experiments, Mori and collaborators (1995) have identified two
(possibly three) neurons as being critical for the memory process. These neurons (AIY
and AIZ) are integrating interneurons, which receive synapses from most head sensory
neurons (including AFD, a sensory neuron specialized in sensing temperature). Ani-
mals in which AIY or AIZ are deleted have peculiar phenotypes: AIY-deleted worms
are cryophilic (they are attracted by colder temperatures) and AIZ-deleted worms are
thermophilic, indicating that these neurons probably have antagonistic effects in wild-
type worms. Since both AIY and AIZ synapse on RIA (another integrating interneuron
downstream of AIY and AIZ), this suggests a mechanism by which the temperature
information could be encoded by the relative strength of AIY and AIZ connections
on RIA. When the sensory neuron AFD is ablated, the worms are either cryophilic,
or atactic (do not feel the temperature). Thanks to the numerous mutants available in
C. elegans, some of these ablation experiments can be reproduced genetically, which
then provide the immense advantage of having unlimited numbers of modified animals
1903_C024.fm Page 365 Thursday, July 27, 2006 8:38 PM
A B
25C 25C
15C 15C
C D
25C 25C
15C 15C
FIGURE 24.5 Associative learning assay and mutant classes. When placed in a temperature
gradient without food (here 15C at the center and 25C at the edge of the dish), worms will
migrate to the temperature they last encountered food (A). Several mutants have been isolated
which are atactic (B), cryophilic (C), or thermophilic (D). Some of these mutations mimic
phenotypes obtained after key neurons have been ablated, disturbing the balance between
cold-driving and heat-driving neurons.
(Figure 24.5). For example, mutations in the ttx-3 gene specifically disable the AIY
neuron due to axonal defects.
How does the worm compare the ambiant temperature (Tamb) to the cultivation
temperature (Tcult)? To test whether AFD has a role in this process, an easy way
would be to record the activity of AFD (and other neurons) in various conditions.
Unfortunately, it is not possible to stick an electrode in a single worm neuron (cell
diameter = a few micrometers) without severely perturbing it. Samuel and collabo-
rators (2003) have elegantly got around this difficulty by using a pH-sensitive green
fluorescent protein localized in the synaptic vesicles of AFD. Since the internal side
of the synaptic vesicles is more acidic than the extracellular environment, the fluo-
rescence levels can be correlated with synaptic release. By doing so on immobilized
live animals, the authors could come to the conclusion that AFD synaptic release is
high if Tamb and Tcult are different, and lower if they are close. Therefore, the single
AFD neuron encodes a direct comparison between actual and memorized tempera-
ture. It is likely that this process involves regulation of gene expression in AFD, but
1903_C024.fm Page 366 Thursday, July 27, 2006 8:38 PM
the molecular targets are still unknown. A schematic view of the neuronal circuitry
associated with this form of learning is shown in Figure 24.6.
The picture is clearer in the AIY interneuron, which naturally expresses the
neuron-specific calcium sensor (NCS-1) protein. In mammals, NCS-1 is involved
in long-term potentiation. When worms are deprived of NCS-1 by a mutation, their
behavior becomes cryophilic and resembles that of AIY-deleted animals, indicating
that NCS-1 is a key component of AIY function. More interestingly, overexpression
of NCS-1 in AIY increases performance levels, accelerates learning, and produces
a memory with slower extinction (Gomez et al., 2001).
Serotonin is likely to be the main inductor of this form of memory in C. elegans
for several reasons: (1) serotonin levels are high when animals are fed and low when
animals are starved; (2) serotonin is released in the head by two neurosecretory cells,
the NSM, which have sensory endings in the pharynx, or the muscle that pumps
food in C. elegans. Thus, NSM release in the body cavity is directly related to the
presence of food; (3) exogenous serotonin can mimic the presence of food in the
sensory input
(temperature)
Temperature
assessment
Serotonin AFD X and comparison
AIY AIZ
Information
storage
RIA
RIB
RIM
Behavior
FIGURE 24.6 Neuronal circuitry underlying associative behavior. Neurons are represented
according to their class (triangle: sensory, hexagon: interneuron, circle: motorneuron). Neuron
X (dashed) has been postulated but is not identified. Only the major connections are drawn.
During the conditioning phase, the signal is memorized by the action of serotonin, which
may act on sensory or interneurons. See main text for details.
1903_C024.fm Page 367 Thursday, July 27, 2006 8:38 PM
learning process: animals grown without food but in the presence of serotonin will
go to this temperature when tested. It is particularly interesting to note that serotonin,
one of the oldest neurohormones in evolution, has conserved its role of memory
inductor from roundworms (C. elegans) to mollusks (Aplysia) and mammals
(reviewed in Mayford and Kandel, 1999).
Over the years, several mutants have been identified in phenotypic screens after
random mutagenesis. These mutants have been picked because they are defective
for the memory assay. Some of them are cryophilic (e.g., ttx-1) and some of them
are thermophilic (e.g., tax-6), suggesting again that the memory process results from
a balance between thermo- and cryophilic inputs. The genes corresponding to these
mutants have been identified in some cases, and this has revealed that proteins
involved in information storage in C. elegans are common to neuronal activity. For
instance, tax-2 and tax-4 encode both subunits of a cyclic nucleotide-gated ion
channel and are expressed in the AFD neuron. Such channels have been implicated
in olfaction and photoreception in mammals. Electrophysiological properties of the
TAX-2/TAX-4 channel suggest that calcium influx is part of the signal transduction
cascade that encodes the temperature information.
24.4 CONCLUSION
Despite its simplicity, the nematode C. elegans possesses in its CNS many of the
neuronal features that have been selected and conserved throughout evolution
because they confer a selective advantage for the bearer. These are the internal
clocks, the possibility to habituate or to sensitize to a stimulus the need for coordi-
nation and integration, and, of course, the faculty to learn from experience, that is
memory. The molecular description of the mechanisms underlying these neuronal
properties is still far from complete in any organism.
As a model, C. elegans offers the advantage of being integrated. This means
that it is possible to study a behavior at the level of the whole animal, and yet to
describe it with a cellular and molecular resolution. The possibility to destroy a
single cell to assay its role in a function is a powerful tool. But the strength of the
C. elegans model resides in the power of genetics, which unambiguously relates
molecules and biological functions, and which is an invaluable way to identify the
genes relevant to a behavior.
In the few examples described in this chapter, it is striking to discover how much
neuronal molecular mechanisms have been conserved through evolution.
REFERENCES
Coates, JC and de Bono, M. (2002) Antagonistic pathways in neurons exposed to body fluid
regulate social feeding in Caenorhabditis elegans. Nature, 419, 925929.
Davies, AG, Pierce-Shimomura, JT, Kim, H, VanHoven, MK, Thiele, TR, Bonci, A, Barg-
mann, CI, and McIntire, SL. (2003) A central role of the BK potassium channel in
behavioral responses to ethanol in C. elegans. Cell., 115, 65566.
1903_C024.fm Page 368 Thursday, July 27, 2006 8:38 PM
Davies, AG, Bettinger, JC, Thiele, TR, Judy, ME, and McIntire, SL. (2004) Natural variation
in the npr-1 gene modifies ethanol responses of wild strains of C. elegans. Neuron.,
42, 73143.
de Bono, M and Bargmann, CI. (1998) Natural variation in a neuropeptide Y receptor homolog
modifies social behavior and food response in C. elegans. Cell., 94, 679689.
de Bono, M, Tobin, DM, Davis, MW, Avery, L and Bargmann, CI. (2002) Social feeding in
Caenorhabditis elegans is induced by neurons that detect aversive stimuli. Nature,
419, 899903.
Fire, A, Xu, S, Montgomery, MK, Kostas, SA, Driver, SE and Mello, CC. (1998) Potent and
specific genetic interference by double-stranded RNA in Caenorhabditis elegans.
Nature, 391, 806811.
Gomez, M, De Castro, E, Guarin, E, Sasakura, H, Kuhara, A, Mori, I, Bartfai, T, Bargmann,
CI, Nef, P. (2001) Ca2+ signaling via the neuronal calcium sensor-1 regulates asso-
ciative learning and memory in C. elegans. Neuron, 30, 241248.
Hedgecock, EM and Russel, RL (1975) Normal and mutant thermotaxis in the nematode
Caenorhabditis elegans. Proc. Nat. Acad. Sci., 72, 40614065.
Hukema, RK, Rademakers, S, Burghoorn, J and Jansen, G. Antagonistic sensory cues generate
gustatory plasticity in Caenorhabditis elegans. Submitted.
Jansen, G, Weinkove, D and Plasterk, RHA. (2002) The G-protein gamma subunit gpc-1 of
the nematode C. elegans is involved in taste adaptation. EMBO J., 21, 986994.
Kubiak, TM, Larsen, MJ, Nulf, SC, Zantello, MR, Burton, KJ, Bowman, JW, Modric, T and
Lowery, DE. (2003) Differential activation of social and solitary variants of the
Caenorhabditis elegans G protein-coupled receptor NPR-1 by its cognate ligand AF9.
J. Biol. Chem., 278, 3372433729.
Lans, H, Rademakers, S and Jansen, G. (2006) A network of stimulatory and inhibitory G
subunits regulates olfaction in Caenorhabditis elegans. Genetics, 167, 16771687.
Mayford, M, and Kandel, ER. (1999) Genetic approaches to memory storage. Trends Genet.,
15, 46370.
Morgan, PG, and Sedensky, MM. (1995) Mutations affecting sensitivity to ethanol in the
nematode, Caenorhabditis elegans. Alcohol Clin Exp Res., 19:14239.
Mori, I, and Oshima, Y. (1995) Neural regulation of thermotaxis in Caenorhabditis elegans.
Nature, 376, 344348.
Mori, I, and Oshima, Y. (1997) Molecular neurogenetics of chemotaxis and thermotaxis in
the nematode Caenorhabditis elegans. Bioessays, 19, 10551064.
Pierce-Shimomura, J.T., Faumont, S., Gaston, M.R., Pearson, B.J. and Lockery, S.R. (2001).
The homeobox gene lim-6 is required for distinct chemosensory representations in
C. elegans. Nature, 410, 694698.
Ranganathan, R, Cannon, SC and Horvitz, HR. (2000). MOD-1 is a serotonin-gated chloride
channel that modulates locomotory behaviour in C. elegans. Nature, 408, 470475.
Ranganathan, R, Sawin, ER, Trent, C and Horvitz, HR. (2001). Mutations in the Caenorhab-
ditis elegans serotonin reuptake transporter MOD-5 reveal serotonin-dependent and
independent activities of fluoxetine. J. Neurosci., 21, 58715884.
Riddle, TBDL, Meyer, BJ and Priess, JR eds. (1997) C. elegans II. Cold Spring Harbor
Laboratory Press, Plainview, New York.
Rogers, C, Reale, V, Kim, K, Chatwin, H, Li, C, Evans, P and de Bono, M. (2003) Inhibition
of Caenorhabditis elegans social feeding by FMRFamide-related peptide activation
of NPR-1. Nat. Neurosci., 6, 11781185.
Samuel, AD, Silva, RA, and Murthy, VN. (2003) Synaptic activity of the AFD neuron in
Caenorhabditis elegans correlates with thermotactic memory. J Neurosci., 23, 3736.
1903_C024.fm Page 369 Thursday, July 27, 2006 8:38 PM
Sawin, ER, Ranganathan, R and Horvitz, HR. (2000) C. elegans locomotory rate is modulated
by the environment through a dopaminergic pathway and by experience through a
serotonergic pathway. Neuron., 26, 619631.
The C. elegans Sequencing Consortium (1998) Genome sequence of the nematode C. elegans:
A platform for investigating biology. Science, 282, 20122018.
Ward, S. (1973) Chemotaxis by the nematode Caenorhabditis elegans: identification of
attractants and analysis of the response by use of mutants. Proc. Natl. Acad. Sci. USA
70, 817821.
Wes, PD and Bargmann, CI. (2001) C. elegans odour discrimination requires asymmetric
diversity in olfactory neurons. Nature, 410, 698701.
Wicks, SR, de Vries, CJ, van Luenen, HGAM and Plasterk, RHA (2000) CHE-3, a cytosolic
dynein heavy chain, is required for sensory cilia structure and function in Caenorhab-
ditis elegans. Dev. Biol,. 221, 295307.
Wood, W.B, ed. (1988) The Nematode Caenorhabditis elegans. Cold Spring Harbor Labora-
tory Press, Plainview, New York.
AUTHORS NOTE
We apologize to the authors whose work could not be cited due to lack of space.
1903_C024.fm Page 370 Thursday, July 27, 2006 8:38 PM
1903_C025.fm Page 371 Wednesday, July 26, 2006 7:09 PM
CONTENTS
25.1 INTRODUCTION
Even for the reader unfamiliar with the topic of this chapter but familiar with
dopamine and behavior, it is obvious that the topic of genes, behavior and brain
dopamine systems cannot be covered in detail. Rather, the goal of the chapter is
to provide the reader with an introduction and some examples, largely extracted
from work in the authors laboratories. Previous reviews (Hitzemann et al., 1995;
Hitzemann, 1998) can be referenced for important background material. The focus
of this chapter will be largely on D2 dopamine receptors. The reasons for this
emphasis include the longstanding emphasis on D2 receptors in the etiology
and/or expression of a variety of behaviors including schizophrenia, substance
abuse and alcoholism (Hitzemann, 1998). The other reason for this emphasis is
the amazing variability among individuals in D2 receptor density and in the response
to drugs which either stimulate or block these receptors. The reasons for this vari-
ability remain unclear, but there is ample evidence to suggest that genetic factors
have an important role. Here we provide a clinical example to illustrate the variation
in receptor density and drug response. Twenty-three normal controls were adminis-
tered 0.5 mg/kg of methylphenidate iv. after being told they would receive either
placebo or methylphenidate; behavioral effects were measured before and after
the injection and included self-ratings of pleasant and unpleasant drug effects.
On another day, the subjects were scanned for D2 /D3 receptor density using positron
371
1903_C025.fm Page 372 Wednesday, July 26, 2006 7:09 PM
3.8
3.6
3.4
3.2
BMax /KD
3.0
2.8
2.6
2.4
2.2
2.0
Pleasant Unpleasant Neutral
Perception of Methylphenidate
FIGURE 25.1 The relationship between D2 dopamine receptor binding potential and the
response to a methylphenidate challenge (data adapted from Volkow et al., 1999). The binding
potential (BMax/KD) for the receptor was estimated in 23 normal controls using positron
emission tomography (PET) and 11C-raclopride as the PET ligand. Note that with the binding
potential calculation a value of 1.0 indicates that there is no specific binding; thus, the range of
binding potentials in this figure varies by >100%. Subsequently, the subjects were challenged
with methylphenidate (0.5 mg/kg, iv.) and were asked to rate their response to the challenge
on a visual analog scale; responses varied from very pleasant to neutral to unpleasant.
1903_C025.fm Page 373 Wednesday, July 26, 2006 7:09 PM
remainder of this chapter focuses on these and related data which have contributed
to our genetic understanding of brain dopamine systems and behavior.
10
14
12 Standard Inbred Strains
10
8
8
6
Haloperidol ED50 - mg/kg
4
6 2
0
A
BALB/c
AKR
DBA/2
PL
CE
SJL
C3H/He
A/He
C57Bl/6
CBA
S129
LP
C57L
P
4
0
21
11
31
1
DBA
C57
5
30
22
24
15
27
13
14
19
28
20
12
6
25
32
16
18
23
29
2
8
BXD RI Strains
FIGURE 25.2 The ED50 values for haloperidol-induced catalepsy among a panel of stan-
dard-inbred mouse strains and among a panel of the BXD recombinant inbred strains (data
adapted from Hitzemann et al. 1995). A full description of the catalepsy response and the
methods used to determine the ED50 value is found in Kanes et al. (1993).
behavioral data (Chesler et al., 2004). The databases focus on the original BXD RI
panel (Taylor 1978), although data for some of the newer BXD RI strains are
available. Important for the immediate discussion is that the database contains more
than 650 published phenotypes, with the largest single segment focusing on drug-
related behavioral phenotypes. The catalepsy data illustrated above have been entered
into the database (ID# 10336). Querying the database for phenotypes that are
correlated with catalepsy revealed significant correlations with cocaine-induced
(15 mg/kg) stereotypy (ID# 10297) and ventral midbrain D2 dopamine receptor
density (ID# 10270) (Jones et al., 1999). These data are illustrated in Figures 25.3
and 25.4. The data in Figure 25.3 show that the RI strains most sensitive to the
haloperidol effect are also the strains most sensitive to the effects of cocaine
(r = 0.68, p < 0.0002); we interpret these data to show that if we could determine
the traits associated with increased haloperidol sensitivity, some of these would also
be associated with increased cocaine sensitivity (or vice versa). Given that both drugs
1903_C025.fm Page 375 Wednesday, July 26, 2006 7:09 PM
100
60
40
20
-20
0 2 4 6 8 10
Haloperidol-Induced Catalepsy
(ED50 mg/kg)
FIGURE 25.3 The relationship between haloperidol-induced catalepsy (ID# 10336) and
cocaine (15 mg/kg)-induced stereotypy (ID# 10297) in a panel of the BXD recombinant
inbred strains. The ID values are those that reference the phenotypes at WebQTL
(www.WebQTL.org); the data entered into the analysis may be downloaded from this site.
The catalepsy data are taken from Kanes et al. (1996) and the cocaine data are taken from
Jones et al. (1999).
exert their pharmacological effects largely via effects on brain dopamine systems,
such an association cannot be unexpected. Figure 25.4 illustrates a potential
mechanism. Here we see that the animals most sensitive to haloperidol are those
that have the highest density of D2 receptorsa result consistent with that reported
by Fink et al. (1982). Data not shown is the correlation between receptor density
and cocaine response (r = 0.68, p < 0.001). Thus, we conclude that the density of
ventral midbrain D2 receptors (presumably autoreceptors) is associated both with
haloperidol and cocaine response. These relationships are not entirely intuitives, but
the data illustrate the power of the approach and the cumulative value of the RI data.
Furthermore, we again see the wide variation in D2 receptor density and the apparent
association with a drug response. It could be argued that the variation in receptor
density is simply an epiphenomenon reflecting a difference in the number of dopam-
ine neurons in the ventral midbrain. Data on the number of neurons in the ventral
tegmental area are also found in the database (ID# 10223). For the RI panel there
is no significant correlation between receptor density and neuron number (p > 0.3).
Finally, it should be noted that for the striatum there are 12 different measures
of D2 receptor density in the database; none of these are significantly correlated
to catalepsy, confirming earlier results from selective breeding studies (Hitzemann
et al., 1991; Qian et al., 1992; 1993).
1903_C025.fm Page 376 Wednesday, July 26, 2006 7:09 PM
80
40
20
0 2 4 6 8 10
Haloperidol-Induced Catalepsy
(ED50 mg/kg)
FIGURE 25.4 The relationship between haloperidol-induced catalepsy (ID# 10336) and
ventral midbrain D2 dopamine receptor density (ID# 10270) in a panel of the BXD recom-
binant inbred strains. Details as in the legend to Figure 3. Receptor density data taken from
Jones et al. (1999).
-2
-4
-6
-8
-10
-12
-14
1 2 3 4 5 6
Dopamine Transporter Density
(BMAX - pmoles/mg protein)
FIGURE 25.5 The relationship between striatal dopamine transporter density (ID# 10234)
and cocaine (30 mg/kg)-induced changes of vertical rearing movements (ID# 10310) in a
panel of the BXD recombinant inbred strains. Transporter density data taken from Janowsky
et al. (2001) and the cocaine data are taken from Jones et al. (1999).
18000
Cocaine - Induced Locomotor Activity
16000
14000
(cm traveled/15min)
12000
10000
8000
6000
0
1 2 3 4 5 6
Dopamine Transporter Density
(BMAX - pmoles/mg protein)
FIGURE 25.6 The relationship between striatal dopamine transporter density (ID# 10234)
and cocaine (40 mg/kg)-induced changes in locomotor activity (ID# 10224) in a panel of the
BXD recombinant inbred strains. Cocaine data are taken from Phillips et al. (1998).
1903_C025.fm Page 378 Wednesday, July 26, 2006 7:09 PM
cocaine (40 mg/kg) response was also indexed as a difference score. Figures 25.5
and 25.6 illustrate a familiar pattern in behavioral researchas some behaviors
increase, others decrease. Figures 25.5 and 25.6 also illustrate another theme of this
chapter, the marked genetic variation within the brain dopamine system; DAT den-
sity, here calculated as the BMax differs by more than 300% among the RI strains.
This variation is not significantly associated with the number of DA neurons in the
ventral tegmental area (r = 0.13) (ID# 10223) or the substantia nigra zona compacta
(r = 0.07) (ID# 10224) (data not shown). Here it should be noted that unlike DAT
density, among the RI strains variation in the number of dopamine neurons is
relatively modest (~40%).
Measures of RI strain striatal D2 receptor density in the WebQTL database were
obtained independently in two studies (Jones et al., 1999; Hitzemann et al., 2003)
that used different techniques and ligands to assess receptor binding. Given that
there were only 14 RI strains in common to both studies, the agreement between
the studies is relatively good. Figure 25.7 illustrates the agreement for the dorsal
striatum (r = 0.67, p < 0.008) (ID# 10254 vs. ID# 10220); the strains with low and
high density were readily identified in both studiesthe strains with intermediate
receptor density were not as reliable. Also note the marked variation in receptor
density among the RI strains: ~100% in Hitzemann et al. (2003) and >300% in Jones
et al. (1999). Using the binding data for the dorsomedial striatum (Hitzemann et al.,
2003), the database was queried as was done for DAT. The most significant behav-
ioral correlation (r = 0.74, p < 0.0004) was for a cocaine (45 mg/kg) response (here
400
D2 Receptor Density - (fmoles/mg)
350
300
r = 0.67 ( p < 0.008)
250
200
150
100
50
0
160 180 200 220 240 260 280 300 320
D2 Receptor Binding (fmoles/mg)
10000
Cocaine - Induced Locomotor Activity
9000 r = 0.74 ( p < 0.004)
(cm traveled/15min)
8000
7000
6000
5000
4000
160 180 200 220 240 260 280 300 320
FIGURE 25.8 The relationship between striatal D2 dopamine receptor density (ID# 10220)
and cocaine (45 mg/kg)-induced locomotor activation (ID# 10318) in a panel of the BXD
recombinant inbred strains. Receptor binding data are taken from Hitzemann et al. (2003);
the cocaine response data are taken from Jones et al. (1999).
a difference score measure of locomotor activity- ID# 10318) (Jones et al., 1999)
(Figure 25.8); the analogous data for Phillips et al. (1998)see abovewas also
significantly correlated to receptor density (r = 0.55, p < 0.03) (data not shown).
Overall, the examples cited above illustrate how it is possible to use the RI
strains to dissect the genetic relationships among dopamine systems and behavioral
phenotypes. Importantly, it was found that individual cocaine responses are associ-
ated with different aspects of the brain dopamine system.
receptor density and gene expression from the WebQTL database are compared, the
results are the same: there appears to be no relationship (see, e.g., Figure 25.9).
Thus we again conclude that most of the variance in receptor density is asso-
ciated with some factor or factors different from Drd2 expression. Mapping all 12
measures of D2 receptor density revealed moderate associations (p < 0.005) on Chr
15 and Chr 16 but none on Chr 9 near the Drd2 locus (data not shown); in contrast,
mapping Drd2 expression reveals a moderate signal on Chr 9 near the gene locus
(LOD ~2) and weaker signals on other chromosomes but none on Chr 15 and 16
(data not shown).
Given the mismatch between receptor density and gene expression, the question
arises as to whether Drd2 gene expression data per se is of some value to our
understanding of the relationships among genetics, brain dopamine systems and
behavior. We answer this question positively for several reasons. First, the possibility
exists that receptor density and gene expression do tract together in some brain
region that has not been measured. Areas that come to mind with potentially very
salient behavioral effects include the prefrontal cortex, the central nucleus of the
amygdala and the bed nucleus of the stria terminalis. Our attempts to measure
receptor binding in these areas with quantitative autoradiography have (to date) not
been successful. Second, as noted above, there is evidence that Drd2 expression is
important for explaining the receptor gradients within the brain and this perhaps is
part of a larger pattern of brain organization that is important to the issue at hand
genetics, dopamine and behavior. Third, there is compelling evidence to show that
320
300
D2 Receptor Binding (fmoles/mg)
280
260
240
220
200
180
r = 0.05 ( p < 0.5)
160
140
100 150 200 250 300 350 400
Drd2 Expression
(relative units)
FIGURE 25.9 The relationship between Drd2 expression and D2 dopamine receptor density
(ID# 10220) in a panel of the BXD recombinant inbred strains. Receptor binding data are
taken from Hitzemann et al. (2003); Drd2 expression data are from the Affymetrix U74A
array and are posted at the WebQTL site (www.WebQTL.org).
1903_C025.fm Page 381 Wednesday, July 26, 2006 7:09 PM
Drd2 expression is strongly associated with behaviors for which there is strong
evidence for dopamine modulation (Hitzemann et al., 2003). A query to the WebQTL
database for phenotypes correlated to Drd2 expression reveals that the strongest
association (r = 0.74, p < 0.00004) is for ethanol-induced conditioned place prefer-
ence (CPP)time spent on the ethanol paired floor (ID# 10090) (Cunningham,
1995). CPP is considered to be a measure of a drugs hedonic or rewarding properties.
These data are illustrated in Figure 25.10. Other behavioral phenotypes within the
top 10 correlations include morphine two-bottle choice (ID# 5968) (r = 0.60,
p < 0.003) and ethanol two-bottle choice (ID# 10479). Although the database is
over-populated with drug-related phenotypes, the chance that 3 drug reward/con-
sumption phenotypes would appear in the top 2% of all possible correlations is very
small. The possibility must be considered that the associations with Drd2 expression
result from the close linkage of Drd2 to other genes. In this regard, we have
previously reported that ethanol two-bottle choice is highly correlated to Ncam
expression (Hitzemann et al., 2003); Ncam is 96 kb from Drd2.
An alternative hypothesis is that Drd2 is coexpressed with a family of genes and
that it is the collective action of the gene family which modulates behavior. To test
the first component of this hypothesis we turned to a C57BL/6J DBA/2J F2 brain
gene-expression dataset. The data set contains expression results (Affymetrix 430
A and B arrays) for 56 F2 animals formed from the reciprocal F1 hybrids and is
balanced for males and females. Genotypic data for this F2 dataset have been obtained
for >300 microsatellite markers (or on average 1 marker/5 cM). Both the genotypic
80
70
65
60
55
50
45
100 150 200 250 300 350 400
Drd2 Expression
(relative units)
FIGURE 25.10 The relationship between Drd2 expression and ethanol conditioned place
preference (ID# 10090) in a panel of the BXD recombinant inbred strains. See legend to
Figure 25.9 for details on expression data. Conditioned place preference data are taken from
Cunningham (1995).
1903_C025.fm Page 382 Wednesday, July 26, 2006 7:09 PM
and phenotypic data have been posted at the WebQTL site. Drd2 expression was
significantly (p < 106) associated with 38 unique transcripts (Table 25.1); assuming
a total of 30,000 unique transcripts on both the A and B arrays, one would predict
only 1 chance association. The SymAtlas database at http://www.gnf.org was queried
to determine if the transcripts listed in Table 25.1 were specifically expressed in the
striatum. For some transcripts data were available only for human brain expression.
Expression in the striatum was rated on a four-point scale ranging from under-
expressed to markedly and uniquely over-expressed in the striatum (compared to
other brain regions). Fourteen of the 38 transcripts (37%) were moderately to sig-
nificantly over-expressed in the striatum and these were by and large the transcripts
most significantly associated with Drd2 expression. Figure 25.11 illustrates the
correlation between Drd2 and preproenkephalin (Penk1) expression and Drd2 and
regulator of G-protein signaling 9 (Rgs9) expression. Drd2 expression varied 49%
across the F2 sample while Penk1 and Rgs9 expression varied 94 and 73%, respec-
tively. It should be noted (Table 25.1) that none of the genes significantly associated
with Drd2 expression were significantly differentially expressed between the B6 and
D2 strains. There are numerous implications to this paradox but importantly it
suggests that screening gene lists for significant differences between the parental
strains will significantly underestimate the number of genes showing differential
expression in a segregating population.
The data in Figure 25.11 were collected from a whole brain sample; the question
arises as to whether similar results would be obtained for striatal gene expression.
Striatal data for an F2 sample are not currently available; however, striatal data are
available for a subset of the BXD RI strains. Figure 25.12 illustrates the correlation
between Drd2 and Penk1 striatal expression in the RI strains; the correlation (r =
0.6) is similar to that seen in the F2 sample, likely reflecting the balance between
the increased sample power in the F2 sample vs. the increased specificity of the
striatal sample.
One could argue that strongest associations seen in Table 25.1 are simply an
artifact related to differences in the size and/or number of cells in the striatum. Since
only half the brain was used for the expression analysis, it was possible to determine
if striatal size covaried with expressionno significant association was detected
(data not shown). To determine if there is a relationship between gene expression
and cell density requires a detailed stereological analysis. However, data are available
in the BXD RI series for the number of cholinergic neurons in the striatum (Dains
et al., 1996). A significant correlation (r = 0.61, p < 0.001) was detected between
Drd2 expression and the number of cholinergic neurons (ID# 10106) but only in the
most rostral aspect of the striatum (data not shown); in more caudal areas, no
significant correlation was detected but it should be noted that gene expression was
not measured in a parallel fashion to measures of the number of cholinergic neurons.
Thus, in lieu of other data, we conclude that cell density may partially account for
the associations seen in Table 25.1.
Walker et al. (2004) have analyzed a rat brain gene expression dataset using a
strategy somewhat similar to that described here. These authors asked the question
what genes are coexpressed with Drd1. The results of their analysis (summarized
in figure 2 of Walker et al., 2004) indicated that there was a large group of genes
1903_C025.fm Page 383 Wednesday, July 26, 2006 7:09 PM
TABLE 25.1
Transcripts Significantly (p < 106) Associated with Drd2 Expressiona
Affymetrix ID# Array Gene Symbol r Striatal Expression
TABLE 25.1
Transcripts Significantly (p < 106) Associated with Drd2 Expressiona
(Continued)
Affymetrix ID# Array Gene Symbol r Striatal Expression
9.2
9.1 r = 0.77 ( p < 10-6)
Rgs9 Expression
9.0
8.9
8.8
8.7
8.6
10.3
10.2
Penk1 Expression
18.0
17.6
(striatum)
17.4
17.2
17.0
16.8
10.0 10.2 10.4 10.6 10.8 11.0 11.2 11.4 11.6
Drd2 Expression
(striatum)
FIGURE 25.12 The relationship between striatal Drd2 expression and striatal Penk1 expres-
sion in panel of the BXD recombinant inbred strains. Expression data were obtained using
the Affymetrix 430 2.0 array; data provided by Rosen G. and Williams R. (unpublished
observations).
with a pre-compiled background set. The background data set was compiled by iden-
tifying all strict one-to-one human/mouse orthologs from the Ensembl database. The
orthologs were aligned and the conserved noncoding regions were identified. The
conserved regions 5 kb upstream and downstream from the transcription start site are
scanned for common transcription factor sites among a group of coexpressed genes.
Some genes from a coexpressed group are excluded from the analysis, e.g., if the
conserved promoter regions could not be identified in a satisfactory manner. Of the
14 striatal coexpressed genes noted above, 7 were included in the analysis: Drd2, Rgs9,
Pde1b, Nptp, Penk1, CopineV and Adora2a. The transcription factor binding site most
enriched in this group (as compared with the reference group) was NF-kappaB [NF-
] (p < 0.0001); four of the top ten enriched motifs were from the Rel/NF-kappaB
family. These data cannot be unexpected. Previous studies have shown that both Drd2
and Penk1 expression are regulated by NF- (ONeill and Kaltschmidt, 1997; Fioren-
tini et al., 2002). Other genes expressed in brain and known to be regulated by NF-
include neural cell adhesion molecule (Simpson et al., 2000), inducible nitric oxide
synthase (Madrigal et al., 2001), -opioid receptors (Kraus et al., 2003), brain derived
neurotrophic factor (Lipsky et al., 2001), calcium/calmodulin dependent protein kinase
II (Kassed et al., 2004) and amyloid presursor protein (Grilli et al., 1995). Thus, we
must conclude that although the striatal coexpressed genes all have the NF- motif,
the cause of the striatal specific expression must be associated with other factors.
However, as noted by Meffert and Baltimore (2005), there is increasing evidence that
the NF- family of transcription factors are involved in the regulation of neuronal
activity-dependent transcription and behavior.
25.6 CONCLUSIONS
The intent of this chapter was to provide the interested reader with an introduction
to the topic of genetics, brain dopamine systems and behavior. The focus was almost
entirely on data collected in mice; however, some data suggest that these results can
be extrapolated to other species, including man. Key findings which have been
replicated numerous times are (a) the marked genetic differences in behavioral
responses to dopamine agonists and antagonists, (b) the marked genetic variation in
at least some elements of the brain dopamine systemsnotably the D2 dopamine
receptor and the dopamine transporter and (c) the relationships between a and b.
The reasons for the receptor and transporter variation(s) are unclear but some strat-
egies for dissecting these remarkable genetic effects are presented.
ACKNOWLEDGMENTS
The author would like to thank the many students, fellows and collaborators that
have contributed in some way to the research described in this chapter. Special thanks
to Glen Rosen, Rob Williams and colleagues for the BXD recombinant inbred striatal
gene expression data. This work was supported in part by grants from the US Public
Health ServiceAA 11034, AA 13484 and MH 51372and support from the
Department of Veterans Affairs Medical Research Program. The Gene Network
WebQTL site featured prominently in the review was funded in part by MH P20-
62009 to Rob Williams.
1903_C025.fm Page 387 Wednesday, July 26, 2006 7:09 PM
REFERENCES
Chesler, E.J. et al., WebQTL: rapid exploratory analysis of gene expression and genetic
networks for brain and behavior, Nat Neurosci, 7, 485, 2004.
Cunningham, C.L., Localization of genes influencing ethanol-induced conditioned place pref-
erence and locomotor activity on BXD recombinant inbred mice, Psychopharmacol,
120, 28, 1995.
Dains, K., Hitzemann, B., and Hitzemann, R., Genetics, neuroleptic response and the orga-
nization of cholinergic neurons in the mouse striatum, J Pharmacol Exp Ther, 279,
1430, 1996.
Fink, J.S., Swerdloff, A., and Reis, D.J., Genetic control of dopamine receptors in the mouse
caudate nucleus: relationship on cataleptic response to neuroleptic drugs, Neurosci
Lett, 32, 301, 1982.
Fiorentini, C. et al., Nerve growth factor regulates dopamine D2 receptor expression of
prolactinoma cell lines via p75NGFR-mediated activation of nuclear factor -B, Molec-
ular Endocrinology, 16, 353, 2002.
Fowler, S.C., Zarcone, T.J., and Vorontsova, E., Haloperiodol-induced microcatalepsy differs
in CD-1, BALB/c, and C57BL/6 mice, Exp Clin Psychopharmacol, 9, 277, 2001.
Grilli, M. et al., Identification and characterization of a kappa B/Rel binding site in the
regulatory region of the amyloid precursor protein gene, J Biol Chem, 270, 26774,
1995.
Hitzemann, R.J. et al., On the selection of mice for haloperidol response and non-response,
Psychopharmacol, 103, 244, 1991.
Hitzemann, R. et al., Genetics and the organization of the basal ganglia, Int Rev Neurobiol,
7, 485, 1995.
Hitzemann, R. The regulation of D2 dopamine receptor expression, Mol Psych, 3, 198, 1998.
Hitzemann, R. et al., Dopamine D2 receptor binding, Drd2 expression and the number of
dopamine neurons in the BXD recombinant inbred series: genetic relationships to
alcohol and other drug associated phenotypes, Alcohol Clin Exp Res, 27, 1, 2003.
Hwang, C.K. et al., Dopamine receptor regulating factor, DDRF: a zinc finger transcription
factor, Proc Natl Acad Sci USA, 98, 7558, 2001.
Janowsky, A. et al., Mapping genes that regulate density of dopamine transporters and
correlated behaviors in recombinant inbred mice, J Pharmacol Exp Ther, 298, 634,
2001.
Jones, B.C. et al., Quantitative-trait loci analysis of cocaine-related behaviors and neurochem-
istry, Pharmacogenetics, 9, 607, 1999.
Kanes, S., Hitzemann, B.A., and Hitzemann R.J., On the relationship between D2 receptor
density and neuroleptic-induced catalepsy among eight inbred mouse strains, J Phar-
macol Exp Ther, 267, 538, 1993.
Kanes, S. et al., Mapping the genes for haloperidol-induced catalepsy, J Pharmacol Exp Ther,
277, 1016, 1996.
Kassed, C.A. et al., Injury-induced NF-B activation in the hippocampus: implications for
neuronal survival, FASEB J, 18, 723, 2004.
Kraus, J. et al., The role of nuclear factor kappaB in tumor necrosis factor-regulated tran-
scription of the human :-opioid receptor gene, Mol Pharmacol, 64, 876, 2003.
Lee, S-H., Minowa, M.T., and Mouradian, M.M., Two distinct promoters derive transcription
of the human D1A dopamine receptor gene, J Biol Chem, 271, 25292, 1996.
Lee, S-H. et al., Tissue-specific promoter usage in the D1A dopamine receptor gene in brain
and kidney, DNA Cell Biol, 16, 1267, 1997.
1903_C025.fm Page 388 Wednesday, July 26, 2006 7:09 PM
Lee, S-H. et al., Genomic organization and promoter characterization of the murine dopamine
receptor regulator factor (DRRF) gene, Gene, 304, 193, 2003.
Lipsky, R.H. et al., Nuclear factor kappaB is a critical determinant in N-methyl-D-aspartate
receptor-mediated neuroprotection, J Neurochem, 78, 254, 2001.
Madrigal, J.L. et al., Inducible nitric oxide synthase expression in brain cortex after acute
restraint stress in regulated by nuclear factor B-mediated mechanisms, J Neurochem,
76, 532, 2001.
Meffert, M.K. and Baltimore, D., Physiological function for brain NF-kappaB, Trends in
Neurosci, 28, 37, 2005.
Minowa, M.T., Characterization of the 5' flanking region of the human D1A dopamine receptor
gene, Proc Natl Acad Sci USA, 89, 3045, 1992a.
Minowa, M.T., Characterization of the 5' flanking region of the human D2 dopamine receptor
gene, Biochemistry, 31, 8389, 1992b.
Minowa, M.T., Minowa, T., and Mouradian, M.M., Activator region analysis of the human
D1A dopamine receptor gene, J Biol Chem, 268, 23544, 1993.
Minowa, M.T., Minowa, M.T., and Mouradian, M.M., Negative modulator of the rat D2
dopamine receptor gene, J Biol Chem, 269, 11656, 1994.
ONeill, L.A.J. and Kaltschmidt, C., NF-kappaB: a crucial transcription factor for glial and
neuronal cell function, Trends Neurosci, 20, 252, 1997.
Patel, N.V. and Hitzemann, R.J., Detection and mapping of quantitative trait loci for halo-
peridol-induced catalepsy in a C57BL/6J x DBA/2J F2 intercross, Behav Genet, 29,
303, 1999.
Phillips, T.J., Huson, M.G., and McKinnon, C.S., Localization of genes mediating acute and
sensitized locomotor responses to cocaine in BXD-By recombinant inbred mice, J
Neurosci, 18, 3023, 1998.
Qian, Y., Hitzemann, B., and Hitzemann, R., D1 and D2 dopamine receptor distribution in
the neuroleptic non-responsive (NNR) and neuroleptic responsive (NR) lines of
micea quantitative receptor autoradiographic study, J Pharmacol Exp Therap, 261,
341, 1992.
Qian, Y. et al., D1 and D2 dopamine receptor turnover and D2 mRNA levels in the neuroleptic
responsive (NR) and neuroleptic non-responsive (NNR) lines of mice, J Pharmacol
Exp Ther, 267, 1582, 1993.
Simpson, C.S. and Morris, B.J., Regulation of neuronal cell adhesion molecule expression
by NF-kappaB, J Biol Chem, 275, 16879, 2000.
Taylor, B.A., Recombinant inbred strains: use in gene mapping, in Origins of Inbred Mice,
Morse, H.C., Ed., Academic Press, New York, 1978, pp 428438.
Volkow, N.D. et al., Decreased dopamine-D(2) receptor availability is associated with reduced
frontal metabolism in cocaine abusers, Synapse, 14, 169, 1993.
Volkow, N.D. et al., Prediction of reinforcing responses to psychostimulants in humans by
brain dopamine D2 receptor levels, Am J Psych, 156, 1440, 1999.
Walker, J.R. et al., Applications of a rat multiple tissue gene expression data set, Genome
Research, 14, 742, 2004.
Yang, Y. et al., Three-amino acid extension loop homeodomain proteins Meis2 and TGI
differentially regulate transcription, J Biol Chem, 275, 20734, 2000a.
Yang, Y. et al., ZIC2 and Sp3 repress Sp1-induced activation of the human D1A dopamine
receptor gene, J Biol Chem, 275, 38863, 2000b.
Zhang, L., Miles, M.F., and Aldape, K.D., A model of molecular interactions in short oligo-
nucleotide array, Nat Biotechnol, 21, 818821, 2003.
1903_C026.fm Page 389 Wednesday, July 26, 2006 7:12 PM
CONTENTS
ABSTRACT
This chapter analyzes the natural genetic variability of the so-called intra/infrapy-
ramidal mossy fiber (IIP-MF) projection in mice and rats, which terminates upon
the basal dendrites of hippocampal pyramidal cells, and its relations to behavior.
The analysis included thus far the following steps: (i) identification of structural
traits sensitive to selective breeding for extremes in two-way avoidance, (ii) testing
389
1903_C026.fm Page 390 Wednesday, July 26, 2006 7:12 PM
the robustness of the associations found by studying individual and genetic corre-
lations between hippocampal traits and behavior, (iii) confirming causal relationships
by manipulating the structural variable in inbred (isogenic) strains, thereby elimi-
nating the possibility of genetic linkage, (iv) further ruling out the possibility of
spurious associations by studying the correlations between the hippocampal trait and
other behaviors known to depend on hippocampal functioning, (v) searching for
behavioral correlates of the IIP-MF variation in behaviors thought to be unrelated
to hippocampal function, (vi) testing the effects of natural selection on differential
IIP-MF distributions in naturalistic environments, and (vii), studying the variability
of the IIP-MF distribution in small mammals with differential lifestyles. Taken
together, the data imply that the hippocampus mediates a considerable number of
behaviorally distinct processes, a view in accordance with the connectivity of this
structure. The common denominator may be that variations of the IIP-MF entail
differences in the stability of parallel hippocampal processing.
The advantages of this approach are: (i) hippocampal traits can be manipulated
simply by controlled breeding (no manipulative approach can match this experimen-
tal finesse thus far); (ii) it identifies natural regulation sites for behavior; and (iii) it
offers an easy way to discover possible behavioral functions of the hippocampus.
The disadvantages are: (i) the study of the underlying causality depends on the
detection of genes responsible for the trait; (ii) the behavioral effects are typically
smaller than after invasive manipulations, and this entails a greater sensitivity to
environmental influences; (iii) correlative studies often face traditionally more skep-
ticism than invasive manipulations; and (iv) the rising costs of maintaining breeding
colonies of mice for classical behavior genetic studies is slowly suffocating this
approach.
26.1 INTRODUCTION
In any mammalian species, no brain is like another, and every individual behaves
differently. Some of this variability obviously is dependent on the environment. On
the other hand, a considerable portion must be caused by biochemical and structural
variations of the brain that originate from both exogenous influences and actions of
genes during brain development. This chapter will deal with the biology of traits,
meaning the study of individual structural or behavioral differences that lie within
the quantified normal range of a population, in contrast to abnormal, i.e., patholog-
ical, differences that are most obvious in so-called neurological mutants.
Since behavioral traits appear to be complex, they have been attributed to many
genes, more precisely, to mutated alleles, which were thought to contribute small
effects adding to a phenotype. More recently, gene-targeting techniques allow for
selective inactivation of genes thought to be important for memory and learning and
other aspects of behavior. Such studies have revealed a remarkable capacity of the
genome and the brain to compensate for targeted mutations. However, such com-
pensation is variable, and this variable penetrance of targeted deletions often results
in changes resembling a quantitative trait. Hence, it appears reasonable to assume
that at least part of the observed behavioral variability of genetic origin is not based
1903_C026.fm Page 391 Wednesday, July 26, 2006 7:12 PM
on the additive action of hundreds of genes but on the effects of a few major genes,
and, perhaps, on single locus effects.
A theoretical problem is whether such single locus effects could be specific for
behavioral talents and motivational traits. Since activation of genes is thought to
precede brain development, this would seem unlikely. Most genes regulating brain
development are probably affecting many different neuronal populations, and their
effects must thus be pleiotropic. Likewise, mutation of genes controlling the basic
physiology of neurons and glial cells must show pleiotropic effects. There is an
exception however. Genes that are activated late during development can modulate
only the final steps of brain differentiation,1 a period during which only a few
forebrain systems undergo their final formation. The known late-forming systems
include portions of the olfactory bulb, the hippocampal formation, in particular the
dentate gyrus, and parts of the cerebellum. Thus, late-acting alleles that influence
the connective organization, receptor regulation, and neurochemistry of these struc-
tures are likely to be fairly specific for learning, cognition, and motivation. Such
mutated alleles will cause behavioral changes not because of a particular nucleic
acid sequence but by virtue of timing. Another mechanism for behavioral selectivity
is possible if brain structures known to modulate behavior show morphological or
biochemical features not found elsewhere in the brain.
FIGURE 26.1 Schematic view of a cross-section of the rodent hippocampal formation. The
terminal fields of the hippocampal mossy fiber projection appear in black. Abbreviations:
CA1, CA3, and CA4, hippocampal subfields; FIM, fimbria fornicis; GC, granule cells; LM,
stratum lacunosum-moleculare; MOL, molecular layer of dentate gyrus; OR, stratum oriens;
PYR, stratum pyramidale; RAD, stratum radiatum; SGL, supragranular layer of dentate gyrus.
appears to be a promising candidate structure for testing whether its variability might
be related to hippocampus-dependent behaviors.
FIGURE 26.2 Selective breeding for extremes in two-way avoidance learning of rats and
differentiation of the mossy fiber projections (modified after Lipp and Wolfer, 1995).3 Note
that the percentage of the IIP-MF in Figs. 26.3, 26.4, and 26.8 have been calculated in relation
to the entire area of CA3/CA4 as measured by manual planimetry used for the older studies.
In more recent publications, the use of digital morphometry permitted to measure only the three
mossy fiber subfields. Hence, Figs. 26.5, 26.6, 26.7, and 26.9 give only the percentage of IIP-
MF in relation to the size of the suprapyramidal mossy fiber layer, correcting in a simple way
for differences in hippocampal size. Since the size of the suprapyramidal MF projection is a
good covariate of the total area of CA3/CA4, the two indices are strongly correlated, however.
either trait. These values can be obtained from publications or can be assessed by
studying a group of individuals. In the case of the IIP-MF and two-way avoidance
learning, it was found that the rank orders of the two variables showed a very strong
negative correlation [r = 0.96, p < 0.01, (Figure 26.3a)]. Again, poorly avoiding
strains showed the largest IIP-MF projections, but there was no correlation with the
size of the other mossy fiber fields. Genetic correlations between strain means are
useful when the behavioral variable shows large epigenetic (environment-dependent)
variability, blurring the influence of the brain variable. Their weakness is that they
depend on a small number of strains that appear as data points in the correlations:
one outlier strain can annihilate an existing real correlation, or an unlucky choice
of strains can fake it.
This problem can be solved by studying phenotypic correlations, that is, corre-
lations using individuals in which both traits are measured. Two extreme strains in
terms of IIP-MF and two-way avoidance learning were intercrossed, namely DBA/2
and C3H and the resulting F2 generation of 51 animals was then tested and their
hippocampi morphometrized.4 In such a sample, other alleles potentially affecting
two-way avoidance or IIP-MF projections are scrambled during meiosis and there-
fore might mask a weak phenotypic correlation. Nonetheless, the analysis revealed
again a strong negative correlation [r = 0.82, p < 0.001, (Figure 26.3b)]. Thus, the
association between avoidance learning and IIP-MF as observed in rats and inbred
strains was confirmed again, and turned out to be insensitive against potentially
confounding alleles. In other words, the correlation was functionally robust. This
step of the analysis clarified also causal relations: the distribution of the IIP-MF is
established in the postnatal period between days 10 to 20 and remains rather stable
afterward. Because the variability of the IIP-MF was obtained experimentally (by
means of meiotic scrambling of alleles), and because the effects of this manipulation
predict adult behavior, one is no longer dealing with a simple correlation without
causal implications (for example length of arm vs. length of leg), but with a regres-
sion of behavior on a structural variable. Nonetheless, phenotypic correlations have
two weak points. One is genetic linkage (more precisely, linkage disequilibrium),
in which case a chromosomal segment may contain both an allele relevant for
avoidance learning and another one for the size of the IIP-MF projection. One may
note that this is also a problem in gene targeting.5 A more subtle point is that
producing hybrids between strains with behavioral extremes often results in hybrid
vigor: the filial generations outperform both parental strains. The reasons for hybrid
vigor are unknown, but it may improve complex learning behavior to a point where
it can mask strain-specific factors. See also Wolfer et al. for an example of this
problem in knockout mice.6
FIGURE 26.3 Variations of the infrapyramidal mossy fiber projection and avoidance learning in mice (modified after Lipp et al., 1989).4 (a) genetic
correlation between the strain means of the IIP-MF projection and avoidance learning at the last day of training (day 5); (b) phenotypic correlation
between IIP-MF projection and avoidance learning in individual mice from an F2 cross between the inbred strains DBA/2 and C3H; (c) developmentally
Natural Genetic Variation of Hippocampal Structures and Behavior
induced variation of the IIP-MF by means of thyroxin injections and adult two-way avoidance learning.
395
1903_C026.fm Page 396 Wednesday, July 26, 2006 7:12 PM
MFs. This treatment was applied to pups of the mouse strain DBA/2, a good
avoidance learner with scanty IIP-MF projections. It resulted in considerable, par-
tially dose-dependent variability of the IIP-MF projection that remained strongly
and negatively correlated with the avoidance learning of the adult individuals [r =
0.75, p < 0. 00014 (Figure 26.3c)]. Because all animals of this strain are isogenic,
chromosomal linkage cannot account for the correlation between the two traits.
Rather, it appears that size variations of the IIP-MF predict two-way avoidance
learning, regardless of whether the structural variation has been caused by genetic
or epigenetic factors. A further confirmation was then found by the group of Crusio
reporting subline differentiation in radial maze learning (that is, other hippocampus-
dependent behavior, see below) in the mouse strain C57BL/6.7 This difference was
found to be associated with a difference in the extent of the IIP-MF projection. Since
the sublines can differ at only a few loci at best, these findings are a strong argument
against chromosomal linkage.
FIGURE 26.4 (a) Genetic correlation between IIP-MF and radial maze learning (strain means S.E.M.); (b) variability of the IIP-MF projection as
observed after postnatal saline (sham) injections. Note that the total error scores are higher than those of the thyroxin-injected mice, and that a correlation
Natural Genetic Variation of Hippocampal Structures and Behavior
can be found; (c) reduced error scores of mice having received postnatal injections of thyroxin showing a similar correlation (modified after Lipp and
Wolfe, 1995).3
397
1903_C026.fm Page 398 Wednesday, July 26, 2006 7:12 PM
FIGURE 26.5 Activity and IIP-MF in mice selectively bred for differential open-field activ-
ity. Highly active mice show larger IIP-MF projections.10 Abbreviations: HI, two mouse lines
selectively bred for high activity in the open field, both of them pigmented; LO, two mouse
lines selectively bred for low activity in the open field, both of them albinotic; CTL, two
randomly bred mouse lines, except that one was selected for pigmented coat color, the other
for albinism.
In the Morris water maze, phenotypic correlations were found, too.11,12 However,
the variation of the IIP-MF was correlated not with the learning of the task, but with
the ability of relearning a new platform location, which was superior in mice with
extended IIP-MF projections (Figure 26.6). Curiously, the correlation appeared to
be better in the left hippocampus. Subsequent studies showed that the asymmetry
of the IIP-MF projection was contributing to this kind of performance as well; mice
with larger IIP-MF projections at left crossed over the former platform position more
during the first day of reversal, but reoriented faster toward the new platform location
the following day. The worst relearners were animals with small symmetrical IIP-
MF projections.11,12
These studies indicated that variations of the IIP-MF projections (or of an
intrahippocampally linked structural covariate) were probably associated with a
physiological process inside the hippocampus causing behavioral differences
in spatial learning and exploration. On the other hand, the results did not blend
well into unitary concepts of hippocampal function, such as spatial reference13 vs.
working memory.14 Depending on the task, they appeared to correlate once with
working memory, once with spatial reference memory, and seemed to be also
involved in behavioral flexibility and perhaps hemispheric lateralization.
1903_C026.fm Page 399 Wednesday, July 26, 2006 7:12 PM
FIGURE 26.6 Correlation between IIP-MF and escape performance from the Morris water
maze in the left hippocampus of 19 mice from a randomly bred stock generated by means of
a 4-way diallel cross. Day 5 is the second day after platform reversal. Note superior perfor-
mance of mice with enlarged IIP-MF projections. Along the ordinate typical swim paths
(modified after Bernasconi-Guastalla et al. 1994).12
FIGURE 26.7 IIP-MF variations and strength of paw preference. (a) Mice with strong paw
preference (HI) show the largest IIP-MF projections, the smallest are found in mice with
weak paw preference. Random-bred controls (RND-2) lie in-between; (b) weak genetic (strain
mean) correlation between IIP-MF and strength of paw preference in inbred strains of mice
(modified after Lipp et al., 1996).15
FIGURE 26.8 IIP-MF variations and attack latencies toward an intruder into the home cage.
(a) Mice selectively bred for long attack latencies show larger IIP-MF projections;16 (b)
percentage of attacking males is higher in mouse strains with scanty IIP-MF projections
(modified after Guillot et al., 1994).17
made carefully when the changes in MF distribution in mutant mice are subtle and
have been assessed only in a limited portion of the hippocampus. The extent of the
IIP-MF decreases gradually from rostral (septal) to ventral (temporal) levels of the
hippocampus, while recurrent MF collaterals in the dentate granule cell layer are
abundant in ventral parts and almost missing in the rostral portion of the hippocam-
pus. Thus, subtle mutation-induced changes in hippocampal size and form can fake
a difference in MF distribution when compared at one hippocampal plane only.
FIGURE 26.9 Reduction of the IIP-MF projection in two demes of feralized mice after 2
years in Russian outdoor pens as compared with control mice kept indoors. The time span
includes 7 to 8 generations of mice. Note that the reduction in pen 2 was observed after 1
year, while it took 2 years in the other pen.
1903_C026.fm Page 405 Wednesday, July 26, 2006 7:12 PM
ecological relevance. The mice had to live in the shelters throughout the year.
Intentional exploration of the pen territory (as expected for carriers with expanded
IIP-MF projections) bears a high risk of predation by owls, while the social structure
in the shelter (a few highly aggressive dominant males and many quickly escaping
subordinates) favors again high behavioral reactivity (either for flight or fight), at
least in males. As the lifestyle in these types of pens is not demanding in terms of
spatial abilities and complex learning, the lacking selection effects on performance
in classic cognitive tests such as the water maze is not unsurprising. Most likely,
other genetically dependent brain traits supporting the observed phenotype were
subject to selection as well. Nonetheless, the speed of genetic adaptation of the IIP-
MF projection implies that selection acted upon a few major loci only.
These results received support from studying the hippocampi of small wild
mammals in Russia and elsewhere. From the results of experimental natural selec-
tion, it could be predicted that small rodent species living in ecological niches
requiring high behavioral reactivity or readiness for flight would show small IIP-
MF projections. An example of such adaptation is given in Figure 26.10, which
shows large IIP-MF projections in the bank vole (Clethrionomys glareolus), a species
adapted to a wide range of different habitats, and extremely reduced IIP-MF pro-
jections in the root vole (Microtus oeconomus), living in homogeneous grassland
habitats with small home ranges. Interestingly, the two species showed about equal
performance in water maze learning but used very different strategies, the bank voles
showing controlled spatial searching behavior, the root voles searching inefficiently
but swimming at high speed, reacting to any change in the setup, such as platform
reversal, with frantic but ill-directed swimming.29 Other studies showed that wood
mice (Apodemus ssp.), known for roaming large territories, have large IIP-MF
projections as well, which were also found in patrolling insectivores (Sorex ssp.).
On the other hand, species living chiefly underground, such as the vole Microtus
subterraneus and the mole (Talpa europea), have shown much reduced or missing
IIP-MF projections.
FIGURE 26.10 IIP-MF projections in wild voles. (a) Root vole (Microtus oeconomus), living
in small monotonic habitats and showing rapid yet chaotic swim strategies in the water maze;
(b) Bank vole (Clethrionomys glareolus), a species capable of adapting to many different
habitats and showing predictable and efficient learning in the water maze (modified after
Pleskacheva et al. 2000).29
1903_C026.fm Page 406 Wednesday, July 26, 2006 7:12 PM
to identify by molecular techniques those major genes that determine the genetic
variation of the IIP-MF projection because these regulatory genes might act locally
and their products occur in low concentrations during a limited developmental
period. But it is not impossible.
Another point is that such a correlation analysis is time consuming, requiring
many control studies. This is because working with brain traits means working
with relatively minor behavioral differences as compared with the effects of struc-
tural destruction. Minor differences are more sensitive to environmental effects and
genetic interactions. One must notice, however, that the natural genetic differences
in behavior as found between mouse strains are often much larger than those
observed after targeted disruption of genes.31 Thus, the most straightforward
genotype-to-behavioral phenotype approach faces similar problems but tends to
overlook them.
Correlative approaches are often dismissed because of the difficulties in disen-
tangling pleiotropic effects of unknown genes causing the traits in brain and behavior.
These difficulties should not be denied. Anyone familiar with the effects of gene
targeting on memory, learning, and cognition is aware of the fact that there is no
way to avoid pleiotropy of the targeted mutation. Even worse, there is a common
belief that elimination of a particular locus is revealing the function of that gene.
What is really observed, however, is the function of the entire genome without the
targeted locus. Again, the basic problems remain always the samethey are just
perceived differently.
The unique strength of this approach is that it starts with individual, within-
species variations of brain traits and behavior that are sensitive to artificial selection
and manipulation by classical breeding and eventually helps predict species-specific
adaptations. This is a fundamental difference to the traditional comparative approach,
which tries to match a special talent of a species with the particular development of
a brain structure. While fruitful in linking sensory abilities to corresponding devel-
opment of brain parts, this strategy faces problems in matching cognitive abilities
with altered patterns of neuronal circuitry. It would seem that the example described
in this chapter is, thus far, the only rational approach in finding natural regulatory
sites for complex behavior that is not depending on luck and speculation.
Last, one would expect that such a simple and relatively cost-effective approach
might win more followers. Unfortunately, the popularity of the mouse as a main
species for genetic engineering has created a ballooning veterinarian industrial and
administrative complex imposing excessive (and often unnecessary) costs of mouse
care that threaten to suffocate traditional behavioral genetic methods in this species.
With steadily rising costs, the insight may grow that neither classical Mendelian
crosses nor artificial selection for behavioral traits needs pathogen-free conditions
and veterinarian management and supervision.
ACKNOWLEDGMENTS
This article was supported by Swiss National Science Foundation and the NCCR
Neural Plasticity and Repair.
1903_C026.fm Page 408 Wednesday, July 26, 2006 7:12 PM
REFERENCES
1. Lipp, H.-P., Genetic variability, individuality and the evolution of the mammalian-
brain, Behavioural Processes 35, 1933, 1995.
2. Barber, R.P., Vaughn, J.E., Wimer, R.E., and Wimer, C.C., Genetically-associated
variations in the distribution of dentate granule cell synapses upon the pyramidal cell
dendrites in mouse hippocampus, Journal of Comparative Neurology 156, 417434,
1974.
3. Lipp, H.-P., and Wolfer, D.P. New paths towards old dreams: microphrenology or the
study of intact brains in intact worlds. In Behavioural Brain Research in Naturalistic
and Semi-Naturalistic Settings: Possibilities and Perspectives, Alleva, E., Fasolo, A.,
Lipp, H.-P., Nadel, L. and Ricceri, L., eds. (Kluwer, Dordrecht, the Netherlands),
pp. 139, 1995
4. Lipp, H.-P., Schwegler, H., Crusio, W.E., Wolfer, D.P., Heimrich, B., Driscoll, P., and
Leisinger-Trigona, M.-C., Using genetically-defined rodent strains for the identifica-
tion of hippocampal traits relevant for two-way avoidance learning: a non-invasive
approach, Experientia 45, 845859, 1989.
5. Gerlai, R., Gene-targeting studies of mammalian behavior: Is it the mutation or the
background genotype? Trends in Neurosciences 19, 177181, 1996.
6. Wolfer, D.P.P., Mller, U., Stagliar-Bozizevic, M., and Lipp, H.-P., Assessing the
effects of the 129/Sv genetic background on swimming navigation learning in trans-
genic mutants: a study using mice with a modified beta-amyloid precursor protein
gene, Brain Research 771, 113, 1997.
7. Jamot, L., Bertholet, J.Y., and Crusio, W.E., Neuroanatomical divergence between
two substrains of C57BL/6J inbred mice entails differential radial-maze learning,
Brain Research 644, 352356, 1994.
8. Crusio, W.E., Schwegler, H., and Brust, I., Covariations between hippocampal mossy
fibres and working and reference memory in spatial and non-spatial radial maze tasks
in mice, European Journal of Neuroscience 5, 14131420, 1993.
9. Prior, H., Schwegler, H., and Dcker, G., Dissociation of spatial reference memory,
spatial working memory, and hippocampal mossy fiber distribution in two rat strains
differing in emotionality, Behavioural Brain Research 87, 183194, 1997.
10. Hausheer-Zarmakupi, Z., Wolfer, D.P., Leisinger-Trigona, M.-C., and Lipp, H.-P.,
Selective breeding for extremes in open-field activity of mice entails a differentiation
of hippocampal mossy fibers, Behavior Genetics 26, 167176, 1996.
11. Schpke, R., Wolfer, D.P., Lipp, H.-P., and Leisinger-Trigona, M.-C., Swimming
navigation and structural variations of the infrapyramidal mossy fibers in the hippoc-
ampus of the mouse, Hippocampus 1, 315328, 1991.
12. Bernasconi-Guastalla, S., Wolfer, D.P., and Lipp, H.-P., Hippocampal mossy fibers
and swimming navigation in mice: correlations with size and leftright asymmetries,
Hippocampus 4, 5364, 1994.
13. OKeefe, J., and Nadel, L. The Hippocampus as a Cognitive Map, Clarendon Press,
Oxford, 1978.
14. Olton, D.S., and Feustle, W.A., Hippocampal function required for nonspatial working
memory, Experimental Brain Research 41, 380389, 1981.
15. Lipp, H.-P., Collins, R.L., Hausheer-Zarmakupi, Z., Leisinger-Trigona, M.-C., Crusio,
W.E., Nosten-Bertrand, M., Signore, P., Schwegler, H., and Wolfer, D.P., Paw later-
alization and intra/infrapyramidal mossy fibers in the hippocampus of the mouse,
Behavior Genetics 26, 167176, 1996.
1903_C026.fm Page 409 Wednesday, July 26, 2006 7:12 PM
16. Sluyter, F., Jamot, L., Van Oortmerssen, G.A., and Crusio, W.E., Hippocampal mossy
fiber distributions in mice selected for aggression, Brain Research 646, 145148,
1994.
17. Guillot, P.-V., Roubertoux, P.L., and Crusio, W.E., Hippocampal mossy fiber distri-
butions and intermale aggression in seven inbred mouse strains, Brain Research 660,
167169, 1994.
18. Belzung, C., Hippocampal mossy fibres: implication in novelty reactions or in anxiety
behaviours? Behavioural Brain Research 51, 149155, 1992.
19. Roullet, P., and Lassalle, J.-M., Genetic variation, hippocampal mossy fibres distri-
bution, novelty reactions and spatial representation in mice, Behavioural Brain
Research 41, 6170, 1990.
20. Prior, H., Schwegler, H., Marashi, V., and Sachser, N., Exploration, emotionality, and
hippocampal mossy fibers in nonaggressive AB/Gat and congenic highly aggressive
mice, Hippocampus 14, 135140, 2004.
21. Dimitrieva, N.I., Gozzo, S., Dimitriev, Y., and Ammassari-Teule, M., Mossy fiber
distribution in four lines of rats: a correlative study with avoidance abilities and
excitability thresholds, Physiological Psychology 12, 3034, 1984.
22. Lipp, H.-P., Schwegler, H., Heimrich, B., Cerbone, A., and Sadile, A.G., Strain-
specific correlations between hippocampal structural traits and habituation in a spatial
novelty situation, Behavioural Brain Research 24, 111123, 1987.
23. Roullet, P., and Lassalle, J.M., Behavioural strategies, sensorial processes and hip-
pocampal mossy fibre distribution in radial maze performance in mice, Behavioural
Brain Research 48, 7785, 1992.
24. Hoffmann, H.J., Wenkel, R., Schicknick, H., Bernstein, H.-G., and Schwegler, H.,
Hippocampal mossy fiber distribution does not correlate with two-way active avoid-
ance performance in backcross lines derived from inbred mouse strains DBA/2 and
C3H, Brain Research 589, 171174, 1992.
25. Isgor, C., Slomianka, L., and Watson, S.J., Hippocampal mossy fibre terminal field
size is differentially affected in a rat model of risk-taking behaviour, Behavioural
Brain Research 153, 714, 2004.
26. Poletaeva, I.I., Pleskacheva, M.G., Markina, N.W., Perepiolkina, O.W., Scheffrahn,
H., Wolfer, D.P.P., and Lipp, H.-P., [Environmental habitat-related pressure: behav-
ioral alterations and morphological changes in the brain of the house mouse] Russian,
Ecologia 3, 231236, 2001.
27. Rissi, S. Moosfasersystem und Verhalten bei der Hausmaus nach Embryonentransfer.
M.D. thesis, University of Zrich, Zrich (2002).
28. Ceschi, A. Effects of experimental natural selection on exploratory activity and
anxiety in mice: assessment by means of three behavioral paradigms. M.D. thesis,
University of Zrich, Zrich (2004).
29. Pleskacheva, M.G., Wolfer, D.P., Kupriyanova, I.F., Nikolenko, D.L., Scheffrahn, H.,
Dell'Omo, G., and Lipp, H.-P., Hippocampal mossy fibers and swimming navigation
learning in two vole species occupying different habitats, Hippocampus 10, 1730,
2000.
30. Amrein, I. Functional and neuroanatomical correlates of adult neurogenesis in the
dentate gyrus of domesticated and wild rodents. Ph.D. thesis, University of Zrich,
Zrich (2004).
31. Lipp, H.-P., and Wolfer, D.P., Genetic background problems in the analysis of
cognitive and neuronal changes in genetically modified mice, Clinical Neuroscience
Research 3, 221232, 2003.
1903_C026.fm Page 410 Wednesday, July 26, 2006 7:12 PM
CONTENTS
411
1903_C027.fm Page 412 Wednesday, July 26, 2006 7:49 PM
27.1 INTRODUCTION
The operation of the adult mammalian brain involves a large fraction of the
genome.1,2 Upward of 50% of the gene complement is expressed in even single
discrete brain regions such as the cerebellum, striatum, and hippocampus. Transcrip-
tional complexity in the brain is unusually high compared with other organs, and
roughly 40% of all genes are transcribed in two or more alternative forms.3 Devel-
opment adds another major axis of complexity. In small mammals such as mice,
assembling a brain is a beautifully choreographed dance that involves the prolifer-
ation, migration, differentiation, and interconnection of 75 million neurons,
25 million glial cells, and a web of 10 million or more blood vessel cells over a
30-day period. In humans, this process involves three orders of magnitude more
neurons and their supporting cast and takes a decade or two.4 It seems conservative
to estimate that 75% or more of the mammalian genome is engaged at some stage
of making or maintaining the brain.
The major topic of this chapter is the study the genetic influences on brain
organization at various levels. Variations in genes affect either transcription or
function, and both of these can influence brain structure. Brain structure, in turn,
affects behavior both in variations in gross organization and variation in the ways
circuits are formed and utilized. Studying each level of organization is valuable and
gives us the opportunity to treat transcription and morphology in the brain as
intermediate phenotypes and not simply a black box between brain and behavior.
prominently when the comparisons are made among species.711 Of course, the more
specific the region, cell population, or synaptic subcircuit being compared, the more
specific the related behavior differences are likely to be. In songbirds, for example,
variation in the size of neuron populations in the system is correlated with various
features of song learning and production.12,13 A more specific relationship is the
variation in size of the infrapyramidal mossy fibers in the hippocampus, which
correlates with differences in open-field exploration and novelty-induced fear,14 as
well as with avoidance learning.15,16 Other aspects of hippocampal morphology also
have behavioral correlates,1720 a subset of which is likely to be causal. We expect
that numerous other structural variants will be linked to behavioral differences.
to study behavioral and pharmacological traits of relevance to drug and alcohol use
and abuse.2127 Complex trait analysis has also been applied to a wide variety of
CNS divisions and cell populations: total brain weight,28,29 neuron and glial cell
number,30 the size and structure of the hippocampus31 and dentate gyrus,32 olfactory
bulb,33 striatum,34 retina,35 neocortex,36 cerebellum,37 as well as cytoarchitectonic
fields including the barrel cortex.36,38
These and other quantitative trait locus (QTL) mapping studies are intermediaries
between purely reductionist tools such as single-background gene knockouts, and
large, human-genetic studies, which may have much lower power (certainly per
participant!) but whose results, hopefully, are more general. A typical mouse QTL
study falls somewhere between these extremes. It addresses the complexity of genetic
variation but limits its scope to improve the power to detect differences. When
applied to a genetic reference population and to molecular phenotypes, however,
QTL mapping provides an invaluable tool for building observations that can bridge
different levels of complexity.
Instead of studying a small number of isogenic animals or using a single animal
model, we usually study a segregating population that consists of 20 to 2,000
genetically unique individuals or strains. Studies of the genetic basis of common
diseases in human populations often exploit this approach. For example, much of
the population of Iceland has been harnessed to find common gene variants that
contribute to pervasive human diseases. Rodent geneticists have their own, much
simpler, version of a population-based approach that usually involves breeding two
inbred strains known to differ significantly in a trait of biological interest. These
parental strains are used to generate heterozygous F1 offspring. The F1s are used
to make an experimental crosseither an intercross (F2) or a backcross to one of
the parental strains (usually abbreviated N2). Both F2 and N2 populations are
genetically complexno two animals are alike. We therefore refer to these popula-
tions as a segregating cross because gene variants segregate following Mendels
laws. For example, there may be a twofold difference between the parental strains
in the number of dopaminergic neurons in the substantia nigra, horizontal cells in
the retina, or the volume of the hippocampal dentate gyrus.32 The question we would
like to answer with the F2 and N2 progeny is what set of gene variants contribute
to the variation among members of the cross. Can we track down the QTLs that are
jointly responsible for the genetic fraction of the variation?
A quantitative trait can generally be measured continuously, often on a ratio
scale. While an essentially dichotomous Mendelian trait potentially can be converted
into a quantitative measurement, most quantitative traits vary more continuously.
Generally, this is because multiple genetic and environmental factors affect the
phenotype. QTL mapping is a process in which we tease apart the individual and
joint contributions. We test whether there is a statistically significant association
between differences in genotype at a particular gene or marker and differences in
the phenotype. The question is, Is there an association between a trait, such as the
volume of the dentate gyrus, and any gene (a gene variant can act as a marker) or
genetic marker in our experimental cross? In the case of a Mendelian trait such as
albinism in mice or Huntingtons disease in humans, we will find an almost perfect
1903_C027.fm Page 415 Wednesday, July 26, 2006 7:49 PM
linkage between a gene variant (tyrosinase and huntingtin, respectively) and variation
in the phenotype.
In the case of QTL mapping when a marker is very near a gene that modulates
the trait, there will also be a comparatively high level of association between the
phenotype and the genotype, though far less than for a Mendelian trait. When a trait
is genuinely polygenic, however, a single QTL may account for only 5 to 10% of
the variance. In this case, no single marker will be highly predictive of phenotype.
A QTL map is a simple graph of the degree of association between chromosomal
intervals (usually on the x-axis) and trait values. These plots are usually presented
with degree of association (often given as likelihood ratio score (LRS), log of odds
(LOD), or log P) plotted on the y-axis as a function of chromosomal location.
A chromosomal region with a high degree of phenotypic association marks each
QTL. While we call this a genetic linkage or association, a significant linkage is
actually a directional claim of causality (though not necessarily of mechanistic
interaction): a QTL map is a search for the set of genuine genetic causes of variation
in phenotype, and in that respect is no different from a study of Mendelian mutations,
although loci with small effects on the phenotype are, of course, harder to detect.
Another way to think of a QTL map is as an association heat map by position,
generated when we fit a model that allows for one marker at a time to have an
influence on the phenotype. Considered this way, it is easy to imagine extending
the model to allow two genes and their interaction to have an effect on the phenotype.
This two-gene model, or pair scan, is valuable for finding pairs of QTLs with
significant interactions, with or without individual effects. Higher-order models are
easy to imagine, but they rapidly become prohibitive due to the large numbers of
cases needed to rigorously test more complex models.
Most of the common tools for QTL mapping, including pair scans, are integrated
into GeneNetwork (GN, www.genenetwork.org), a Web-accessible resource for sys-
tems genetics. Many of the examples in the rest of this chapter are drawn from this
resource and can be easily replicated and extended from any browser. GN includes
tools for QTL mapping and analysis of large-scale phenotypes and array data, a
variety of analyses utilizing genetic correlation, clustering, network building and
other exploratory tools, and extensive connections to external systems genetics
related tools. The remainder of this chapter will focus on the capabilities of this tool
set (and associated external data sets built by other investigators) and approaches to
analyzing morphological phenotypes.
FIGURE 27.1 Recombinant inbred (RI) breeding scheme. Breeding (RI) lines simply
requires sequential intercrosses of offspring, usually starting with inbred parental strains, until
the resulting offsprings are fully inbred (approximately F20).
40,000 messages. Array costs continue to decline rapidly even as array quality
improves, making mRNA abundance the cheapest phenotype to acquire.39
is real (given that it has mapped to the location of the transcript itself and that such
occurrences are at a rate much higher than chance in the population of transcripts)
is much higher than would be the case for a trans-acting QTL.
2250
1500
750
FIGURE 27.2 cis- and trans-acting QTLs in the BXD brain. This figure plots QTL position
(x-axis) vs. transcript genomic position (y-axis) using QTLs from the INIA M430 Brain data
set (PDNN transform, Jan. 06 version). The diagonal line indicates cis-acting QTLs (the QTL
maps to the same position as the transcript) while dots off of the vertical line indicate trans-
acting QTLs. The precise cause of the light, vertical bands of trans-acting QTLs (often called
trans-bands) is unknown, but caution should be taken interpreting them since they do not
replicate well between data sets.
1903_C027.fm Page 419 Wednesday, July 26, 2006 7:49 PM
the diagonal line of cis-acting QTLs, there are several vertical lines where there
appear to be a high concentration of trans-acting QTLs. Exactly what causes
these trans-bands is unclear, but they may relate to the action of one or more
transcription factors with many targets. It is interesting to note, however, that while
many cis-acting QTLs replicate between our RI array data and F2 array data in the
same tissue (whole brain, cerebellum), fewer trans-acting QTLs and only one trans-
band replicates.54
on, the deepest data sets, especially where CNS morphology is concerned, are for
these strains. While the crown jewels of the GN data sets are obviously the extensive
microarray data, a valuable tool GN provides is easy access to published phenotype
data in a useful analysis environment for many common GRPs. Again, the BXDs
are the most thoroughly characterized with 830 published phenotypes in the database
at time of writing.
For the BXD strain set there are many CNS-related array data sets (all generated
using Affymetrix arrays, though GN easily supports data from other array platforms)
including data for the forebrain, hippocampus, striatum, cerebellum, and eye. (There
is also a data set for liver mRNA abundance generated using the Agilent array and
an Affymetrix array-based study of hematopoietic stem cells.) Each of these data
sets was generated using at least one array per sex per strain and from 2 to 4 age-
and sex-matched littermates pooled per array. We are currently in the midst of
expanding on this range of CNS-related tissues. Since the BXD strain set has recently
been expanded55 we can also considerably improve the power of our inquiries by
simply expanding our phenotyping effort to the new strains.
For the reader following along on the GN site, these data sets are named
according to six characteristics: the institution or group that funded them, tissue the
array was applied to, array type, array version, date this version of the data was
released, and normalization method. We generally provide pre-normalized and qual-
ity controlled array data using at least three normalization procedures: RMA, PDNN,
and MAS5. Choosing a good normalization to address with your research question
is well beyond the scope of this chapter, but normalization can substantially shift
the results of an array analysis, so care is necessary.
A primary advantage of GRPs is never having to genotype again. Over 10,000
SNP genotypes were assayed in many of our mouse GRPs using the Illumina SNP
genotyping platform, resulting in very high-density genotypes generated from those
SNPs that differed between strains. In the BXD strain set, for instance, there are
now 3,795 combined SNP and traditional simple sequence length polymorphism
(SSLP) markers that differ between the B6 and D2 strains!
While GN is already set up to analyze data for the AKXD, BXH, CXB,
AXB/BXA, and LXS RI strain sets, currently there are only two array data sets
available for non-BXD mouse RI strain sets. The first, of considerably more interest
for brain structure research, is an array data set for the CXB hippocampus. The
second is for mammary tumors in the AKXD strain set. The others are currently
limited to genotypes and published phenotypes. In addition, GN also has genotypes
and published phenotype data for the HXB/BXH rat strain set.
GN is an easily extensible framework, and adding the most simple data sets to
the system is simply a matter of importing phenotypes and genotypes. There are
currently several F2 crosses resident on the site, for instance. While they are not an
effectively immortal inbred population, like the RIs, integrating them into the GN
analytical framework affords us the opportunity to compare and confirm results
between crosses and to apply the GN tool set to their analysis. As we continue to
expand the variety of data sets available in GN, we look forward to the integration
of new tissues, crosses, and treatment conditions into the sites repertoire.
1903_C027.fm Page 421 Wednesday, July 26, 2006 7:49 PM
8.0
NET mRNA abundance (Striatum, RMA transform, 04/05)
+
7.8 27 +
+ 06
+ 15
19 + +
13+ 09
+ + 28
7.7 16 DBA/2J +
+
+ 08 11
22 +
31
7.6 +
+ C57BL/6J
+ 01
05
7.5
+
29
7.4
20 10 0.0 10.0 20.0 30.0 40.0 50.0
Prepulse Inhibition (PPI, BXD Published Phenotype 104399)
FIGURE 27.3 Prepulse inhibition and NET mRNA abundance. This scatterplot shows the
genetic correlation between prepulse inhibition (percent inhibition, BXD published phenotype
10399; x-axis) and expression of the norepinephrine transporter mRNA (NET, Slc6a2;
RMA transform 4/05, probe set 1447311; y-axis) in the striatum. Each point in this scatterplot
is the average value for an entire strain (BXD24 is labeled 24, etc.). This covariance is quite
significant. (p = 5 10-6) Relations like these can lead to novel hypotheses or be used to
test existing ones.
1903_C027.fm Page 422 Wednesday, July 26, 2006 7:49 PM
Let us examine some possibilities. The correlation between PPI and NET may
be caused by a true molecular and mechanistic interaction. If the transporter gene
has a strong allelic variant (e.g., a premature stop codon in one of the parents of the
BXDs), and if variation in PPI just happens to map near the NET locus on Chr 8
at 91.6 Mb, then we would have a good causal hypothesisthat variation in NET
protein causes correlation between these two traits (there are at least 73 B vs. D
SNPS in NET). In this case, however, there is absolutely no evidence that PPI is
modulated by a QTL on Chr 8.
An alternative explanation is that a cascade of protein intermediaries, some of
which happen to overlap, generates the correlation between these two traits. While
PPI and NET may overlap in this molecular network space, they may not be directly
associated. The correlation may be generated by common gene variants that are far
upstream of both PPI and NET. If this is the case, then NET and PPI are still in
some sense mechanistically coupled, but they do not have a direct causal relationship.
If we understood the molecular cascades and networks controlling both traits, we
could provide a compelling reason for the correlation, but it might not be possible
to induce a change in the PPI simply by changing NET expression as implied by
the graph. Over-expression of NET would not improve PPI scores or cure schizo-
phrenia.
A final alternative is that PPI and NET co-vary tightly but that this covariation
is entirely due to linkage disequilibrium. This source of correlation is a bit harder
to grasp. Linkage disequilibrium is a biological phenomenon, but it usually does not
interest nongeneticists. To understand this source of correlation, consider a pair of
very different QTLs that both just happen to be located on Chr 11 at 98 Mb. One
of these QTLs is actually NeuroD2, a gene apparently essential for the normal
development of the basolateral amygdala.57 Mutations in NeuroD2 can produce
fearless, mean mice. Right next to NeuroD2 is another key developmental gene,
namely growth hormone (GH). Imagine that strain X just happened to have alleles
at NeuroD2 and GH that lead to the production of small and unusually aggressive
mice and that strain Y just happened to have alleles that lead to gentle and large
mice. Imagine also that NeuroD2 and GH had many other independent downstream
molecular targets (which they do). Because these two genes are physically tied
together on Chr 11, these two mechanistically separate downstream networks
will also be tied together in any experimental cross between strains X and Y.
Linkage produces secondary effects that ripple through data sets. Disequilibrium
between modifiers that control separate biochemical networks will cause these
networks to appear to co-vary, leading to a mixture of mechanistic covariance and
linkage covariance.
Fortunately, at least for individual hypotheses where upstream modulators are
known, it is easy to control for linkage disequilibrium effects by stratifying the
correlation analysis by allele or genotype at the upstream modulator. In the example
above, we would stratify our analysis by allele at a marker near NeuroD2 and GH
to disambiguate correlation due to linkage disequilibrium from correlation due to
variation in individual networks.
1903_C027.fm Page 423 Wednesday, July 26, 2006 7:49 PM
organizational task for microarray results. For most GN-derived gene lists, however,
the primary interest is functional relatedness.
WebGestalt provides four similar tools for gene set analysis by function. The
most commonly used is the gene ontology tree tool (GO Tree). GO Tree uses gene
ontology categories and compares the observed and expected fraction of genes
represented in the gene set across category, returning the gene set in a tree form
with over-represented areas marked. It is worth exploring broad categories that are
not over-represented because narrower subcategories may still have over-represen-
tation, especially in a large gene list. This can be easily done via the Enriched GO
DAG tool, which visualizes the GO Tree results as a directed acyclic graph. For
convenience this option is directly linked from GN but can also be accessed from
WebGestalt.
WebGestalt also provides three other tools for gene set analysis by function.
Users can fit gene sets to and perform over-representation analyses on KEGG (Kyoto
Encyclopedia of Genes and Genomes; www.genome.jp/kegg) or BioCarta pathways
(www.biocarta.com). In these cases over-representation means more genes (in the
gene set) that participate in a given pathway than would be expected by chance.
These options are particularly good for correlation analyses because genes in the
same pathway would be expected to have correlated expression levels. In addition,
users can look for over-representation of PFAM (www.sanger.ac.uk/Software/Pfam)
protein domains in their gene list.
If you are interested in collecting your own volumetric data using the MBL data
set, the MBL includes metadata on the age, sex, body, and brain weight for each
animalin addition to strain identificationand each of these terms can be searched
to select animals according to the users requirements. Images are available at a
variety of resolutions, depending on the slide. All slides are available at the base
resolution of 24.5 0.5 m per pixel in the XY plane. Along the z-axis, sections
were taken at a 150 m interval (300 m between sections on each slide, 2 slides
per case). Significantly higher resolution images of single brain sections have been
acquired at 4.5 m per pixel for more than 100 cases (look for the blue, hi-res
button), and 1 m/pixel images for the neocortex, hippocampus, and dorsal lateral
geniculate nucleus are underway.
Images can be used to calculate volumes using several free programs: NIH
Image, Scion Image, or Image/J. (Adobe Photoshop can also be used.) Details on
processing, imaging, and calibration using MBL images can be found on procedures,
pages, or in papers written based on this data, such as our analysis of hippocampal
volume.31
regions without the need to purify single cell types for array analysisan expensive
and difficult proposition.
27.22 GENSAT
One other mouse imaging tool worth mentioning is the GENSAT database59
(www.gensat.org), which contains a mouse gene expression database for embryos
(E15.5), neonates (P7), and adult animals. GENSATs expression model is based on
bacterial artificial chromosomes (BACs) in which coding sequences were replaced
by the enhanced green fluorescent protein (EGFP). Gene expression here is a relative
and somewhat more indirect measure, since the EGFP mRNA and protein may have
different stability than the gene it replaces. GENSAT uses FVB mice to generate its
transgenic embryos, so, like ABA, it is not designed to investigate a genetic axis.
The ability to introduce multiple copies of the BAC enhances the ability to detect
genes at low levels, however, and may be more sensitive than in situ methods. When
adult expression is of primary interest, GENSAT is a useful resource to check, either
to confirm expression observed in the ABA or for when the ABA shows no expres-
sion and there is reason to suspect that the actual expression level may simply be
low. Of course, when embryonic or neonatal expression is crucial, as will often be
the case with morphological characteristics, GENSAT is an extremely important
resource.
1903_C027.fm Page 427 Wednesday, July 26, 2006 7:49 PM
4.2 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 x
2.0
Trait ID: BXDPublish: 10460 LOD
Interval Mapping for Dataset: BXD, mapping on All Chromosome Addititive Effect Frequency of the
Significant LOD = 3.94 Peak LRS
Using Haldane mapping function with no control for other QTLs Suggestive LOD = 2.29
3.0
LOD
2.0
1.0
25
50
75
100
125
150
175
25
50
75
100
125
150
175
25
50
75
100
125
150
25
50
75
100
125
150
25
50
75
100
125
150
25
50
75
100
125
150
25
50
75
100
125
25
50
75
100
125
25
50
75
100
25
50
75
100
125
25
50
75
100
25
50
75
100
25
50
75
100
25
50
75
100
25
50
75
100
25
50
75
25
50
75
25
50
75
25
50
25
50
75
100
125
150
Megabases
FIGURE 27.4 Whole genome QTL map of dentate gyrus volume. This figure shows a typical
QTL map for a morphological phenotype. Dentate gyrus volume has one significant QTL on
Chr 13. The line near the top indicates genome-wide adjusted p = 0.05 threshold.
1903_C027.fm Page 428 Wednesday, July 26, 2006 7:49 PM
Click to view the corresponding section of the qenome in an 8x expanded WebOTL map
Click to view the corresponding section of the qenome in the UCSC Genome Browser
Click to view the corresponding section of the qenome in theEnsemnl Genome Browser
Chr 13
4.2
4.0
3.0
LOD
2.0
1.0
0 10 20 30 40 50 60 70 80 90 100 110
Megabases
FIGURE 27.5 Chr 13 QTL map of dentate gyrus volume QTL DGV13a. This is a closeup
of DGV13a. Features are discussed further in the text and include the SNP density map at
the bottom, bars indicating the frequency with which the best QTL position was in that position
in a bootstrapped population, and the light trace indicating the additive effect of the marker
on the phenotype. A B6 allele increases dentate gyrus volume.
There are many genes within the 18 Mb DGV13a interval, and searching for the
gene or genes responsible for the QTLs effect is more art than science. There are
two major complementary approaches: narrowing the interval and building evidence
for a particular gene. Narrowing the QTL interval reduces the number of potential
candidates and generally involves more genetic approaches, such as building con-
genics or phenotyping additional animals that are recombinant across the QTL
interval. Building evidence for a particular candidate is trickier given the formidable
number of possibilities and the capacity of the human mind for constructing plausible
stories.
That said, our favorite candidates for DGV13a are Ror2 and Msx2. Ror2 is a
receptor tyrosine kinase expressed in the developing nervous system62 as well as
the branchial arches, heart, and limb/tail bud.63 The gene is located from 51.673 to
51.845 Mb, right in the center of the DGV13a interval, and, crucially, in a region
with a considerable number of SNPs. A glance at the SNP track at the bottom of
Figure 27.5 will indicate how unevenly the SNPs between B6 and D2 are distributed.
This is a fairly typical distribution, and if we make the simplifying assumption that
a gene with no SNPs is not a likely candidate, we can immediately eliminate a
substantial fraction of the genes in the interval. In this case, Ror2 has 289 SNPs,
though none of these are non-synonymous (the apparent mis-sense mutation,
mCV23266532, is actually in intron 1).
Ror2 regulates transcription of MSH-like homeo box 2 (Ms 2) by sequestering
Maged1 (Dlxin-1) from the membrane.64 Since Ms 2 is located very near Ror2,
we might expect this to result in what would appear to be a cis-acting QTL. While
Ror2 seems to be expressed in adult tissue according to the ABA, it has extremely
low expression, which is hard to distinguish from noise, in our hippocampus data.
This suggests that the effect of the gene may occur during development. This
1903_C027.fm Page 429 Wednesday, July 26, 2006 7:49 PM
7.4
6.8
3.0 3.5 4.0 4.5 5.0 5.5 6.0
Dentate Gyrus Volume (mm3)
FIGURE 27.6 Correlation of Ms 2 mRNA abundance and dentate gyrus weight. This
scatterplot shows the genetic correlation between Ms 2 (probe set 1449559, BXD hippoc-
ampus data) mRNA abundance and dentate gyrus weight. (BXD published phenotype 10460).
suggestion is bolstered by the expression of Ror2 in the early nervous system.62 The
peri-natal lethality of a Ror2 knockout63 and the association of Ror2 with dwarfism,
cyanosis, and short limbs and tails also underscore its importance in development.
Ms 2 interacts with Ror2, and its expression is strongly correlated with dentate
gyrus volume (Figure 27.6; Pearsons r = 0.53; p = 0.0018), which make it a
compelling candidate in the same pathway. Ms 2 is located from 52.026 to 52.031
Mb on Chr 13 and has an even higher SNP density per Kb than Ror2. Maged1,
which mediates the interaction between Ms 2 and Ror2, is located on Chr X and
has 0 SNPs, so while it interacts with our favorite candidates it is not itself a
candidate.
Also, at least with the smaller BXD RI strain set, (this is to some extent
ameliorated by the larger, extended BXD strain set) GN experiments are underpow-
ered for detection of QTLs of small to medium effect. This is not to say that one
should not identify a QTL by the designation of the least restrictive population it
effects or by the region in which it has the largest effect size if that is where it
demonstrated significance. This still seems appropriate in the absence of biological
understanding of the QTLs effect. It does, however, suggest liberal interpretation
of candidate-gene evidence regarding the genes effect in larger areasand cautious
interpretation of the importance of having selectively identified a subregion QTL.
27.24 CONCLUSION
The major purpose of understanding biological processes is to be able to make
reliable predictions and to modulate outcome in a logical and controlled way. To
achieve this level of understanding, we must understand complex systems and
networks of phenotypes. We need to understand how combinations of gene variants,
networks of molecules, and neuronal circuits work together in different individuals
and in many different environments to achieve functional equilibrium. We need to
adopt holistic and integrative approaches to experimental design that would have
been completely impractical a decade ago. The advent of sophisticated statistical
and computational methods and the advent of high-throughput methods to acquire
hundreds of thousands of genotypes and phenotypes are rewriting the rules of design.
The scientific and statistical chaos that followed the introduction of microarrays is
just the technical spearhead of this new experimental order. We should not be
surprised if there has been a backlash on the part of those scientists more accustomed
to juggling one ball at a time. We are now living through an awkward transition that
is moving us in the direction of systems biology for good reason: only this approach
will enable us to make reliable predictions of complex outcomes. Animals and
humans are not machines. We are stochastic, fluid, willful, and richly redundant
cybernetic systems.
REFERENCES
1. Adams MD, Soares MB, Kerlavage AR, Fields C, Venter JC: Rapid cDNA sequencing
(expressed sequence tags) from a directionally cloned human infant brain cDNA
library. Nat Genet 1993, 4(4):373380.
2. Sutcliffe JG: mRNA in the mammalian central nervous system. Annu Rev Neurosci
1988, 11:157198.
3. Yeo G, Holste D, Kreiman G, Burge CB: Variation in alternative splicing across
human tissues. Genome Biol 2004, 5(10):R74.
4. Williams RW, Herrup K: The control of neuron number. Annu Rev Neurosci 1988,
11:423453.
5. Gautvik KM, de Lecea L, Gautvik VT, Danielson PE, Tranque P, Dopazo A, Bloom
FE, Sutcliffe JG: Overview of the most prevalent hypothalamus-specific mRNAs, as
identified by directional tag PCR subtraction. Proc Natl Acad Sci USA 1996,
93(16):87338738.
1903_C027.fm Page 431 Wednesday, July 26, 2006 7:49 PM
6. Usui H, Falk JD, Dopazo A, de Lecea L, Erlander MG, Sutcliffe JG: Isolation of
clones of rat striatum-specific mRNAs by directional tag PCR subtraction. J Neurosci
1994, 14(8):49154926.
7. Lashley K: Persistent problems in the evolution of mind. Quart Rev Biol 1949,
24:2842.
8. Wimer C, Prater L: Some behavioral differences in mice genetically selected for high
and low brain weight. Psychol Rep 1966, 19(3):675681.
9. Keverne EB, Martel FL, Nevison CM: Primate brain evolution: genetic and functional
considerations. Proc Biol Sci 1996, 263(1371):689696.
10. Aboitiz F: Does bigger mean better? Evolutionary determinants of brain size and
structure. Brain Behav Evol 1996, 47(5):225245.
11. Rensch B: Increase of learning capability with increase of brain size. Amer Natur
1956, 90(8195).
12. Ward BC, Nordeen EJ, Nordeen KW: Individual variation in neuron number predicts
differences in the propensity for avian vocal imitation. Proc Natl Acad Sci U S A
1998, 95(3):12771282.
13. Devoogd TJ, Krebs JR, Healy SD, Purvis A: Relations between song repertoire size
and the volume of brain nuclei related to song: comparative evolutionary analyses
amongst oscine birds. Proc Biol Sci 1993, 254(1340):7582.
14. Crusio WE, Schwegler H, van Abeelen JH: Behavioral responses to novelty and
structural variation of the hippocampus in mice. II. Multivariate genetic analysis.
Behav Brain Res 1989, 32(1):8188.
15. Schwegler H, Lipp HP: Hereditary covariations of neuronal circuitry and behavior:
correlations between the proportions of hippocampal synaptic fields in the regio
inferior and two-way avoidance in mice and rats. Behav Brain Res 1983, 7(1):138.
16. Lipp HP, Schwegler H, Heimrich B, Driscoll P: Infrapyramidal mossy fibers and two-
way avoidance learning: developmental modification of hippocampal circuitry and
adult behavior of rats and mice. J Neurosci 1988, 8(6):19051921.
17. Lipp HP, Collins RL, Hausheer-Zarmakupi Z, Leisinger-Trigona MC, Crusio WE,
Nosten-Bertrand M, Signore P, Schwegler H, Wolfer DP: Paw preference and intra-
/infrapyramidal mossy fibers in the hippocampus of the mouse. Behav Genet 1996,
26(4):379390.
18. Lipp HP, Schwegler H, Crusio WE, Wolfer DP, Leisinger-Trigona MC, Heimrich B,
Driscoll P: Using genetically-defined rodent strains for the identification of hippoc-
ampal traits relevant for two-way avoidance behavior: a non-invasive approach. Expe-
rientia 1989, 45(9):845859.
19. Schwegler H, Crusio WE, Lipp HP, Heimrich B: Water-maze learning in the mouse
correlates with variation in hippocampal morphology. Behav Genet 1988, 18(2):153165.
20. Lipp HP, Schwegler H, Heimrich B, Cerbone A, Sadile AG: Strain-specific correla-
tions between hippocampal structural traits and habituation in a spatial novelty situ-
ation. Behav Brain Res 1987, 24(2):111123.
21. Patel NV, Hitzemann RJ: Detection and mapping of quantitative trait loci for halo-
peridol-induced catalepsy in a C57BL/6J x DBA/2J F2 intercross. Behav Genet 1999,
29(5):303310.
22. Kanes S, Dains K, Cipp L, Gatley J, Hitzemann B, Rasmussen E, Sanderson S,
Silverman M, Hitzemann R: Mapping the genes for haloperidol-induced catalepsy.
J Pharmacol Exp Ther 1996, 277(2):10161025.
23. Bucan M, Abel T: The mouse: genetics meets behaviour. Nat Rev Genet 2002,
3(2):114123.
1903_C027.fm Page 432 Wednesday, July 26, 2006 7:49 PM
24. Melo JA, Shendure J, Pociask K, Silver LM: Identification of sex-specific quanti-
tative trait loci controlling alcohol preference in C57BL/ 6 mice. Nat Genet 1996,
13(2):147153.
25. Peirce JL, Derr R, Shendure J, Kolata T, Silver LM: A major influence of sex-specific
loci on alcohol preference in C57Bl/6 and DBA/2 inbred mice. Mamm Genome 1998,
9(12):942948.
26. Takahashi JS, Pinto LH, Vitaterna MH: Forward and reverse genetic approaches to
behavior in the mouse. Science 1994, 264(5166):17241733.
27. Plomin R, McClearn GE, Gora-Maslak G, Neiderhiser JM: Use of recombinant inbred
strains to detect quantitative trait loci associated with behavior. Behav Genet 1991,
21(2):99116.
28. Belknap JK, Phillips TJ, OToole LA: Quantitative trait loci associated with brain
weight in the BXD/Ty recombinant inbred mouse strains. Brain Res Bull 1992,
29(34):337344.
29. Williams RW: Mapping genes that modulate mouse brain development: a quantitative
genetic approach. In: Mouse Brain Development. Edited by Goffinet A, Rakic P.
New York: Springer; 2000: 2149.
30. Williams RW, Strom RC, Goldowitz D: Natural variation in neuron number in mice
is linked to a major quantitative trait locus on Chr 11. J Neurosci 1998, 18(1):138146.
31. Lu L, Airey DC, Williams RW: Complex trait analysis of the hippocampus: mapping
and biometric analysis of two novel gene loci with specific effects on hippocampal
structure in mice. J Neurosci 2001, 21(10):35033514.
32. Peirce JL, Chesler EJ, Williams RW, Lu L: Genetic architecture of the mouse hip-
pocampus: identification of gene loci with selective regional effects. Genes Brain
Behav 2003, 2(4):238252.
33. Williams RW, Airey DC, Kulkarni A, Zhou G, Lu L: Genetic dissection of the
olfactory bulbs of mice: QTLs on four chromosomes modulate bulb size. Behav Genet
2001, 31(1):6177.
34. Rosen GD, Williams RW: Complex trait analysis of the mouse striatum: independent
QTLs modulate volume and neuron number. BMC Neurosci 2001, 2:5.
35. Zhou G, Williams RW: Eye1 and Eye2: gene loci that modulate eye size, lens weight,
and retinal area in the mouse. Invest Ophthalmol Vis Sci 1999, 40(5):817825.
36. Airey DC, Robbins AI, Enzinger KM, Wu F, Collins CE: Variation in the cortical
area map of C57BL/6J and DBA/2J inbred mice predicts strain identity. BMC Neu-
rosci 2005, 6(1):18.
37. Airey DC, Lu L, Williams RW: Genetic control of the mouse cerebellum: identifica-
tion of quantitative trait loci modulating size and architecture. J Neurosci 2001,
21(14):50995109.
38. Li CX, Wei X, Lu L, Peirce JL, Williams RW, Waters RS: Genetic analysis of barrel
field size in the first somatosensory area (SI) in inbred and recombinant inbred strains
of mice. Somatosens Mot Res 2005, 22(3):141150.
39. Williams RW: Expression Genetics and the Phenotype Revolution. Mamm Genome
2006, in press.
40. Seecharan DJ, Kulkarni AL, Lu L, Rosen GD, Williams RW: Genetic control of
interconnected neuronal populations in the mouse primary visual system. J Neurosci
2003, 23(35):1117811188.
41. Damerval C, Maurice A, Josse JM, de Vienne D: Quantitative trait loci underlying
gene product variation: a novel perspective for analyzing regulation of genome expres-
sion. Genetics 1994, 137(1):289301.
1903_C027.fm Page 433 Wednesday, July 26, 2006 7:49 PM
42. Jansen RC, Nap JP: Genetical genomics: the added value from segregation. Trends
Genet 2001, 17(7):388391.
43. Brem RB, Yvert G, Clinton R, Kruglyak L: Genetic dissection of transcriptional
regulation in budding yeast. Science 2002, 296(5568):752755.
44. Chesler EJ, Lu L, Shou S, Qu Y, Gu J, Wang J, Hsu HC, Mountz JD, Baldwin NE,
Langston MA et al: Complex trait analysis of gene expression uncovers polygenic
and pleiotropic networks that modulate nervous system function. Nat Genet 2005,
37(3):233242.
45. Bystrykh L, Weersing E, Dontje B, Sutton S, Pletcher MT, Wiltshire T, Su AI, Vellenga
E, Wang J, Manly KF et al: Uncovering regulatory pathways that affect hematopoietic
stem cell function using genetical genomics. Nat Genet 2005, 37(3):225232.
46. Schadt EE, Monks SA, Drake TA, Lusis AJ, Che N, Colinayo V, Ruff TG, Milligan
SB, Lamb JR, Cavet G et al: Genetics of gene expression surveyed in maize, mouse
and man. Nature 2003, 422(6929):297302.
47. Hubner N, Wallace CA, Zimdahl H, Petretto E, Schulz H, Maciver F, Mueller M,
Hummel O, Monti J, Zidek V et al: Integrated transcriptional profiling and linkage
analysis for identification of genes underlying disease. Nat Genet 2005, 37(3):243253.
48. Stranger BE, Forrest MS, Clark AG, Minichiello MJ, Deutsch S, Lyle R, Hunt S,
Kahl B, Antonarakis SE, Tavare S et al: Genome-Wide Associations of Gene Expres-
sion Variation in Humans. PLoS Genet 2005, 1(6):e78.
49. Cheung VG, Spielman RS, Ewens KG, Weber TM, Morley M, Burdick JT: Mapping
determinants of human gene expression by regional and genome-wide association.
Nature 2005, 437(7063):13651369.
50. Morley M, Molony CM, Weber TM, Devlin JL, Ewens KG, Spielman RS, Cheung
VG: Genetic analysis of genome-wide variation in human gene expression. Nature
2004, 430(7001):743747.
51. Kirst M, Basten CJ, Myburg AA, Zeng ZB, Sederoff RR: Genetic architecture of
transcript-level variation in differentiating xylem of a eucalyptus hybrid. Genetics
2005, 169(4):22952303.
52. Benjamini Y, Hochberg Y: Controlling the False Discovery Rate: a practical and
powerful approach to multiple testing. J Royal Statistical Society B 1995, 57:289300.
53. Doss S, Schadt EE, Drake TA, Lusis AJ: Cis-acting expression quantitative trait loci
in mice. Genome Res 2005, 15(5):681-691.
54. Peirce J, Lu L, Hongqiang L, Wang J, Manly KF, Hitzemann B, Belknap JK, Rosen
GD, Williams RW: How replicable are mRNA expression QTLs. Mamm Genome
2006, in press.
55. Peirce JL, Lu L, Gu J, Silver LM, Williams RW: A new set of BXD recombinant
inbred lines from advanced intercross populations in mice. BMC Genet 2004, 5(1):7.
56. McCaughran J, Jr., Bell J, Hitzemann R: On the relationships of high-frequency
hearing loss and cochlear pathology to the acoustic startle response (ASR) and
prepulse inhibition of the ASR in the BXD recombinant inbred series. Behav Genet
1999, 29(1):2130.
57. Lin CH, Hansen S, Wang Z, Storm DR, Tapscott SJ, Olson JM: The dosage of the
neuroD2 transcription factor regulates amygdala development and emotional learning.
Proc Natl Acad Sci U S A 2005, 102(41):1487714882.
58. Homayouni R, Heinrich K, Wei L, Berry MW: Gene clustering by latent semantic
indexing of MEDLINE abstracts. Bioinformatics 2005, 21(1):104115.
59. Heintz N, Hatten M: The Gene Expressio Nervous System Atlas (GENSAT)
Project, NINDS Contract #N01NS02331 to The Rockefeller University, New York.
February 1, 2006. www.gensat.org.
1903_C027.fm Page 434 Wednesday, July 26, 2006 7:49 PM
60. Churchill GA, Doerge RW: Empirical threshold values for quantitative trait mapping.
Genetics 1994, 138(3):963971.
61. Dupuis J, Siegmund D: Statistical methods for mapping quantitative trait loci from
a dense set of markers. Genetics 1999, 151(1):373386.
62. Oishi I, Takeuchi S, Hashimoto R, Nagabukuro A, Ueda T, Liu ZJ, Hatta T, Akira
S, Matsuda Y, Yamamura H et al: Spatio-temporally regulated expression of receptor
tyrosine kinases, mRor1, mRor2, during mouse development: implications in devel-
opment and function of the nervous system. Genes Cells 1999, 4(1):4156.
63. Takeuchi S, Takeda K, Oishi I, Nomi M, Ikeya M, Itoh K, Tamura S, Ueda T, Hatta
T, Otani H et al: Mouse Ror2 receptor tyrosine kinase is required for the heart
development and limb formation. Genes Cells 2000, 5(1):7178.
64. Matsuda T, Suzuki H, Oishi I, Kani S, Kuroda Y, Komori T, Sasaki A, Watanabe K,
Minami Y: The receptor tyrosine kinase Ror2 associates with the melanoma-associ-
ated antigen (MAGE) family protein Dlxin-1 and regulates its intracellular distribu-
tion. J Biol Chem 2003, 278(31):2905729064.
1903_C028.fm Page 435 Wednesday, July 26, 2006 8:40 PM
28 Synaptic Mechanisms
Involved in Cognitive
Function: Cues from
Mental Retardation Genes
Guntram Borck, Florence Molinari, Birgit Dreier,
Peter Sonderegger, and Laurence Colleaux
CONTENTS
435
1903_C028.fm Page 436 Wednesday, July 26, 2006 8:40 PM
28.1 INTRODUCTION
Mental retardation (MR) affects about 2% of the general population and is the most
frequent handicap in children and young adults. It is characterized by a broad range
of deficits in higher brain functions that result in significant limitations in adaptive
and cognitive capacities required for competence in daily living, communication,
social interaction and integration, self-direction, and work (DSM-IV). An onset of
the symptoms in the developmental period is an essential diagnostic trait. The severity
of mental retardation is commonly classified on the basis of the intelligence quotient
(IQ) although other criteria have also been used. Based on a population mean of 100
and a standard deviation of 15, MR is usually classified as mild when the IQ ranges
between 50 to 70 and as severe when the IQ value is below 50.1,2
The causes of MR are diverse and include environmental factors, teratogens,
numerical or structural anomalies of chromosomes, gene defects, and metabolic
diseases. An estimated fraction of 25 to 40% of the severe MRs and most mild MRs
remain unexplained.35 They are currently thought to be due to monogenic or mul-
tigenic defects or to a combination of genetic and environmental factors. Only
25 to 50% of the severe forms of MR are estimated to be genetically determined.
They include metabolic diseases impairing neuronal function in a nonspecific man-
ner, conditions which alter the normal patterning of the brain, neuromuscular dis-
orders, as well as MRs that do not show any clinical features besides cognitive
deficits. The last category is termed nonsyndromic MR (NSMR).
Although NSMR conditions are much difficult for geneticist studies, they are
best suited for identifying the molecular basis of cognitive functions, and analysis
of these disorders will very likely give new insight into the neurobiology of human
cognitive processing.
NS-XLMRs were first investigated. Since 1996, positional and functional candidate
gene approaches led to the identification of 11 X-linked nonsyndromic MR genes.7,8
Despite the extreme variety of processes in which genes responsible for these
disorders are involved, several common cellular processes have emerged.
Extracellular signals
(Adhesion molecules, growth factors, electric/synaptic activity)
Rho GDI
Rho-GDP
Rac-GDP
Cdc42-GDP
Actin cytoskeleton
of dendritic spines
FIGURE 28.1 (SEE COLOR INSERT FOLLOWING PAGE 236) One cluster of nonsyn-
dromic MR genes relates to the Rho family of small GTPases. Rho family GTPases transduce
extracellular signals into adaptive responses of the actin cytoskeleton. Actin-dependent pro-
cesses at the synapse include the regulation of the morphology and the dynamics of the
dendritic spines. The Rho family comprises three members, termed Rho, Rac, and Cdc42.
They shuttle between an inactive (red) and an active (green) state under the control of three
types of regulatory proteins (blue). GEFs (guanine nucleotide exchange factors) promote the
release of GDP and its replacement by GTP and, thereby, mediate the transition of Rho, Rac,
and Cdc42 from the inactive into the active state. GAPs (GTPase-activating proteins) activate
the endogenous GTPase function and, thus, the self-inactivation of Rho, Rac, and Cdc42.
GDIs (GDP dissociation inhibitors) bind to the GDP form of the GTPases and prevent their
premature activation as long as the GTPase is not in the correct place and situation for a new
round of activation. Further downstream activators (yellow) eventually act directly or indi-
rectly on regulatory components of the actin cytoskeleton. The genes affected in MRs are
printed red.
1903_C028.fm Page 438 Wednesday, July 26, 2006 8:40 PM
cytoskeleton and regulate gene expression. They are involved in neurite outgrowth,
axon guidance, dendrite maturation, synapse formation, and the morphogenesis and
dynamics of dendritic spines.912
The family comprises three members, termed Rho, Rac, and Cdc42. They exert
their diverse cellular functions via distinct effects on the actin cytoskeleton that are
mediated by PAK and ROCK kinases as immediate downstream signaling molecules.
Both PAK and ROCK kinases in turn activate LIM kinase, which acts upon and
regulates the dynamics of the actin cytoskeleton. Three major classes of upstream
regulators control the signaling activity of the Rho family GTPases. First, the guanine
nucleotide exchange factors (RhoGEFs) mediate the release of GDP and its replace-
ment by GTP and, thus, drive the conversion from the inactive into the active form.
Second, the GTPase activating proteins (RhoGAPs) enhance the conversion of bound
GTP into GDP and, thus, drive the conversion from the active into the inactive state.
And third, the GDP dissociation inhibitors (RhoGDIs) bind to the GDP-bound form
of the GTPase and prevent their activation as long as other prerequisites, such as
the correct location for reinitiation of the activation cycle, are not fulfilled.
Over the past years, two regulators and one downstream effector of Rho family
GTPases have been found among the genes causing X-linked nonsyndromic MR,
described as follows.
28.2.1.1 Oligophrenin-1
The OPHN1 gene, encoding oligophrenin-1, maps to Xq12 and is expressed in fetal
brain at high levels. It is disrupted in a mildly mentally retarded female carrier of
a balanced X;12 translocation. Furthermore, it has been shown to be mutated in
affected males of a large X-linked MR family by Billuart et al.13 Oligophrenin
contains a domain typical for Rho-GTPase-activating proteins (RhoGAP). In vitro
assays demonstrated that it can regulate the activity of the Rho GTPases RhoA,
Rac1 and Cdc42Hs.
Recent studies show that oligophrenin-1 is present in neuronal and astroglial
cells and that it colocalizes with actin at the tip of growing neurites.14 In addition,
using siRNAs in organotypic hippocampal slices, Govek and co-workers15 demon-
strated that knocking down oligophrenin-1 significantly decreased dendritic spine
length in CA1 pyramidal neurons. Spine morphological changes of the same mag-
nitude have been reported for a mouse model of fragile X,16 indicating that such
changes can compromise synaptic plasticity and potentially lead to learning and
memory deficits.
PAK3 is highly expressed in fetal human brain. In the newborn mouse brain it
localizes mainly to cortical and hippocampal dendrites and axons. LIM kinase 1, a
downstream effector kinase regulated by PAK kinase, was recently reported to
regulate the morphology of dendritic spines of hippocampal neurons.18
Vesicle release from the presynaptic nerve terminal is an essential process for
synaptic transmission. For a long time it has been known that synaptic vesicle release
is subject to modulations that depend on previous activity patterns of a given synapse.
It is known that this modulatory mechanism, termed synaptic facilitation, depends
on intracellular calcium, but its exact molecular basis is still under investigation.
Neurotransmitter release is essential for basic synaptic function. Its activity-
dependent regulation is of critical importance for the coordinated and dynamic
1903_C028.fm Page 440 Wednesday, July 26, 2006 8:40 PM
Neurotransmitter
synthesis/uptake
Rab GEF
Rab-GTP
Rab-GTP Rab GDI1
Rab GDI
Vesicle
IL1RAPL1 recycling
(endocytosis)
Rab-GTP Rab-GTP Rab-GDP
NCS-1
Docked Fused Restricting
FACL4
Rab GAP
vesicle Ca2+ vesicle ilin
bud oph
End
FIGURE 28.2 (SEE COLOR INSERT) A second cluster of nonsyndromic MR genes relates
to the release of neurotransmitters from the presynaptic nerve terminal and the regeneration
of new releasable vesicles by the budding and endocytosis of vesicles from the presynaptic
membrane. The release of neurotransmitters from storage vesicles occurs in several consec-
utive steps. First, the vesicles are docked to the release site, the presynaptic active zone, a
process that involves the interaction of several proteins of the vesicle and the active zone.
Docked vesicles fuse with the presynaptic membrane in a process regulated by transiently
increased intracellular calcium, and release their content into the synaptic cleft. In order to
maintain a constant supply of releasable neurotransmitters, vesicles recycle by budding off
the presynaptic membrane and endocytosis. The budding occurs in a clathrin-dependent
manner. The restriction of the bud neck, a process that prepares the bud for scission, requires
endophilin. Endophilin is a lysophosphatidic acid acyltransferase, which introduces arachi-
donic acid into the cytoplasmic face of the bud neck. It, thereby, facilitates the formation of
a strong negative curvature of the membrane required for the restriction of the bud neck in
order to allow the scission process to begin. It is conceivable that the supply of arachidonoyl-
CoA requires FACL4. The genes affected in MR are printed in red.
machinery, such as those thought to be necessary for learning and memory, may
be deficient.
The members of the GDI family (Rab GDP-dissociation inhibitors) play an essential
role in the recycling of Rab GTPases required for vesicular transport. Using a
candidate gene approach, DAdamo et al.26 identified RabGDI1 mutations in two
X-linked MR families linked to the Xq28 region. RabGDI1 binds Rab3a, a small
GTP-binding protein highly enriched in the synapse and involved in neurotransmitter
release. First evidence for a role of RabGDI1 protein in neurite outgrowth and
synaptic function came from in vitro studies showing that RabGDI1-antisense
oligonucleotides impaired neurite extension in cultured rat hippocampal neurons.
Further support for this role came from histological, behavioral and electrophysio-
logical studies of RabGDI1-deficient mice. These mice, which do not have any gross
morphological or neuropathological anomalies, display an altered plasticity of hip-
pocampal neurotransmission.27 In addition, behavioral studies have revealed defects
in short-term memory, a reduced aggression level, and altered social behavior.28 A
recent ultrastructural analysis revealed pronounced presynaptic changes in several
brain regions. Most strikingly, a marked reduction of the number of synaptic vesicles
and a clustering of the residual vesicles were observed. These ultrastructural alter-
ations are in perfect accordance with a role of RabGDI1 in synaptic vesicle recycling.
The IL1RAPL1 gene has been identified in the analysis of non-overlapping Xp22
deletions and point mutations in X-linked MR families by Carrie et al.29 The
IL1RAPL1 protein is a member of the IL-1/Toll receptor family. It shows homologies
with the IL-1 receptor accessory proteins (IL1RACPs), and localizes to the plasma
membrane. Recent functional studies showed that IL1RAPL1 is not a receptor for
IL-1 but a specific interaction partner of neuronal calcium sensor 1.30 The NCS-1
protein has been shown to be involved in synaptic facilitation at excitatory hippoc-
ampal synapses.31 Synaptic facilitation is a mechanism of short-term plasticity
that enhances transmitter release from the presynaptic terminal and increases postsyn-
aptic activation as a consequence of recurrent stimulation. Because synaptic facili-
tation has long been known to depend on calcium and because NCS-1 was demon-
strated to be a sensor of presynaptic calcium, NCS-1 is put into a crucial position
for the activity-dependent adaptation of presynaptic transmitter release. By interact-
ing with NCS-1 in the inhibition of exocytosis, IL1RAPL-1 might exert a regulatory
role on the activity-dependent dynamics of presynaptic transmitter release.
By deletion mapping, FACL4, which encodes fatty acid-CoA ligase 4, was identified
as an X-linked nonsyndromic MR gene.32 Acyl-CoA ligases (or synthases) form a
1903_C028.fm Page 442 Wednesday, July 26, 2006 8:40 PM
family of enzymes that catalyze the formation of acyl-CoA esters from fatty acids,
ATP and coenzyme A. Both MR-associated mutations lead to a drastic decrease of
enzymatic activity. That fatty acid-CoA ligase 4 has a strong preference for arachi-
donic acid as a substrate raises interesting speculations about its role in the recycling
of synaptic vesicles. Recent reports suggest an essential role of arachidonic acid in
the endocytosis of synaptic vesicles from the presynaptic membrane, a process
required for the regeneration of releasable neurotransmitter vesicles. A crucial step
in this process is the formation of a bud which then constricts and separates from
the presynaptic plasma membrane to form a new vesicle. This process requires the
remodeling of the lipids on the cytosolic side of the bud neck by endophilin A1.33
Endophilin A1 acts as a lysophosphatidic acid acyltransferase, i.e., it binds lyso-
phosphatidic acid and fatty acyl-coenzyme A and condenses them to phosphatidic
acid. The formation of phosphatidic acid promotes the negative curvature required
at the bud neck for bud restriction. Unsaturated fatty acyl-CoAs, such as arachi-
donoyl-CoA, are most effective in promoting negative curvature when inserted into
the cytosolic leaflet. Thus, a deficiency in unsaturated fatty acyl-CoAs, such as
arachidonoyl-CoA, that may result from deficient FACL4 is likely to reduce the
efficiency of bud restriction and thus results in reduced synaptic vesicle recycling.
suitable for linkage analyses have hitherto hampered identification of the genes
responsible for these diseases. The hope that a careful study of individuals with
chromosomal anomalies would lead to the rapid identification of autosomal MR
genes has been largely disappointed. However, linkage studies and chromosomal
breakpoint analyses recently led to the identification of the first autosomal NS-MR
genes.
REFERENCES
1. Penrose, L.S., A clinical and genetic study of 1280 cases of mental defects (The
Colchester Survey). Special report series, Medical Research Council No. 229, Her
Majestys Stationary Office, London, 1938.
2. Roeleveld, N., Zielhuis, G.A., and Gabreels, F. The prevalence of mental retardation:
a critical review of recent literature, Dev Med Child Neurol, 39, 125, 1997.
3. Turner, G. An etiological study of 1,000 patients with an I.Q. assessment below 51.
Med J Aust, 2, 927, 1975.
4. Lamont, M.A. and Dennis, N.R., Etiology of mild mental retardation. Arch Dis Child,
63, 1032, 1988.
5. Kahler, S.G. and Fahey, M.C., Metabolic disorders and mental retardation. Am J Med
Genet, Part C (Sem Med Genet), 117C, 31, 2003.
6. Stevenson, R.E., Schwartz, C.E., and Schroer, R.J., Emergence of the concept of X-
linked mental retardation, In X-Linked Mental Retardation, Stevenson, R.E.,
Schwartz, C.E., and Schroer, R.J.. New York: Oxford University Press, 2000:2367.
7. Chelly, J. and Mandel, JL. Monogenic causes of X-linked mental retardation. Nature
Rev Genet, 2, 669, 2001.
8. Ramakers, G.J.A. Rho proteins, mental retardation and the cellular basis of cognition.
Trends in Neurosci, 25, 191, 2002.
9. Mackay, D.J., Nobes, C.D., and Hall A. The Rhos progress: A potential role during
neuritogenesis for the Rho family of GTPases. Trends Neurosci, 18, 496, 1995.
10. Luo, L. Rho GTPases in neuronal morphogenesis. Nat Rev Neurosci, 1, 173, 2000.
11. Ishizaki, H. et al. Role of rab GDP dissociation inhibitor alpha in regulating plasticity
of hippocampal neurotransmission. Proc Natl Acad Sci USA, 97, 11587, 2000.
12. Negishi, M. and Katoh, H. Rho family GTPases as key regulators for neuronal network
formation. J Biochem. (Tokyo) 132, 157, 2002.
13. Billuart, P. et al. Oligophrenin-1 encodes a rhoGAP protein involved in X-linked
mental retardation. Nature, 392, 923, 1998.
14. Fauchereau, F. et al. The RhoGAP activity of OPHN1, a new F-actin-binding protein
is negatively controlled by its amino-terminal domain. Mol. Cell. Neurosci., 23, 574,
2003.
15. Govek, E.E. et al. The X-linked mental retardation protein oligophrenin-1 is required
for dendritic spine morphogenesis. Nat Neurosci., 7, 364, 2004.
16. Comery, T.A. et al. Abnormal dendritic spines in fragile X knockout mice: Maturation
and pruning deficits. Proc. Natl. Acad. Sci. USA, 94, 5401, 1997.
17. Allen, K.M. et al. PAK3 mutation in nonsyndromic X-linked mental retardation. Nat
Genet., 20: 2530, 1998
18. Meng, Y. et al. Abnormal spine morphology and enhaced LTP in LIMK-2 knockout
mice. Neuron 35: 121133, 2002.
19. Kutsche, K. et al. Mutations in ARHGEF6, encoding a guanine nucleotide exchange
factor for Rho GTPases, in patients with X-linked mental retardation. Nat Genet. 26,
24750, 2000.
20. Rosenberger, G. et al. Interaction of alphaPIX (ARHGEF6) with beta-parvin
(PARVB) suggests an involvement of alphaPIX in integrin-mediated signaling. Hum
Mol Genet 12: 155167, 2003.
21. Penzes, P. et al. Rapid induction of dendritic spine morphogenesis by trans-synaptic
ephrinB-EphB receptor activation of the Rho-GEF kalirin. Neuron 37: 263274, 2003.
1903_C028.fm Page 446 Wednesday, July 26, 2006 8:40 PM
22. Matus, A., Brinkhaus, H., and Wagner, U. Actin dynamics in dendritic spines: a form
of regulated plasticity at excitatory synapses. Hippocampus, 10, 555, 2000.
23. Matus, A. Postsynaptic actin and neuronal plasticity. Curr Opin Neurobiol, 9, 561,
1999.
24. Yuste, R. and Bonhoeffer, T. Morphological changes in dendritic spines associated
with long-term synaptic plasticity. Annu Rev Neurosci, 24, 1071, 2001.
25. Purpura, D.P. Dendritic spine dysgenesis and mental retardation. Science, 186, 1126,
1974.
26. DAdamo, P. et al. Mutations in GDI1 are responsible for X-linked non-specific
mental retardation. Nat Genet, 19,134, 1998.
27. Ishizaki, H. et al. Role of rab GDP dissociation inhibitor alpha in regulating plasticity
of hippocampal neurotransmission. Proc Natl Acad Sci USA, 97, 11587, 2000.
28. DAdamo, P. et al. Deletion of the mental retardation gene Gdi1 impairs associative
memory and alters social behavior in mice. Hum Mol Genet, 11, 2567, 2002.
29. Carrie, A. et al. A new member of the IL-1 receptor family highly expressed in
hippocampus and involved in X-linked mental retardation. Nat Genet 23 , 25, 1999.
30. Bahi, N. et al. IL1 receptor accessory protein like, a protein involved in X-linked
mental retardation, interacts with Neuronal Calcium Sensor-1 and regulates exocyto-
sis. Hum Molec Genet, 12, 1425, 2003.
31. Sippy, T. et al. Acute changes in short-term plasticity at synapses with elevated levels
of neuronal calcium sensor-1. Nature Neurosci, 6, 1031, 2003.
32. Meloni, I. et al. FACL4, encoding fatty acid-CoA ligase 4, is mutated in nonspecific
X-linked mental retardation. Nat Genet, 30, 436, 2002.
33. Huttner, W.B. and Schmidt, A. Lipids, lipid modification and lipid-protein interaction
in membrane budding and fission insights from the roles of endophilin A1 and
synaptophysin in synaptic vesicle endocytosis. Curr Opin Neurobiol, 10, 543, 2000.
34. Merienne, K. et al. A missense mutation in RPS6KA3 (RSK2) responsible for non-
specific mental retardation. Nat Genet, 22, 13, 1999.
35. Couvert, P. et al. MECP2 is highly mutated in X-linked mental retardation. Hum Mol
Genet, 10, 941, 2001.
36. Lebel, R.R. et al. Non-syndromic X-linked mental retardation associated with a
missense mutation (P312L) in the FGD1 gene. Clin Genet., 61, 139, 2002.
37. Philip, N. et al. Mutations in the oligophrenin-1 gene (OPHN1) cause X linked
congenital cerebellar hypoplasia. J Med Genet., 40, 441, 2003.
38. Bergmann, C. et al. Oligophrenin 1 (OPHN1) gene mutation causes syndromic X-
linked mental retardation with epilepsy, rostral ventricular enlargement and cerebellar
hypoplasia. Brain, 126, 1537, 2003.
39. Ropers, HH. et al. Nonsyndromic X-linked mental retardation: where are the missing
mutations? Trends Genet., 19, 316, 2003.
40. Molinari, F. et al. Truncating neurotrypsin mutation in autosomal recessive nonsyn-
dromic mental retardation. Science, 298, 1779, 2002.
41. Gschwend, T.P. et al. Neurotrypsin, a novel multidomain serine protease expressed
in the nervous system. Mol Cell Neurosci, 9, 207, 1997.
42. Fukami, M. et al. A member of a gene family on Xp22.3, VCX-A, is deleted in
patients with X-linked nonspecific mental retardation. Am J Hum Genet, 67, 563,
2000.
43. Zemni, R. et al. A new gene involved in X-linked mental retardation identified by
analysis of an X;2 balanced translocation. Nat Genet, 24, 167, 2000.
1903_C028.fm Page 447 Wednesday, July 26, 2006 8:40 PM
44. Freude, K. et al. Mutations in the FTSJ1 Gene Coding for a Novel S-Adenosylme-
thionine-Binding Protein Cause Nonsyndromic X-Linked Mental Retardation. Am J
Hum Genet. 2004.
45. Kleefstra, T. et al. Zinc finger 81 (ZNF81) mutations associated with X-linked mental
retardation. J Med Genet., 41, 394, 2004.
46. Shoichet, S.A. et al. Mutations in the ZNF41 gene are associated with cognitive
deficits: identification of a new candidate for X-linked mental retardation. Am J Hum
Genet., 73, 1341, 2003.
47. Gecz, J. The FMR2 gene, FRAXE and non-specific X-linked mental retardation:
clinical and molecular aspects. Ann Hum Genet., 64, 95106, 2000.
1903_C028.fm Page 448 Wednesday, July 26, 2006 8:40 PM
1903_C029.fm Page 449 Wednesday, July 26, 2006 7:51 PM
29 Pharmacogenetics
Byron C. Jones
CONTENTS
29.1 INTRODUCTION
Pharmacogenetics is the study of individual differences in drug response. The field
has been recognized since the early 1950s, but the biomedical research devoted to
it has been relatively slow in development. Nevertheless, pharmacogenetics has
manifold, important implications for health. For some individuals, for example, what
would otherwise be termed a drugs side effect may in fact be the main drug effect.
Alternatively, an individual may be predisposed to insensitivity to a particular drug,
with implications for dose or to make administration of the drug futile. In this chapter
we will explore the genetic bases for individual differences in drug actions (although
there are environmental effects as well).
449
1903_C029.fm Page 450 Wednesday, July 26, 2006 7:51 PM
distinctions are somewhat arbitrary but focus the attention to different aspects of
drug action.
Pharmacogenetics 451
Drug metabolizing enzymes are classified into Phase I and Phase II enzymes. Phase
I enzymes by oxidation, reduction or hydrolysis increase the polarity of a drug and
in some cases make the drug more active for example, changing codeine to morphine.
Phase II enzymes act by conjugation, adding for example acetyl group, to further
increase polarity and thus enable elimination by the kidneys. Genetic polymorphisms
may occur in either type, however, some of the most important occur in the Phase
I, especially in the major group of oxidizing enzymes, the Cytochrome P 450
enzymes (CYP). The CYP enzymes consist of more than 150 classified into sub-
families. The CYP2 family is particularly interesting for psychopharmacology,
because many of the drugs used to treat neuropsychiatric illnesses are metabolized
by CYP2D6. Drugs so affected include tricyclic and specific serotonin uptake inhibitor
antidepressants (fluoxetine, paroxetine) and haloperidol, thoridizine, perphenazine
and similar antipsychotics. It is further estimated that in all therapeutic categories,
CYP2D6 is involved in the metabolism of more than 20% of prescribed drugs.6
More than 50 allelic variants of this enzyme are known and are used to categorize
individuals into four classes of metabolizers: poor, intermediate, efficient, and ultra-
rapid. Of particular interest is that the development of Parkinsonian symptoms in
individuals taking antipsychotic medication is associated with the poor metabolizing
phenotype. It is obvious, therefore, that drug dosing recommendations should be
made based on the allelic configuration of the individual and one researcher believes
that the differential distribution of the polymorphisms would make genotyping
worthwhile prior to prescription of medications metabolized by CYP2D6.6
were injected intraperitoneally with 3.3 g/kg ethanol in saline and then observed for
loss of righting response by placing the animals in a supine position. The time from
loss of righting response to time of regain was recorded as the dependent variable.
Large individual differences were observed among these animals and mass selection
over several generations produced relatively rapid differentiation between the two
lines. The more sensitive line, as determined by sleep time was named long sleep
(LS) and the less sensitive line was named short sleep (SS). The pharmacogenetics
of alcohol is more completely covered in the next chapter by J.C. Crabbe. The
example of the LS and SS mice is used here to illustrate two important principles
in pharmacogenetics. First, what is the nature of the phenotype? In the example of
the LS and SS mice, does sleep time reflect a pharmacodynamic effect (i.e., target
tissue sensitivity) or is the effect a result of different rates of alcohol disappearance
between the two lines, i.e., do SS mice eliminate alcohol more rapidly than LS?
Tabakoff and Ritzmann addressed this question directly by measuring blood alcohol
concentrations (BEC) at loss and regain of the righting response9 and they reported
that the BEC for the SS was nearly twice that of the LS both at loss and at regain
of righting. Thus, this gives us good evidence that pharmacodynamic processes
account for most of the differential effect of ethanol in these two lines of mice. The
second principle illustrated by the LS-SS example is that differential response pro-
gressed at nearly steady rate across generations, rather than within one or two
generations of selection. This gives good evidence for hypnotic sensitivity to ethanol
as being influenced by multiple genes (alleles) with algebraic additive effects (some
increase effect, some decrease) and that across generations, each line accumulates
more of the alleles that push the phenotype in the direction of selection. In the
pharmacokinetic approach, most of the attention is focused on single genes and their
alleles that alter the function of metabolizing enzymes. In the pharmacodynamic
approach, most of the phenotypes measured show continuous variation, respond to
selection very much like the LS-SS, thus indicating polygenic influence.
Pharmacogenetics 453
that individuals who carry at least one long allele of 5-HTTLPR fare better in long-
term antidepressant treatment than those who are homozygous for the short allele.14
In animals, approaches to single-gene effects on drug sensitivity include induced
mutations by irradiation, chemical mutagenesis, and creation of targeted null
mutants. Genes may also be amplified, i.e., by insertion of multiple copies of desired
alleles into the genome. These models can be useful for elucidating the role of
specific proteins, receptors, etc. in drug action. For example, in the serotonin 1b
receptor null mutant mouse, treatment with paroxetine, a selective serotonin reuptake
inhibitor used to treat depression causes a twofold increase in extracellular serotonin
compared with treatment with the same drug in wild-type mice. In humans, allelic
variants of the 5HT1B receptor gene promoter region have been described.15 The
allelic variants are associated with transcriptional activity of the receptor gene,
showing a more than twofold difference between the highest efficiency and lowest
efficiency alleles. Although the allelic configuration may not affect the action of
specific serotonin reuptake inhibitors, the fact that they do exist, in light of the work
in the null mutant mice should give cause for further investigation into possible
ramifications for possible individual differences in drug treatment for depression.
Pharmacogenetics 455
29.7 SUMMARY
Pharmacogenetics is a small, but important component of the study of individual
differences in response to environmental challenges to homeostatic biological sys-
tems. It is not different in principle from inborn errors of metabolism, dietary
problems such as lactose intolerance, disease resistance, or individual differences in
sensitivity to toxins. What makes pharmacogenetics worthy of study is the great
importance of its impact on human health and the eventual development of drug
treatment strategies based on an individuals genetic constitution.
REFERENCES
1. Garrod, A.E. The incidence of alkaptonuria: a study in chemical individuality. The
Lancet ii:16161620, 1902.
2. Flling, A. Excretion of phenylpyruvic acid in urine as a metabolic. anomaly in
connection with imbecility. Nord Med Tidskr. 8: 10541059, 1934
3. Kalow, W. Familial incidence of low pseudocholinesterase level. The Lancet
2:576577, 1956
4. Crabb, D.W., Edenberg, H. J., Bosron, W. F., Li, T.-K. Genotypes for aldehyde
dehydrogenase deficiency and alcohol sensitivity: the inactive ALDH2*2 allele is
dominant. J. Clin. Invest. 83: 314316, 1989.
5. Yokoyama, A., Kato, H., Yokoyama, T., Tsujinaka, T., Muto, M., Omori, T., Haneda,
T., Kumagai, Y., Igaki, H., Yokoyama, M., Watanabe, H., Fukuda, H., Yoshimizu, H.
and Ingelman-Sundberg, M. Genetic polymorphisms of cytochrome P4502D6
(CYP2D6): clinical consequences, evolutionary aspects and functional diversity. The
Pharmacogenetics Journal 5:613, 2005.
6. Mardones, J. On the relationship between deficiency of B vitamins and alcohol intake
in rats. Q J Stud Alcohol 12:563575, 1951.
7. McClearn, G.E. and Kakihana, R. Selective breeding for ethanol sensitivity: short-
sleep and long-sleep mice. In Development of Animal Models as Pharmacogenetic
Tools G.E. McClearn, R. A. Deitrich and V.G. Erwin, eds., USDHHS-NIAAA
Research Monographs No 6, Washington, pp. 147159, 1981.
8. Tabakoff, B. and Ritzmann, R.F. Acute tolerance in inbred and selected lines of mice
Drug Alcohol Depend 4:8790, 1979.
9. Jones, B., Goldstine, R., Kegel, M., Gurley, M., and Reyes, E. The influence of
infantile handling, age and strain on alcohol selection in mice. Alcohol, 2:327331,
1985.
10. Jones, B. C., Connell, J. M. and Erwin, V. G. Isolate housing affects ethanol sensitivity
in long-sleep and short-sleep mice. Pharmacol Biochem Behav, 35:469472, 1990.
11. Erikson, K.M., Jones, B.C., and Beard, J.L. Iron deficiency alters dopamine trans-
porter functioning in rat striatum. Journal of Nutrition, 130:28312837, 2000.
12. Jones, B.C., Wheeler, D.S., Beard, J.L. and Grigson, P.S. Iron deficiency decreases
acquisition of and suppresses responding for cocaine. Pharmacol Biochem Behav,
73:813819, 2002.
13. De Groote, L., Olivier, B. and Westenberg, G.M. The effects of selective serotonin
uptake inhibitors on extracellular 5-HT levels in the hippocampus of 5-HT1B receptor
knockout mice. Eur J Pharm 439:93100, 2002.
1903_C029.fm Page 456 Wednesday, July 26, 2006 7:51 PM
14. Duan, J., Vander Molen, J.E., Martinolich, L., Mowry, B.J., Levinson, D.F., Crowe,
R.R., Silverman, J.M. and Gejman, P.V. Polymorphisms in the 5-untranslated region
of the human serotonin receptor 1B (HTR1B) gene affect gene expression. Mol
Psychiat 8:901910, 2003.
15. Jones, B.C., Tarantino, L.M., Rodriguez, L.A., Reed, C.L., McClearn, G.E., Plomin,
R. and Erwin, V.G. Quantitative trait loci analysis of cocaine-related behaviours and
neurochemistry. Pharmacogenetics 9:607617, 1999.
16. DiPetrillo, K., Wang, X., Stylianou, I.M. and Paigen, B. Bioinformatics toolbox for
narrowing rodent quantitative trait loci. TRENDS in Genetics, 21:68392, 2005.
17. Shirley, R.L., Walter, N.A., Reilly, M.T., Fehr, C. and Buck, K.J. Mpdz is a quanti-
tative trait gene for drug withdrawal seizures. Nat Neurosci 7:699700, 2004.
18. Kalow, W., Meyer, U.A. and Tyndale, R.F., eds. Pharmacogenomics, Marcel Decker,
New York, 1991.
1903_C030.fm Page 457 Wednesday, July 26, 2006 7:51 PM
30 Alcohol
Psychopharmacogenetics
John C. Crabbe
CONTENTS
30.1 INTRODUCTION
Of all drugs of abuse studied to determine genetic contributions to susceptibility to
their effects, alcohol has been by far the most frequent focus. Both animal model
and human genetic studies will be mentioned here. A historically rich literature with
genetic animal models has explored the general contribution of genetics to alcohol
responsiveness, and has been helpful in elucidating the drugs mechanism of action
on the nervous system. Some studies using genetic animal models have attempted
to identify specific genes that increase or decrease responsiveness for a number of
alcohols effects. Most human studies have been of alcoholics and their relatives,
and have compared relative risks for alcohol dependence disorders in twins, adoptees,
or other relatives. More recently, human studies have also addressed the goal of
identifying individual genes that might contribute to alcoholism risk, or to individual
differences in endophenotypes, which also are associated with alcoholism risk.
457
1903_C030.fm Page 458 Wednesday, July 26, 2006 7:51 PM
A comprehensive review of these studies is beyond the scope of this chapter, but a
few examples will be discussed below, and readers will be referred to relevant
reviews.
40 common inbred strains. This effort, the Mouse Phenome Project, can be assessed
to search for evidence of pleiotropic gene effects.48
Most alcohol QTL studies have been carried out using the BXD RI strains of mice
as a starting point for rough QTL identification, followed by verification tests in
1903_C030.fm Page 461 Wednesday, July 26, 2006 7:51 PM
other populations, such as the F2 cross of C57BL/6J and DBA/2J, lines selectively
bred from the F2 cross, or other selected populations. The experiments that have
progressed the furthest have shown that three loci for acute alcohol withdrawal
severity have been definitively mapped with very high LOD scores to regions of
mouse chromosomes 1, 4, and 11.10 Genes encoding the "1, "6, and (2 subunits of
the neuroinhibitory GABAA receptor map near the chromosome 11 QTL. Studies
using inbred strains, lines selectively bred from the B6D2F2 population, and a
number of so-called congenic strains were then used to reduce the area of genome
surrounding the chromosome 4 QTL until it contained only a handful of candidate
genes.28 The remaining candidate genes were gradually eliminated because they did
not differ in sequence, expression, or function in ways that systematically co-varied
with alcohol withdrawal severity. In the end, a single gene remained, Mpdz, which
must be the QT Gene for this QTL. MPDZ is a multiple PDZ domain protein that
affects the coupling of neurotransmitters with their receptors, including the binding
of serotonin with the 5-HT2 receptor. This gene also affects the severity of with-
drawal from pentobarbital, another nervous system depressant that interacts with
GABA receptors. This provides another example of pleiotropic gene effects, and
Mpdz may indeed also play a role in susceptibility to seizures more generally, in
addition to its specific role in alcohol and pentobarbital withdawal.28,63
For sensitivity to high-dose ethanol, five significant QTLs with high LOD scores
have been identified on chromosomes 1, 2, 8, 11, and 15.41 Starting with the LS and SS
selected lines and using RI strains and crosses derived from them, the QTLs were nom-
inated and subsequently verified in a multistage procedure much like that used for acute
alcohol withdrawal.24 Some interesting potential candidate genes near the QTL regions
include Ntsr on chromosome 2, which codes for the high-affinity neurotensin receptor.
Several different research groups have reported mapping QTLs for ethanol
preference drinking and related traits.3,44,45,52,65 The significant QTLs are found on
chromosomes 2, 9, 11 and 15, and several other QTLs have been suggested on
chromosomes 1, 3, 4, 5, 7, and 13. The chromosome 9 QTL is near genes encoding
both the dopamine D2 and serotonin 5-HT1B receptor, Drd2, and Htr1b, respectively.
The results of these studies are not currently in complete agreement, probably
because different crosses, sexes, strains, and mapping strategies were employed, as
well as different measures of preference-related drinking. However, a meta-analysis
revealed a remarkable degree of convergence of findings for several QTLs.6
A number of other alcohol-related traits are also being mapped, although the
projects are in earlier stages. One set of studies mapping the effect of ethanol to
stimulate locomotor activity in mice has used a variety of mapping methods.20,34 Other
traits being mapped include: ethanol-conditioned place preference and taste aversion;
hypothermia; traits related to ataxia; tolerance to some of these effects; and ethanols
anxiolytic effects. Progress in ethanol QTL mapping has been reviewed recently.9,15,30,49
such as: other psychiatric disorders (e.g., antisocial personality, depression, attention
deficit hyperactivity disorder); personality characteristics (e.g., impulsivity, aggres-
siveness); and substance abuse.54 Disentangling the chains of cause and effect among
these different diagnostic aspects makes this a very difficult area of research to
interpret. Finally, each typology for conceptualizing alcohol dependence disorders
leads to emphasis on different interactions between heritable risk factors and the
modulating effects of environmnental factors that are also important for turning risk
into diagnosis.33,38
Recently, much attention has turned to analyzing identifiable aspects of alco-
holism and alcohol responses for their separable genetic determinants. By using
other heritable traits as potential indicators, it is hoped that the genetic basis for
such traits will make it easier to determine genebehavior relationships than is trying
to study the diagnoses directly. Such studies are meeting with some success.21,26
30.5 CONCLUSIONS
The field of alcohol pharmacogenetics offers an interesting viewpoint on the explo-
sive impact modern molecular biological methods have had on genetic analyses. The
rich history of standard genetic animal model research into alcohols effects has
provided many cases where clear genetic influence was known to exist. Thus, the
move to identification of specific candidate genes and markers occurred quite rapidly
after the advent of dense genetic maps and the ability to target specific genes for
deletion and overexpression. Because there are readily available populations of
human alcoholics with well-defined pedigrees, it will be possible for the animal
model and human genetic findings to be related rather directly. This offers the hope
that in addition to improving our understanding of how alcohol works, genetic studies
will be able to provide novel methods for identifying at-risk individuals and for
developing new pharmacotherapies to attack this major disease.
ACKNOWLEDGMENTS
Preparation of this chapter was supported by a grant from the Department of Veterans
Affairs and by NIH Grants from NIAAA and NIDA.
REFERENCES
1. Allan AM, Harris RA. Neurochemical studies of genetic differences in alcohol action,
In: Crabbe JC, Harris RA, editors. The Genetic Basis for Alcohol and Drug Actions.
New York: Plenum; 1991. pp. 10552.
2. Anthenelli RM, Tabakoff B. The search for biological markers. Alcohol Health and
Research World 1996;19:17681.
3. Bachmanov AA, Reed DR, Li X, Li S, Beauchamp GK, Tordoff MG. Voluntary
ethanol consumption by mice: genome-wide analysis of quantitative trait loci and
1903_C030.fm Page 465 Wednesday, July 26, 2006 7:51 PM
22. Dick DM, Nurnberger J, Jr., Edenberg HJ, Goate A, Crowe R, Rice J et al. Suggestive
linkage on chromosome 1 for a quantitative alcohol-related phenotype. Alcohol Clin.
Exp. Res. 2002;26:145360.
23. Ehringer MA, Thompson J, Conroy O, Goldman D, Smith TL, Schuckit MA et al.
Human alcoholism studies of genes identified through mouse quantitative trait locus
analysis. Addiction Biology 2002;7:36571.
24. Ehringer MA, Thompson J, Conroy O, Yang F, Hink R, Bennett B et al. Fine mapping
of polymorphic alcohol-related quantitative trait loci candidate genes using interval-
specific congenic recombinant mice. Alcohol Clin. Exp. Res. 2002;26:160308.
25. Enoch MA, Goldman D. The genetics of alcoholism and alcohol abuse. Curr
Psychiatry Rep 2001;3:14451.
26. Enoch MA, Schuckit MA, Johnson And BA, Goldman D. Genetics of alcoholism
using intermediate phenotypes. Alcohol Clin. Exp. Res. 2003;27:16976.
27. Fan L, Bellinger F, Ge YL, Wilce P. Genetic study of alcoholism and novel gene
expression in the alcoholic brain. Addict. Biol. 2004;9:1118.
28. Fehr C, Shirley RL, Belknap JK, Crabbe JC, Buck KJ. Congenic mapping of alcohol
and pentobarbital withdrawal liability loci to a <1 centimorgan interval of murine
chromosome 4: Identification of Mpdz as a candidate gene. J Neurosci 2002;22:
373038.
29. Files FJ, Samson HH, Denning CE, Marvin S. Comparison of alcohol-preferring and
nonpreferring selectively bred rat lines. II. Operant self-administration in a continu-
ous-access situation. Alcohol Clin. Exp. Res. 1998;22:214758.
30. Flint J. Analysis of quantitative trait loci that influence animal behavior. J Neurobiol.
2003;54:4677.
31. Gerlai R. Molecular genetic analysis of mammalian behavior and brain processes:
caveats and perspectives. Seminars in the Neurosciences 1996;8:15361.
32. Heath AC. Genetic influences on alcoholism risk: a review of adoption and twin
studies. Alcohol Health and Research World 1995;19:16671.
33. Heath AC, Todorov AA, Nelson EC, Madden PAF, Bucholz KK, Martin NG. Gene-
environment interaction effects on behavioral variation and risk of complex disorders:
The example of alcoholism and other psychiatric disorders. Twin Research 2004;5:3037.
34. Hitzemann R, Demarest K, Koyner J, Cipp L, Patel N, Rasmussen E et al. Effect of
genetic cross on the detection of quantitative trait loci and a novel approach to
mapping QTLs. Pharmacol. Biochem. Behav. 2000;67:76772.
35. Holmes RS. Genetic variants of enzymes of alcohol and aldehyde metabolism.
Alcohol Clin. Exp. Res. 1985;9:53538.
36. Homanics GE, Ferguson C, Quinlan JJ, Daggett J, Snyder K, Lagenaur C et al.
Gene knockout of the alpha6 subunit of the gamma-aminobutyric acid type A recep-
tor: lack of effect on responses to ethanol, pentobarbital, and general anesthetics.
Mol. Pharmacol. 1997;51:58896.
37. Homanics GE, Xu Y, Tang P. Integrated approaches to the action of general anesthetics
and alcohol. Physiol Behav. 2002;77:49599.
38. Johnson EO, van den Bree M, Gupman AE, Pickens RW. Extension of a typology
of alcohol dependence based on relative genetic and environmental loading. Alcohol
Clin. Exp. Res. 1998;22:142129.
39. Lewohl JM, Wang L, Miles MF, Zhang L, Dodd PR, Harris RA. Gene expression in
human alcoholism: microarray analysis of frontal cortex. Alcohol Clin. Exp. Res.
2000;24:187382.
1903_C030.fm Page 467 Wednesday, July 26, 2006 7:51 PM
40. Lumeng L, Murphy JM, McBride WJ, Li T-K. Genetic influences on alcohol prefer-
ence in animals, In: Begleiter H, Kissin B, eds., The Genetics of Alcoholism. New
York: Oxford University Press; 1995. p. 165201.
41. Markel PD, Bennett B, Beeson M, Gordon L, Johnson TE. Confirmation of quanti-
tative trait loci for ethanol sensitivity in long-sleep and short-sleep mice. Genome
Res. 1997;7:9299.
42. Mayfield RD, Lewohl JM, Dodd PR, Herlihy A, Liu J, Harris RA. Patterns of gene
expression are altered in the frontal and motor cortices of human alcoholics.
J. Neurochem. 2002;81:80213.
43. McClearn GE, Rodgers DA. Differences in alcohol preference among inbred strains
of mice. Quarterly Journal of Studies on Alcohol 1959;20:69195.
44. McClearn GE, Tarantino LM, Rodriguez LA, Jones BC, Blizard DA, Plomin R.
Genotypic selection provides experimental confirmation for an alcohol consumption
quantitative trait locus in mouse. Mol. Psychiat 1997;2:48689.
45. Melo JA, Shendure J, Pociask K, Silver LM. Identification of sex-specific quantitative
trait loci controlling alcohol preference in C57BL/6 mice. Nat. Genet. 1996;13:14753.
46. Metten P, Crabbe JC. Dependence and withdrawal, In: Deitrich RA, Erwin VG, eds.
Pharmacological Effects of Ethanol on the Nervous System. Boca Raton FL: CRC
Press; 1996. p. 26990.
47. Metten P, Crabbe JC. Genetic determinants of severity of acute withdrawal from
diazepam in mice: commonality with ethanol and pentobarbital. Pharmacol Biochem
Behav 1999;63:47379.
48. Grubb SC, Churchill GA, Boque MA. A collaborative database of infred mouse strain
characteristics. Bioinformatics 2004; 20: 28572859.
49. Palmer AA, Phillips TJ. Quantitative trait locus (QTL) mapping in mice, In: Liu Y,
Lovinger DM, eds. Methods in Alcohol-Related Neuroscience Research. Boca Raton
FL: CRC Press; 2002. p. 130.
50. Phillips TJ. Behavior genetics of drug sensitization. Crit Rev. Neurobiol. 1997;11:2133.
51. Phillips TJ, Crabbe JC. Behavioral studies of genetic differences in alcohol action,
In: Crabbe JC, Harris RA, editors. The Genetic Basis of Alcohol and Drug Actions.
New York: Plenum Press; 1991. p. 25104.
52. Phillips TJ, Crabbe JC, Metten P, Belknap JK. Localization of genes affecting alcohol
drinking in mice. Alcohol Clin. Exp. Res. 1994;18:93141.
53. Phillips TJ, Hen R, Crabbe JC. Complications associated with genetic background
effects in research using knockout mice. Psychopharmacology 1999;147:57.
54. Pickens RW, Svikis DS, McGue M, LaBuda MC. Common genetic mechanisms in
alcohol, drug, and mental disorder comorbidity. Drug and Alcohol Dependence
1995;39:12938.
55. Reich T. A genomic survey of alcohol dependence and related phenotypes: results
from the Collaborative Study on the Genetics of Alcoholism (COGA). Alcohol Clin.
Exp. Res. 1996;20:133A7A.
56. Rodd ZA, Bell RL, Melendez RI, Kuc KA, Lumeng L, Li TK et al. Comparison of
intracranial self-administration of ethanol within the posterior ventral tegmental area
between alcohol-preferring and wistar rats. Alcohol Clin. Exp. Res. 2004;28:121219.
57. Rodd-Henricks ZA, McKinzie DL, Melendez RI, Berry N, Murphy JM, McBride
WJ. Effects of serotonin-3 receptor antagonists on the intracranial self-administration
of ethanol within the ventral tegmental area of Wistar rats. Psychopharmacology
2003;165:25259.
1903_C030.fm Page 468 Wednesday, July 26, 2006 7:51 PM
58. Rout UK, Holmes RS. Alcohol dehydrogenases and aldehyde dehydrogenases among
inbred strains of mice: multiplicity, development, genetic studies and metabolic roles.
Addiction Biology 1996;1:34962.
59. Samson HH, Files FJ, Denning C, Marvin S. Comparison of alcohol-preferring and
nonpreferring selectively bred rat lines. I. Ethanol initiation and limited access operant
self-administration. Alcohol Clin. Exp. Res. 1998;22:213346.
60. Schuckit MA. Self-rating of alcohol intoxication by young men with and without
family histories of alcoholism. J. Stud. Alcohol 1980;41:24249.
61. Schuckit MA, Smith TL. An 8-year follow-up of 450 sons of alcoholic and control
subjects. Archives of General Psychiatry 1996;53:20210.
62. Schumann G, Spanagel R, Mann K. Candidate genes for alcohol dependence: animal
studies. Alcohol Clin Exp Res 2003;27:88088.
63. Shirley RL, Walter NA, Reilly MT, Fehr C, Buck KJ. Mpdz is a quantitative trait
gene for drug withdrawal seizures. Nat. Neurosci. 2004;7:699700.
64. Song J, Koller DL, Foroud T, Carr K, Zhao J, Rice J et al. Association of GABAA
receptors and alcohol dependence and the effects of genetic imprinting. Am. J. Med.
Genet. 2003;117B:3945.
65. Vadasz C, Saito M, Balla A, Kiraly I, Gyetvai B, Mikics E et al. Mapping of
quantitative trait loci for ethanol preference in quasi-congenic strains. Alcohol
2000;20:16171.
66. Wafford KA, Burnett D, Harris RA, Whiting PJ. GABAA receptor subunit expression
and sensitivity to ethanol. Alcohol Suppl 1993;2:32730.
67. Wall TL, Ehlers CL. Genetic influences affecting alcohol use among Asians. Alcohol
Health and Research World 1995;19:18489.
68. Wehner JM, Bowers BJ. Use of transgenics, null mutants, and antisense approaches
to study ethanol's actions. Alcohol Clin. Exp. Res. 1995;19:81120.
69. Weitemier AZ, Woerner A, Backstrom P, Hyytia AP, Ryabinin AE. Expression of
c-Fos in Alko Alcohol rats responding for ethanol in an operant paradigm. Alcohol
Clin. Exp. Res. 2001;25:70410.
1903_Index.fm Page 469 Wednesday, July 26, 2006 7:52 PM
Index
A historical research developments, 11
humans, 286287
Aarskog-Scott syndrome, 442 interim solution, 25
ABA (Allen Brain Atlas), 426 mice, 282285, 283, 285286, 287
Abiola studies, 116 phenotypes, 3031
Abkevich studies, 237 rodents, 9
Absence, genetic activity, 23 types, 282
Acceptance of association, 175 voles, 12
Accessory olfactory bulb (AOB), 2223 AILS (advanced intercross lines), 60, 133
Acetaldehyde toxicity, 451, 459 Akaikes information criterion (AIC), 256
Acoustic startle reflex, 295 AKR/J strains, 134
Acoustic stimuli, 320321 Albino strains and albinism, 138, 414
Actin cytoskeleton, 437, 437439 Alcohol, alcohol-related disorders, and abuse,
Activators, regulation of transcription, 81 see also Ethanol and ethanol resistance
Active avoidance paradigms, 295, 298 bipolar disorders, 234
Active genotype-environment correlation, 190 brain dopamine systems, 371372
Addictions, 12, 87, see also specific substance Caenorhabditis elegans, 10
Additive genetic effects and variations, 3334, 34, expression variation sources, 87
40, 184 false positives, 175
Additive model, gene-environment interactions, gene-environment interactions, 35, 191
179 historical research, 7
ADHD (attention-deficit hyperactivity disorder), inbreeding, 136
269 metabolism, 450
Adobe Photoshop, 425 multi-gene disorder, 30
Adoption studies, see also Family and twin pharmacodynamics, 451452
studies; Human studies; Twins studies QTLs, 301
alcohol psychopharmacogenetics, 462463 recurrent unipolar disorders, 236
bipolar disorders, 229 response and sensitivity, 451452
historical research developments, 4, 12 rodents, 9, 11
schizophrenia, 214216 serotonin, 2425
Advanced intercross lines (AILS), 60, 133 Alcohol psychopharmacogenetics
Affected cohorts requirement, 174 adoption studies, 462463
Affective disorders, major animal studies, 458462
bipolar disorders, 228236 candidate gene studies, 463
family studies, 228, 236 fundamentals, 457458, 464
fundamentals, 227, 237238 gene expression analysis, 464
linkage disequilibrium studies, 231236, gene identification, 460462
233 human studies, 462464
molecular linkage studies, 230231, inbred strains studies, 458459
232233, 237 QTLs identification, 463464
recurrent unipolar disorders, 236237 risk marker identification, 463464
twins studies, 228229, 229, 237 selectively bred lines studies, 459460
Affymetrix system, 9091, 381, 416, 418, 420 targeting studies, 460462
Aggression twins studies, 462463
fundamentals, 281282 ALDH2 variants, 450451, 463
gene-environment interactions, 195 Algerian family study, 443
genetic vulnerability, 178 Alkaptonuria, 4, 450
469
1903_Index.fm Page 470 Wednesday, July 26, 2006 7:52 PM
Index 471
Index 473
Index 475
Index 477
Index 479
Genetic quality, JAX rodents, 143 Hall, Calvin S., 46, 46, 292
Genetic reference populations (GRPs), 101104, Hall, Jeff, 11
102 Haloperidol, 374375, 451
The Genetics and Social Behavior of the Dog, 7 Hamer, Dean, 12
The Genetics of Behavior, 8 Hamilton Station for Behavioral Research, 4
Genetic variation, schizophrenia, 211216 Hamster prion protein, 138
GenMapp, 109 Hancock and Keverne, Brennan studies, 22
GenNetwork, 419420, see also GeneNetwork Haplotype and haplotype mapping
(GN) alternative strategies, 60
Genome Institute of Novartis, 107 bioinformatics, 96
Genome integration, 109110 family-based association studies, 171
Genome location, 100, 101 linkage, 32
Genome sequence importance, 131 positional cloning, 6667
Genome size, human, 169 schizophrenia, 221
Genome-tagged mice, 122 Haplotype Trend Regression, 272
Genotype-environment interactions HAPMAP, 173
animal studies, 195197, 196197 Hariri studies, 194
human studies, 191192, 191194 Harm avoidance
Genotype-phenotype relationships, 31 epistasis, 202
GENSAT, 412, 425429 fibromyalgia, 267
Geotaxes, 6 linkage studies, 271
Gerlai studies, 43 phenotypical heterogeneity, 177
Gershenfeld studies, 49 theoretical personality approach, 265
Gershon, Badner and, studies, 231 Harris, R.Adron, 11
Ginsburg, Benson, 47, 910, 14 Hart, Demant and, studies, 59
Giot studies, 334 Hayman studies, 46
Gomez studies, 366 HDL (high density lipoprotein) cholesterol, 24
Gorwood studies, 169180 Health, 183184, 187
Gottesman and Bertelsen studies, 214 Heath, Andrew, 13
Gottesman and McGue studies, 219 Hedgecock and Russel studies, 364
Gottesman and Shields studies, 217 Hendersons mixed model equation, 254255
Gottesman studies, 12, 209222 Henderson studies, 9, 12, 195
Gottliebs approach, 22 Hendley studies, 299
Gottlieb studies, 1726 Hereditary Genius, 3
Govek studies, 438 Heritability
GRAIL, 64 additive-genetic effects, 41
Greenspan, Ralph, 11 aggression, 287
GRIF (Gene Reference Into Function), 104, 109 alcoholism typologies, 191
Grilli studies, 386 altruism, 268
Grossfield, K., 8 Alzheimers disease, 65, 138
Gross morphological changes, 20 antisocial behaviors, 190
GSL (generalized single locus) model, 217218 dominance, 256
Guenet studies, 119 fundamentals, 185
Gurling studies, 219 genetic reference populations, 102
Gustatory receptors, 324325 misapplication, quantitative genetics, 5051
Gutierrez studies, 235 phenotypes, 266
quantitative genetic analysis, 299
rearing behavior, 48
H recurrent unipolar disorders, 237
schizophrenia, 219
HAB rats, 300 SIP-IQ relationship, 49
Hahn, Martin, 9, 12 theoretical background, 41
Haldane, J.S., 18 Heroin addiction, 287, 372
Haldane studies, 4 Heston studies, 7, 215
Half-life, association, 174 Heterochromatin, gene silencing, 86
1903_Index.fm Page 481 Wednesday, July 26, 2006 7:52 PM
Index 481
Index 483
Keverne, Brennan, Hancock and, studies, 22 LD, see Linkage and linkage disequilibrium (LD)
KID (kinase inducible domain), 83 Learned helplessness, 295
Kimura studies, 328 Learning and memory
Kinase inducible domain (KID), 83 Caenorhabditis elegans, 10, 364367,
King studies, 247258 365366
Kirov studies, 235 gene expression, 74
Kitamoto studies, 343 historical research developments, 9
Klinefelters condition, calico cats, 31 rodents, 11
Klinefelters Syndrome, 12 Learning and memory, Drosophila melanogaster
Knockout mice (fruit flies)
congenic strains and substrains, 118, brain activity imaging, 339340
119126, 121122 fundamentals, 333334, 335337
fundamentals, 139 historical research development, 11
gene-gene interactions, 35 localization, olfactory memory, 338,
historical research developments, 1112 343345
Konopka, Ronald, 8 mutagensis, 139
Korstanje studies, 56 mutants, 340341
Kraemer and Thiemann studies, 156157 neuronal circuits, 338339, 338339
Kraepelin, Emil, 210211 olfactory memory, 343347
Kraus studies, 386 phases, olfactory memory, 345347,
Krech, D., 7 346347
Kruglyak, Lander and, studies, 153 temporal control, gene expression, 341343,
Kubiak studies, 363 342
Kunugi studies, 236 tools, anatomo-functional maps, 343
Kutsche studies, 439 Web resources, 335337
Kyoto University, 141 Least squares estimation, 253
Kyriacou, Charalambos P., 11 Lee studies, 385
Lennox and Wolfe studies, 270
Lesch studies, 191193, 202
L Levinson studies, 231
Lewis rats (LEW), 132, 300301
Laboratory animals, see Animals; Sample size Li, Bausell and, studies, 157, 165
requirements Limbic-system lesions, 43
Laboratory Information Management Systems LIMS (Laboratory Information Management
(LIMS), 97 Systems), 97
Laboratory mice strains and substrains, 141, 142, Lindzey, Gardner, 8
see also specific strain or substrain Linkage and linkage disequilibrium (LD)
LAB rats, 300 association studies, 172175
Lack, genetic activity, 23 bipolar disorders, 230236, 232233
Lac repressor, 83 fundamentals, 32
Lactation, aggression, 282 genetic correlations, 44
Lagerspetz, K., 7 IIPMF, 394, 395, 396
LAL (long attack latency), 12 law of independent assortment, 32
Landau, Virginia, 258 phenotypes, 271272
Lander, Eric, 13 recurrent unipolar disorders, 233, 237
Lander and Kruglyak studies, 153 schizophrenia, 219221
Lans studies, 357 theoretical background, 4647
Lassalle, Roullet and, studies, 401 Lipp and Schwegler studies, 403
Lassalle studies, 129146 Lipp and Wolfer studies, 392
Latent semantic indexing (LSI), 423 Lipp studies, 12, 44, 389407
Latent semantic relationships, 109 Lipsky studies, 386
Latent variable modelo, 251, 252 Lissencephaly, 23
Laumonnier studies, 178 Locality assumption, 4344
Law of Independent Assortment (Mendel), 32 Localization, olfactory memory, 338,
Law of Segregation (Mendel), 3132 343345
1903_Index.fm Page 484 Wednesday, July 26, 2006 7:52 PM
Index 485
Index 487
Index 489
Index 491
Index 493
Transcription, see also Expression profiling; University of Tennessee Health Sciences Center,
Gene expression 12, 136137
factors, 76 Unpacked model, 20, 21
fundamentals, 7680, 7778 3 untranslated region (3UTR), 82
regulation, 8085, 84
signal transduction, 8385, 84
Transgenesis, 144, 303 V
Transgenic rodents and studies
alcohol psychopharmacogenetics, 461462 Vale, Jack, 9
fundamentals, 138 Van Abeelen, Crusio and, studies, 43
historical research developments, 1112 van Abeelen studies, 8, 38, 52
Translocations, 139 Vandenberg, S., 8
Transmission disequilibrium test (TDT), 171 van Oortmerssen, Geert, 7, 9
TRANSMIT program, 172 Van Tol, Wong, Buckle and, studies, 24
Transporter and receptor density, 376, 377379, Variable number of tandem repeats (VNTRs),
378379 235
Tridimensional Personality Questionnaire (TPQ), Variances
265, 267268 Hendersons mixed model equation,
Trullas and Skolnick studies, 297 254255
Tryon effect, 46 latent variable modelo, 251, 252
Tryon studies, 4, 46 least squares estimation, 253
Tryptopham hydroxylase, aggression, 287 multiple regression, 252
Tsien, J.Z., 11 negative variance components, 253
Tully, Tim, 11 pedigree analyses, 254255
Turners Syndrome, 7, 12 proportions estimation, 252
Turri studies, 301 theoretical background, 4546
Twins studies, see also Adoption studies; Family Variations, genetic, 8788, 211216
and twin studies; Human studies Vasopressin systems, voles, 12
alcohol psychopharmacogenetics, 462463 Vaughn, C.H., 11
bipolar disorders, 228229, 229 Venard and Jallon studies, 321
general behavior genetic methods, 184185 Venard studies, 324
historical research developments, 12 Vincent studies, 235
Lousiville Twin Study, 7, 10 Visual stimuli, 320321
nerve conduction velocity, 49 VNTRs (variable number of tandem repeats), 235
recurrent unipolar disorders, 237 Volavka studies, 287
schizophrenia, 212, 213214 Voles, 12
22 design, 165166, 166167 Volkow studies, 372
Two independent groups, 158159, 158159 Von Knorring studies, 229
Two-way avoidance learning, 394, 403 von Teschermak studies, 4
Two-way factorial designs, 162165
Tyrosine, 450
W
U Waddington, Conrad H., 30, 34
Wahlsten, Cohen and, studies, 157
Ultrasonic pup vocalization, 295, 298 Wahlsten studies, 12, 149167
Unconditioned stress responses, 293, 294, 295, Wald, George, 4
296 Waldman, Irwin, 12
Unique environmental effects, 184 Walker studies, 382
Univariate analysis, 45, 185186 Ward studies, 357
Univariate case foundation, 254255 Water maze, 163, 403404, see also Mazes;
University of California San Diegos Biology Morris water maze
Workbench, 105 Watson, James, 5, 75
University of Illinois Biology Student Watson-Crick base pairing, 76, 86, 88
Workbench, 106 Watson-Crick complementarity, 77
1903_Index.fm Page 495 Wednesday, July 26, 2006 7:52 PM
Index 495
497
FIGURE 16.2 Confirmed linkage loci for bipolar disorder (B), schizophrenia (S), and both
disorders (*) on a diagram of the human genome.
KC KC
Ca Ca
Ca
LH
' '
' LH
AGT AGT
' '
AL AL
FIGURE 23.1 Drosophila olfactory memory system. Olfactory sensory neurons project to
the antennal lobes (ALs). From there, projection neurons project through the antenno-glom-
erular tract (AGT) and connect mushroom body (MB) dendrites localized in the calyx (Ca),
as well as the lateral horn (LH). Each MB is composed of about 2,500 neurons, the Kenyon
cells (KC). Three types of KC project in five lobes: /, / and .
1903_Color Insert.fm Page 498 Thursday, July 27, 2006 6:39 PM
AAAAA
GAL4-mRNA
GFP expressed
within MB
AAAAA
GFP-mRNA
GAL4
+
UAS
UAS
UAS
UAS
UAS
GFP
5IR 3IR
FIGURE 24.3 DIC photograph of the head of a worm. The outline of the pharynx (p) can
be seen and the beginning of the gut (g). The approximate positions of the amphid sensory
neuron cell bodies are indicated with colored circles. The dendrite (d) and an axon (a) of one
of the cells are drawn. Below, the names of the corresponding cells are given, and the
compounds perceived by these cells.
1903_Color Insert.fm Page 499 Thursday, July 27, 2006 6:39 PM
499
Extracellular signals
(Adhesion molecules, growth factors, electric/synaptic activity)
Rho GDI
Rho-GDP
Rac-GDP
Cdc42-GDP
Actin cytoskeleton
of dendritic spines
FIGURE 28.1 One cluster of nonsyndromic MR genes relates to the Rho family of small
GTPases. Rho family GTPases transduce extracellular signals into adaptive responses of the
actin cytoskeleton. Actin-dependent processes at the synapse include the regulation of the
morphology and the dynamics of the dendritic spines. The Rho family comprises three
members, termed Rho, Rac, and Cdc42. They shuttle between an inactive (red) and an active
(green) state under the control of three types of regulatory proteins (blue). GEFs (guanine
nucleotide exchange factors) promote the release of GDP and its replacement by GTP and,
thereby, mediate the transition of Rho, Rac, and Cdc42 from the inactive into the active state.
GAPs (GTPase-activating proteins) activate the endogenous GTPase function and, thus, the
self-inactivation of Rho, Rac, and Cdc42. GDIs (GDP dissociation inhibitors) bind to the
GDP form of the GTPases and prevent their premature activation as long as the GTPase is
not in the correct place and situation for a new round of activation. Further downstream
activators (yellow) eventually act directly or indirectly on regulatory components of the actin
cytoskeleton. The genes affected in MRs are printed red.
1903_Color Insert.fm Page 500 Thursday, July 27, 2006 6:39 PM
Neurotransmitter
synthesis/uptake
Rab GEF
Rab-GTP
Rab-GDP Rab GDI1
Rab GDI
Vesicle
IL1RAPL1 recycling
(endocytosis)
Rab-GTP
Rab-GTP Rab-GDP
NCS-1
Docked Fused
FACL4
Rab GAP Restricting
vesicle Ca2+ vesicle ilin
oph
V
bud End
FIGURE 28.2 A second cluster of nonsyndromic MR genes relates to the release of neu-
rotransmitters from the presynaptic nerve terminal and the regeneration of new releasable
vesicles by the budding and endocytosis of vesicles from the presynaptic membrane. The
release of neurotransmitters from storage vesicles occurs in several consecutive steps. First,
the vesicles are docked to the release site, the presynaptic active zone, a process that involves
the interaction of several proteins of the vesicle and the active zone. Docked vesicles fuse
with the presynaptic membrane in a process regulated by transiently increased intracellular
calcium, and release their content into the synaptic cleft. In order to maintain a constant
supply of releasable neurotransmitters, vesicles recycle by budding off the presynaptic mem-
brane and endocytosis. The budding occurs in a clathrin-dependent manner. The restriction
of the bud neck, a process that prepares the bud for scission, requires endophilin. Endophilin
is a lysophosphatidic acid acyltransferase, which introduces arachidonic acid into the cyto-
plasmic face of the bud neck. It, thereby, facilitates the formation of a strong negative curvature
of the membrane required for the restriction of the bud neck in order to allow the scission
process to begin. It is conceivable that the supply of arachidonoyl-CoA requires FACL4. The
genes affected in MR are printed in red.