3: Genes and Behavior

• Genes, Genetic Analysis, and Heritability in Behavior

• The Nature of the Gene

• Genes Are Arranged on Chromosomes

• The Relationship Between Genotype and Phenotype

• Genes Are Conserved Through Evolution

• The Role of Genes in Behavior Can Be Studied in Animal Models

  • Circadian Rhythm Is Generated by a Transcriptional Oscillator in Flies, Mice, and Humans

  • Natural Variation in a Protein Kinase Regulates Activity in Flies and Honeybees

  • The Social Behaviors of Several Species Are Regulated by Neuropeptide Receptors

• Genetic Studies of Human Behavior and Its Abnormalities

  • Neurological Disorders in Humans Suggest That Distinct Genes Affect Different Brain Functions

  • Autism-Related Disorders Exemplify the Complex Genetic Basis of Behavioral Traits

• Psychiatric Disorders and the Challenge of Understanding Multigenic Traits

  • Complex Inheritance and Genetic Imprinting in Human Genetics

  • Multigenic Traits: Many Rare Diseases or a Few Common Variants?

• An Overall View

• Glossary

All behaviors are shaped by the interplay of genes and the environment. The most stereotypic behaviors of simple animals are influenced by the environment, and the highly evolved behaviors of humans are constrained by innate properties specified by genes. Genes do not control behavior directly, but the RNAs and proteins encoded by genes act at different times and at many levels to affect the brain. Genes specify the developmental programs that assemble the brain and are essential to the properties of neurons and synapses that allow neuronal circuits to function. Genes that are stably inherited over generations create the machinery that allows new experiences to change the brain during learning.

In this chapter we ask how genes contribute to behavior. We begin with an overview of the evidence that genes do influence behavior, and then review basic principles of molecular biology and genetic transmission. We then provide a few examples of the way that genetic influences on behavior have been documented. Many persuasive links between genes and human behavior have emerged from the analysis of human brain development and function. However, a deep understanding of the ways that genes regulate behavior has emerged primarily from studies of worms, flies, and mice, animals whose genomes are accessible to experimental manipulation. Despite formidable challenges in studying complex genetic traits in humans, recent progress has revealed the genetic basis of some developmental and psychiatric disorders, and holds the promise of future understanding.

Many human psychiatric disorders and neurological diseases have a genetic component. The relatives of a patient are more likely than the general population to have the disease. The extent to which genetic factors account for traits in a population is called heritability. The strongest case for heritability is based on twin studies, first used by Francis Galton in 1883. Identical twins develop from a single fertilized egg that splits into two soon after fertilization; such monozygotic twins share all genes. In contrast, fraternal twins develop from two different fertilized eggs; these dizygotic twins, like normal siblings, share on average half their genetic information. Systematic comparisons over many years have shown that identical twins tend to be more similar (concordant) for neurological and psychiatric traits than fraternal twins, providing evidence of a heritable component of these traits (Figure 3–1A).

In an extension of the twin study model, Thomas Bouchard and colleagues in the Minnesota Twin Study examined identical twins that were separated early in life and raised in different households. Despite sometimes great differences in their environment, twins shared predispositions for the same psychiatric disorders and even tended to share personality traits, like extraversion. Thus genetic variation contributes to normal human differences, not just to disease states.

Heritability for human diseases and behavioral traits is usually substantially less than 100%, demonstrating that the environment is an important factor in acquiring diseases or traits. Estimates of heritability for many neurological, psychiatric, and behavioral traits from twin studies range around 50%, but heritability can be higher or lower for particular traits (Figure 3–1B). Although studies of identical twins and other kinships provide support for the idea that human behavior has a hereditary component, they do not tell us which genes are important, let alone how specific genes affect behavior. These questions can be addressed more easily by genetic studies in experimental animals in which both the gene and the environment are strictly controlled.

The related scientific areas of molecular biology and transmission genetics are central to our modern understanding of genes. The following primer introduces these ideas; a glossary at the end of the chapter defines commonly used terms.

Genes are made of DNA, and it is DNA that is passed on from one generation to the next. Through DNA replication exact copies of each gene are provided to all cells in an organism as well as to succeeding generations. DNA is made of two strands, each of which has a deoxyribose-phosphate backbone attached to a series of four subunits: the nucleotides adenine (A), guanine (G), thymine (T), and cytosine (C). The two strands are paired so that an A on one strand is always paired with a T on the complementary strand, and a G with a C (Figure 3–2). This complementarity ensures accurate copying of DNA during DNA replication and drives transcription of DNA into RNA. RNA differs from DNA in that it is single-stranded, has a ribose rather than a deoxyribose backbone, and uses the nucleoside base uridine (U) in the place of thymine.

Most genes encode protein products, which are generated by translation of the linear messenger RNA (mRNA) sequence into a linear polypeptide (protein) sequence composed of 20 different amino acids. A typical gene consists of a coding region, which is translated into a protein, and noncoding regions (Figure 3–3). The coding region is usually broken into small segments called exons that are separated by noncoding stretches called introns. The introns are deleted from the mRNA before its translation into protein.

Some functional RNAs do not encode protein. These include ribosomal RNAs (rRNAs) and transfer RNAs (tRNAs), essential components of the machinery for mRNA translation; small nuclear RNAs (snRNAs), which guide mRNA splicing; and microRNAs (miRNAs), small RNAs that pair with complementary sequences in specific mRNAs to inhibit their translation.

Each cell in the body contains the DNA for every gene but only expresses a specific subset of the genes as RNAs. The part of the gene that is transcribed into RNA is flanked by noncoding DNA regions, called promoters and enhancers, that allow accurate expression of the RNA in the right cells at the right time. Promoters and enhancers are typically found close to the beginning of the region to be transcribed, but can also act at a distance. A unique complement of DNA-binding proteins within each cell interacts with promoters and enhancers to regulate gene expression and resulting cellular properties.

The brain expresses a greater number of genes than any other organ in the body, and within the brain diverse populations of neurons express different groups of genes. The selective gene expression controlled by promoters, enhancers, and the DNA-binding proteins that interact with them permits a fixed number of genes to generate a vastly larger number of neuronal cell types and connections in the brain.

Although genes specify the initial development and properties of the nervous system, the experience of an individual and the resulting activity in specific neural circuits can itself alter the expression of genes. In this way environmental influences are incorporated into the structure and function of neural circuits. The goal of genetics is to unravel the ways that individual genes affect a biological process, the ways that networks of genes affect each other's activity, and the ways that genes interact with the environment.

The genes in a cell are arranged in an orderly fashion on long, linear stretches of DNA called chromosomes. Each gene in the human genome is reproducibly located at a characteristic position (locus) on a specific chromosome, and this genetic "address" can be used to associate biological traits with a gene's effects. Most multicellular animals (including worms, fruit flies, mice, and humans) are diploid; every somatic cell carries two complete sets of chromosomes, one from the mother and the other from the father.

Humans have about 25,000 genes but only 46 chromosomes: 22 pairs of autosomes (chromosomes that are present in both males and females) and 2 sex chromosomes (2 X chromosomes in females, 1 X and 1 Y chromosome in males) (Figure 3–4). Each parent supplies one copy of each autosome to the diploid offspring. Each parent also supplies one X chromosome to female (XX) offspring, but XY males inherit their single X chromosome from their mothers and their single Y chromosome from their fathers. Sex-linked inheritance was discovered in fruit flies by Thomas Hunt Morgan in 1910. This pattern of sex-linked inheritance associated with the single X chromosome has been highly significant in human genetic studies, where certain X-linked genetic diseases are commonly observed only in males but are genetically transmitted from mothers to their sons.

In addition to the genes on chromosomes, a very small number of an organism's genes are transmitted through cytoplasmic organelles called mitochondria, which carry out metabolic processes. Mitochondria in all children come from the ovum and therefore are transmitted from the mother to the child. Certain human disorders, including some neuromuscular degenerative diseases, mental retardation, and some forms of deafness, are caused by mutations in the mitochondrial DNA.

Because an individual has two copies of each autosomal gene, it is important to distinguish the genotype of an organism (its genetic makeup) and the phenotype (its appearance). The two copies of a particular autosomal gene in an individual are called alleles. If the two alleles are identical, the individual is said to be homozygous at that locus. If the alleles vary because of mutations, the individual is heterozygous at that locus. Males are hemizygous for genes on the X chromosome. A population can have a large number of alleles of a gene; for example, a single gene that affects human eye color called OCA2 can have alleles that encode shades of blue, green, hazel, or brown. In the broad sense a genotype is the entire set of alleles forming the genome of an individual; in the narrow sense it is the specific alleles of one gene. By contrast, a phenotype is a description of a whole organism, and is a result of the expression of the organism's genotype in a particular environment.

The difference between genotype and phenotype is evident when considering the consequences of having one normal (wild-type) allele and one mutant allele of the same gene. Most proteins are able to function even if only half their wild-type levels are present, so inactivating one copy of most genes does not lead to a change in phenotype. Thus, two organisms with different genotypes (two wild-type alleles, or one wild-type and one mutant allele) can have the same phenotype. The organism with two mutant alleles has a different phenotype. The phenotype of an animal can change over its lifetime, whereas the genotype remains constant (except for sporadic mutations in cancer cells and DNA rearrangements in the immune system).

If a mutant phenotype is expressed only when both alleles of a gene are mutated (that is, only individuals homozygous for the mutant allele exhibit the phenotype) the phenotypic trait as well as the mutant allele are said to be recessive. Recessive mutations usually result from loss or reduction of a functional protein. Recessive inheritance of mutant traits is commonly observed in humans and experimental animals.

If a mutant phenotype results from a combination of one mutant and one wild-type allele, the phenotypic trait and mutant allele are said to be dominant. In general, dominant mutations are more rare than recessive mutations. Some mutations are dominant because 50% of the gene product is not enough for a normal phenotype. Other dominant mutations lead to the production of an abnormal protein by the mutant gene or to expression of the wild-type gene product at an inappropriate time or place.

Genes and proteins are resilient in the face of small changes, so not every sequence change in a gene is deleterious. Most mutations are simply allelic polymorphisms that are silent; that is, they do not have any effect on the phenotype. Some are not silent but are expressed in ways that are benign, for example in eye color. Only rare mutations cause disturbance in development, cell function, or overt behavior. Some mutations are pathogenic, however, and these lead to disease (Box 3–1).

The nearly complete nucleotide sequence of the human genome was reported in 2000, and the complete nucleotide sequences of many animal genomes have also been decoded. Comparisons between these genomes leads to a surprising conclusion: the unique human species did not result from the invention of unique new human genes.

Humans and chimpanzees are profoundly different in their biology and behavior, yet they share 99% of their protein-coding genes. Moreover, most of the approximately 25,000 genes in humans are also present in other mammals like mice, and over half of all human genes are very similar to genes in invertebrates such as worms and flies (Figure 3–5). The conclusion from this surprising discovery is that ancient genes that humans share with other animals are regulated in new ways to produce novel human properties, like the capacity to generate complex thoughts and language.

Because of this conservation of genes throughout evolution, insights from studies of one animal can often be applied to other animals with related genes, an important fact as animal experiments are often possible when experiments on humans are not. For example, a gene from a mouse that encodes an amino acid sequence similar to a human gene usually has a similar function to the orthologous human gene.

Approximately one-half of the human genes have functions that have been demonstrated or inferred from orthologous genes in other organisms (Figure 3–6). A set of genes shared by humans, flies, and even unicellular yeasts encodes the proteins for intermediary metabolism, synthesis of DNA, RNA, and protein, cell division, and for cytoskeletal structures, protein transport, and secretion.

The evolution from single-celled organism to multicellular animals was accompanied by an expansion of genes concerned with intercellular signaling and gene regulation. The genomes of multicellular animals, such as worms, flies, mice, and humans, typically encode thousands of transmembrane receptors, many more than are present in unicellular organisms. These transmembrane receptors are used in cell-to-cell communication in development, in signaling between neurons, and as sensors of environmental stimuli. The genome of a multicellular animal also encodes 1,000 or more different DNA-binding proteins that regulate the expression of other genes. Many of the transmembrane receptors and DNA-binding proteins in humans are related to specific orthologous genes in other vertebrates and invertebrates. By enumerating the shared genetic heritage of the animals, we can infer that the basic molecular pathways for neuronal development, neurotransmission, electrical excitability, and gene expression were present in the common ancestor of worms, flies, mice, and humans. Moreover, studies of animal and human genes have demonstrated that the most important genes in the human brain are those that are most conserved throughout animal phylogeny. Differences between mammalian genes and their invertebrate counterparts most often result from gene duplication in mammals or subtle changes in gene expression and function, rather than the creation of entirely new genes.

Because of the conservation between human and animal genes, the relationships between the genes, proteins, and neural circuits that underlie behavior can be studied in animal models that are experimentally tractable. Two important strategies have been applied with great success in the study of gene function.

In classical genetic analysis organisms are first subjected to mutagenesis with a chemical or irradiation that induces random mutations, and then screened for heritable changes that affect the behavior of interest, say, sleep. This approach does not impose a bias as to the kind of gene involved; it is a random search of all possible mutations that cause detectable changes. Genetic tracing of heritable changes allows the identification of the individual genes that are altered in the mutant organism. Thus the pathway of discovery in classical genetics moves from phenotype to genotype, from organism to gene. In reverse genetics a specific gene of interest is targeted for alteration, a genetically modified animal is produced, and the animals with these altered genes are studied. This strategy is both focused and biased—one begins with a candidate gene—and the pathway of discovery moves from genotype to phenotype, from gene to organism.

These two experimental strategies and their more subtle variations form the basis of experimental genetics. Gene manipulation by classical and reverse genetics is conducted in experimental animals, not in humans.

Circadian Rhythm Is Generated by a Transcriptional Oscillator in Flies, Mice, and Humans

The first large-scale studies of the influence of genes on behavior were initiated by Seymour Benzer and his colleagues around 1970. They used random mutagenesis and classical genetic analysis to identify mutations that affected learned and innate behaviors in the fruit fly Drosophila melanogaster: circadian (daily) rhythms, courtship behavior, movement, visual perception, and memory (Boxes 3–2 and 3–3). These induced mutations have had an immense influence on our understanding of the role of genes in behavior.

We have a particularly complete picture of the genetic basis of the circadian control of behavior. An animal's circadian rhythm couples certain behaviors to a 24-hour cycle linked to the rising and setting of the sun. The core of circadian regulation is an intrinsic biological clock that oscillates over a 24-hour cycle. Because of the intrinsic periodicity of the clock, circadian behavior persists even in the absence of light or other environmental influences.

The clock can be reset, so that changes in the day-night cycle eventually result in a matching shift in the intrinsic oscillator, a phenomenon that is familiar to any traveler recovering from jet lag. Light-driven signals are transmitted by the eye to the brain to reset the clock. Finally, the clock drives output pathways for specific behaviors, such as sleep and locomotion.

Benzer's group searched through thousands of mutant flies to look for rare flies that failed to follow circadian rhythms because of mutations in the genes that direct circadian oscillation. From this work emerged the first insight into the molecular machinery of the circadian clock. Mutations in the period, or per, gene affected all circadian behaviors generated by the fly's internal clock.

Interestingly, per mutations could change the circadian clock in several ways (Figure 3–9). Arrhythmic per mutant flies, which exhibited no discernible intrinsic rhythms in any behavior, lacked all function of the per gene, so per is essential for rhythmic behavior. Per mutations that maintained some function of the gene resulted in abnormal rhythms. Long-day alleles produced 28-hour behavioral cycles, whereas short-day alleles produced a 19-hour cycle. Thus per is not just an essential piece of the clock, it is actually a timekeeper whose activity can change the rate at which the clock runs.

The per mutant has no major adverse effects other than the change in circadian behavior. This observation is important because, prior to the discovery of per, many had questioned whether there could be true "behavior genes" that were not required for the other physiological needs of an animal. Per does seem to be such a "behavior gene."

How does per keep time? The protein product PER is a transcriptional regulator that affects the expression of other genes. Levels of PER are regulated throughout the day. Early in the morning PER and its mRNA are low. Over the course of the day the PER mRNA and protein accumulate, reaching peak levels after dusk and during the night. The levels then decrease, falling before the next dawn. These observations provide an answer to the circadian rhythm puzzle—a central regulator appears and disappears throughout the day. But they are also unsatisfying because they only push the question back one step—what makes PER cycle? The answer to this question required the discovery of additional clock genes, which were discovered in flies and also in mice.

Emboldened by the success of the fly circadian rhythm screens, Joseph Takahashi began similar but far more labor-intensive genetic screens in mice in the 1990s. He screened hundreds of mutant mice for the rare mice with alterations in their circadian locomotion period, and found a single gene mutation that he called clock (Figure 3–10). When mice homozygous for the clock mutation are transferred to darkness, they initially experience extremely long circadian periods and later a complete loss of circadian rhythmicity. The clock gene therefore appears to regulate two fundamental properties of the circadian rhythm: the length of the circadian period and the persistence of rhythmicity in the absence of sensory input. These properties are conceptually identical to the properties of the per gene in flies.

The mouse clock gene, like the per gene in flies, encodes a transcriptional regulator whose activity oscillates through the day. The mouse CLOCK and fly PER proteins also shared a domain called a PAS domain, characteristic of a subset of transcriptional regulators. This observation suggests that the same molecular mechanism—oscillation of PAS-domain transcriptional regulation—controls circadian rhythm in flies and mice.

More significantly, parallel studies of flies and mice showed that similar groups of transcriptional regulators affected the circadian clock in both animals. After the mouse clock gene was cloned, a fly circadian rhythm gene was cloned and found to be closely related to mouse clock, even more so than per. In a different study a mouse gene similar to fly per was identified and inactivated by reverse genetics. The mutant mouse had a circadian rhythm defect, like fly per mutants. In other words, both flies and mice use both clock and per genes to control their circadian rhythms. A group of genes, not one gene, are conserved regulators of the circadian clock.

Characterization of these genes has led to an understanding of the molecular mechanisms of circadian rhythm, and a dramatic demonstration of the similarity of these mechanisms in flies and mice. In both flies and mice the CLOCK protein is a transcriptional activator. Together with a partner protein, it controls the transcription of genes that determine output behaviors such as locomotor activity levels. CLOCK and its partner also stimulate the transcription of the per gene. However, PER protein represses CLOCK's ability to stimulate per expression, so as PER protein accumulates, per transcription falls (Figure 3–11). The 24-hour cycle comes about because the accumulation and activation of PER protein is delayed by many hours after the transcription of per, a result of PER phosphorylation, PER instability, and interactions with other cycling proteins.

The molecular properties of per, clock, and related genes generate all properties essential for circadian rhythm. The key elements of the regulatory process are:

1. The transcription of circadian rhythm genes varies with the 24-hour cycle: PER activity is high at night, CLOCK activity is high during the day.

2. The circadian rhythm genes are transcription factors that affect each other's mRNA level, generating the oscillations. CLOCK activates per transcription and PER represses CLOCK function.

3. The circadian rhythm genes also control the transcription of output genes that affect many downstream responses. For example, in flies the neuropeptide gene pdf controls locomotor activity levels.

4. The oscillation of these genes can be reset by light.

Recent work has shown that the same genetic network controls circadian rhythm in humans. People with advanced sleep-phase syndrome have short 20-day cycles and an extreme early-to-bed, early-to-rise "morning lark" phenotype. Louis Ptáček and Ying-hui Fu found that these individuals have mutations in a human per gene. These results demonstrate that genes for behavior are conserved from insects to humans. Advance sleep-phase syndrome is discussed in detail in the chapter on sleep (see Chapter 51).

Natural Variation in a Protein Kinase Regulates Activity in Flies and Honeybees

In the genetic studies of circadian rhythm described above, random mutagenesis was used to identify genes of interest in a biological process. All normal individuals have functional copies of per, clock, and the related genes; only after mutagenesis were different alleles generated. Another, more subtle question about the role of genes in behavior is to ask which genetic changes make normal individuals behave differently from each other. Work by Marla Sokolowski and her colleagues led to the identification of the first gene associated with variation in behavior among normal individuals in a species.

Larvae of Drosophila vary in activity level and locomotion. Some larvae, called rovers, move over long distances (Figure 3–12). Others, called sitters, are relatively stationary. Drosophila larvae isolated from the wild can be either rovers or sitters, indicating that these are natural variations and not laboratory-induced mutations. These traits are heritable; rover parents have rover offspring and sitter parents have sitter offspring.

Sokolowski used crosses between different wild flies to investigate the genetic differences between rover and sitter larvae. These crosses showed that the difference between rover and sitter larvae lies in a single major gene, the for (forager) locus. The for gene encodes a signal transduction enzyme, a protein kinase activated by the cellular metabolite cyclic guanosine 3′-5′monophosphate (cGMP). Thus this natural variation in behavior arises from altered regulation of signal transduction pathways. Many neuronal functions are regulated by protein kinases such as the cGMP-dependent kinase encoded by the for gene. Molecules such as protein kinases are particularly significant at transforming short-term neural signals into long-term changes in the property of a neuron or circuit.

Why would variability in signaling enzymes be preserved in wild populations of Drosophila, which typically include both rovers and sitters? The answer is that variations in the environment favor alternative genetic forms: There is balanced selection for both behaviors. Crowded environments favor the rover larva, which is more effective at moving to new, unexploited food sources in advance of competitors, whereas sparse environments favor the sitter larva, which exploits the current source more thoroughly.

The for gene also affects honeybee behavior. Honeybees exhibit different behaviors at different stages of their life; in general, young bees are nurses, while older bees become foragers that leave the hive. The for gene is expressed at high levels in the brains of active foraging honeybees, and at low levels in the younger and more stationary nurse bees. Activation of cGMP signaling in young bees can cause them to enter the forager stage prematurely. Presumably, this change is normally programmed by an environmental stimulus or the bee's advancing age. Thus the same gene controls variation in a behavior in two different insects, the fruit fly and the bee. However, in the fruit fly the variation is expressed in different individuals, whereas in the honeybee it is expressed in one individual at different ages. The difference illustrates how an important regulatory gene can be recruited to different behavioral strategies in different species.

The Social Behaviors of Several Species Are Regulated by Neuropeptide Receptors

Many aspects of behavior are associated with an animal's social interactions with other animals. Social behaviors are highly variable between species, yet have a large innate component within a species that is controlled genetically. A simple form of social behavior has been analyzed in the roundworm C. elegans. These animals live in soil and eat bacteria. Different wild-type strains exhibit profound differences in feeding behavior. Some strains are solitary, dispersing across a lawn of bacterial food and failing to interact with each other. Other strains have a social feeding pattern, joining large feeding groups of dozens or hundreds of animals (Figure 3–13). The difference between these strains is genetic, as both feeding patterns are stably inherited.

The difference between social and solitary worms is caused by a single amino acid substitution in a single gene, a member of a large family of genes involved in signaling between neurons. This gene, npr-1, encodes a neuropeptide receptor. Neuropeptides have long been appreciated for their roles in coordinating behaviors across networks of neurons. For example, a neuropeptide hormone of the marine snail Aplysia stimulates a complex set of movements and behavior patterns associated with a single behavior, egg laying. Mammalian neuropeptides have been implicated in feeding behavior, sleep, pain, and many other behaviors and physiological processes. The existence of a mutation in the neuropeptide receptor that alters social behavior suggests that this kind of signaling molecule is important both for generating the behavior and for generating the variation between individuals.

Neuropeptide receptors have also been implicated in the regulation of mammalian social behavior. The neuropeptides oxytocin and vasopressin stimulate mammalian affiliative behaviors such as pair bonding and parental bonding with offspring. Oxytocin is required in mice for social recognition, the ability to identify a familiar individual. Oxytocin and vasopressin have been studied in depth in prairie voles, rodents that form lasting pairs to raise their young. Oxytocin released in the brain of female prairie voles during mating stimulates pair-bond formation. Likewise, vasopressin released in the brain of male prairie voles during mating stimulates pair-bond formation and paternal behavior.

The extent of pair-bonding varies substantially between mammalian species. Male prairie voles form pair-bonds with females and help them raise their offspring, but the closely related male montane voles breed widely and do not engage in paternal behavior. The difference between the behaviors of males in these species correlates with differences in the expression of the V1a class of vasopressin receptors in the brain. In prairie voles V1a vasopressin receptors are expressed at high levels in a specific brain region, the ventral pallidum (Figure 3–14). In montane voles the levels are much lower in this region, although high in other brain regions.

The importance of oxytocin and vasopressin and their receptors has been confirmed and extended by reverse genetic studies in mice, which are genetically easier to manipulate than voles. For example, introducing the V1a vasopressin receptor gene from prairie voles into male mice, which behave more like montane voles, increases the expression of the V1a vasopressin receptor in the ventral pallidum and increases the affiliative behavior of the male mice toward females. Thus differences between species in the pattern of expression of the vasopressin receptor can contribute to differences in social behaviors.

The analysis of vasopressin receptors in different rodents provides insight into the mechanisms by which genes and behaviors can change during evolution. Genetic changes in evolution alter the pattern of expression of the V1a vasopressin receptor in the ventral forebrain. These changes in turn alter the activity of a neural circuit, so that the function of the ventral forebrain is linked to the function of the vasopressin-secreting neurons that are activated by mating. As a result, social behaviors are altered.

The importance of oxytocin and vasopressin in human social behavior is not known, but the central role of pair-bonding and pup rearing in mammalian species suggests that these molecules might play a role in our species as well.

Human genetics is studied by observing traits that vary between individuals, analyzing pedigrees to trace the transmission of genetic traits across generations, and identifying functional variants using molecular biology. Human genetics is more challenging than experimental genetics because one cannot do inbreeding or crossbreeding experiments in people as one can in worms, flies, or mice.

Nonetheless, careful pedigree tracing has identified a variety of mutations that affect the human brain and behavior, including mutations that lead to neurodegenerative diseases, neurological disorders, and developmental disorders. In some cases mutations in a single gene have large effects on a neurological or behavioral trait. However, most neurological or psychiatric diseases are believed to reflect interactions between multiple genes and the environment, and our ability to tease apart these influences is still limited.

The first gene discovered to be important for a neurological disease in humans was the gene responsible for phenylketonuria (PKU), described by Asbørn Følling in Norway in 1934. This rare disease affects 1 in 15,000 children and results in severe impairment of cognitive function. Children with this disease have two abnormal copies of the gene that codes for phenylalanine hydroxylase, the enzyme that converts the amino acid phenylalanine to tyrosine. The mutation is recessive; heterozygous carrier individuals have no symptoms. Children who lack normal function in both copies of the gene accumulate high blood concentrations of phenylalanine from dietary proteins, which in turn leads to the production of toxic metabolites that interfere with neuronal function. The specific biochemical processes by which phenylalanine adversely affects the brain are still not understood.

The PKU phenotype (mental retardation) results from the interaction of the genotype (the homozygous pku mutation) and the environment (the diet). The treatment for PKU is thus simple and effective: Mental retardation can be prevented by a low-protein diet. The molecular and genetic analysis of gene function in PKU has led to dramatic improvements in the life of PKU patients. Since the early 1960s the United States has instated mandatory testing for PKU in newborns. By identifying children with the genetic disorder and modifying their diet before the disease appears many aspects of the disorder can be prevented.

Later chapters of this book describe many examples of single-gene traits that, like PKU, have led to insights into brain function and dysfunction. Certain themes have emerged from these studies. A number of rare neurodegenerative disorders such as Huntington disease and spinocerebellar ataxia result from the pathological, dominant expansion of glutamate residues within proteins. The discovery of these polyglutamine repeat disorders highlighted the danger to the brain of unfolded and aggregated proteins. The discovery that epileptic seizures can be caused by a variety of mutations in ion channels led to the realization that these disorders are primarily disorders of neuronal excitability.

The single-gene disorders mentioned earlier are rare compared to the total burden of neurodegenerative and psychiatric disease. One might wonder why we study rare single-gene variants if they represent a tiny fraction of the total disease burden. The reason is that rare, relatively simple disorders can provide insight into the biological processes affected in more common, complex forms of the disease. For example, among the prominent successes of human genetics has been the discovery of rare genetic variants that lead to early-onset Alzheimer disease or Parkinson disease. Individuals with these severe rare variants represent a tiny subset of all individuals with Alzheimer disease or Parkinson disease, but the rare variants uncovered cellular processes that are also disrupted in the larger patient pool, pointing the way to general therapeutic avenues.

In the remainder of this chapter we discuss the genetics of autism and schizophrenia, two psychiatric disorders that manifest themselves in childhood or adolescence, respectively. Compared to the polyglutamine repeat diseases or ion channel-related seizure disorders, the biological causes of autism and schizophrenia are still largely unknown. Nonetheless, in some cases we can begin to see how certain genes, and the associated biological processes, affect complex brain functions such as language, learning, and emotion.

Neurological Disorders in Humans Suggest That Distinct Genes Affect Different Brain Functions

Autism is a common, devastating developmental disorder characterized by deficiencies in language acquisition, difficulties in social interactions, and stereotyped interests. About 1 in 200 children are diagnosed with autism (the exact frequency varies with diagnostic criteria), with about three times as many boys affected as girls. The clinical symptoms of autism typically emerge in the first 3 years of life and often include a regression phase in which children lose the language skills they mastered when younger.

There is considerable variability between autistic individuals. Children affected with autism have a higher frequency of seizures and cognitive problems than the general population, and some are severely disabled. However, many autistic individuals have normal or above-normal intelligence, and can, with appropriate care, goes on to lead highly successful lives.

Autism has a very strong heritable component (see Figure 3–1A), which should increase the chances of identifying the underlying genes. Autism is also a disorder of broader significance because it provides insight into behaviors that are unique to humans: language, complex intelligence, and interpersonal interactions. The fact that autistic defects in social communication can coexist with normal intelligence in other domains suggests that the brain is modular, with distinct cognitive functions that can vary independently.

The idea that distinct genes affect different cognitive domains is supported by the contrasting effects of a rare genetic syndrome called Williams syndrome. Children with Williams syndrome acquire language late, but they ultimately overcome their early deficits to develop strong language skills and normal social behavior. Indeed, these children exhibit extreme sociability; for example, they lack the shyness children typically have in the presence of strangers. However, Williams children are profoundly defective in spatial processing, and they score as poorly as, or even worse than, autistic children on IQ tests. The different patterns of impairments in autism and Williams syndrome suggest that language and social skills can be separated from some other brain functions. Brain areas concerned with language are impaired in autistic children but active or accented in Williams syndrome children. By contrast, general and spatial intelligence is more impaired in Williams syndrome than in many autistic children.

Williams syndrome is genetically simple; it is caused by a heterozygous deletion of chromosome region 7q11.23. The simplest interpretation of this defect is that the level of expression of one or more genes within this interval is reduced because there is only one copy instead of two of each gene in the region. The precise genes in this region that affect sociability and spatial processing are not yet known, but they are of great interest because of their potential to provide insight into the genetic regulation of human behaviors.

Another rare syndrome has shed light on genetic requirements for human language. A study of a human family with defects in speech articulation, as well as grammatical and linguistic impairment, led to the identification of a specific mutation in the transcriptional regulator FOXP2. This gene is not uniquely human—it is present in all mammals and birds—but several changes in the gene have appeared since the divergence of humans and other primates. A human-specific pattern of development of the brain, larynx, or mouth under the control of FOXP2 might be one of the genetic adaptations that made human speech possible.

Autism-Related Disorders Exemplify the Complex Genetic Basis of Behavioral Traits

Autism is a common disorder, yet there is no single genetic location for autism as there is for the much rarer Williams syndrome. At least part of the explanation for this difference is that autism is not really one disorder caused by a single kind of genetic lesion, but rather a cluster of related disorders caused by a variety of different genetic changes. The recognition of this genetic complexity is providing inroads into autism and related syndromes that affect social communication and language.

The symptoms of classical autism partly overlap with those of a number of genetic disorders whose genetic basis is understood. For example like autism, fragile X syndrome affects mostly boys, and patients have poor social cognition, high social anxiety, and repetitive behavior. However, fragile X syndrome is associated with broader cognitive defects, along with certain physical characteristics, like an elongated face and protruding ears. Fragile X syndrome has been genetically associated with a mutation that reduces expression of a single gene on the X chromosome. This gene regulates the translation of mRNAs into proteins in neurons, a process that is regulated during synaptic plasticity and learning. The male preponderance of fragile X syndrome is explained because males have only a single X chromosome, and therefore lose all expression of the relevant gene when it is mutated. Since females have two X chromosomes, they can be carriers of the disease without being strongly affected.

Another autism-like syndrome whose genetic basis is understood is Rett syndrome, which we consider in detail in Chapter 64. Rett syndrome is an X-linked, progressive neurodevelopmental disorder and one of the most common causes of mental retardation in females. The disease is confined to females because Rett syndrome is lethal in the developing male embryo, which has a single X chromosome. Affected girls seem to develop normally until they are 6 to 18 months of age, when they fail to acquire speech and purposeful hand use is replaced by compulsive, uncontrolled hand-wringing.

Huda Zoghbi and her colleagues found that the major cause of this disease is a mutation in the methyl CpG binding protein 2 (MeCP2) gene. Methylation of specific CpG sequences in DNA alters expression of nearby genes, and it is thought that MeCP2 binds methylated DNA as part of a process that regulates mRNA transcription. This discovery shows that acquisition of complex developmental properties such as language is associated with specific patterns of gene expression, most likely in the brain. Current research on Rett syndrome focuses on identifying the affected brain regions and the inappropriately transcribed genes in these children.

The observation that common diseases such as autism are less well understood than rare diseases such as Rett syndrome applies broadly to psychiatric disorders. Schizophrenia affects about 1% of all young adults, causing a pattern of thought disorders and emotional withdrawal that profoundly impairs life. It is strongly heritable (see Figure 3–1B), but its genetic basis is poorly understood. Despite evidence of heritability, and despite progress in research on individual diseases using experimental animal models, it has been difficult to identify the genes associated with many common human psychiatric diseases like chronic depression and anxiety disorders, which affect a substantial fraction of people. In this final section of the chapter we survey the approaches that are being developed to address these shortfalls.

Complex Inheritance and Genetic Imprinting in Human Genetics

Understanding Mendelian inheritance is not always sufficient for understanding human genetic disorders. For example, studies of human developmental disorders have shown important effects of an unusual form of gene regulation called parental imprinting. Although humans have two copies of all autosomes, they sometimes express mRNA from only one of the two copies of a gene, either the copy from the mother or the copy from the father. Angelman syndrome and Prader-Willi syndrome are examples of genetic diseases with a complex heritability caused by parental imprinting. To understand these diseases, one must not only know the DNA lesion associated with the disease but also understand whether the DNA lesion was inherited from the father or the mother (Figure 3–15).

Angelman syndrome is an inherited disorder that can include severe mental retardation, epilepsy, absence of speech, hyperactivity, and inappropriate laughter. The syndrome results from deletion of multiple genes on one copy of the chromosome region 15q11-q13. In this region some genes are expressed only from the maternally derived chromosome, and a different set of genes are expressed only from the paternally derived chromosome. The genes implicated in Angelman syndrome are expressed only from the maternally derived chromosome. Therefore, if a child receives a normal chromosome from her mother and an abnormal chromosome with a deletion from her father, she will not have Angelman syndrome. However, if she receives a normal chromosome from her father and an abnormal chromosome from her mother, she will have Angelman syndrome.

For Angelman syndrome the most important gene in this region appears to be the gene that encodes UBE3A, a ubiquitin-protein ligase that is expressed only from the maternally derived chromosome. Ubiquitin ligases stimulate the degradation and turnover of other proteins. The important targets of UBE3A are thought to regulate general brain plasticity, perhaps by regulating the activity of neurotransmitter receptors.

A different neurodevelopmental behavioral disorder, Prader-Willi syndrome, is also caused by a deletion of multiple genes on chromosome 15q11-q13, but the genes implicated in Prader-Willi syndrome are expressed only from the paternally derived chromosome. Therefore, Prader-Willi syndrome results from the reciprocal of the pattern of inheritance in Angelman syndrome: A child must receive the abnormal chromosome from his father to develop Prader-Willi syndrome. Prader-Willi syndrome is associated with obesity, and psychiatric symptoms include obsessive- compulsive behavior, bipolar disorder, and pervasive developmental disorders.

Parental imprinting is an example of the complexity that can confound expectations in human gene mapping. Recent studies in mice by Catherine Dulac's group suggest that parental imprinting is more common in the brain than it is in other tissues in the body. If this observation holds true in humans, it could help explain why mapping human behavioral traits is difficult.

Multigenic Traits: Many Rare Diseases or a Few Common Variants?

Genetic mapping methods that follow linkage across generations are highly effective in identifying single-gene traits (Box 3–4). However, most common genetic disorders in humans cannot be followed through pedigrees as single-gene traits, and they are suspected or known to involve multiple genes as well as genetic interactions with the environment. Schizophrenia, autism, and depression fall into this category, as do diabetes, coronary artery disease, and asthma.

In contrast to single-locus Mendelian traits, multigenic traits do not have a simple recognizable pattern of inheritance (autosomal dominant, recessive, or X-linked), and thus the relative contributions of several genes to one trait are difficult to analyze. Nevertheless, determining which genes contribute to complex human traits has profound implications for the care and treatment of human disease and therefore is an area of intense study.

To understand complex psychiatric disorders it will be important to distinguish between various models of transmission. According to one (monogenic) model, many genes in the population can contribute to a disorder, but each gene mutation is individually rare and has a strong effect. Taking the examples described earlier in the chapter, it is known that MeCP2 mutations (Rett syndrome) and fragile X mutations each account for 1–2% of children with autism-related disorders. In a rare monogenic model there might be 100 other mutations, each individually rare, that can result in autism.

A second (polygenic) model assumes that a small number of relatively common mutations, each of which has only a small effect on risk, interact together to cause a disorder. The latter model is sometimes called the "common variant– common disease" hypothesis. Common human variants are known to increase the risk of some common diseases such as macular degeneration, which leads to adult-onset blindness. Thus, there are well-understood examples of both the monogenic and polygenic models in human genetics. An important question is the extent to which each model contributes to the burden of psychiatric diseases.

In both the "rare monogenic" model and the "common variant – common disease" model, it is challenging to identify the effects of individual mutations. Therefore, current psychiatric genetic studies usually involve international consortia cooperating in the diagnosis of very large clinical groups of patients. These studies can be analyzed using both monogenic and polygenic models of inheritance, and an increasingly sophisticated array of genotyping methods to assess correlations between regions of the genome and traits of interest (Box 3–4). Promising results have emerged from a variety of studies. Verification of these results will require independent replication of results and very large numbers of patients. With that caveat, some of the first fruits of these studies are described below.

A successful single-gene mapping study of a large Scottish family with a high rate of schizophrenia led to the identification of the risk gene Disc1 (for Disrupted in Schizophrenia). In mice Disc1 has effects on neuronal development, neuronal signaling, and the incorporation of newborn neurons into neural circuits. In humans a Disc1 mutation greatly increases the risk of schizophrenia, but it is mutated only in a small fraction of schizophrenic patients. Other candidate genes for schizophrenia have been identified from the alternative approach of looking for statistical associations between the disorder and specific loci in the genome in a sample of thousands of schizophrenic patients from different families. One risk gene identified in this way is neuregulin, a developmental signaling gene that affects cell migration and synapse formation. Together the studies of Disc1 and neuregulin begin to suggest that a developmental defect in the brain can underlie schizophrenia.

Similar large-scale studies of thousands of autistic children and their families have led to the repeated identification of mutations in certain autism risk genes. Several mutations in autistic patients affect transmembrane signaling proteins called neurexins and neuroligins, which affect the strength of synapses. These results and other genetic results suggest that subtle alterations in synaptic transmission play a role in autism. Sophisticated genetic analysis suggests that an increased risk for autism exists at many locations in the genome (Figure 3–17). Some of these loci may be rare monogenic high-risk variants and others may be more common, low-risk variants.

One remarkable discovery made by Michael Wigler's group while trying to understand the genetic basis of autism is that a fraction of the mutations associated with autism are not actually inherited from either parent. They are new mutations in the autistic child that arose in the sperm or the oocyte prior to fertilization. Identical monozygotic twins that separate after fertilization will both have the trait, the classical definition of heritability. Moreover, once the mutation arises, the affected child has a trait that can be passed on to his or her children, so it is indeed a heritable trait. However, classical genetic linkage studies rely on the assumption that virtually all disease-causing mutations are preexisting mutations inherited across many generations, and the finding of new mutations in autism violate that expectation. Human biology is a constant source of surprises.

New tools are being developed to help address the complexities of multigenic traits. For example, brain imaging techniques are being used to pinpoint anatomical loci underlying clinical syndromes that might be caused by specific mutations. At this writing, it will soon be possible to sequence the complete genomes of individual psychiatric patients to discover genetic variants. Even when these sequences are available, however, we will still need sophisticated approaches to ask which of the many variations that distinguish one human being from every other human being represent the risk factors for disease.

Genes affect many aspects of behavior. There are remarkable similarities in personality traits and psychiatric illnesses in human twins, even those raised separately; domestic and laboratory animals can be bred for particular, stable behavioral traits; a few specific genes are associated with human behaviors.

In recent years we have seen great success in identifying the genes associated with neurological diseases such as Alzheimer disease, Parkinson disease, Huntington disease, and spinocerebellar ataxias, as well as those associated with developmental disorders such as Rett syndrome and Angelman syndrome. What we have learned about the genetic basis of these diseases may help us fashion approaches to researching the more difficult psychiatric diseases such as schizophrenia, depression, and autism.

Human behavioral traits are multigenic in origin. Although alterations in a single gene account for individual differences in social behavior in nematodes and activity levels in Drosophila, only rarely will genetic alteration at a single gene fully explain a psychiatric disease or a behavior in humans. More frequently, several different genes will each act independently to affect a trait, or have effects only in combination with each other. Model organisms such as the fly, the worm, and the mouse are important for genetics because they are experimentally tractable. By studying genes in these simple animals, where the genome and environment can be rigorously controlled, it is possible to identify the conserved biological pathways that mediate behavior in humans. Several successes have emerged from this approach. The per gene in the fruit fly led to an understanding of human advanced sleep phase syndrome, and the gene for narcolepsy in dogs has led to an understanding of the underlying biology of sleep in humans (see Chapter 51).

There is still much to learn about the ways that genes affect the brain. Biologists are becoming more skilled at delineating the role of individual genes within particular neurons and circuits. There are important insights to be gained about the genetics of psychiatric disease and the genetic influences on normal psychological, physiological, and cognitive traits. Yet it is the interplay of genetics, environment, chance, and individual choice that ultimately determines the behavioral differences between individuals. The challenge of genetics is to understand the undeniable effects of genes, while acknowledging that many factors influence human mental processes and behavior.

¹Allele. Humans carry two sets of chromosomes, one from each parent. Equivalent genes in the two sets might be different, for example, because of single nucleotide polymorphisms. An allele is one of the two (or more) forms of a particular gene.

Centromere. Chromosomes contain a compact region known as a centromere, where sister chromatids (the two exact copies of each chromosome that are formed after replication) are joined.

Cloning. The process of generating sufficient copies of a particular piece of DNA to allow it to be sequenced or studied in some other way.

Complementary DNA (cDNA). A DNA sequence made from a messenger RNA molecule, using an enzyme called reverse transcriptase. cDNAs can be used experimentally to determine the sequence of messenger RNAs after their introns (nonprotein-coding sections) have been spliced out.

Conservation of genes. Genes that are present in two distinct species are said to be conserved, and the two genes from the different species are called orthologous genes. Conservation can be detected by measuring the similarity of the two sequences at the base (RNA or DNA) or amino acid (protein) level. The more similarities there are, the more highly conserved the two sequences.

Copy number variation (CNV). A deletion or duplication of a limited genetic region that results in an individual having more or fewer than the usual two copies of some genes. Copy number variations are observed in some neurological and psychiatric disorders.

Euchromatin. The gene-rich regions of a genome (see also heterochromatin).

Eukaryote. An organism with cells that have a complex internal structure, including a nucleus. Animals, plants, and fungi are all eukaryotes.

Genome. The complete DNA sequence of an organism.

Genotype. The set of genes that an individual carries; usually refers to the particular pair of alleles (alternative forms of a gene) that a person has at a given region of the genome.

Haplotype. A particular combination of alleles (alternative forms of genes) or sequence variations that are closely linked—that is, are likely to be inherited together—on the same chromosome.

Heterochromatin. Compact, gene-poor regions of a genome, which are enriched in simple sequence repeats.

Introns and exons. Genes are transcribed as continuous sequences, but only some segments of the resulting messenger RNA molecules contain information that encodes a protein product. These segments are called exons. The regions between exons are known as introns and are spliced from the RNA before the product is made.

Long and short arms. The regions on either side of the centromere are known as arms. As the centromere is not in the center of the chromosome, one arm is longer than the other.

Messenger RNA (mRNA). Proteins are not synthesized directly from genomic DNA. Instead, an RNA template (a precursor mRNA) is constructed from the sequence of the gene. This RNA is then processed in various ways, including splicing. Spliced RNAs destined to become templates for protein synthesis are known as mRNAs.

Mutation. An alteration in a genome compared to some reference state. Mutations do not always have harmful effects.

Phenotype. The observable properties and physical characteristics of an organism.

Polymorphism. A region of the genome that varies between individual members of a population. To be called a polymorphism, a variant should be present in a significant number of people in the population.

Prokaryote. A single-celled organism with a simple internal structure and no nucleus. Bacteria and Archaebacteria are prokaryotes.

Proteome. The complete set of proteins encoded by the genome.

Recombination. The process by which DNA is exchanged between pairs of equivalent chromosomes during egg and sperm formation. Recombination has the effect of making the chromosomes of the offspring distinct from those of the parents.

Restriction endonuclease. An enzyme that cleaves DNA at a particular short sequence. Different types of restriction endonuclease cleave at different sequences.

RNA interference (RNAi). A method for reducing the function of a specific gene by introducing into a cell small RNAs with complementarity to the targeted mRNA. Pairing of the mRNA with the small RNA leads to destruction of the endogenous mRNA.

Single nucleotide polymorphism (SNP). A polymorphism caused by the change of a single nucleotide. SNPs are often used in genetic mapping studies.

Splicing. The process that removes introns (noncoding portions) from transcribed RNAs. Exons (protein- coding portions) can also be removed. Depending on which exons are removed, different proteins can be made from the same initial RNA or gene. Different proteins created in this way are splice variants or alternatively spliced.

Transcription. The process of copying a gene into RNA. This is the first step in turning a gene into a protein, although not all transcripts lead to proteins.

Transcriptome. The complete set of RNAs transcribed from a genome.

Translation. The process of using a messenger RNA sequence to build a protein. The messenger RNA serves as a template on which transfer RNA molecules, carrying amino acids, are lined up. The amino acids are then linked together to form a protein chain.