Книги по МРТ КТ на английском языке / Functional Neuroimaging in Child Psychiatry Ernst 1 ed 2000

324D. J. Vandenbergh

already requires a high degree of volunteer participation. DNA that is collected from any part of the body is the same as that from the brain. Additionally, the DNA can be collected many years after the imaging study because the DNA remains unchanged. Once puriWed, the DNA is very stable and can be saved for years as volunteers are recruited to a neuroimaging study.

The most common source for DNA is lymphocytes from the blood, which can be taken in a 10ml or smaller volume at the time of the imaging study and can generate as much as 500-g DNA. In addition to the large amounts of DNA generated, blood samples can be used to establish cell lines by viral transformation of the leukocytes. These cell lines provide a permanent source of DNA because the cells can be stored frozen and thawed and regrown at a later date. The establishment of cell lines requires a dedicated tissue culture facility to infect, grow, and store the cells under liquid nitrogen, but it may be possible to Wnd commercial sources to provide this service.

Another source of DNA is buccal cells taken by a cheek swab (Richards et al., 1993; Freeman et al., 1997). Buccal cells do not provide as much DNA as lymphocytes (up to 50-g) but can be sent by mail, a particularly useful way to undertake genetic studies in individuals who already participated in neuroimaging investigations. Methods to grow buccal cells in culture as a means to increase the DNA yield are being explored in the author's laboratory (unpublished observations). Blood spotted on Wlter paper (McCabe, 1991), buccal cells from a saline mouth wash (Hayney et al., 1995), or cells from hair follicles (Higuchi et al., 1988) are additional sources of DNA, but the yield of DNA is so small that they are only practical for studies of very limited scope.

Well-established techniques for purifying DNA from eukaryotic cells are found in several molecular biology manuals (e.g., Sambrook et al., 1989; Ausubel et al., 1997). These techniques apply equally well to lymphocytes or buccal cells and use enzymatic digestion (proteinase K and RNAase A) in the presence of detergent followed by organic solvent (phenol/chloroform) extraction and ethanol precipitation. Several other techniques that use solvent extraction or resins or aYnity gels have been described (Lahiri et al., 1993).

In ongoing experiments in the author's laboratory, the following method is being used for routine isolation of high-purity DNA from buccal swabs. Research volunteers swab the inside of their mouth for 1min with a cotton swab and then swab again for 1min with a second swab. Swabs with a plastic stick produce less residue than wooden sticks. The tip of the swab is cut oV into a 1.5ml microfuge tube, and 500ml DNA lysis solution (10mmol/l Tris, pH 8.0,

100mmol/l NaCl, 10mmol/l EDTA, 0.5% sodium dodexylsulfate (SDS), 200g/l proteinase K, and 20g/l RNAase A) is added. The solution is incubated at 55°C for 2h; after the swab is removed, the solution is extracted once with 500-l Sevag (phenol/chloroform/isoamyl alcohol, 25:24:1). The upper, aqueous phase is removed carefully to a fresh microfuge tube containing 50-l 3mol/l sodium acetate pH5.3. After gentle mixing of the two solutions, 1ml ethanol is added and the solution is mixed again. The DNA is precipitated by storing at 220°C for 1h or more (at this point the DNA can be stored indeWnitely). The DNA is then pelleted by centrifugation for 15min in a microfuge; the pellet is dried brieXy after removing the supernatant, and the DNA is dissolved in 100-l TE (10mmol/l Tris, 1mmol/l ethylenediamine tetraacetic acid (EDTA) pH 8.0). The DNA can be stored in TE at 270°C for many years but should not be thawed and refrozen repeatedly.

Summary and future directions

Although molecular genetics shows great promise in helping to break the genetic ªcodesº underlying complex disorders, still much hard work lies ahead. There should also be a note of caution in continuing work in this young Weld of complex genetics. Genetic results require replication and careful scrutiny with systematic approaches to ensure that the results are reliable. Advances in genetic methodology need to be paralleled with advances in the identiWcation of phenotypes (see Chapter 19). Brain imaging Wndings are expected to help to deWne consistent phenotypes to be used in genetic studies. These brainimaging Wndings need to be the direct consequences of the genetic characteristics and not secondary manifestations of neural adaptive changes. This requirement underlines the importance of studying childhood disorders early when adaptive changes have not yet taken place that may mask the primary deWcits.

In addition, it is critical to reWne the mathematical modeling of the contribution of genetics to psychiatric disorders. Indeed, the complexity of the task of identifying the functional relationships between genetic markers and phenotypes is highlighted by the fact that a given genetic marker may be associated with several diVerent conditions, and a given phenotype with several genetic markers. As already mentioned, it is also unlikely that identiWed genes will be necessary and suYcient for the expression of a clinical phenotype. The understanding of the contribution of the genes will require one to dissect the clinical phenotype into elements that can be reliably and precisely quantiWed.

Ultimately, the interface of genetics and neuroimaging in the investigation of childhood psychiatric disorders is expected to further our understanding of basic neurobiologic mechanisms, which will help in the design of eVective preventive and therapeutic interventions.

Appendix: glossary of genetic terms

Allele One of several alternative forms of a gene or any other DNA sequence. Alleles may be detected at the level of a gene product, usually as a protein (e.g., two or more forms of a receptor measured by diVerences in binding aYnity) or at the level of DNA (e.g., a fragment of DNA that diVers in length with no associated functional diVerence). An individual has only two alleles at any site (locus), one maternal and one paternal; however, many diVerent alleles may exist at a site in a population.

Candidate gene Any gene whose characteristics suggest that it may be the trait-causing gene. The characteristic can be based on relevant function of its protein, or by physical proximity to markers that are tightly linked to a disease.

Complex genetics Patterns of inheritance suggesting that the sum of the eVects of at least two, and usually more, genes are required to produce a trait. This pattern is sometimes referred to as nonMendelian.

Co-segregation Inheritance of two, or more, genetic factors together. These factors could be markers, genes, or traits.

Expression In the context of a particular gene, expression refers to the process of transcribing the gene into RNA so that a protein can be synthesized. In the context of a phenotype, it is the presence of the phenotype.

Genotype The two alleles of an individual at any site in the genome.

Haplotype A combination of two or more alleles that are found on a chromosome. For example, a haplotype A1B3 indicates a chromosome with allele A1 at site A and B3 at site B. Haplotype is sometimes used even if it is not known that both alleles are on one chromosome, or one on each sister chromosome.

Heterozygous Having two diVerent alleles at a genetic marker (i.e., an individual's maternal and paternal alleles can be discerned).

Homozygous Having two alleles of the same type at a locus.

Linkage A measure of the tendency of two sites along a chromosome to be inherited as a unit.

LOD Log of odds, measure used to indicate the likelihood

that a phenotype is located at a particular site. A LOD of 3 indicates odds of 1000/1.

Marker Any site (locus) on a chromosome at which a diVerence can be detected. The diVerent forms of a marker are alleles and can be measured by diVerences in length of DNA, or diVerences in enzyme function for example.

Mendelian genetics Genetic patterns of inheritance that show characteristics of being caused by a mutation at a single gene. Mendelian traits show dominant or recessive characteristics.

Nonparametric A genetic test that does not require describing the mode of inheritance (dominant, recessive) is termed nonparametric.

p Symbol for the short (petit) arm of a chromosome, usually as part of a symbol for a speciWc site based on Giemsa stain banding of chromosomes (e.g., 5p15.3 is chromosome 5, short arm, band 15, sub-band 3) (see also q).

Parametric Genetic model with parameters describing the expected mode of inheritance, such as, dominant, recessive, sex-chromosome linked.

Penetrance A measure of how frequently a diseasecausing mutation is inherited without expressing the disease. Complete penetrance, when the trait is always seen when the mutation is inherited, is contrasted with incomplete penetrance, when there is some dissociation of the trait and mutation.

Phenotype An observed characteristic of an individual (see also qualitative and quantitative phenotypes).

Polymerase chain reaction (PCR) A method of synthesizing multiple copies of a small region of DNA in a short period of time, usually less than 3h. PCR is used to generate suYcient quantities of DNA containing a polymorphism to allow rapid analysis.

Polymorphic Having multiple forms.

Polymorphism An allele that is found with a frequency greater than or equal to 1%.

Positional cloning IdentiWcation and puriWcation of a gene based on its position on a chromosome relative to genetic markers that have been mapped to nearby sites. q Symbol for the long (not petit) arm of a chromosome (see

p for more information).

Qualitative phenotypes Dichotomous, such as the presence or absence of a disease or presence or absence of an enzyme.

Quantitative phenotypes A normal distribution among a group of individuals, such as height. Neurotransmitter densities, patterns of cerebral blood Xow activation, and brain nucleus volumes are likely to show quantitative characteristics.

326D. J. Vandenbergh

Qualitative trait A trait deWned as having only two possibilities (aVected/unaVected, present/absent, above/ below threshold value); also known as discontinuous or dichotomous traits.

Quantitative trait A trait that has a continuous range of possible values (e.g., receptor binding aYnity, percentage change in blood oxygenation levels).

Quantitative trait locus (QTL) A region, or site, on a chromosome that accounts for some part of a quantitative trait.

Recombination Exchange of large fragments of DNA between sister chromosomes that serves to generate a new combination of alleles along a chromosome. The recombination fraction (,) is the frequency with which recombination occurs between two sites; this increases with increasing distance between the sites.

Restriction fragment length polymorphism (RFLP) A marker detected by the presence or absence of the correct DNA sequence that allows the DNA to be cut by a restriction endonuclease. Absence of the correct sequence generates a larger fragment of DNA that can be detected by gel electrophoresis.

Trait See phenotype.

Variant Any allele detected, regardless of the frequency of the allele (see polymorphism).

iReferencesi

Ausubel, F., Brent, R., Kingston, R. E. et al. (eds.) (1997). Short Protocols in Molecular Biology, 3rd edn. New York: Wiley.

Benjamin, J., Greenberg, B., Murphy, D. L., Lin, L., Patterson, C. and Hamer, D. H. (1996). Population and familial association

between the D4 dopamine receptor gene and measures of Novelty Seeking. Nat. Genet., 12, 81±4.

Castellanos, F. X. (1997). Toward a pathophysiology of attentiondeWcit/hyperactivity disorder. Clin. Pediatr., 36, 381±93.

DeFries, J. C., Fulker, D. W. and La Buda, M. C. (1987). Evidence for a genetic aetiology in reading disability in twins. Nature, 329, 537±9.

de Vries, B. B., Halley, D. J., Oostra, B. A. and Niermeijer, M. F. (1998). The fragile X syndrome. J. Med. Genet., 35, 579±89.

Eaves, L. and Meyer, J. (1994). Locating human quantitative trait loci: guidelines for the selection of sibling pairs for genotyping.

Behav. Genet., 24, 443±55.

Ebstein, R. P., Novick, O., Umansky, R. et al. (1996). Dopamine D4 receptor (DRD4) exon III polymorphism associated with human

personality trait of Novelty Seeking. Nat. Genet., 12, 78±80. Ernst, M., Zametkin, A. J., Matochik, J. A. et al. (1996). Presynaptic

dopaminergic deWcits in Lesch±Nyhan Disease. N. Engl. J. Med.,

334, 1568±1604.

Ernst, M., Zametkin, A. J., Matochik, J. A., Jons, P. H. and Cohen, R. M. (1998). Presynaptic dopaminergic activity in ADHD adults. A [Xuorine-18]Xuorodopa positron emission tomographic study.

J. Neurosci., 18, 5901±7.

Falconer, D. S. (1989). Introduction to Quantitative Genetics, 3rd. edn. New York: Longman (Wiley).

Feil, R. and Kelsey, G. (1997). Insights from model systems. Genomic imprinting: a chromatin connection. Am. J. Hum. Genet., 61, 1213±19.

Freeman, B., Powell, J., Ball, D., Hill, L., Craig, I. and Plomin, R. (1997). DNA by mail: an inexpensive and noninvasive method for collecting DNA samples from widely dispersed populations.

Behav. Genet., 27, 251±7.

Freimer, N. B., Service, S. K. and Slatkin, M. (1997). Expanding on population studies. (Comment on Laan and Paabo, 1997). Nat. Genet., 17, 371±3.

Fulker, D. W. and Cherny, S. S. (1996). An improved multipoint sibpair analysis of quantitative traits. Behav. Genet., 26, 527±32.

Fulker, D. W., Cherny, S. S. and Cardon, L. R. (1995). Multipoint interval mapping of quantitative trait loci using sib pairs. Am. J. Hum. Genet., 56, 1224±33.

Gelerter, T. D., Collins, F. S. and Ginsburg, D. (1998). Principles of Medical Genetics, 2nd edn. Baltimore, MD: WIlliams & Wilkins.

Grant, S., London, E. D., Newlin, D. B. et al. (1996). Activation of memory circuits during cue-elicited cocaine craving. Proc. Natl. Acad. Sci. USA, 93, 12040±5.

Grigorenko, E. L., Wood, F. B., Meyer, M. S. et al. (1997). Susceptibility loci for distinct components of developmental dyslexia on chromosomes 6 and 15. Am. J. Hum. Genet., 60, 27±39.

Hallgren, B. (1950). SpeciWc dyslexia (`congenital word blindness'): a clinical and genetic study. Acta Psychiatr. Neurol., 65, 2±289.

Haseman, J. K. and Elston, R. C. (1972). The investigation of linkage between a quantitative trait and a marker locus. Behav. Genet., 2, 3±19.

Hayney, M. S., Dimlanlig, P., Lipsky, J. J. and Poland, G. A. (1995). Utility of a `Swish and Spit' technique for the collection of buccal cells for TAP haplotype determination. Mayo Clin. Proc., 70, 951±4.

Higuchi, R., von Beroldingen, C. H., Sensabaugh, G. F. and Erlich, H. A. (1988). DNA typing from single hairs. Nature, 332, 543±6.

Jason, G. W., Suchowersky, O., Pajurkova, E. M. et al. (1997). Cognitive manifestations of Huntington disease in relation to genetic structure and clinical onset. Arch. Neurol., 54, 1081±8.

Kruglyak, L. and Lander, E. (1995). Complete multipoint sib-pair analysis of qualitative and quantitative traits. Am. J. Hum. Genet.,

1, 121±4.

Laan, M. and Paabo, S. (1997). Demographic history and linkage disequilibrium in human populations. Nat. Genet., 17, 435±8.

LaBuda, M. C., Gottesman, I. I. and Pauls, D. L. (1993). Usefulness of twin studies for exploring the etiology of childhood and adolescent psychiatric disorders. Am. J. Med. Genet., 48, 47±59.

Lahiri, D. K., Bye, S. and Nurnberger, J. I. Jr (1993). A non-organic and non-enzymatic extraction method gives higher yields of genomic DNA from whole-blood samples than do nine other methods tested. J. Biochem. Biophys. Meth., 25, 193±205.

Lander, E. S. and Schork, N. J. (1994). Genetic dissection of complex traits. Science, 265, 2037±48.

Leboyer, M., Bellivier, F., Nosten-Bertrand, M., Jouvent, R., Pauls, D. and Mallet, J. (1998). Psychiatric genetics: search for phenotypes.

TINS, 21, 102±5.

Lewin, B. (1997). Genes VI. Cambridge, MA: MIT Press.

Malhotra, A. K., Virkkunen, M., Rooney, W., Eggert, M., Linnoila, M. and Goldman, D. (1996). The association between the dopamine D4 receptor (DRD4) 16 amino acid repeat polymorphism and Novelty Seeking. Mol. Psychiatry, 1, 388±91.

McCabe, E. R. B. (1991). Utility of PCR for DNA analysis from dried blood spots on Wlter paper blotters. PCR Meth. Appl., 1, 99±106.

McKusick, V. A. (1990). Mendelian Inheritance in Man, 9th edn. Baltimore, MD: Johns Hopkins University Press.

Ott, J. (1991). Analysis of human genetic linkage. Baltimore, MD: Johns Hopkins University Press.

Pianezza, M. L., Sellers, E. M. and Tyndale, R. F. (1998). Nicotine metabolism defect reduces smoking. Nature, 393, 750.

Rao, D. C. (1998). CAT scans, PET scans, and genomic scans. Genet. Epidemiol., 15, 1±18.

Richards, B., Bkoletsky, J., Shuber, A. P. et al. (1993). Multiplex PCR ampliWcation from the CFTR gene using DNA prepared from buccal brushes/swabs. Hum. Mol. Genet., 2, 159±63.

Risch, N. (1997). Evolving methods in genetic epidemiology. II. Genetic linkage from an epidemiologic perspective. Epidemiol. Rev., 19, 24±32.

Risch, N. and Merikangas, K. (1996). The future of genetic studies of complex human diseases. Science, 273, 1516±17.

Risch, N. and Zhang, H. (1995). Extreme discordant sib pairs for mapping quantitative trait loci in humans. Science, 268, 1584±9.

Risch, N. and Zhang, H. (1996). Mapping quantitative trait loci with extreme discordant sib pairs: sampling considerations. Am. J. Hum. Genet., 58, 836±43.

Saiki, R., Gelfand, D. J., StoVel, S. et al. (1988). Primer-directed

enzymatic ampliWcation of DNA with a thermostable DNA polymerase. Science, 239, 487±91.

Sambrook, J., Fritsch, E. F. and Maniatis, T. (1989). Molecular Cloning, a Laboratory Manual, 2nd edn. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press.

Spielman, R. S. and Ewens, W. J. (1998). A sibship test for linkage in the presence of association: the sib transmission/disequilibrium test. Am. J. Hum. Genet., 62, 450±8.

Spielman, R. S., McGinnis, R. E. and Ewens, W. J. (1993). Transmission test for linkage disequilibrium: the insulin gene region and insulin-dependent diabetes mellitus (IDDM). Am. J. Hum. Genet., 52, 506±16.

Strachan, T. and Read, A. P. (1996). Human Molecular Genetics. New York: Wiley.

Terwilliger, J. D. and Ott, J. (1992). A haplotype-based `haplotype relative risk' approach to detecting allelic association. Hum. Hered., 42, 337±46.

TishkoV, S. A., Dietzsch, E., Speed, W. et al. (1996). Global patterns of linkage disequilibrium at the CD4 locus and modern human origins. Science, 271, 1380±7.

Vandenbergh, D. J., Rodriguez, L. A., Miller, I., Uhl, G. R. and Lachman, H. M. (1997a). A high-activity catechol-O-methyl- transferase allele is more prevalent in polysubstance abusers.

Am. J. Med. Genet., 74, 439±42.

Vandenbergh, D. J., Zonderman, A. B., Huang, J., Uhl, G. R. and Costa, P. T. Jr (1997b). No association between Novelty Seeking and dopamine D4 receptor (D4DR) exon III seven repeat alleles in Baltimore Longitudinal Study of Aging participants. Mol. Psychiatry, 2, 417±19.

Weeks, D. E. and Lange, K. (1988). The aVected-pedigree-member method of linkage analysis. Am. J. Hum. Genet., 42, 315±26.

Wong, D. F., Harris, J. C., Naidu, et al. (1996). Dopamine transporters are markedly reduced in Lesch±Nyhan disease in vivo. Proc. Natl. Acad. Sci. USA, 93, 5539±43.

Introduction

Several diVerent methodologic approaches are useful to establish that genes are important for the development and manifestation of complex disorders. Historically, twin, adoption, and family studies have been the methods of choice. It is assumed that genetic factors are important in the expression of a disorder if (i) monozygotic twins have a higher concordance rate than do dizygotic twins, (ii) adopted children resembled their biological parents more often than they resemble their adoptive parents, or (iii) a particular condition is more likely to occur among biological relatives of the patient than would be expected by chance. If there is compelling evidence from these types of study, then genetic linkage studies can be initiated to identify regions of the genome that harbor susceptibility genes.

Genetic linkage has long been recognized as one of the most powerful methods for clarifying the role of genetics in the expression of human disorders. Historically, the method has had limited applicability because of the small number of suYciently polymorphic genetic markers available for study in humans. This situation has changed dramatically. Advances in DNA technology have made it possible to detect many highly polymorphic genetic markers. These genetic markers, based on DNA sequence polymorphisms, have stimulated a renewed interest in linkage approaches to the study of human disorders. As a result, extensive linkage maps of all human chromosomes are available for use in genome wide scans (Gyapay et al., 1994). Theoretical and empirical work suggests that linkage studies can identify the location and thereby verify the existence of genetic loci important in the expression of complex human disorders (Lander and Kruglyak, 1995). However, there remain methodologic issues that need to be addressed so that linkage studies can be more eVectively applied to the study of these conditions.

The focus of this volume is on functional neuroimaging in children. Methods are described that should help to delineate the structure and function of the human brain and how both might be related to variation in human behavior and it is expected that further methodologic development will allow Wner description of the function and anatomy of the brain. Chapter 18 described the techniques of genetics that will facilitate the use of neuroimaging data in genetic studies. In this chapter, some of the issues that need to be considered in the genetic study of complex phenotypes are discussed and some examples are given of how neuroimaging data might be helpful in addressing some of these concerns.

Delineation of phenotype

Arguably, the most critical issue in any genetic study is the accurate delineation of the inherited phenotype. Before genes can be reliably identiWed, it is necessary to understand the range of expression of the phenotype under study. In most studies of neurobehavioral disorders done to date, the approach has been to use categorical diagnoses as the unit of analysis. Furthermore, in most instances, genetic studies begin with a narrow deWnition of aVected status that is rigorously applied when probands are selected. However, when relatives are examined, wider and wider deWnitions are used in deciding whether relatives are aVected. In most studies, considerable eVort has been made to ensure that the probands are as clinically homogeneous as possible. However, when relatives are studied, most investigators have not applied the same strict classiWcation to document who in the family is aVected. It is essential to have additional assessments and evaluation criteria to help to decide which relatives will be considered to have the phenotype of interest.

As noted, to examine adequately hypotheses about the importance of genetic factors, it is necessary to know which individuals have the inherited phenotype. In genetic terms, we need to know something about the variable expressivity of the trait being studied. Variable expressivity is deWned as the range of phenotypes resulting from a speciWc genotype (here genotype refers to the genetic makeup of the individual that is relevant to the expression of the condition being studied). Variable expressivity is not to be confused with comorbidity. To determine whether there is variable expressivity, it is essential to distinguish between comorbidity of two conditions and a spectrum of symptoms that may be part of the inherited trait. Data collected for genetic studies can be helpful in the examination of hypotheses about expression of the phenotype (see Pauls et al., 1993, 1994). If relatives of a proband with a narrowly deWned phenotype are more likely to express other behaviors than would be expected by chance, this can be taken as evidence that those behaviors might be etiologically related to the phenotype of the proband. Furthermore, if twin studies suggest that monozygotic twins are more likely to exhibit those same behaviors as seen in relatives of probands, then it can be concluded that those behaviors are part of the inherited phenotype.

The issue of phenotypic classiWcation is particularly critical in genetic linkage studies. In twin, family, and adoption studies, it is possible to demonstrate the importance of genetic factors even when the deWnition is not exactly correct. However, in genetic linkage studies, false-positive diagnoses can be fatal. The traditional approach in assigning aVected status to relatives in a genetic linkage study has been to start with a very narrow rigorous deWnition of the phenotype and then progressively expand it. Another approach that has been applied in the study of speciWc reading disability (Grigorenko et al., 1997) is to examine component parts of the condition that might represent separately inherited phenotypes. (These components do not represent subtypes of reading disability, but rather processes that are important in learning to read (see Chapter 15).) In this study, the evidence for linkage was strongest with phenotypes that represented processes that put individuals at risk for reading problems (e.g., phonemic awareness and phonologic decoding).

Another example comes from work on obsessive-com- pulsive disorder (OCD). In several studies of OCD, it has been demonstrated that the wide array of symptoms observed among individuals can be reduced to a smaller number of factors that account for a signiWcant proportion of the variance observed in a sample of patients (Baer, 1994; Leckman et al., 1997). These factors form diVerent obsessive-compulsive behaviors. The factor that accounts

for most of the variance in the OCD samples studied is characterized by aggressive, sexual, religious, and somatic obsessions and related checking compulsions. The behaviors that are included in the second factor are obsessions and compulsions related to symmetry, evening up, and having things feel and look just right. The third factor is characterized by concerns about dirt and germs and compulsions that include excessive washing and cleaning. The fourth factor includes obsessions and compulsions that have to do with hoarding. Furthermore, Alsobrook and colleagues (J. P. Alsobrook II et al., unpublished data) have shown that these factors appear to have separate genetic mechanisms. The results of complex segregation analyses suggest that there might be unique genes of major eVect for the aggressive and symmetry factors, while the underlying genetic mechanisms for the other behaviors appear to be more complex. Both the reading and OCD work suggests that alternative strategies might be helpful in identifying phenotypes that are closer to the underlying biological mechanisms which are more likely to be under genetic control.

These two approaches represent ways in which the behavioral phenotype can be partitioned into more meaningful heritable components. Neuroimaging can also characterize individuals and identify diVerent potential biological phenotypes that can be submitted to genetic analyses. By examining the speciWc functional and anatomic diVerences that might be associated with a speciWc neurobehavioral condition, it should be possible to delineate either subtypes of the disorder or components of the larger phenotype that might be under separate genetic control. If a relationship could be established between a behavioral phenotype and a speciWc region of the brain that is either structurally or functionally diVerent, then that brain phenotype could become the phenotype of interest in subsequent genetic investigations, even in the absence of the behavioral phenotype. That is, it could be possible that brain changes could be observed in family members who did not show the behavioral phenotype. Those individuals might represent cases of ªincomplete penetranceº for the behavioral phenotype but not the brain image.

Phenotypes of known genetic disorders

Another approach that is currently being used to understand the relationship between genes and behavior is to examine the so-called behavioral phenotypes of known genetic disorders. Nyhan (1972) ®rst introduced the term ªbehavioral phenotypeº to suggest that there might be

330D. L. Pauls

speci®c clusters of symptoms that are associated with speci®c disorders known to have a speci®c genetic lesion. In fact, when Nyhan ®rst proposed the notion of a behavioral phenotype, he proposed that the speci®c phenotype was chemically determined. At the present time, it is proposed that the genes responsible for the syndrome being studied determine the speci®c phenotype. Or, in the case of deletion syndromes, it is hypothesized that the behavioral phenotype results from one of the genes in the deleted segment of the chromosome that harbors the gene responsible for the speci®c genetic disorder being investigated. One of the best examples of a behavioral phenotype is the self-injurious behavior seen in patients with the Lesch±Nyhan syndrome. Another genetic syndrome that is being investigated in hopes of understanding part of the behavioral phenotype is Williams syndrome. Most individuals with Williams syndrome have a deletion in the short arm of chromosome 7. Hence, several genes are deleted and it is not yet known which one(s) is(are) responsible for the disorder.

Individuals with Williams syndrome are generally mildly to moderately retarded (IQ range 40±100, average 60) and usually have poor abilities in reading, writing, and arithmetic. Yet, despite these lower than average cognitive abilities, they often display remarkable strengths in several other domains. They often have language abilities far above what is expected given their IQs, a facility for recognizing faces, and unusual musical talent. They also tend to be sociable, loquacious, and empathetic. Researchers have been interested in Williams syndrome because it is expected that the unique abilities of these individuals may be related to the genes that are deleted in the region on chromosome 7 (see LenhoV et al. (1997) for a review). The deleted piece of DNA can contain 15 genes or more. It is expected that as the deleted genes are identiWed, it will be possible to determine how their absence may be related to the proWle of behaviors observed in typical individuals with Williams syndrome. For example, some of the genes involved in the deletion have been identiWed; these include those for elastin and LIM-kinase 1, and FZD3, WSCR1, and RFC2. The elastin gene is likely to be responsible for the physical characteristics of Williams syndrome (e.g., cardiac defects, hernias, premature wrinkling) but not for the cognitive aspects of the disorder. The gene for LIMkinase 1 and FZD3 and WSCR1 are known to be expressed in the brain. LIM-kinase 1 is now thought to be involved in visuospatial processes. Concurrent to the search for responsible genes, functional neuroimaging can explore the functional neuroanatomy underlying the unique cognitive characteristics of the disorder and help to map the functional organization and adaptability of the normal

brain. Ultimately, the linkage of genotype to behavioral phenotype in brain imaging studies can enhance our understanding of the neuroimaging phenotype. Several studies are currently underway to delineate the neuropsychology and neurobiology of this disorder, including brain imaging studies of patients. This approach, connecting gene function to neurobiology and Wnally to behavior, may become one way of identifying the eVects of speciWc genes on the developing brain and subsequent behavior. It must be said, however, that while this approach may be useful in identifying the behavioral eVects of genes that are deleted in speciWc genomic regions, it will most likely not be helpful in identifying genes of major eVect for speciWc neurobehavioral syndromes.

Assessment of phenotype

Another critical issue in the search for genes important for the manifestation of behavior is the assessment of the phenotype. Most neurobehavioral genetic research (especially in psychiatric genetics) has been limited by reliance on categorical deWnitions of phenotype. Currently, much eVort is directed to collecting data necessary to make diagnoses. In so doing, much of the richness of the information is lost when categories of illness are deWned. This taxonomy has limited the power in many of the analyses and may have led to incomplete conclusions regarding transmission of traits within families. Given the variability in the manifestation of a neurobehavioral diagnosis, it might be useful to have assessments that would lead to continuous deWnitions or multidimensional deWnitions. Categorical deWnitions may be quite useful when considering treatment and outcome. However, for research purposes, it is helpful to have information about the range of symptom expression that might occur in ªunaVectedº relatives. Although some attempts have been made to include continuous assessment of phenotype, these have not been used extensively in genetic research.

Neuroimaging data could be very helpful in elucidating more quantitative biological aspects of the phenotype. Examining the variability in function or structure should allow a more comprehensive evaluation of the underlying genetic mechanism. A caveat needs to be mentioned. Most studies using quantitative phenotypes require sample sizes that are larger than the typical imaging study. If neuroimaging data are to be used in quantitative genetic studies, samples of at least 75±100 individuals will be necessary. This will require collaborative studies to collect suYciently large samples to include in genetic studies. It should also be stated that imaging data would be required

for all critical members of the family rather than just aVected individuals. Therefore, for sib-pair studies it will be necessary to image at least both siblings and it would be preferable to get image data for both parents as well. For linkage studies of large multigenerational families, all members of the family for which DNA samples have been collected should be imaged. This will also add to the expense and complexity of the research.

Methodologic approaches

A third issue has to do with methodologic approaches. Until recently, there has been an almost exclusive commitment to one approach (i.e., genetic linkage studies of large multigenerational families and data analytic methods requiring speciWcation of genetic model parameters, socalled parametric linkage studies). When it was not possible to replicate the initial genetic linkage Wndings for bipolar disorder and schizophrenia, those working in this Weld were faced with the question of whether this approach would really provide the desired results. In fact, there was considerable discussion in the literature expressing doubt about the possibility of Wnding any major genes responsible for the expression of any neurobehavioral disorder (see Pauls, 1994; Risch and Botstein, 1996). However, with continued development of methodologies (both laboratory and computational), it has been possible to use other genetic linkage paradigms to study more genetically complex disorders (see Chapter 18).

Considerable work has been done (and is ongoing) to clarify the limitations of the genetic linkage approach for the study of complex disorders. Current approaches include a variety of diVerent kinds of family and analytic strategy that have been shown to be more appropriate for the study of complex conditions. Risch and Merikangas (1996) suggested that association studies might be the method of choice for identifying the actual susceptibility locus. Their argument was based on the assumption that a saturated map of the genome would be available. At the present time, such a map is not available and, even if it were, it would be too costly to undertake a genome-wide screen using this strategy. However, association studies can be used to assess the importance of speciWc candidate genes for speciWc traits. Genetic association studies are an eYcient way to identify genotype/phenotype relationships, but in application, they are perilous. Association studies are highly susceptible to false-positive results (Gelernter et al., 1994). In general, false-positive results are most likely when marker polymorphisms, rather than coding region polymorphisms aVecting structure, are used

and when selection of aVected and unaVected groups does not control for ethnicity.

It is possible to reduce or eliminate these problems. First, by studying only polymorphisms in the coding region of a gene, the Wrst problem is minimized. Second, by making allele frequency comparisons only between individuals in the same racial groups (unless it has been demonstrated that racial groups do not diVer in allele frequency for a particular marker), it is possible essentially to eliminate the second problem. One way of perfectly matching for ethnicity is by applying the haplotype relative risk (HRR) method (Falk and Rubinstein, 1987; Terwilliger and Ott, 1992). This method controls for variation in allele frequency owing to ethnicity by constructing a control group of nontransmitted parental alleles (see Chapter 18). The nontransmitted parental alleles (determined by subtracting the set of the oVspring's two alleles from the set of the parent's four alleles) are considered to be in the ªcontrolº group. Since the two parents each donate one allele to the ªillº group and one allele to the comparison group, it is clear that both groups are perfectly matched for ethnicity, as contributions to the two sets of alleles are completely balanced. The strength of the HRR method lies in the parental alleles that are not transmitted to the proband: these alleles form an independent control sample, thereby avoiding problems related to ascertaining control individuals appropriately matched for ethnicity. Neuroimaging studies could become increasingly important in future association studies. If speciWc regions of the brain are shown to be highly correlated with speciWc behavioral phenotypes, genes that are preferentially expressed in those regions would be candidate genes for the behavior being investigated. Of course, considerable work is needed Wrst to demonstrate that speciWc genes are uniquely expressed in the brain region of interest. However, once it has been established which genes are preferentially expressed in the regions of the brain implicated by imaging studies for a particular trait, it is possible with association studies to determine if functional variants of genes are associated with diVerences in brain images. This could become a very powerful technique for future studies.

Risk factors other than genetic ones

A fourth issue in the study of the genetics of behavior, is the relative lack of attention to nongenetic risk factors. While most genetic investigators acknowledge the need to consider environmental factors in genetic studies, few studies have actually been able to measure environmental factors adequately. The same can be said for those studies focused

332D. L. Pauls

on the impact of detrimental environments on the development of psychopathology. Researchers who focus on these eVects also acknowledge the importance of genetic factors but most have failed to account for individual diVerences that might be a consequence of underlying genetic variability. There is a need for a concerted eVort to combine strategies to identify all risk factors important for the expression of the phenotype. In fairness, it may not have been appropriate to include comprehensive assessment of the environment in the initial genetic studies of neurobehavioral conditions. In fact, it may not have been possible to the extent that it is today. At the present time, however, it is possible to combine the methods of neurobehavioral genetics with those from developmental psychology. If that were done, the investigator would have a better opportunity to learn more about the interaction of genes and environment. With appropriate experimental designs, it should be possible to document the role of both genetic and nongenetic factors in the development of abnormal behavior.

Developmental aspects

Finally, a Wfth issue that needs to be considered is the lack of a developmental focus in genetic studies of behavior. For research focused on understanding the underlying genetics of neurobehavioral phenotypes, it is necessary to reshape our thinking in terms of human behavior across the lifespan. As is exempliWed from the focus of this book, it is critical to learn more about the childhood manifestations that might lead to later adult behavior as a consequence of developmental and adaptive changes in the brain. In that regard, it is important to know about possible sensitive periods in development and their relationship to later normal/abnormal behavior. At the present time, the assumption is that some psychopathology is homotypic. That is, it is assumed that the phenotype for a speciWc disorder is invariant over time. For example, childhood depression is diagnosed with criteria for adults and is treated as the same condition in studies examining the familial transmission of major aVective disorders. This is also true for OCD. While it may be possible that some disorders are homotypic, it is highly unlikely that childhood illness will mimic exactly what is seen in adult patients. New research designs are needed that will allow an examination of the possibility of a heterotypic phenotype. It is necessary to know how related behaviors are manifested over diVerent stages of development.

Attention to development is particularly important if the ultimate goal is an understanding of the genetics of a speciWc condition. It is vital to know how genes function

through all stages of development and what impact that function has on the ultimate expression of the phenotype throughout the lifespan. Gene function may be observed through speciWc behaviors, levels of neurochemicals, response to speciWc environmental stressors, interaction with family members or peers, speciWc brain structure or function, actual gene function at the cellular level, or some yet to be determined phenotype. New research paradigms will need to be developed that include assessments of a number of these domains.

It is imperative in future research directed at the elucidation of genetic factors that great care be taken to collect data so that it will be possible to examine both genetic and environmental contributions to the expression of neurobehavioral disorders. While it may be possible to Wnd evidence that genes are important in the etiology of illness, it is unlikely that genetic factors will be both necessary and suYcient for the expression of any illness. It is also possible that it will only be possible to identify the genetic contribution to a speciWc condition when we have adequately documented the impact of nongenetic factors on the expression of the illness. It is also important that this work take place in a developmental context. Work in developmental psychology has demonstrated that early life experiences can have a signiWcant impact on later mental health. For example, it has been clearly demonstrated that the quality of an infant's attachment to her or his mother has strong predictive power for later childhood behavior.While attachment theorists want to attribute much of deviant behavior to deviant attachments (Bowlby, 1988), it is becoming apparent that the genetic endowment of both parent and child may inXuence the attachment (Goldsmith and Alansky, 1987). Therefore, it is not unreasonable to deduce that the attachment of a child to the parent may inXuence how genetic factors may be manifested. This may be part of the unique environment that each child experiences and it may shape the phenotypic expression of the underlying genotype.

Work investigating the impact of environmental factors needs to be done in the context of a genetic design. Prospective longitudinal studies of children at risk need to be undertaken that take into account developmental stages throughout the lifespan. The goal of these studies should not just be to document early signs and symptoms of the syndrome (although this is a worthy goal). The studies should document the early experience of the individual. This experience is not just some global measure of the home and family life. Attempts should be made to document speciWc interactions within and outside the home (i.e., school experiences). In addition, biological assessments should be made including neurochemical measures

and neuroimaging measures. Attention should also be given to diet, illness, and other ªcommonº events in the child's life. While this sounds intrusive, research can be designed to obtain excellent data with minimal intrusion.

As suggested, these studies need to follow entire families, not just single children in families. Furthermore, the assessment of the individuals in these families needs to include various domains of behavior as well as other measures related to those behaviors (e.g., images of the brain). Furthermore, particular attention needs to be paid to critical developmental periods when these measures are obtained. It is important to document whether speciWc stressful events or neurochemical or neuroanatomic changes took place at a particularly critical time in development.

To facilitate this type of prospective longitudinal study, more descriptive studies of development should be initiated. It is necessary to know what normal development (both behavioral and neuroanatomic) is so that it will be possible to assess the impact of abnormal behavior on development (and vice versa) at each diVerent stage. Behaviors are not invariant over time, certainly not from childhood to adulthood. So it is quite possible that the observed phenotype for a speciWc genotype will change over time; consequently, in cross-sectional studies it may not be possible to determine what the complete phenotype is. Furthermore, without longitudinal data, it may not be possible to determine exactly what the impact of speciWc genetic factors might be on an observed phenotype.

Finally, better studies of environmental risk factors are needed. They need to be designed so that the eVects can be examined in the context of diVerent genetic risk. Studies of environmental risk need to be carried out for all ages but, most importantly, we need to learn more about the work being done with very young people (infants). Attachment research has shown that the parental environment is critical for emotional development and the expression of emotions in young children. Some research also suggests that not all children respond in the same way to speciWc events. What is not known is whether this diVerence in response is in some way related to diVerent genes. Certainly, it is not known whether it is in some way related to diVerences in brain structure and function. Consequently, it is important to incorporate more sensitive measurements of unique environments.

Summary and future directions

Genetic studies are done at the molecular level and, with the proposed use of neuroimaging techniques, some aspects of

the phenotype are being described in a more molecular way. It is important to examine the environment at the molecular level as well. Careful examination of parent±child dyads, sib±sib dyads, and peer±peer dyads is important so that the unique environment of individuals can be estimated. It is also important to measure the response to environment. This response should be evaluated in a number of ways: behaviorally, psychologically, neurochemically, and neuroanatomically. These data combined with comprehensive quantitative phenotypic data and evaluated to determine if there is any relationship with genotypic data should help us to understand more fully the etiology of neurobehavioral disorders. The ability to utilize genotypic data and other aspects of genetic studies to design and carry out a study of nongenetic etiologic factors of a neurobehavioral trait is a signiWcant methodological advancement that has not been possible heretofore. Data from prospective studies should make it possible to examine individuals with speciWc genotypes to determine which factors are important for the development of speciWc phenotypes.

iReferencesi

Baer, L. (1994). Factor analysis of symptom subtypes of obsessive compulsive disorder and their relation to personality and tic disorders. J. Clin. Psychiatry, 55 (Suppl), 18±23.

Bowlby, J. (1988). Developmental psychiatry comes of age. Am. J. Psychiatry, 145, 1±10.

Falk, C. T. and Rubinstein, P. (1987). Haplotype relative risks: an easy reliable way to construct a proper control sample for risk calculations. Ann. Hum. Genet., 51, 227±33.

Gelernter, J., Pauls, D.L., Leckman, J. et al. (1994). D2 dopamine receptor (DRD2) alleles do not inXuence severity of Tourette's syndrome: results from four large kindreds. Arch. Neurolog., 51, 397±400.

Grigorenko, E. L., Wood, F. B., Meyer, M. S. et al. (1997). Susceptibility loci for distinct components of developmental dyslexia on chromosomes 6 and 15. Am. J. Hum. Genet., 60, 27±39.

Goldsmith, H. H. and Alansky, J. A. (1987). Maternal and infant temperamental predictors of attachment: a meta-analytic review. J Consult. Clin. Psychol., 55, 805±15.

Gyapay, G., Morissette, J., Vignal, A. et al. (1994). The 1993±94 Genethon human genetic linkage map. J. Nat. Genet., 7, 246±339.

Lander, E. S. and Kruglyak, L. (1995). Genetic dissection of complex traits: guidelines for interpreting and reporting linkage results.

Nat. Genet., 11, 241±7.

Leckman, J. F., Grice, D. E., Boardman, J. et al. (1997). Symptoms of obsessive compulsive disorder. Am. J. Psychiatry, 154, 911±17.

LenhoV, H. M., Wand, P. P., Greenberg, F. and Bellugi, U. (1997). Williams syndrome and the brain. Sci. Am., Dec., 68±73.