BEHAVIORAL GENETICS: Theory, Substance & Methods Michael Stallings, Ph.D.

Transcription

1 BEHAVIORAL GENETICS: Theory, Substance & Methods Michael Stallings, Ph.D. Institute for Behavioral Genetics University of Colorado

2 Behavioral Genetics Quantitative Genetic Studies Gene Mapping Studies Linkage Studies Association Studies

3 Quantitative Genetics Experimental animal studies Twin, adoption, and family studies How can we use these tools to quantify the extent to which genetic and environmental factors contribute to individual differences in behavior?

4 Understanding Individual Differences Observed phenotypic variance = individual differences

5 Primary Goals Decompose observed phenotypic variation into genetic and environmental sources, or variance components Do these influences operate similarly, or differently, - for males and females - across developmental stages - in different contexts

6 Why do we think genes influence behavior? Animal models: selection, strain differences, cross-breeding, gene manipulations, e.g. knock outs Inherited disorders Patterns of family resemblance, e.g. twin studies, adoption studies, family studies Linkage and association studies

7

8 Inbreed Within High and Low Responders Bottom 10% Short Sleep Top 10% Long Sleep Observed sleep time at each generation

9

10 DeFries et al, 1978

11 DeFries et al, 1978

12 Averaged Parent-Offspring Correlations Biological Adoptive Control Age of Child Parent-offspring correlations between first principal component scores derived from tests of specific cognitive abilities (Plomin( et al., 1997)

13 Monozygotic (MZ) and dizygotic (DZ) twin correlations for individual IQ scores in the Colorado Longitudinal Twin Study (Data from: Bishop et al, 2003, Intelligence ; Friedman et al, 2006, Psychological Science) Correlation MZ DZ Age of assessment (years)

14 Twin studies of common phenotypes Phenotype rmz rdz BMI (age 20 yrs) [Fabsitz et al, 1992] IQ (age 7 years) [Bishop et al, 2001] IQ (age 16 years) [Friedman et al, 2006] Any drug, ever use [Rhee et al, 2003] Any drug, problem use [Rhee et al, 2003] Depression (Finns, female) [Wamboldt et al, 2000] Atopy (Finns, female) [Wamboldt et al, 2000] Heart rate, resting (age 7) [VanHulle et al, 2000] HDL cholesterol (14 years) [Nance et al, 1998] Neuroticism (females,( Australia) [Keller et al, 2005] Extraversion (females,( Australia) [Keller et al, 2005]

15 Twin studies of common phenotypes Phenotype Genetic Environmental Shared Non-shared BMI (age 20 yrs) IQ (age 7 years) IQ (age 16 years) Any drug, ever use Any drug, problem use Depression (Finns, female) Atopy (Finns, female) Heart rate, resting (age 7) HDL cholesterol (14 years) Neuroticism (females,( females, Australia) Extraversion (females,( females, Australia)

16 Important Caveat About Heritability Heritability is not an absolute property of a physical or behavioral characteristic. It is a function of the genetic and environmental variation for a given population in particular circumstances, and at a particular developmental stage.

17 Elaborations of the simple model: sex-limitation G-E E correlation GxE interaction assortative mating non-additive genetic variation social interaction observer effects pleiotropy patterns of development measured environments measured genes

18 Averaged Parent-Offspring Correlations Biological Adoptive Control Age of Child Parent-offspring correlations between first principal component scores derived from tests of specific cognitive abilities (Plomin( et al., 1997)

19 Monozygotic (MZ) and dizygotic (DZ) twin correlations for individual IQ scores in the Colorado Longitudinal Twin Study (Data from: Bishop et al, 2003, Intelligence ; Friedman et al, 2006, Psychological Science) Correlation MZ DZ Age of assessment (years)

20

21 Human twin, family, and adoption studies of substance use/abuse/dependence Rhee, Hewitt, Young, Corley, Crowley, & Stallings (2003) Genetic and environmental influences on substance initiation, use, and problem use in adolescents. Arch Gen Psychiatry. 60: Young, Corley, Stallings, Rhee, Crowley, Hewitt (2002) Substance use, abuse and dependence in adolescence: prevalence, symptom profiles and correlates. Drug Alcohol Depend. 68:

22 Percent Percentage of community sample adolescents experimenting with substances Age in years alcohol, males alcohol, females tobacco, males tobacco, females marijuana, males marijuana, females `other', males `other', females

23 Percentage of community sample adolescents dependent on substances 25 Percent alcohol, males alcohol, females tobacco, males tobacco, females marijuana, males marijuana, females `other', males `other', females any substance, males any substance, females Age in years

24 Substance use and the development of dependence symptoms are common in adolescence. By age eighteen, 70% of adolescents are experimenting with alcohol, 1 in 3 with marijuana, 1 in 3 with tobacco, 1 in 10 with other illicit drugs. 1 in 5 show dependence on some substance (often nicotine). There are few marked sex differences in substance use or dependence during adolescence. There is a marked, almost linear, increase in substance use and then dependence during the high school years.

25 Family resemblance for `any drug in the Colorado Drug Research Center (Rhee et al, Arch. General Psychiat., 2003) Correlation Use Problem use MZ twins DZ twins Full Siblings Adoptive Siblings Behavior

26 All aspects of substance use show family similarity. Much of the family similarity appears to be genetic, especially for abuse and dependence. The shared environmental influences are more marked for close age siblings, underscoring the likely importance of peer influences.

27 Problem Substance Use

28 Comorbidity and genetic pleiotropy Are there common risk factors across different substances of abuse? To what extent is substance dependence comorbid with other psychiatric symptoms? To what extent are comorbidity or correlations among traits due to common genetic influences?

29 Mulitivariate Genetic Analysis

30 Univariate Analysis G E P Square = observed phenotype/trait Circles = latent influences

31 Univariate Analysis G C E P C = common, or shared environment E = unique, or nonshared environment

32 Univariate Analysis A D C E P A = additive genetic influences D = nonadditive (dominant) genetic influences

33 Multivariate Analysis A 1 A 2 P 1 P 2 E 1 E 2

34 Multivariate Analysis r A A 1 A 2 P 1 P 2 E 1 E 2 r E

35 Genetic and Environmental Vulnerabilities Underlying Adolescent Substance Use and Problem Use (Young et al., 2006) Substance Use: repeated use as defined by CIDI-SAM A T A C A M T T T A T M TOB ALC MAR TOB ALC MAR C T C A C M E T E A E M

36 Note the substantial shared environmental influences on substance use, both specific to twins (t 2 ) and for non-twin siblings (c 2 ) T T T A T M TOB ALC MAR TOB ALC MAR C T C A C M

37 And relatively modest intercorrelations among the genetic and environmental influences for measures of substance use A T A C A M T T T A T M TOB ALC MAR TOB ALC MAR C T C A C M E T E A E M

38 Problem Use: endorsement of 1+ DSM dependence criteria A T A C A M T T T A T M TOB ALC MAR TOB ALC MAR C T C A C M E T E A E M

39 Correlations among substance dependence symptoms and other psychiatric symptoms in our adolescent community sample GAD MDD ADHD Conduct Disorder Nicotine Alcohol Marijuana Other Dependence Vulnerability

40 Correlations among substance dependence symptoms and other psychiatric symptoms in our adolescent community sample GAD MDD ADHD Conduct Disorder Nicotine Alcohol Marijuana Other Dependence Vulnerability

41 Correlations among substance dependence symptoms and other psychiatric symptoms in our adolescent community sample GAD MDD ADHD Conduct Disorder Nicotine Alcohol Marijuana Other Dependence Vulnerability

42 Other Structural Models A 1 A 2 A 3 ASB ALC DRG

43 Cholesky Decomposition A 1 A 2 A 3 ASB ALC DRG

44 Cholesky Decomposition A 1 A 2 A 3 ASB ALC DRG

45 Cholesky Decomposition A 1 A 2 A 3 ASB ALC DRG

46 Cholesky Decomposition A 1 A 2 A 3 ASB ALC DRG

47 Common Factor Model A C E LT ASB ALC DRG

48 Button et al., 2007 (Drug & Alcohol Dependence) A 1 A ALC DRG E 1 E 2

49 Button et al., 2007 (Drug & Alcohol Dependence) A 1 A 2 A ASB ALC DRG

50 Button et al., 2007 (Drug & Alcohol Dependence) A 1 A 2 A ASB ALC DRG

51 Substance Abuse is Part of a Spectrum of Problem Behaviors (Young et al., 2000) A Beh Disinhib E.47 CD SUB ADHD NS

52 There is marked comorbidity both across substances and between substance use and other psychiatric problems, especially externalizing problem behavior, such as conduct disorder. In adolescence, underlying genetic factors may have general influences on vulnerability to substance use and dependence and, possibly, conduct problems.

53 Gene-Environment Interaction

54 the Caspi hypothesis. Science Aug 2;297(5582): Role of genotype in the cycle of violence in maltreated children. Caspi A, McClay J, Moffitt TE, Mill J, Martin J, Craig IW, Taylor A, Poulton R. Medical Research Council Social, Genetic, and Developmental Psychiatry Research Centre, Institute of Psychiatry, King's College, London SE5 8AF, UK.

55 Caspi et al., Science, 297:

56 Haberstick, Lessem, Hopfer, Smolen, Ehringer,, Timberlake, Hewitt (2005) Monoamine oxidase A (MAOA) and antisocial behaviors in the presence of childhood and adolescent maltreatment. American Journal of Medical Genetics Part B: Neuropsychiatric Genetics, 135: 59-64

57 Huizinga et al, Biological Psychiatry 2005

58 Percent with Conduct Disorder Foley et al, Archives of General Psychiatry, Level of exposure to childhood adversity Low MAO-A High MAO-A

59 Percent with Conduct Disorder Foley et al, Level of exposure to childhood adversity N = 1 Low MAO-A High MAO-A

60 Foley et al, 2004 changing two cases Percent with Conduct Disorder Level of exposure to childhood adversity Low MAO-A High MAO-A

61 Foley et al, 2004 binning adversity to have >5 subjects per cell Percent with Conduct Disorder Level of exposure to childhood adversity Low MAO-A High MAO-A

62 Young et al, American Journal of Psychiatry, 2006

63 Genotype Environment Interaction in Psychopathology: Fact or Artifact? Lindon J. Eaves, Twin Research, 2006 ``The main outcome of this study was simulated liabilities and diagnoses of major depression and antisocial behavior. Analysis of the dichotomized data by logistic regression frequently detected significant G E interaction even though none was present for liability. There is therefore reason to question the biological significance of published findings.

64 Behavior genetic studies suggest that both genetic and environmental influences are ubiquitous. Gene by environment interactions are a given --- heritability may be different in different environments, at different ages, at different stages of development. Genes can create both stability and change during development. Genes often influence multiple aspects of behavior (pleiotropy( pleiotropy).