Manner of genetic effects on autistic traits: suggestions from twin studies

Shinji Ijichi, M.D. & Naomi Ijichi, M.D.
July, 1998

Autism is a behaviorally defined syndrome with a subtle and complex etiology, and little is known about its pathological basis (1). It is suggested that this life-long static developmental disorder may be associated with a combination of several genetic factors (1). Both serotonin transporter gene variant in possible subgroups of patients (2,3) and a gene on chromosome 7q in UK patients (4) are so far just the candidates.

Autism twin studies consistently demonstrate that none of studies reported 100% concordance in monozygotic twins and that the monozygotic twins are not necessarily affected to the same degree (1). This strongly suggests the presence of environmental modulation of the genetic influence (1). Such balanced view that considers both nature (genetics) and nurture (environment) has been implicated in the basis of human cognitive activities and personality (5,6). In this article, published results of twin studies are summarized to speculate the position of autistic traits in human behavior.

Basic concepts
If genetics is important, identical twin cases must be more concordant than fraternal (dizygotic) cases. To scale the heritability of human traits in twin studies, Holzinger's formula for % concordance data had been established (7)(Table 1). In twin studies including measures of correlation of variables in each class of twins, modified Holzinger's formula (7) and doubling the difference between monozygotic correlation and dizygotic correlation (5,6) are utilized as formulae for heritability (Table 1). For instance, if monozygotics were 90% concordant and dizygotics 10% concordant, heritability is 0.89 by the Holzinger's formula. In case of autosomal dominant hereditary diseases exclusively resulted from single genetic abnormality such as Huntington's disease (CAG repeat/polyglutamine disease, 4p16.3), the theoretical concordances are 100% for monozygotic pairs and less than 50% (probandwise) or less than 33.3% (pairwise) for dizygotic pairs with the heritability of 1.0. If twin intraclass correlations were 0.87 for monozygotics and 0.61 for dizygotics, the heritability is 0.67 by modified Holzinger's formula and 0.52 by the doubling method.

Table 1. Formulae for heritability (H) in twin studies.
Holzinger's: H =
(Monozygotic Concordance % - Dizygotic Concordance %)
(100 - Dizygotic Concordance %)
Modified Holzinger's: H =
(rMZ - rDZ)(1 - rDZ)
Doubling method: H =
2 ~ (rMZ - rDZ)
rMZ: correlation coefficient in monozygotic pairs, rDZ: correlation coefficient in dizygotic pairs.

Autism and twin studies
It is pointed out that the earliest study (8) underestimated the autism concordance in monozygotic twin pairs (less than 50%) (1). If one accepts as affected co-twins with somewhat impaired sociability, cognition, or language, autistic traits concordance in monozygotic twins is over 90% and the concordance in dizygotic twins is less than 10% (1). Therefore the heritability of autistic traits can be approximately estimated as 0.89 as shown in Table 3. Even if standard diangnostic criteria were applied for non-probands, the monozygotic concordance should be more than 60% and heritability should be approximately 0.6 (Table 3). A large part of the variance in autism heritability data can be explained by problems in diagnoses of autism or judgement of autistic traits. Autism is just a tail of the consecutive distribution of behavioral parameters and standard criteria treat autistic individuals just above or below the diagnostic cutoff (threshold) (9). Autistic-spectrum disorders merge into what can be called eccentric normality and there is no absolute cutoff point. As compared with neurological diseases and other behavioral disorders, autism is characterized by the concomitance of high monozygotic concordance and low dizygotic concordance with a high heritability index in twin studies (Table 2, 3), suggesting the presence of powerful polygenic influences and a possible hyposensitiveness to the empirical modulation.

Table 2. Twin studies in neurological diseases with apparent pathological processes.
Diseases
Methods
Monozygotic concordance
Dizygotic concordance
Heritability
References
Huntington's disease
probandwise
100%
50%
1.0
theoretical
Idiopathic epilepsy
probandwise
(85%)
(32%)
(0.78)
6
Epilepsy
pairwise
61%
10%
0.57
10
Mental retardation
pairwise
60%
9%
0.56
10
Alzheimer's disease
probandwise
(60%)
(27%)
(0.45)
6
Cerebral palsy
pairwise
40%
0%
0.40
10
Multiple sclerosis
pairwise
27%
2%
0.26
10
Parkinson's disease
probandwise
(5%)
(9%)
(-0.04)
6
( ): approximately

Table 3. Twin studies in behavioral disorders without consistent pathological change.
Disorders
Methods
Monozygotic concordance
Dizygotic concordance
Heritability
References
Autistic traits
pairwise?
more than 90%
less than 10%
(0.89)
1
Specific language impairment
?
90%
50%
0.80
11
Autism
probandwise
(64%)
(9%)
(0.60)
6
Major affective disorder
probandwise
(65%)
(23%)
(0.55)
6
Bipolar disorder
pairwise
56%
14%
0.49
10
Reading disability
?
68%
38%
0.48
11
Schizophrenia
probandwise/pairwise
40% / 28%
15% / 6%
0.29 / 0.23
10
Alcoholism (males)
probandwise
(40%)
(23%)
(0.22)
6
Alcoholism (females)
probandwise
(30%)
(25%)
(0.07)
6
( ): approximately

Cognitive abilities and twin studies
For cognitive abilities, people differ throughout the range from disability to genius. The frequency distribution of individual differences follows the normal (Gaussian) and bell-shaped curve that characterizes quantitative traits. Cognitive abilities are recognized as multifactorial traits resulting from the interplay of multiple environmental factors with multiple genes, and are one of the most heritable domains of human behaviour (11). Quantitative traits are defined as polygenic traits and quantitative trait loci (QTLs) may be largely responsible for the variance accounted for by genetic influences in quantitative traits such as cognitive abilities (11). From a QTL perspective, cognitive disabilities (disorders) are not defined as discontinuous traits (disease) but defined as just the extremes of quantitative traits caused by the same genetic and environmental factors responsible for variation throughout the dimension (11). A threshold (the diagnostic cutoff) is utilized to distinguish disabilities from normal dimensions. QTLs associated with cognitive disabilities are usually by no means necessary or sufficient for the development of the extremes (disabilities) (11). Consequently, twin studies on cognitive abilities are much informative for the understanding of genetic influences on cognitive disabilities (disorders). Table 4 shows twin study data for cognitive abilities and related achievements. A recent study concerning IQ showed that the effects of maternal (intrauterine) environment account for 20% of the statistical covariance of IQ scores between twins, and the heritability was less than 50% (12).

Table 4. Twin intraclass correlations for cognitive abilities and achievements.
Dimensions
Correlation
monozygotic
Correlation
dizygotic
Heritability
(doubling method)
References
Verbal ability
0.76
0.43
0.66
11
Spatial ability
0.62
0.34
0.56
11
General inteligence
(0.86)
(0.60)
0.52
6
Verbal reasoning
(0.77)
(0.52)
0.50
6
Spatial reasoning
(0.65)
(0.45)
0.40
6
Scholastic achievement
(adlescence)
(0.68)
(0.49)
0.38
6
Processing speed
(0.70)
(0.59)
0.22
6
Memory
0.45
0.34
0.22
6
( ): approximately

Parsonality and twin studies
Currently there is a modest consensus that five broad traits or "super factors" are necessary to describe personality; extraversion, neuroticism, conscientiousness, agreeableness, and openness (5). Table 5 shows twin intraclass correlation data for extraversion, neuoticism, and vocational interests in adolescence. Recent twin studies, whose subjects include adopted cases, yield mean estimates of genetic influence of just over 40% (Table 6) and modest estimates of shared environmental influence (7%) (5). It should be emphasized that a sizable contribution from nonadditive genetic factors and quite small estimates of shared environmental influence were described in those studies. Surprisingly, a recent twin study demonstrated a much higher heritability of bad behaviour in adolescents with intraclass co-efficients of 0.814 for monozygotic and 0.293 for dizygotic twins (13). Personality variations can be recognized as two-dimensional frequency distributions of each parameters (super factors) with a manner of quantitative traits. Many genetic factors associated with the personality variation will be revealed as QTLs in the near future.

Table 5. Twin intraclass correlations for personality and vocational interests in adolescence.
Dimensions
Correlation
monozygotic
Correlation
dizygotic
Heritability
(doubling method)
Reference
Extraversion
(0.51)
(0.26)
0.51
6
Neuroticism
(0.47)
(0.24)
0.46
6
Vocational interests
(0.48)
(0.27)
0.42
6
( ): approximately

Table 6. Genetic influence as a source of variation in personality.
Dimensions
MISTRA
twin studies
Loehlin
twin studies
Earlier
twin studies
Reference
Extraversion
(0.38)
(0.48)
(0.54)
5
Neuroticism
(0.50)
(0.41)
(0.58)
5
Conscientiousness
(0.42)
(0.38)
(0.40)
5
Agreeableness
(0.30)
(0.35)
(0.51)
5
Openness
(0.43)
(0.45)
(0.51)
5
( ): approximately

Discussion
After the birth of a first affected child, recurrence risk in families with autism without specific neurological abnormalities is as high as 3 to 7% (1). This 5% risk is substantial, at least 50 times the risk in the population at large, but much lower than 25% (in autosomal recessive cases) or 50% (in autosomal dominant cases) risk of recurrence of single mendelian traits. Therefore, it is suggested that autism may be caused by a combination of several genetic factors (1). As summarized above, twin studies also suggest the presence of powerful polygenic influences for the development of autism. Importantly, as well as cognitive abilities and personality, autistic traits are characterized by the polygenic model and the quantitative trait distribution in the population. The variance accounted for by genetic influences in quantitative traits is usually resulted from influences of QTLs. In human behavioral disorders, QTL linkage was first applied to reading disability and significant QTL linkage was found for markers on the short arm of chromosome 6 (6p21) (11). The serotonin transporter gene variant in possible subgroups of autistic patients (2,3), a gene on chromosome 7q in UK autisitic patients (4), and GABAA receptor subunit gene or adjacent genes (15q11-13) (14) should be considered as QTLs. A functional DNA marker for dopamine D4 receptor (in personality, attention deficit-hyperactivity disorder, opioid dependence, smoking, major depressive disorder, and Tourette syndrome) and a functional polymorphism in a regulatory sequence of the serotonin transporter gene (in neuroticism, depression, and anxiety) are also QTLs (11).

Autistic traits are segrigated from other traits by the characteristics including high (but not 100%) monozygotic concordance and extremely low dizygotic concordance. Autism is one of the most heritable behavioral disorder and the low concordance in dizygotic twin cases may be explained by the polygenic model and a possible hyposensitiveness to the empirical modulation in autistic individuals. The manner of genetic influences in autistic traits, cognitive activities, and personality may be common and the genes should be searched as QTLs.


References
1. Rapin I, Katzman R. Neurobiology of autism. Ann Neurol 43: 7-14, 1998.
2. Cook EH et al. Evidence of linkage between the serotonin transporter and autistic disorder. Mol Psychiat 2: 247-250, 1997.
3. Klauck SM, et al. Serotonin transporter (5-HTT) gene variants associated with autism. Hum Mol Genet 6: 2233-2238, 1997.
4. International Molecular Genetic Study of Autism Consortium. A full genome screen for autism with evidence for linkage to a region on chromosome 7q. Hum Mol Gen 7: 571-578, 1998.
5. Bouchard Jr T.J. Genes, environment, and personality. Science 264: 1700-1701, 1994.
6. Plomin R, et al. The genetic basis of complex human behaviors. Science 264: 1733-1739, 1994.
7. Polygenic inheritance. In: Textbook of Human Genetics (3rd. Edition). Edited by Levitan M. Oxford University Press, New York, p333-p362, 1988.
8. Folstein SE, Rutter M. Infantile autism: a genetic study of 21 twin pairs. J Child Psychol Psychiatry 18: 297-321, 1977.
9. Wing L. The autistic spectrum. Lancet 350: 1761-1766, 1997.
10. Schizophrenia and Manic-depressive Disorder: The Biological Roots of Mental Illness as Revealed by The Landmark Study of Identical Twins. Edited by Torrey EF, et al. BasicBooks, A Division of Harper Collins Publishers, Inc., 1994.
11. Plomin R & Craig I. Human behavioural genetics of cognitive abilities and disabilities. BioEssays 19: 1117-1124, 1997.
12. Devlin B, et al. The heritability of IQ. Nature 388, 468, 1997.
13. McGuffin PM and Thapar A. Genetic basis of bad behaviour in adolescents. Lancet 350: 411-412, 1997.
14. Cook EH, et al. Linkage-desequilibrium mapping of autisitc disorder, with 15q11-13 markers. Am J Hum Genet 62: 1077-1083, 1998.


(If you have any suggestions or questions, feel free to e-mail me !)



Back to AUT