The teratology of autism
Tara L. Arndt, Christopher J. Stodgell, Patricia M. Rodier*
Department of Obstetrics and Gynecology, University of Rochester Medical Center, 601 Elmwood Avenue, Rochester, NY 14642, USA
Received 9 September 2004; received in revised form 22 November 2004; accepted 22 November 2004
Autism spectrum disorders affect behaviors that emerge at ages when typically developing children become increasingly social and
communicative, but many lines of evidence suggest that the underlying alterations in the brain occur long before the period when symptoms
become obvious. Studies of the behavior of children in the first year of life demonstrate that symptoms are often detectable in the first 6
months. The environmental factors known to increase the risk of autism have critical periods of action during embryogenesis. Minor
malformations that occur frequently in people with autism are known to arise in the same stages of development. Anomalies reported from
somatic that originate early in the first trimester. In addition, it is possible to duplicate a number of anatomic and behavioral features
characteristic of human cases by exposing rat embryos to a teratogenic dose of valproic acid at the time of neural tube closure.
# 2004 ISDN. Published by Elsevier Ltd. All rights reserved.
Keywords: Autism; Autism spectrum disorders (ASDs); Valproic acid; Eyeblink conditioning; Congenital malformations
Teratology is the study of congenital anomalies and their
causes, whether they are genetic or environmental in origin.
Over most of the period since autism was first described
there has been debate over whether the disorder was
acquired in utero or whether it was acquired closer to the
time when symptoms become obvious, typically around age
two. Hintsthat a change in the course of development occurs
long before the diagnostic symptoms appear have been
available for many years, but it is only in the last decade that
many lines of evidence have come together to indicate that
autism spectrum disorders (ASDs) have their origins in early
prenatal life. This paper summarizes several different kinds
of evidences that address the time when development is
altered and focuses on the anatomical and behavioral
parallelism between human cases and animals exposed in
utero to valproic acid (VPA).
1. How early can symptoms be observed?
It is natural to suspect recent events when symptoms of
any disorder appear, but there are many examples of
neurological disorders in which symptoms begin long after
the precipitating event. It is easy to understand why this
would be true in disorders caused by a gradual loss of
neurons (e.g., Parkinsonism and amyotrophic lateral
sclerosis) or the buildup of some injurious product (e.g.,
phenylketourea). However, even with a single discreet
injury, symptoms may appear long after the fact. A classic
example is the role of dorsolateral prefrontal cortex in
delayed response tasks in primates. Using cryogenic
(1977) demonstrated that the reversible lesion had no effect
on performance when animals were 9 to 16-month-old. In
contrast, a significant impairment of performance was
subjected to cooling between 34 and 36 months of age the
impairment was much greater. Presumably, the late-
developing dorsolateral prefrontal cortex plays little or no
role in the performance of the task until some time in the
second year of life, and its full contribution is not
Int. J. Devl Neuroscience 23 (2005) 189–199
Abbreviations: VPA, valproic acid; ASD, autism spectrum disorder;
SLOS, Smith–Lemli–Opitz syndrome; RARE, retinoic acid responsive
element; RA, retinoic acid; CS, conditioned stimulus; US, unconditioned
stimulus; CR, conditioned response; UR, unconditioned response; PKCg,
protein kinase C–gamma isoform knockout
* Corresponding author. Tel.: +1 585 275 4789; fax: +1 58 5244 2209.
E-mail address: Patricia_Rodier@urmc.rochester.edu (P.M. Rodier).
0736-5748/$30.00 # 2004 ISDN. Published by Elsevier Ltd. All rights reserved.
demonstrable until shortly before puberty. Thus, the onset of
symptoms is a far from perfect guide to when a disorder
began. However, the onset of symptoms cannot precede the
initial injury, so the time when symptoms of autism can be
recognized is an important issue.
An early study to determine when the first symptoms of
autism are expressed used responses of parents to a
questionnaire. Ornitz et al. (1977) compared the responses
of parents of 74 children with autism to the responses of
parents of 38 age-matched typically-developing children.
The questionnaire queried many aspects of early develop-
ment, including motor function, speech, language, and
perception. The results indicated that children who were
later diagnosed with autism exhibited developmental delays
in every area of function, with some landmarks delayed as
early as the second or third month of life. About half of
families were concerned that something was wrong by the
time their child was 14-month-old. Obviously, having
recently received their child’s diagnosis could have
influenced parents as they tried to remember their child’s
early development, but soon investigators came up with a
way around the problems of retrospective recall. Rosenthal
et al. (1980) reported a study of home movies from families
with children later diagnosed with psychoses and movies of
children with typical development. They used a scale of
sensorimotor stages described by Piaget to rate the age
appropriateness of behavioral development in tapes from the
first two years of life. Children in the group with psychoses
differed significantly from controls in showing fewer age-
Better-controlled studies have examined movies of
specific events, such as first birthday parties, and shown
that most children later diagnosed with autism can be
distinguished from controls at one year of age by such
anomalies asfailuretopointtoobjects andfailure torespond
to their name (Osterling and Dawson, 1994). In a more
recent study, investigators blind to diagnosis evaluated
movies from the first six months of life (Maestro et al.,
2002). Even at this age, children who would later be
diagnosed with autism differed significantly from controls
on eight items related to social attention (e.g., looking at
people) and social behavior (e.g., smiling at people and
vocalizing to people), while they did not differ from controls
on most measures related to objects (e.g., looking at objects
and smiling at objects). Several groups are now studying
infant siblings of children with autism, and their preliminary
reports suggest that many symptoms are present at 12
months and a few as early as six months in siblings who later
receive a diagnosis (e.g., Bryson et al., 2004; Zwaigenbaum
et al., 2004). Taken together, studies of behavior prior to
diagnosis suggest that most, if not all, children who will be
diagnosed with autism show symptoms long before the age
when symptoms become obvious.
In a study of heel-stick blood collected from newborns,
would later be diagnosed with either autism or mental
retardation were distinguished by anomalous concentrations
of neuron-related products in blood. Recycling immunoaf-
finity chromatography was used to measure the neuropep-
tides substance P (SP), vasoactive intestinal peptide (VIP),
calcitonin gene-related peptide (CGRP), and the neurotro-
phins, nerve growth factor (NGF), brain-derived neuro-
trophic factor (BDNF), neurotrophin 3 (NT3), and
neurotrophin 4/5 (NT4/5). Of these, VIP, CGRP, BDNF,
and NT4/5 were elevated in neonates who would later be
diagnosed with autism or mental retardation. Each group
differed significantly from a group of typically-developing
children, but the two clinical groups did not differ from each
other. A fourth group, later diagnosed with cerebral palsy,
resembled the controls. It is difficult to interpret why such
differences would be present in blood, where the bulk of the
products measured must have been generated in the enteric
nervous system. Nonetheless, the results suggest that
children with autism are already different from typically-
developing children at the time of birth.
2. What is the exposure period when teratogens that
increase the risk of autism act?
The critical period for exposure to teratogens shown to
increase the risk of autism is early in the first trimester. Five
teratogens related to autism risk have been identified in
? maternal rubella infection (Chess, 1971);
? ethanol (Nanson, 1992);
? thalidomide (Stro ¨mland et al., 1994);
? valproic acid (Moore et al., 2000);
? misoprostol (Bandim et al., 2003).
Of these, thalidomide and misoprostal have been disc-
ussed extensively in the review by Miller et al. (2005). The
timing of the thalidomide critical period was deduced from
accompanying somatic defects to be days 20–24 postcon-
ception (Stro ¨mland et al., 1994). The timing of the miso-
prostol exposures was determined by questioning mothers
and was in the sixth week postconception (Bandim et al.,
Fortunately, the critical period for the other three
teratogens can be estimated from existing data, as well.
The epidemiological sample used to identify the increased
risk for autism after rubella infection did not include data on
time of onset of the rash that heralds rubella, but the
investigators did note that all the children with an autism
outcome had multiple symptoms of rubella injury (Chess
and Fernandez, 1980). In a study specifically designed to
identify the critical periods for eye defects, deafness, mental
retardation, and heart malformations after rubella exposure,
Ueda et al. (1979) found that cases with multiple symptoms
came mainly from those exposed within the first eight weeks
T.L. Arndt et al./Int. J. Devl Neuroscience 23 (2005) 189–199190
Ueda, K., Nishida, Y., Oshima, K., Shepard, T.H., 1979. Congenital rubella
syndrome: correlations of gestational age at time of maternal rubella
with type of defect. J. Pediatr. 94, 763–765.
Van Lint, C., Emiliani, S., Verdin, E., 1996. The expression of a small
fraction of cellular genes is changed in response to histone hyperace-
tylation. Gene Exp. 5, 245–253.
Vorhees, C.V., 1987. Behavioral teratogenicity of valproic acid: selective
effectson behaviorafter prenatalexposure to rats.Psychopharmacology
92 (2), 173–179.
Williams, R.S., Hauser, S.L., Purpura, D.P., DeLong, G.R., Swisher, C.N.,
1980. Autism and mental retardation: neuropathologic studies per-
formed in four retarded persons with autistic behavior. Arch. Neurol.
37 (12), 749–753.
autism. Dev. Med. Child Neurol. 39, 632–634.
Williams, P.G., King, J., Cunningham, M., Stephan, M., Kerr, B., Hersh,
J.H., 2001. Fetal valproate syndrome and autism: additional evidence of
an association. Dev. Med. Child Neurol. 43, 202–206.
Winter, R.M., Donnai, D., Burn, J., Tucker, S.M., 1987. Fetal Valproate
Syndrome: is there a recognizable phenotype? J. Med. Genet. 24, 692–
Woodruff-Pak, D.S., Papka, M., 1996a. Alzheimer’s disease and eyeblink
conditioning: 750 ms trace versus 400 ms delay paradigm. Neurobiol.
Aging 17 (3), 397–404.
Woodruff-Pak, D.S., Papka, M., 1996b. Huntington’s disease and eyeblink
classical conditioning: normal learning but abnormal timing. J. Int.
Neuropsychol. Soc. 2 (4), 323–334.
Woodruff-Pak, D.S., Papka, M., Simon, E., 1994. Eyeblink classical con-
ditioning in Down’s Syndrome, Fragile X syndrome and normal adults
over and under age 35. Neuropsychology 8, 14–24.
Yachnis, A.T., Rorke, L.B., 1999. Neuropathology of Joubert syndrome. J.
Child Neurol. 14, 655–659.
Zwaigenbaum, L., Bryson, S.E., Brian, J., Roberts, W., McDermott, C.,
olds. Presentation at International Meeting for Autism Research,
T.L. Arndt et al./Int. J. Devl Neuroscience 23 (2005) 189–199 199