Autism Neuroscience
Autism Neuroscience
Autism Neuroscience
Related article
IMPORTANCE Autism spectrum disorder (ASD) is a highly prevalent disorder, and community
psychiatrists are likely to treat many individuals with ASD during their clinical practice. This
clinical case challenge describes a routine evaluation of irritability and self-injury in a
preschool-aged child who meets the criteria for ASD. The case also illustrates the importance
of known risk factors for ASD, such as chromosomal deletion and prematurity. This clinical
neuroscience article seeks to educate the clinician of current avenues of research that can
inform and may already affect clinical practice for this patient, while providing a preview of
research that may yield biological treatments for ASD within the next decade.
T
he Clinical Challenge in this issue of JAMA Psychiatry occurringsymptoms.ThiscaseillustrateswhereweareinthearcofASD
describes a young boy who presented with irritability and assessment and care: at the cusp of incorporating an understanding
self-injury in the context of developmental delay, social of autism neurobiology into treatment, but not quite there yet.
deficits, and restricted, repetitive behaviors.1 Evaluation confirmed Parallelingrapidadvancesinautismresearch,wehaveseenasurge
a diagnosis of autism spectrum disorder (ASD), along with intellectual in the recognition and diagnosis of ASD, and current prevalence rates
and language impairments per DSM-5 criteria (Table 1). 16p11.2 Dele- are estimated at 1% to 2%, with 4 to 5 males receiving a diagnosis for
tion syndrome can also be added to the diagnosis, to indicate relevant every female who receives a diagnosis of ASD. This surge, coupled with
neurodevelopmental and medical co-occurring conditions. However, fraudulent research, led to concerns that vaccines could contribute to
this molecular genetic diagnosis did not change the process of evalu- the risk of ASD, which has clearly been disproven.2 Most data instead
ating him in the psychiatric clinic, and there are currently no pharma- point to increased awareness and resources for screening, diminished
cologic treatment recommendations based on genotype. Research- stigma, and changes in diagnostic criteria as explanations of today’s
ers are just beginning to test treatments for core symptoms based on higher recognition of ASD. A few known risk factors, such as advanced
biologicalfindingsinsubgroupsofchildrenwithASD,andcurrenttreat- parental age and surviving extreme preterm birth, are also increasing
ments therefore remain either behavioral or applied only to co- in parallel with ASD prevalence.
Genetic variants associated with autism spectrum disorder (ASD) Clinical examples Genetic test
Nucleotide insertion and/or deletion (indel) Single-nucleotide variant (SNV) ASD associated with Whole exome
CHD8 mutation sequencing
Normal Normal
ASD with Targeted gene
DNA DNA macrocephaly sequencing by PCR
associated with
Protein Protein
PTEN mutation
A, Pedigrees of people affected with autism spectrum disorder (ASD) and their families. Left, pedigree of a family demonstrating multiple inherited genetic risk factors in
the affected children of 2 nonaffected parents. Right, pedigree of a family showing the development of a new mutation in the offspring that contributed to the develop-
ment of ASD in that child. B, Large- and small-scale genetic changes associated with increased risk of ASD. The different classes of genetic variants are presented by size,
from large-scale alterations (entire chromosomes or large regions missing or duplicated) at the top, to substitutions of just 1 DNA nucleotide at the bottom. Examples of
DNA sequence changes in the coding region of the CHD8 gene associated with ASD are presented in the bottom panel. Sequence corresponds to chromosome 14 (hg19),
21 878 138 to 21 878 130 base pairs (left) and 21 870 179 to 21 870 160 base pairs (right). Variants may be inherited from a carrier parent or develop de novo in the person
with ASD. Variants that are more damaging to biological function are less likely to be inherited and may interact with other genetic, environmental, and developmental risk
factors to yield a clinical diagnosis of ASD. Single-nucleotide polymorphisms (SNPs) (not depicted) are common variants that contribute to ASD risk, but at this time, indi-
vidual SNPs have not been linked to risk for ASD to justify clinical screening. Whole-genome sequencing detects copy number and nucleotide sequence variants, but it is not
yet clinically available for molecular genetic diagnostic testing owing to increased expense and data storage limitations. PCR indicates polymerase chain reaction.
Table 2. Recurrent De Novo Copy Number Variants (CNVs) Associated With Autism Spectrum Disordera
CNV Locus Copy Number Change Other Associated Conditions Syndrome Name
1q21.1 Duplication ID, SCZ, ADHD, seizures, macrocephaly, CHD, scoliosis, and short 1q21.1 Duplication syndrome
stature23,24
2p16.3 Deletion ID, speech delay, SCZ, seizures, macrocephaly, dysmorphology, CHD, and 2p16.3 Deletion syndrome
hypotonia25-27
3q29 Deletion ID, speech delay, SCZ, anxiety disorder, dysmorphology, dental 3q29 Microdeletion syndrome
abnormalities, FTT, gastrointestinal problems, CHD, and recurrent ear
infections28,29
7q11.23 Duplication ID, speech delay, ADHD, anxiety disorder, seizures, macrocephaly, Williams-Beuren syndrome
dysmorphology, CHD, and renal anomalies30-32
15q11.2-q13.1 Duplication ID, ADHD, seizures, hypotonia, CHD, and short stature33-38 15q11.2-q13.1 Duplication syndrome
15q13.2-q13.3 Deletion ID and ADHD33,39 15q13.3 Microdeletion syndrome
15q13.2-q13.3 Duplication ID, speech delay, epilepsy, urogenital anomalies, and recurrent 15q11-q13 Duplication syndrome
infections33,36
16p11.2 Deletion ID, ADHD, DCD, seizures, macrocephaly, obesity, hypotonia, and 16p11.2 Deletion syndrome
articulation abnormalities40-42
16p11.2 Duplication ID, SCZ, seizures, microcephaly, hypotonia, tremor, and articulation 16p11.2 Duplication syndrome
abnormalities40,41,43
22q11.2 Deletion ID, learning disorder, speech delay, SCZ, ADHD, anxiety disorder, Velocardiofacial syndrome, DiGeorge
dysmorphology, palate abnormalities, CHD, renal anomalies, hypocalcemia, syndrome
and recurrent infections44
22q11.2 Duplication ID, ADHD, learning disorder, hypotonia, hearing loss, palate abnormalities, 22q11.2 Microduplication syndrome
CHD, urogenital anomalies, and growth retardation45-47
22q13.3 Deletion ID, seizures, macrocephaly, dysmorphology, strabismus, CHD, renal Phelan-McDermid syndrome
anomalies, hypotonia, and scoliosis48,49
a
Abbreviations: ADHD, attention-deficit/hyperactivity disorder; CHD, congenital Copy number variants associated with autism spectrum disorder sourced from
heart disease; DCD, developmental coordination disorder; FTT, failure to thrive; Sanders et al.20
ID, intellectual disability; SCZ, schizophrenia.
More recently, geneticists have examined the protein-coding re- The patient carried a microdeletion of 16p11.2, a region that en-
gions of the genome (the exome) using whole-exome sequencing, compasses 29 genes. There are now several large clinical cohorts with
which allows detection of single-nucleotide variants. Compared with 16p11.2 deletions (del16p11.2) and duplications (dup16p11.2), and as-
their unaffected siblings, children with ASD had an elevated rate of sociations have been reported with ASD, schizophrenia, intellec-
rare, de novo, single-nucleotide variants and insertions and/or de- tual disability, and seizures.20,58-60 People carrying 16p11.2 CNVs ex-
letions that likely disrupt genes (Figure 1).20,50-53 Combined analy- hibit mirror phenotypes: deletion is associated with obesity and
sis of multiple family cohorts found that individual genes could be macrocephaly,60-63 while duplication is associated with lower weight
confidently assigned ASD risk based on being disrupted in affected and microcephaly.19,59,64 Some people with a 16p11.2 CNV do not
individuals more frequently than could be expected by chance have a clinical disorder, but performance IQ and general function-
(Table 3).20 These genes cluster into the following 2 groups: 1 en- ing may be lower compared with the general population.65,66 There
riched for proteins involved in brain function (eg, synaptic function- are likely additional genetic and environmental modifiers that ulti-
ing and neuron projection) and the other enriched for gene regula- mately dictate whether the carrier has a particular disorder and to
tion (eg, chromatin organization and transcription). Using what degree.67
computational approaches incorporating a gene expression atlas of
the human brain across the lifespan, researchers identified en- Environmental Risk Factors
riched coexpression of ASD risk genes in specific cortical regions from As noted previously, twin studies suggest that both genetic and en-
early gestation to midgestation.54,55 vironmental risk factors are associated with ASD (Figure 2). Poten-
As remarkable as progress has been, the known variants are just tial nongenetic risk factors include maternal conditions, complica-
the tip of the iceberg. Modeling predicts that approximately 200 de tions of pregnancy, medications, and exposure to toxic materials.
novo CNVs and hundreds of genes contribute to ASD risk,20,53 likely Most of these risk factors act on brain development during the pre-
interacting with environmental risk factors to result in ASD. Al- natal and perinatal stages. Establishing causal roles for environmen-
though each of these variants may be individually rare, collectively tal factors is complicated by retrospective reporting and overlap-
they are found in a large minority of the clinical population with ASD. ping risk factors, but several nongenetic risk factors for ASD have
Currently, medical associations recommend molecular diagnostic now been established.
testing using chromosomal microarray and fragile X testing in all The most well-supported environmental risk factor for ASD is
people diagnosed as having ASD, and adjunctive clinical whole- advanced parental age,68-71 with most evidence suggesting that this
exome sequencing may be helpful in those with congenital abnor- contributes to genetic risk. Maternal and paternal age appear to be
malities or unexplained co-occurring medical conditions.56,57 The independent risk factors, leading to an 18% (for mothers) and 21%
identification of a molecular diagnosis can provide clinical care guide- (for fathers) increase in ASD risk for every 10-year increase in the age
lines and may warn of potential comorbidities that would other- of the parent.71 The risk of chromosomal abnormalities is well known
wise go unmonitored. in mothers of advanced maternal age, and prenatal genetic screen-
ing is routinely offered in this population. Older fathers contribute Exogenous risk factors, such as medication exposure during preg-
more de novo mutations as they age.50,52,72 The total elevated risk nancy, also play a role in the risk of ASD. The prenatal use of anti-
is likely owing to other factors as well, including an increase in com- convulsants has been associated with an increased risk of ASD, with
plications of pregnancy, epigenetic modifications of the genome, and valproic acid use during pregnancy showing the clearest
social factors that may contribute to having children later in life.71 association.70,77 Asthma medications, particularly β2-adrenergic re-
Prematurity, even up to 37 weeks, is another well-established ceptor agonists, have also been linked to autism risk.70 Although still
risk factor for ASD.70,73 The preterm brain has vulnerability to cer- controversial, there is emerging evidence to support prenatal ex-
tain forms of gray and white matter brain injury that affect subcor- posure to air pollution as a risk factor for ASD.70,71
tical and cortical connectivity (theme 2); however, preterm birth is Environmental risk factors for ASD have significant implica-
associated with ASD even in the absence of overt neuronal injury.74 tions for public policy and prenatal health care. Discerning the mecha-
The elevated risk of ASD is likely compounded by prematurity- nisms behind environmental risks may be particularly difficult, how-
associated complications, such as anoxia and hypoglycemia. The risk ever, because they may act directly on neurons and supporting cells
for ASD is greatest in extremely preterm infants. In those with very or may act indirectly by influencing hormones or inflammatory re-
low birth weight and extreme preterm birth, ASD is more likely to sponses. Alternatively, environmental risks may act by altering the
be associated with major cognitive and motor impairments.75 In the expression of relevant genes that affect ASD risk. Despite the diver-
case presented, the patient was born at a gestational age of 32 weeks, sity of genetic and environmental risk factors in ASD, it is important
a time when the brain is still in the migration phase and has begun to understand how they may interact and lead to downstream ef-
synaptic pruning and myelination (Figure 2). This child therefore ap- fects on brain development, and thereby indicate common path-
pears to have 2 independent risk factors for ASD. ways underlying the core deficits of ASD.
Dozens of other environmental factors have been linked with
ASD, although further evidence is needed to consider them as es-
tablished risk factors for ASD. These risk factors can generally be cat-
Theme 2: Altered Brain Development in ASD
egorized as either endogenous, originating from within the maternal-
fetal unit, or exogenous, originating from outside of the mother.76 Neuroimaging studies increasingly support a picture of early atypi-
The endogenous risk factors with the best support, outside of pre- cal brain development and widespread alterations in neural con-
maturity, are short-interval pregnancies, defined as pregnancies oc- nectivity in individuals with ASD. Some of the most promising find-
curring fewer than 12 months apart70; birth complications involv- ings are in infants and toddlers who have an older sibling with ASD,
ing hypoxia and ischemia or trauma; maternal obesity; and diabetes.71 and are therefore at increased risk for ASD. Prospective neuroim-
aging studies consistently point to abnormal trajectories of brain de- owing to a surplus of neurons that did not undergo typical apoptosis
velopment, rather than static changes or localized lesions. Effec- and pruning earlier in development.
tive brain development depends on the balance between cellular and From the ages of 2 to 4 years, brain volumes in children with ASD
synaptic growth and appropriately timed pruning of neurons and syn- remain enlarged compared with those of their peers.85,86 The amyg-
apses. In at least some children with ASD, this balance appears to dala, frontal cortex, and temporal cortex, which are important for
be disrupted. social cognition, language, and regulation of emotions, are particu-
Neuroimaging studies find that changes occur well before the larly affected.79,85 By school age, brain growth has slowed, with brain
onset of clear symptoms of ASD. The association of risk with pre- volumes of typically developing children catching up to those in chil-
natal exposures suggests that many of these disruptions begin dur- dren with ASD.85,86 Connectivity studies find ongoing disruptions
ing the prenatal period. Recent neuroimaging studies have exam- in how brain regions communicate throughout school age, adoles-
ined brain development in high-risk infants from 6 months of age, cence, and adulthood. A large pooled analysis of resting-state data
suggesting that infants who will later develop ASD already exhibit from functional magnetic resonance imaging performed in individu-
abnormal connectivity by this time.78-81 Connections involved in low- als with ASD found a predominance of hypoconnectivity of long-
level sensory processing appear particularly affected at this early range cortical-cortical and interhemispheric projections compared
stage.81 In some studies, the degree of atypical connectivity at 6 with age-matched typical controls.87 In contrast, subcortical re-
months correlates with future symptom severity.79,81 gions exhibited hyperconnectivity of local connections. As a whole,
Starting between the ages of 6 and 12 months, children who will these data suggest that higher-order brain functions requiring com-
go on to have ASD show an expansion of cortical surface area, starting munication between separate brain regions are depressed in favor
insensorydomainsunderlyingauditoryandvisualprocessing,followed of local circuits that may be overactive and difficult to disrupt.
by more global overgrowth from the ages of 12 to 24 months.82,83 This The 16p11.2 microdeletion illustrates the strong influence that
trajectory mirrors typical development, but in an exaggerated manner genetic variants can have on brain development. Individuals with
(Figure 2). In typical development, brain growth during this period is a deletion (1 copy of this region) show significantly larger brain
largely attributed to the expansion of glia and myelination.84 Similarly, volumes than do unaffected individuals who carry 2 copies.88
in individuals with ASD, this observed overgrowth likely does not rep- Similar to results from general studies of ASD, both gray matter
resent new neuron development, although it could reflect overgrowth and white matter are affected by 16p11.2 deletion, particularly in
regions implicated in reward, language, and social cognition.89 exhibit hyperactivity, repetitive behaviors, and learning deficits remi-
More important, these associations are seen at a group level, with niscent of the human phenotype.96-100
considerable variability across individuals. Individuals with a Animal models are also crucial in the development of poten-
duplication (3 copies) of 16p11.2 tend to have microcephaly, which tial treatments. For example, mice with del16p11.2 showed dimin-
is also associated with other ASD risk factors (eg, fetal alcohol ished long-term potentiation, similar to a mouse model of fragile
syndrome and DYRK1A [OMIM 600855] disruption) (Table 3). X syndrome.98,101 These animals were unable to acquire or extin-
This dichotomy demonstrates the heterogeneity underlying ASD. guish a memory of an unpleasant stimulus, which could be res-
Although neuroimaging research is advancing our under- cued by treatment with a glutamate receptor mGluR5 modulator.
standing of the biology of ASD, there is no evidence to support More recently, researchers used R-baclofen, a γ-aminobutyric
routine neuroimaging in autistic individuals.90,91 In 16p11.2 dele- acid–B receptor agonist, to rescue deficits in object or spatial rec-
tion syndrome, the recommendation is the same, with the caveat ognition in mice with del16p11.2.102
that individuals with symptoms suggestive of Chiari I malforma- In addition to animal models, new techniques using human
tion or syringomyelia, conditions that are more common in this cell lines can illuminate the molecular consequences of genetic
syndrome, should undergo neuroimaging.92 More research is risk variants. For example, using lymphocyte cell lines obtained
needed to better understand and standardize the trajectory of from carriers of 16p11.2 CNVs, researchers found that expression
brain development in ASD before neuroimaging could be consid- of genes within this region rose or fell with duplications or dele-
ered for a clinical role. These findings do emphasize the utility of tions, as expected, but some genes outside of the region were
early diagnostic screening, however, and early behavioral inter- also affected, potentially pointing to critical pathways that medi-
vention is critical to improve social and communication function- ate the risk of ASD.103,104 Alternatively, reprogramming cells into
ing. Fine-tuned behavioral screening, such as the use of eye track- induced pluripotent stem cells and differentiating them into indi-
ing to monitor attention to social scenes, could eventually be vidual neuronal lineages may provide better consistency with
used to identify children at risk of ASD before overt symptoms brain cells.105 These cultures have thus far been limited to early
emerge, but this finding has not yet been demonstrated in clinical stages of brain development and manifest fewer specific cell
settings.93 types, but if deficits are identified in cells derived from patients,
therapeutic interventions could be evaluated and provide test-
able targets and hypotheses for clinical research trials.106
Theme 3: Model Systems of ASD
Computational, cellular, and animal model systems can help to
dissect the molecules, cells, and circuits that connect risk factors
Future Directions and the Promise
for ASD to biology. Model systems are particularly important for
of Precision Medicine
understanding how and when these risk factors affect brain
development and lead to downstream changes in behavior. Ani- Based on available evidence, a diagnosis of ASD should be followed as
mal models provide a developmentally informed context in which soon as possible by implementation of early, intensive behavioral in-
to study the consequences of a genetic variant on brain struc- tervention,whichimprovescognitiveandadaptiveoutcomesformany
tures and cell types. Even relatively simple animals, such as (but not all) preschoolers with ASD.107 Behavioral and educational in-
zebrafish, have substantial conservation of ASD risk genes. For terventions for school-age children are tailored to individual needs but
example, of the 29 genes in the 16p11.2 CNV in humans, 24 genes have much less research support.107 Evidence-based pharmacologic
are conserved in zebrafish.94 Rodent and primate models have treatments in ASD are limited to co-occurring symptoms, such as agi-
more biological similarity to humans and may exhibit behaviors tationorhyperactivity.108 Somecliniciansandparentsalsochoosefrom
that may be more analogous to human social or repetitive behav- apanoplyoftreatmentswithnoevidence,drivenbyadesiretoaddress
ior. Researchers recently created several monkeys with disruption the root causes of ASD.109 This use of treatments with no evidence is
of the Rett syndrome gene, MECP2 [OMIM 300005], which understandable when none of our available biological treatments are
exhibited decreased social contact and more stereotyped treating the core neurobiological determinants of the disorder.
behaviors.95 Although primate models are closest in behavior and Thediscoveryofraresingle-genevariantsandCNVsraisesthepos-
biology to humans, ethical issues and their high cost have limited sibility of ASD treatments rooted in biology, rather than in behavioral
their widespread use. principlesorpharmacologicobservations.Thepresenceofaknownge-
The mouse brain is less complex than the human brain, but netic risk variant in the described child is quite likely to have tipped the
mouse models that recapitulate a human gene variant can be cost- scales toward an ASD outcome. This motivates investigations for treat-
efficient and time-saving while still providing a developmental con- ments targeted to molecular genetic diagnoses rather than to an ASD
text to observe the effect of risk factors on specific cells, tissues, and diagnosis per se. With the advance of designer gene editing technol-
organs. For example, several altered cellular mechanisms and path- ogy, it is possible that interventions could be targeted to specific ge-
ways have been identified in mice with the 16p11.2 deletion, includ- netic loci (to repair a genetic sequence or to correct aberrant gene
ing premature exiting of neuronal progenitor cells from the cell cycle activity),110 althoughthepossibilityofunintendedgeneticchanges(off-
and increased dopamine receptor expression on striatal γ-amino- target effects) is currently too large to suggest such interventions for
butyric acid–ergic neurons.96,97 While it must be stated that behav- all but the most devastating syndromes.111
ioral tests in mice may not accurately reflect symptom domains of Precision medicines in ASD-related genomic disorders will not
ASD (or human social cognition more generally), these animals did be as straightforward to test as in cancer, in which clinical response
can be measured by tumor regression or normalization of cell counts. ture. Our neuroscience themes are focused on future clinical effect,
Tests of neurobiological hypotheses in developmental disorders re- rather than what makes a difference for treatment right now. For the
quire careful consideration of our standard protocols. We almost cer- boy and his family described in the Clinical Challenge,1 knowing his ge-
tainly need to study treatments at an earlier point in childhood de- nomic risk variant allows for family planning, provides insight into one
velopment, rather than binding ourselves to the standard timeline cause of his disorder, and connects the family to a network of others
of efficacy studies in adults, then in adolescents, and finally in withthesamegenomicfinding.Theidentificationofawell-definedbio-
children.112 We also need to define and refine outcome measures that logical risk factor can relieve parental self-blame and reduce the pres-
are more sensitive to change and that are informed by the experi- sure to pursue alternative treatments based on false promises to ad-
ences and voices of the autistic community.113 dress the causes of ASD. This scenario represents benefit to the patient
The DSM-5 defines ASD by behavioral symptoms and man- and his family right now, even while we wait for the development and
dates the diagnosis of separate disorders to indicate language im- implementation of precision medicines.
pairment or inattention. Unlike the DSM-5 approach, with well-
defined splitting between categories, we do not imagine that
independent biological pathways separately lead to ASD symp-
Conclusions
toms, language impairment, and inattention. Instead, it is likely that
neurodevelopmental consequences cascade from risk factors to mul- Autism spectrum disorder is not a singular disorder, but rather de-
tiple domains of difficulty, depending on the genetic or environ- scribes a common set of behavioral features that unite a very hetero-
mental insult and when it is incurred. Here, we have largely de- geneous population reflecting diverse genetic and environmental
scribed ASD as a disorder, but ASD is instead made up of many risk factors. Genetic testing, including a chromosomal microarray and
different disorders that have core ASD symptoms in common but fragile X testing, is recommended for all individuals with ASD. While
many disparate symptoms as well. we await precision medicine treatments based on neurobiology, cli-
Simultaneously,whilewethinkthatallpsychiatristsneedtobepre- nicians should use behavioral approaches for core symptoms and
pared for a future in which biological insights lead to transformative consider medications only for co-occurring symptoms in addition to
treatments, we must acknowledge that we do not yet live in that fu- an ongoing behavioral program.
ARTICLE INFORMATION 3. Kanner L. Autistic disturbances of affective stoppage from large UK ASD research family
Accepted for Publication: December 21, 2017. contact. Nerv Child. 1943;2:217-250. databases. Autism Res. 2015;8(1):73-81.
Published Online: March 28, 2018. 4. Bettelheim B. The Empty Fortress: Infantile 14. Gaugler T, Klei L, Sanders SJ, et al. Most genetic
doi:10.1001/jamapsychiatry.2017.4685 Autism and the Birth of the Self. New York, NY: risk for autism resides with common variation. Nat
The Free Press; 1967. Genet. 2014;46(8):881-885.
Author Contributions: Drs Muhle and Reed were
co-first authors and contributed equally to this 5. Folstein S, Rutter M. Infantile autism: a genetic 15. Klei L, Sanders SJ, Murtha MT, et al. Common
work. Drs Muhle and Reed had full access to all the study of 21 twin pairs. J Child Psychol Psychiatry. genetic variants, acting additively, are a major
data in the study and takes responsibility for the 1977;18(4):297-321. source of risk for autism. Mol Autism. 2012;3(1):9.
integrity of the data and the accuracy of the data 6. Bailey A, Le Couteur A, Gottesman I, et al. 16. Anney R, Klei L, Pinto D, et al. Individual
analysis. Autism as a strongly genetic disorder: evidence common variants exert weak effects on the risk for
Study concept and design: All authors. from a British twin study. Psychol Med. 1995; autism spectrum disorders. Hum Mol Genet. 2012;
Acquisition, analysis, or interpretation of data: 25(1):63-77. 21(21):4781-4792.
Muhle, Reed. 7. Ritvo ER, Freeman BJ, Mason-Brothers A, Mo A, 17. Betancur C. Etiological heterogeneity in autism
Drafting of the manuscript: All authors. Ritvo AM. Concordance for the syndrome of autism spectrum disorders: more than 100 genetic and
Critical revision of the manuscript for important in 40 pairs of afflicted twins. Am J Psychiatry. genomic disorders and still counting. Brain Res.
intellectual content: Muhle, Reed, Veenstra- 1985;142(1):74-77. 2011;1380:42-77.
VanderWeele.
Administrative, technical, or material support: 8. Sandin S, Lichtenstein P, Kuja-Halkola R, Larsson 18. Levy D, Ronemus M, Yamrom B, et al. Rare de
Muhle, Reed. H, Hultman CM, Reichenberg A. The familial risk of novo and transmitted copy-number variation in
Study supervision: Veenstra-VanderWeele. autism. JAMA. 2014;311(17):1770-1777. autistic spectrum disorders. Neuron. 2011;70(5):
9. Steffenburg S, Gillberg C, Hellgren L, et al. A twin 886-897.
Conflict of Interest Disclosures: Dr Veenstra-
VanderWeele reported serving on advisory boards study of autism in Denmark, Finland, Iceland, 19. Sanders SJ, Ercan-Sencicek AG, Hus V, et al.
or as a consultant for Roche, Novartis, and Norway and Sweden. J Child Psychol Psychiatry. Multiple recurrent de novo CNVs, including
SynapDx; receiving research funding from Roche, 1989;30(3):405-416. duplications of the 7q11.23 Williams syndrome
Novartis, Seaside Therapeutics, and Forest; and 10. Rosenberg RE, Law JK, Yenokyan G, McGready region, are strongly associated with autism. Neuron.
receiving editorial stipends from Springer and J, Kaufmann WE, Law PA. Characteristics and 2011;70(5):863-885.
Wiley. No other conflicts were reported. concordance of autism spectrum disorders among 20. Sanders SJ, He X, Willsey AJ, et al; Autism
277 twin pairs. Arch Pediatr Adolesc Med. 2009; Sequencing Consortium. Insights into autism
REFERENCES 163(10):907-914. spectrum disorder genomic architecture and
1. Muhle RA, Reed HE, Aguayo P, 11. Constantino JN, Todorov A, Hilton C, et al. biology from 71 risk loci. Neuron. 2015;87(6):
Veenstra-VanderWeele J. Evaluation and Autism recurrence in half siblings: strong support 1215-1233.
management of self-injury in a child with a rare for genetic mechanisms of transmission in ASD. Mol 21. Pinto D, Delaby E, Merico D, et al. Convergence
genomic variant [published online March 28, 2018]. Psychiatry. 2013;18(2):137-138. of genes and cellular pathways dysregulated in
JAMA Psychiatry. doi:10.1001/jamapsychiatry.2017 12. Werling DM, Geschwind DH. Recurrence rates autism spectrum disorders. Am J Hum Genet. 2014;
.4677 provide evidence for sex-differential, familial 94(5):677-694.
2. Taylor LE, Swerdfeger AL, Eslick GD. Vaccines are genetic liability for autism spectrum disorders in 22. Pinto D, Pagnamenta AT, Klei L, et al. Functional
not associated with autism: an evidence-based multiplex families and twins. Mol Autism. 2015;6:27. impact of global rare copy number variation in
meta-analysis of case-control and cohort studies. 13. Wood CL, Warnell F, Johnson M, et al. Evidence autism spectrum disorders. Nature. 2010;
Vaccine. 2014;32(29):3623-3629. for ASD recurrence rates and reproductive 466(7304):368-372.
23. Bernier R, Steinman KJ, Reilly B, et al; Simons individuals with features of autism and other 55. Parikshak NN, Luo R, Zhang A, et al. Integrative
VIP consortium. Clinical phenotype of the recurrent neuropsychiatric disorders. J Med Genet. 2009;46 functional genomic analyses implicate specific
1q21.1 copy-number variant. Genet Med. 2016;18 (4):242-248. molecular pathways and circuits in autism. Cell.
(4):341-349. 40. D’Angelo D, Lebon S, Chen Q, et al; Cardiff 2013;155(5):1008-1021.
24. Dolcetti A, Silversides CK, Marshall CR, et al. University Experiences of Children With Copy 56. Jeste SS, Geschwind DH. Disentangling the
1q21.1 Microduplication expression in adults. Genet Number Variants (ECHO) Study; 16p11.2 European heterogeneity of autism spectrum disorder through
Med. 2013;15(4):282-289. Consortium; Simons Variation in Individuals Project genetic findings. Nat Rev Neurol. 2014;10(2):74-81.
25. Dabell MP, Rosenfeld JA, Bader P, et al. (VIP) Consortium. Defining the effect of the 16p11.2 57. Tammimies K, Marshall CR, Walker S, et al.
Investigation of NRXN1 deletions: clinical and duplication on cognition, behavior, and medical Molecular diagnostic yield of chromosomal
molecular characterization. Am J Med Genet A. comorbidities. JAMA Psychiatry. 2016;73(1):20-30. microarray analysis and whole-exome sequencing
2013;161A(4):717-731. 41. Steinman KJ, Spence SJ, Ramocki MB, et al; in children with autism spectrum disorder. JAMA.
26. Rujescu D, Ingason A, Cichon S, et al; GROUP Simons VIP Consortium. 16p11.2 deletion and 2015;314(9):895-903.
Investigators. Disruption of the neurexin 1 gene is duplication: Characterizing neurologic phenotypes 58. Cooper GM, Coe BP, Girirajan S, et al. A copy
associated with schizophrenia. Hum Mol Genet. in a large clinically ascertained cohort. Am J Med number variation morbidity map of developmental
2009;18(5):988-996. Genet A. 2016;170(11):2943-2955. delay. Nat Genet. 2011;43(9):838-846.
27. Ching MS, Shen Y, Tan WH, et al; Children’s 42. Hanson E, Bernier R, Porche K, et al; Simons 59. Shinawi M, Liu P, Kang SH, et al. Recurrent
Hospital Boston Genotype Phenotype Study Group. Variation in Individuals Project Consortium. The reciprocal 16p11.2 rearrangements associated with
Deletions of NRXN1 (neurexin-1) predispose to a cognitive and behavioral phenotype of the 16p11.2 global developmental delay, behavioural problems,
wide spectrum of developmental disorders. Am J deletion in a clinically ascertained population. Biol dysmorphism, epilepsy, and abnormal head size.
Med Genet B Neuropsychiatr Genet. 2010;153B(4): Psychiatry. 2015;77(9):785-793. J Med Genet. 2010;47(5):332-341.
937-947. 43. McCarthy SE, Makarov V, Kirov G, et al; 60. Weiss LA, Shen Y, Korn JM, et al; Autism
28. Mulle JG, Gambello MJ, Cook EH, Rutkowski Wellcome Trust Case Control Consortium. Consortium. Association between microdeletion
TP, Glassford M. 3q29 Recurrent deletion. Microduplications of 16p11.2 are associated with and microduplication at 16p11.2 and autism. N Engl J
GeneReviews. https://www.ncbi.nlm.nih.gov/books schizophrenia. Nat Genet. 2009;41(11):1223-1227. Med. 2008;358(7):667-675.
/NBK385289/. Updated October 19, 2017. Accessed 44. McDonald-McGinn DM, Sullivan KE, Marino B, 61. Fernandez BA, Roberts W, Chung B, et al.
January 17, 2017. et al. 22q11.2 deletion syndrome. Nat Rev Dis Primers. Phenotypic spectrum associated with de novo and
29. Glassford MR, Rosenfeld JA, Freedman AA, 2015;1:15071. inherited deletions and duplications at 16p11.2 in
Zwick ME, Mulle JG; Unique Rare Chromosome 45. Firth HV. 22q11.2 Duplication. GeneReviews. individuals ascertained for diagnosis of autism
Disorder Support Group. Novel features of 3q29 https://www.ncbi.nlm.nih.gov/books/NBK3823/. spectrum disorder. J Med Genet. 2010;47(3):195-203.
deletion syndrome: Results from the 3q29 registry. Updated November 21, 2013. Accessed January 17, 62. Walters RG, Jacquemont S, Valsesia A, et al.
Am J Med Genet A. 2016;170A(4):999-1006. 2017. A new highly penetrant form of obesity due to
30. Mervis CB, Morris CA, Klein-Tasman BP, 46. Van Campenhout S, Devriendt K, Breckpot J, deletions on chromosome 16p11.2. Nature. 2010;
Velleman SL, Osborne LR. 7q11.23 Duplication et al. Microduplication 22q11.2: a description of the 463(7281):671-675.
syndrome. GeneReviews. https://www.ncbi.nlm.nih clinical, developmental and behavioral 63. Zufferey F, Sherr EH, Beckmann ND, et al;
.gov/books/NBK327268/. Published November 25, characteristics during childhood. Genet Couns. Simons VIP Consortium; 16p11.2 European
2015. Accessed January 17, 2017. 2012;23(2):135-148. Consortium. A 600 kb deletion syndrome at 16p11.2
31. Van der Aa N, Rooms L, Vandeweyer G, et al. 47. Portnoï MF. Microduplication 22q11.2: a new leads to energy imbalance and neuropsychiatric
Fourteen new cases contribute to the chromosomal syndrome. Eur J Med Genet. 2009; disorders. J Med Genet. 2012;49(10):660-668.
characterization of the 7q11.23 microduplication 52(2-3):88-93. 64. Jacquemont S, Reymond A, Zufferey F, et al.
syndrome. Eur J Med Genet. 2009;52(2-3):94-100. 48. Soorya L, Kolevzon A, Zweifach J, et al. Mirror extreme BMI phenotypes associated with
32. Morris CA, Mervis CB, Paciorkowski AP, et al. Prospective investigation of autism and gene dosage at the chromosome 16p11.2 locus.
7q11.23 Duplication syndrome: Physical genotype-phenotype correlations in 22q13 deletion Nature. 2011;478(7367):97-102.
characteristics and natural history. Am J Med Genet syndrome and SHANK3 deficiency. Mol Autism. 65. Stefansson H, Meyer-Lindenberg A, Steinberg
A. 2015;167A(12):2916-2935. 2013;4(1):18. S, et al. CNVs conferring risk of autism or
33. Vorstman JAS, Parr JR, Moreno-De-Luca D, 49. Phelan K, McDermid HE. The 22q13.3 deletion schizophrenia affect cognition in controls. Nature.
Anney RJL, Nurnberger JI Jr, Hallmayer JF. Autism syndrome (Phelan-McDermid syndrome). Mol 2014;505(7483):361-366.
genetics: opportunities and challenges for clinical Syndromol. 2012;2(3-5):186-201. 66. Moreno-De-Luca A, Evans DW, Boomer KB,
translation. Nat Rev Genet. 2017;18(6):362-376. 50. Neale BM, Kou Y, Liu L, et al. Patterns and rates et al. The role of parental cognitive, behavioral, and
34. Battaglia A. The inv dup (15) or idic (15) of exonic de novo mutations in autism spectrum motor profiles in clinical variability in individuals
syndrome (Tetrasomy 15q). Orphanet J Rare Dis. disorders. Nature. 2012;485(7397):242-245. with chromosome 16p11.2 deletions. JAMA Psychiatry.
2008;3:30. 51. O’Roak BJ, Vives L, Girirajan S, et al. Sporadic 2015;72(2):119-126.
35. Thomas JA, Johnson J, Peterson Kraai TL, et al. autism exomes reveal a highly interconnected 67. Girirajan S, Rosenfeld JA, Coe BP, et al.
Genetic and clinical characterization of patients protein network of de novo mutations. Nature. Phenotypic heterogeneity of genomic disorders
with an interstitial duplication 15q11-q13, 2012;485(7397):246-250. and rare copy-number variants. N Engl J Med. 2012;
emphasizing behavioral phenotype and response to 52. Sanders SJ, Murtha MT, Gupta AR, et al. 367(14):1321-1331.
treatment. Am J Med Genet A. 2003;119A(2):111-120. De novo mutations revealed by whole-exome 68. Hultman CM, Sandin S, Levine SZ, Lichtenstein
36. Berg JS, Potocki L, Bacino CA. Common sequencing are strongly associated with autism. P, Reichenberg A. Advancing paternal age and risk
recurrent microduplication syndromes: diagnosis Nature. 2012;485(7397):237-241. of autism: new evidence from a population-based
and management in clinical practice. Am J Med 53. De Rubeis S, He X, Goldberg AP, et al; DDD study and a meta-analysis of epidemiological
Genet A. 2010;152A(5):1066-1078. Study; Homozygosity Mapping Collaborative for studies. Mol Psychiatry. 2011;16(12):1203-1212.
37. Battaglia A. The inv dup(15) or idic(15) Autism; UK10K Consortium. Synaptic, 69. Sandin S, Hultman CM, Kolevzon A, Gross R,
syndrome: a clinically recognisable neurogenetic transcriptional and chromatin genes disrupted in MacCabe JH, Reichenberg A. Advancing maternal
disorder. Brain Dev. 2005;27(5):365-369. autism. Nature. 2014;515(7526):209-215. age is associated with increasing risk for autism:
38. Kalsner L, Chamberlain SJ. Prader-Willi, 54. Willsey AJ, Sanders SJ, Li M, et al. Coexpression a review and meta-analysis. J Am Acad Child
Angelman, and 15q11-q13 duplication syndromes. networks implicate human midfetal deep cortical Adolesc Psychiatry. 2012;51(5):477-486.e1.
Pediatr Clin North Am. 2015;62(3):587-606. projection neurons in the pathogenesis of autism. 70. Lyall K, Croen L, Daniels J, et al. The changing
39. Miller DT, Shen Y, Weiss LA, et al. Cell. 2013;155(5):997-1007. epidemiology of autism spectrum disorders. Annu
Microdeletion/duplication at 15q13.2q13.3 among Rev Public Health. 2017;38(1):81-102.
71. Modabbernia A, Velthorst E, Reichenberg A. 86. Redcay E, Courchesne E. When is the brain phenotypes in mouse models of 16p11.2 deletion
Environmental risk factors for autism: an enlarged in autism? a meta-analysis of all brain size and duplication syndromes. PLoS Genet. 2016;12(2):
evidence-based review of systematic reviews and reports. Biol Psychiatry. 2005;58(1):1-9. e1005709.
meta-analyses. Mol Autism. 2017;8:13. 87. Di Martino A, Yan C-G, Li Q, et al. The autism 101. Michalon A, Sidorov M, Ballard TM, et al.
72. Kong SW, Lee IH, Leshchiner I, et al; MedSeq brain imaging data exchange: towards a large-scale Chronic pharmacological mGlu5 inhibition corrects
Project. Summarizing polygenic risks for complex evaluation of the intrinsic brain architecture in fragile X in adult mice. Neuron. 2012;74(1):49-56.
diseases in a clinical whole-genome report. Genet autism. Mol Psychiatry. 2014;19(6):659-667. 102. Stoppel LJ, Kazdoba TM, Schaffler MD, et al.
Med. 2015;17(7):536-544. 88. Qureshi AY, Mueller S, Snyder AZ, et al; Simons R-Baclofen reverses cognitive deficits and improves
73. Guy A, Seaton SE, Boyle EM, et al. Infants born VIP Consortium. Opposing brain differences in social interactions in two lines of 16p11.2 deletion
late/moderately preterm are at increased risk for a 16p11.2 deletion and duplication carriers. J Neurosci. mice. Neuropsychopharmacology. 2018;43(3):
positive autism screen at 2 years of age. J Pediatr. 2014;34(34):11199-11211. 513-524.
2015;166(2):269-275.e3. 89. Maillard AM, Ruef A, Pizzagalli F, et al; 16p11.2 103. Blumenthal I, Ragavendran A, Erdin S, et al.
74. Padilla N, Eklöf E, Mårtensson GE, Bölte S, European Consortium. The 16p11.2 locus modulates Transcriptional consequences of 16p11.2 deletion
Lagercrantz H, Ådén U. Poor brain growth in brain structures common to autism, schizophrenia and duplication in mouse cortex and multiplex
extremely preterm neonates long before the onset and obesity. Mol Psychiatry. 2015;20(1):140-147. autism families. Am J Hum Genet. 2014;94(6):
of autism spectrum disorder symptoms. Cereb Cortex. 90. Johnson CP, Myers SM; American Academy of 870-883.
2017;27(2):1245-1252. Pediatrics Council on Children With Disabilities. 104. Migliavacca E, Golzio C, Männik K, et al;
75. Movsas TZ, Pinto-Martin JA, Whitaker AH, et al. Identification and evaluation of children with 16p11.2 European Consortium. A Potential
Autism spectrum disorder is associated with autism spectrum disorders. Pediatrics. 2007;120(5): contributory role for ciliary dysfunction in the
ventricular enlargement in a low birth weight 1183-1215. 16p11.2 600 kb BP4-BP5 pathology. Am J Hum Genet.
population. J Pediatr. 2013;163(1):73-78. 91. Filipek PA, Accardo PJ, Ashwal S, et al. Practice 2015;96(5):784-796.
76. Mandy W, Lai M-C. Annual research review: the parameter: screening and diagnosis of autism: 105. Paşca AM, Sloan SA, Clarke LE, et al.
role of the environment in the developmental report of the Quality Standards Subcommittee of Functional cortical neurons and astrocytes from
psychopathology of autism spectrum condition. the American Academy of Neurology and the Child human pluripotent stem cells in 3D culture. Nat
J Child Psychol Psychiatry. 2016;57(3):271-292. Neurology Society. Neurology. 2000;55(4):468-479. Methods. 2015;12(7):671-678.
77. Christensen J, Grønborg TK, Sørensen MJ, et al. 92. Miller DT, Chung W, Nasir R, et al. 16p11.2 106. Paşca SP, Portmann T, Voineagu I, et al. Using
Prenatal valproate exposure and risk of autism Recurrent microdeletion. GeneReviews. iPSC-derived neurons to uncover cellular
spectrum disorders and childhood autism. JAMA. https://www.ncbi.nlm.nih.gov/books/NBK11167/. phenotypes associated with Timothy syndrome.
2013;309(16):1696-1703. Updated December 10, 2015. Accessed January 17, Nat Med. 2011;17(12):1657-1662.
78. Emerson RW, Adams C, Nishino T, et al; IBIS 2017. 107. Weitlauf AS, McPheeters ML, Peters B, et al.
Network. Functional neuroimaging of high-risk 93. Jones W, Klin A. Attention to eyes is present Therapies for children with autism spectrum
6-month-old infants predicts a diagnosis of autism but in decline in 2-6-month-old infants later disorder: behavioral interventions update:
at 24 months of age. Sci Transl Med. 2017;9(393): diagnosed with autism. Nature. 2013;504(7480): comparative effectiveness review no. 137.
eaag2882. 427-431. (Prepared by the Vanderbilt Evidence-based
79. Wolff JJ, Jacob S, Elison JT. The journey to 94. Golzio C, Willer J, Talkowski ME, et al. KCTD13 Practice Center under Contract No.
autism: insights from neuroimaging studies of is a major driver of mirrored neuroanatomical 290-2012-00009-I.) AHRQ Publication No.
infants and toddlers. Dev Psychopathol. 2017:1-17. phenotypes of the 16p11.2 copy number variant. 14-EHC036-EF. Rockville, MD: Agency for
Nature. 2012;485(7398):363-367. Healthcare Research and Quality; 2014.
80. Wolff JJ, Gu H, Gerig G, et al; IBIS Network.
Differences in white matter fiber tract development 95. Chen Y, Yu J, Niu Y, et al. Modeling Rett 108. Fung LK, Mahajan R, Nozzolillo A, et al.
present from 6 to 24 months in infants with autism. syndrome using TALEN-edited MECP2 mutant Pharmacologic treatment of severe irritability and
Am J Psychiatry. 2012;169(6):589-600. cynomolgus monkeys. Cell. 2017;169(5):945-955.e10. problem behaviors in autism: a systematic review
and meta-analysis. Pediatrics. 2016;137(S2)
81. Lewis JD, Evans AC, Pruett JR Jr, et al; Infant 96. Portmann T, Yang M, Mao R, et al. Behavioral (suppl 2):S124-S135.
Brain Imaging Study Network. The emergence of abnormalities and circuit defects in the basal
network inefficiencies in infants with autism ganglia of a mouse model of 16p11.2 deletion 109. Levy SE, Hyman SL. Complementary and
spectrum disorder. Biol Psychiatry. 2017;82(3): syndrome. Cell Rep. 2014;7(4):1077-1092. alternative medicine treatments for children with
176-185. autism spectrum disorders. Child Adolesc Psychiatr
97. Pucilowska J, Vithayathil J, Tavares EJ, Kelly C, Clin N Am. 2015;24(1):117-143.
82. Hazlett HC, Gu H, Munsell BC, et al; IBIS Karlo JC, Landreth GE. The 16p11.2 deletion mouse
Network; Clinical Sites; Data Coordinating Center; model of autism exhibits altered cortical progenitor 110. Hsu PD, Lander ES, Zhang F. Development and
Image Processing Core; Statistical Analysis. Early proliferation and brain cytoarchitecture linked to applications of CRISPR-Cas9 for genome
brain development in infants at high risk for autism the ERK MAPK pathway. J Neurosci. 2015;35(7): engineering. Cell. 2014;157(6):1262-1278.
spectrum disorder. Nature. 2017;542(7641):348-351. 3190-3200. 111. Peng R, Lin G, Li J. Potential pitfalls of
83. Shen MD, Nordahl CW, Young GS, et al. Early 98. Tian D, Stoppel LJ, Heynen AJ, et al. CRISPR/Cas9-mediated genome editing. FEBS J.
brain enlargement and elevated extra-axial fluid in Contribution of mGluR5 to pathophysiology in a 2016;283(7):1218-1231.
infants who develop autism spectrum disorder. Brain. mouse model of human chromosome 16p11.2 112. Erickson CA, Davenport MH, Schaefer TL, et al.
2013;136(pt 9):2825-2835. microdeletion. Nat Neurosci. 2015;18(2):182-184. Fragile X targeted pharmacotherapy: lessons
84. Tau GZ, Peterson BS. Normal development of 99. Horev G, Ellegood J, Lerch JP, et al. learned and future directions. J Neurodev Disord.
brain circuits. Neuropsychopharmacology. 2010; Dosage-dependent phenotypes in models of 2017;9:7.
35(1):147-168. 16p11.2 lesions found in autism. Proc Natl Acad Sci 113. Budimirovic DB, Berry-Kravis E, Erickson CA,
85. Courchesne E, Campbell K, Solso S. Brain U S A. 2011;108(41):17076-17081. et al. Updated report on tools to measure outcomes
growth across the life span in autism: age-specific 100. Arbogast T, Ouagazzal AM, Chevalier C, et al. of clinical trials in fragile X syndrome. J Neurodev
changes in anatomical pathology. Brain Res. 2011; Reciprocal effects on neurocognitive and metabolic Disord. 2017;9:14.
1380:138-145.
E10 JAMA Psychiatry Published online March 28, 2018 (Reprinted) jamapsychiatry.com