Autism: from phenotype to genotype and back

ASDs are debilitating mental disorders2, characterized by defects in
communication and social interaction. Together with rigidity and repetitive
behaviours, these symptoms often remain manifest throughout the entire life of
affected individuals3. ASD includes autism, Asperger syndrome and Pervasive
Developmental Disorder, not otherwise specified (PDD-NOS). The spectrum of
clinical impairment is broad, ranging from patients with mild symptoms and
relatively good language skills, a.k.a. Asperger syndrome, to patients with
severe symptoms often accompanied by low IQ and epilepsy. Our current
understanding of the neurobiological abnormalities underlying ASD is limited;
partly as a consequence, there are at present no adequate treatments available
that can restore or ameliorate the core deficits that characterize this
disorder.

In recent years, whole genome sequencing studies have revealed that each human
individual carries a high number of genetic variants that are very rare (minor
allele frequency <0.1%4) or virtually unique and that a substantial fraction of
these variants has a relevant impact on protein function5.
Even though the heritability of autism is very high, i.e. 90%12, the three
largest published Genome Wide Association Studies (GWAS) so far have failed to
provide convincing evidence for the role of common genetic variants in ASD13
(although the role of common variants cannot be ruled out). In contrast, an
increasing number of rare variants are discovered in autism14,15,16. These rare
variants have modest to relatively large effect sizes and typically occur at
rates of 0.3 * 0.6% in the autistic population17. Importantly, when taken
together, the risk variants known to date are estimated to explain
approximately 5 * 15% of all ASD cases18.


In clinical practice three important problems stand out, against which the
relevance of this proposal can be expressed:

First, an etiologic diagnosis can only be provided to a minority of ASD
patients. An explanatory diagnosis is, amongst others, highly relevant with
respect to prognosis, acceptance and life planning19,20. The current study
aims to identify genetic variants associated with increased ASD risk.
Identifying novel genetic risk variants for ASD will enhance the yield of
clinical genetic testing, providing an etiological diagnosis for an increased
number of ASD patients.
Second, the essential reason for the current lack of effective treatments for
ASD is that the underlying biological processes are not understood. The
discovery of ASD risk genes offers a unique *bottom up* approach into its
biological underpinnings. Once identified, the precise functions of the encoded
proteins can be further studied. This information is crucial to further our
biological understanding of ASD and is needed to open new venues for the
development of treatment.
Third, the heterogeneity of research findings indicates that ASD is not the
consequence of one unifying etiology, but separable into different etiological
subgroups21. Indeed, a pattern starts to emerge, indicating that autism risk
genes may converge into a limited number of different neurobiological
pathways22. However, the number of different risk-contributing genetic variants
is likely to be high (~200-300)23,24,25, with only a small proportion of
alleles identified so far. The lack of insight into the different ASD-subtypes
is currently one of the greatest obstacles towards the development of
etiology-specific treatments26. Etiological subtypes of bleeding disorders are
defined based on genetic causes and require type-specific treatments.
Similarly, it is likely that genetically defined subtypes exist in ASD27 which,
consequently may benefit from subtype-specific treatments. The identification
of causative genetic variants, the aim of this proposal, is a necessary step
towards the definition of such genetic ASD-subtypes28 which in turn will allow
the development of tailored treatments.

In conclusion, ASD is a group of severely handicapping mental disorders. Our
current insight into the biology of autism is limited, which is a major
obstacle towards the development of novel, more efficient treatment strategies.
Recently, an increasing number of rare causal genetic variants for ASD have
been discovered. Although this represents a true breakthrough in our
understanding of autism, these findings still represent only a small proportion
of all genetic variants that are thought to contribute to ASD risk in the
population. The current proposal aims to identify novel genetic ASD risk
variants (first stage, core cohort study) and in the second stage studies,
confirm statistical association and elucidate the associated ASD phenotype via
the extended family and deep phenotyping studies respectively.

Hypotheses:
1. Novel, thus far unidentified genetic variants that are rare on the
population level can contribute substantially to the vulnerability of ASD on an
individual level. Such variants can be either de novo or inherited.
a. In case of a de novo variant a dominant model for genetic impact is
relevant.
b. In case of inherited variants a recessive model of impact and / or
interaction model involving the effect of genetic modifiers elsewhere on the
genome can both be relevant.
2. Specific genetic variants may be associated with specific neuropsychiatric
ASD phenotypical profiles.

Given that the essence of this study is the comparison of genetic variants
between affected and unaffected individuals, the participation of ASD patients
and unaffected subjects is required. Since the genetic variants that are the
focus of this study are rare, or in some cases may even be unique, the analysis
of family members is needed in order to establish patterns of segregation.
Hence, this study can only be carried out with the participation of autistic
subjects and their family members.

1. To identify genetic variants associated with an increased risk of ASD using
different genetic models (dominant, recessive, interaction).

2. To determine the relationship between different genetic variants and the
associated neuropsychiatric ASD phenotype.

This study is a non-intervention study conducted in two stages. The setting of
the study is the department of psychiatry at the UMC Utrecht. There is a
regular flow of children who are referred to this clinic because of behavioral
problems and a suspicion of the possibility of ASD. For these children a
regular assessment procedure is in place which includes the use of standardized
measures and the withdrawal of blood for clinically motivated genetic analyses
including the standard screening for Fragile X.
The duration of this study is estimated to be eight years; given the rate of
new ASD patients in our department we expect to have recruited 40% of the core
cohort in three years, this will provide sufficient data to start selection for
the stage 2 studies, while continuing the completion of the stage 1 core cohort.


The first stage consists of the recruitment of a large sample (n=1000) of ASD
probands and their parents through the clinic (psychiatry clinic and outpatient
clinic at the University Medical Center Utrecht). With regard to the assessment
(phenotyping), participation at this stage does not imply extra tests or
interviews in addition to what is required for normal good clinical practice
(i.e. probands who do not wish to participate in the study will undergo the
same assessments). Similarly, laboratory analyses using patient blood are part
of the standard clinical assessment in patients assessed for ASD in the
Netherlands29. For those patients who agree to participate in the study the
blood draw for clinical purposes is taken advantage of by drawing two extra
tubes of blood for the purpose of the study. At stage 1 the parents will be
asked to provide two tubes of blood for the purpose of the study.

All participants and their parents will be explicitly informed at the start
that, provided their consent, they may be re-contacted for stage 2:
- Based on family structure (i.e. patients who have two or more second degree
relatives affected by ASD) and/or severity of the ASD in the proband,
approximately 150 families will be contacted and asked to participate in the
stage 2 *family subsamples a,b,c*.
- Based on the genetic findings approximately 200 patients and their parents
will be asked to participate in a more thorough phenotypic assessment in the
stage 2 follow-up *deep phenotyping subsample*. Please note that for all
participants feed-back on the findings of the genetic study is provided, since
it is part of the standard clinical assessment.

Risk of participation in the core cohort (stage 1) study is negligible; the
burden is minimal for participating probands, for parents small. Participation
in one of the stage 2 studies does not increase any risk but requires
additional efforts; however taken together the burden remains relatively small.

Total research burden is dependent on study participation; participation in the
core cohort study alone represents a relatively small burden; however the
burden increases when selected to participate in any of the stage 2 studies,
and varies with the specific stage 2 study. In the protocol, on pages 22, 28
and 32 - 40, the research burden is summarized and detailed for each study
component.