Billet publié sur Twitter/X le 28/10/2021 12:53

Résumé

Originally described by both Leo Kanner and Hans Asperger in the early 1940s, autism is a behaviourally-defined condition, characterized by social-communication difficulties and restricted repetitive behaviours and interests. Until today, the description of the condition retains much of the essence of how it was defined early on. However, in other ways, we are somewhat at sea compared with earlier times in terms of understanding autism. For example, we now know that autism rarely occurs on its own: the vast majority of individuals with autism show other conditions too, such as other neurodevelopmental or psychiatric conditions. For many forms of autism, the condition exists on an etiological continuum with typical variation in autistic traits within the general population. Finally, the heterogeneity within autism itself may be the greatest bottleneck to progress. In the 1960s, the possibility of genetic influences on autism was generally dismissed. A damaging belief was that autism was caused by a cold parenting style. However, the recognition that sibling rates of autism were 50- to 100-fold increased when compared to the 2–4/10,000 population base rate of the time, prompted a first small twin study (21 same-sex twin pairs) around the same time (Folstein & Rutter, 1977). The concordance rates observed were markedly different for monozygotic (MZ, 4 out of 11; 36%) and dizygotic (DZ, 0 out of 10; 0%) pairs. Moreover, the study showed that what was inherited was not only severe autism but also a liability to a broader set of developmental differences involving social and communication development alongside atypical patterns of interests, now referred to as the ‘broader autism phenotype’. Like that for narrowly defined autism, concordance for this broader autism phenotype was much higher in MZ vs DZ pairs (82% vs 10%), indicating strong genetic influence on ASD. These findings translated into heritability estimates above 90%, making autism one of the most heritable behaviourally-defined conditions. High heritability estimates were subsequently observed in multiple large twin and family studies. With evidence for genetic factors in autism (spectrum disorder; ASD) accumulating, and following the technological progress in the field of molecular genetics, first molecular genetic studies of ASD were performed. Initial investigations relied on genetic linkage analysis of affected-relative pairs and multiplex family designs. The first genetic locus for ASD from such studies was published in 1998 (International Molecular Genetic Study of Autism Consortium, 1998). Following the advances of the Human Genome Project and the development of large-scale sequencing and genome-wide association study technology, molecular studies led to the identification of large, rare copy number variants as well as de novo mutations linked to ASD. Especially with the advent of genome-wide association studies (GWAS), the ASD genetic field diverged into groups studying rare genetic variants of high penetrance and groups investigating the role of common genetic variants of small effect. Gaugler et al. (2014) estimated that the common variants account for approximately 50% of the liability to ASD, while de novo and inherited rare variants might explain around 10%. The first GWAS of single nucleotide polymorphisms (SNPs) in ASD were published in 2009, but sample sizes were too limited to yield reproducible findings. Research in this field has accelerated enormously in the last two years. The most recent GWAS by the Psychiatric Genomics Consortium and iPSYCH included over 18,000 individuals with ASD and nearly 28,000 controls and reported five significant loci as well as a SNP-heritability estimate of 12% (Grove et al., 2019). The most recent whole exome sequencing study of rare variants in nearly 12,000 individuals with ASD and over 23,000 controls identified 102 ASD-linked genes (Satterstrom et al., 2020). Beyond increasing our understanding of the genetic architecture of ASD, findings from these recent studies now also allow first insights into the molecular mechanisms contributing to the aetiology of ASD. JCPP chose genetics of ASD as the subject for its Annual Special Issue 2021 to signify that we are reaching an exciting new era in genetic research on this condition - and the publications in our issue very clearly demonstrate this. JCPP chose genetics of ASD as the subject for its Annual Special Issue 2021 to signify that we are reaching an exciting new era in genetic research on this condition – and the publications in our issue very clearly demonstrate this. Team science and big data availability are transforming the field, polygenic scores are finding a place in epidemiological research, developmental research on ASD is beginning to embrace genetics within its designs, and the availability of datasets with larger numbers of patients with specific monogenic causes of ASD now enhances research into potential (endo)phenotypes. In this Special Issue, it is gratifying to see authors applying clever strategies to operationalize genetics, including family studies in large population registries, studies of (presumably) monogenic forms of ASD and studies of polygenic scores. In this way, the new genetics of ASD is enabling the field to tackle important unanswered questions about this highly heterogeneous condition. Several contributions to this Special Issue use genetics to inform their studies of ASD comorbidity with other neurodevelopmental conditions: The family study of nearly 2.4 million people from the Swedish National Patient Register by Taylor and co-workers (2021) considerably strengthens our understanding of the associations between ASD and other neurodevelopmental and psychiatric conditions. They report significant genetic correlations between ASD and a wide range of both neurodevelopmental and psychiatric conditions. Of particular note, though, was the finding that familial associations with other neurodevelopmental and psychiatric conditions were less marked for subtypes of ASD that co-occurred with either intellectual disability, epilepsy, chromosomal abnormalities or congenital malformations. Well-powered, robust findings on the inheritance patterns of ASD subtypes such as those by Taylor et al. will help remove the bottleneck in progress in research caused by heterogeneity in ASD. Peyre and colleagues (2021) used multivariate genomic analyses to better understand the mechanisms underlying the comorbidity of ASD and ADHD. Using data from the currently largest published GWAS for each of the conditions, they found that a substantial proportion of ASD- and ADHD-associated genomic loci co-localize; their study is also the first to identify loci and genes specific to each of the individual conditions. Using Mendelian randomization analyses, they found bi-directional effects, with each condition increasing the risk for the other conditions. A second set of contributions includes genetics in analyses of developmental trajectories predictive of ASD-relevant outcomes. Thurm et al. (2021) focused on individuals from the large SPARK cohort with well-characterised aetiological genetic conditions associated with ASD and other neurodevelopmental conditions and compared this group to idiopathic ASD in terms of their associations with motor and language acquisition milestones. In comparison with idiopathic ASD, children with the genetically-defined conditions were more likely to be delayed in walking and to show elevated rates of delay in other motor and language milestones. The authors concluded that milestone delays, particularly within the motor domain, may be an important early flag that differentiates ASD associated with a known genetic condition from other subtypes of ASD. This may signal another step in our ability to differentiate subtypes of ASD early on in development and thus tailor clinical support accordingly. In the longitudinal, prospective study presented by Fish and co-workers (2021), eye-tracking and the development of the pupillary light reflex was investigated as an early predictor of ASD-related behaviour. The authors included in this work a polygenic score for ASD. Specific features of the development of the pupillary light reflex, especially latency and amplitude between 9 and 24 months, were shown to correlate with an ASD diagnosis at age 3 years and/or with dimensional traits of social affect and repetitive/restrictive behaviour. Moreover, the ASD polygenic score was associated with altered light reflex development. Boivin and coauthors (2021) set out to identify high-risk trajectories for social wariness and preference for solitude during childhood and to examine whether the assignment to these high-risk trajectories was associated with polygenic scores for ASD and other mental health conditions and related traits. Using data from two prospective longitudinal studies, the Quebec Newborn Twin Study and the Quebec Longitudinal Study of Child Development, the authors found that a polygenic score for loneliness predicted an increasing and elevated trajectory for social wariness, whereas a general mental health polygenic risk score derived from data on different conditions and traits was significantly associated with the persistence of a preference for solitude. Adding to the work by Fish and co-workers (2021) on the development of the pupillary light reflex, innovative studies have also incorporated genetics to investigate intermediate phenotypes for ASD and autistic traits. Reed and colleagues’ work on emotion recognition moves us a step closer to understanding how emotion recognition connects with autistic traits and symptoms. First, they observed phenotypic associations between social autistic traits and lower scores on emotion recognition tests assessed both in childhood and in adulthood. Then, they employed an ASD polygenic score analysis and linkage disequilibrium (LD) score regression analysis to investigate genetic associations between ASD and emotion recognition ability. The lack of a significant genetic association between ASD polygenic score and emotion recognition in these analyses is of interest; as the authors note, this initial negative finding should be followed up when a larger GWAS of ASD is published. In a separate, carefully executed evaluation of brain structure as a candidate intermediate phenotype for ASD, Frewer and colleagues (2021) performed a systematic review of the existing literature on brain magnetic resonance imaging (MRI) studies of monogenic conditions associated with ASD. These authors identified neuroimaging studies carried out on samples of participants who had mutations in 13 of 20 preselected ASD genes. While a wide range of different brain features were found to be affected, white matter alterations, suggestive of altered brain connectivity, were the most consistent finding. The authors caution wisely, however, that more research is needed before firm conclusions can be drawn, as most source studies described small case series or cohorts. While ASD heritability is high, environmental factors also play important roles in the aetiology and severity of the condition. Gomez Vallejo and colleagues (2021) used a genetically-sensitive (twin) design to study the role of obstetric optimality as a potential risk factor. In their population-based twin sample, they employed an obstetric optimality sum-score based on the presence/absence of pre-, peri- and neonatal factors. While the children with ASD and their unaffected cotwins showed more obstetric complications than controls, importantly the authors did not find differences between children with ASD and their unaffected cotwins. A subsequent review of the literature and meta-analysis of existing studies on individual obstetric factors confirmed findings on the association of obstetric factors with ASD and identified a number of the most relevant factors. The authors conclude that shared familial liability explains associations between obstetric complications and ASD rather than a causal association. The present Special Issue of JCPP provides a cross section of the many ways the new genetics of ASD can enhance our insight into mechanisms and aetiology, early predictors, as well as heterogeneity and comorbidity. Societal impact of ASD-related genetic research is already seen in the support to parents and patients through genetic counselling services, through early intervention services and targeted therapeutic drug development. The field of ASD research is also expanding in the search for brain biomarkers of ASD, identification of early trajectories and development of pre-emptive interventions. Such progress has raised a number of ethical issues that have been debated in both the professional and lay literature (e.g. see Hens, Peeters, & Dierickx, 2016). This Special Issue therefore concludes with a thoughtful examination of the new ethical challenges that are posed by early neurodevelopmental research. Manzini et al. (2021) examine how new models of understanding ASD, the probabilistic nature of diagnostic prediction combined with the risk and benefits of early identification and intervention require a new ethical scrutiny that ought to draw on effective partnership between researchers and people with ASD and their families and to accommodate the views and diversity of the ASD community. We would like to conclude by saying that we enjoyed our task as Editors for this Special Issue very much! We thank all authors for their excellent contributions, and we wish you – the reader – lots of new insights while reading their interesting and far-reaching work. Barbara Franke has received educational speaking fees from Medice.