Autism Spectrum Disorder from the Womb to Adulthood: Suggestions for a Paradigm Shift

Abstract
...
The epidemiological and clinical findings in ASD cannot be explained by the traditional linear genetic model, hence the need to move towards a more fluid conception, integrating genetics, environment, and epigenetics as a whole. The embryo-fetal period and the first two years of life (the so-called ‘First 1000 Days’) are the crucial time window for neurodevelopment. In particular, the interplay and the vicious loop between immune activation, gut dysbiosis, and mitochondrial impairment/oxidative stress significantly affects neurodevelopment during pregnancy and undermines the health of ASD people throughout life. Consequently, the most effective intervention in ASD is expected by primary prevention aimed at pregnancy and at early control of the main effector molecular pathways. We will reason here on a comprehensive and exhaustive pathogenetic paradigm in ASD, viewed not just as a theoretical issue, but as a tool to provide suggestions for effective preventive strategies and personalized, dynamic (from womb to adulthood), systemic, and interdisciplinary healthcare approach.
...

Introduction
...
Fetal neural programming, occurring during the ontogenesis, and early live neuroplasticity are crucial events in neurodevelopment and identify the time window of maximum brain opportunity in the embryo-fetal period and in the first two years of life (the so-called ‘First 1000 Days’) [4].
...
In this period, exogenous insults and changes in the maternal milieu are expected to have the maximum disturbing effect and lifelong consequences on health. The dynamic molecular machinery involved in the ontogenesis transforms early life inputs into long-term programmatic outcomes, influencing the enzymatic and immuno-neuroendocrine pathways, which define the basis of the homeostasis during intrauterine and postnatal life [5].

Several biological abnormalities are involved in the etiopathogenesis of ASD. Our study reviews the main topics addressed in biological research—genetics, epigenetics, environmental issues, immunogenetics, immunology, microbiology, and metabolic and electrophysiological impairments—with particular reference to their synergistic interactions and their links with the clinical phenotype. Therefore, a dynamic (from the womb to adulthood) perspective and the concept of multisystem disorder seems to be the most plausible framework for the study of neurodevelopmental disorders.
...
ASD is associated with abnormal neural connectivity [257,258,259,260,261,262], and some abnormalities in brain development might be already detectable at birth.

The EEG can measure neural activity and may provide a useful tool to early detect ASD in children, thus allowing the opportunity for early intervention. The potential usefulness of EEG in ASD has been reviewed almost ten years ago [259], in order to examine evidence for the utility of three methods of EEG signal analysis in the ASD diagnosis and subtype delineation. All studies identified significant differences between ASD and non-ASD subjects, confirming the presence of specific EEG abnormalities. However, due to the high heterogeneity in the results, findings could not be generalized and none of the methods, if taken alone, has been proposed as a new diagnostic tool [259].

Recent studies on this topic open new avenues and might represent a turning point for the early diagnosis of ASD based on the analysis of electroencephalographic tracing (EEG) supported by new adaptive artificial systems (ANNs). It was hypothesized that the atypical organization of the cerebral cortex in ASD might translate into an EEG signature detectable through powerful analytical systems such as ANNs [265,266,267].

Using particularly advanced machine learning systems, it has been possible to build a software able to distinguish almost perfectly the EEG from subjects with ASD from those of neurotypical controls or with different neuropsychiatric disorders.

The new system, called MS-ROM/I-FAST, belongs to the family of systems developed by the Semeion Research Centre. MS-ROM/I-FAST is a new and complex algorithm for the blind classification of the original EEG trace of each subject, through the recording and analysis of a few minutes of their EEG without any preliminary pre-processing [266]. A first pilot study assessed the discriminatory power of the methodology in distinguishing subjects with ASD from neurotypical controls. After the MS-ROM/I-FAST pre-processing, the overall predictive capacity of the different automatic learning systems in distinguishing autistic cases from the controls was constantly 100% [266]. Notably, these results were obtained at different times and in separate experiments performed on the same training and testing subsets. Furthermore, the similarities between the weight matrices of the neural networks measured with appropriate algorithms were not influenced by the age of the subjects, suggesting that the networks read invariant characteristics related to the disconnection signature in the brain [266]. The results of the pilot study have been recently confirmed. EEG data from ASD children were compared with EEG from controls affected by other neuropsychiatric disorders. With the training-testing protocol, the overall predictive capacity of the machine learning system used to distinguish between ASD and controls was constantly over 90% [267]. Along this research area, it would be of utmost interest to extend EEG tracks recording within the first year of life, with the purpose to use this technique as a specific, sensitive, non-invasive, non-expensive tool for early detection of the signature predictive for ASD. The potential usefulness of this methodology might be extended to find out possible different EEG signature in ASD subgroups with different onset (early/regressive autism) and different phenotypes. Furthermore, this tool could monitor the evolution of EEG abnormalities, find hidden links with clinical and laboratory biomarkers and monitor the effect of therapeutic interventions.
...

Discussion
...
Embryo-fetal brain development is profoundly influenced by numerous interacting environmental factors, named ‘exposome’ as a whole. Both in intrauterine and in postnatal life, environmental information converges on three major interacting/overlapping pathways: dysbiosis, mitochondrial impairment/oxidative stress, and immune activation (named MIA during pregnancy). As a whole, the three above-mentioned effector pathways—as a pathogenetic trio-impact epigenetic machinery. The matrilinear transmission of both microbiota and mitochondria [268,269] further enforces the need for effective women’s health programs, which are even more important in the presence of known risk factors for ASD, such as of the occurrence of neurodevelopmental disorders in previous offspring.
...


Link zur Studie