Autism spectrum disorder (ASD) is a neurodevelopmental disorder characterized by impairments in social-communicative skills and repetitive behaviors, affecting at least 1% of the population. While genetic factors significantly contribute to ASD, the complex genetic architecture and heterogeneity of etiological factors hinder the understanding of its pathophysiology and the identification of therapeutic targets. Although numerous ASD-candidate genes have been identified, they appear to converge on a few common molecular pathways, suggesting that diverse genetic variants might lead to similar functional consequences at the transcriptional or protein levels. Previous whole transcriptome studies using postmortem brain tissue have revealed expression alterations related to immune response, neurotransmission, and neurodevelopment. However, the postmortem brain may not capture transient dysregulated gene expression during prenatal development. Induced pluripotent stem cell (iPSC)-derived neuronal models offer an advantage by potentially recapitulating features of the developing brain. Existing studies utilizing iPSC-derived neurons have used relatively small sample sizes or focused on specific ASD endophenotypes (e.g., macrocephaly). This study aimed to address these limitations by investigating the transcriptome profiles of iPSC-derived neuronal cells from a cohort of mostly high-functioning ASD individuals, comparing them to controls, and examining the correlation between in vitro findings and postmortem brain data.

The literature review section extensively cites previous research on ASD genetics, focusing on the heterogeneity of genetic factors and their convergence on common molecular pathways. It also highlights previous transcriptome studies using postmortem brain tissue and iPSC-derived neuronal cells, noting inconsistencies and limitations in sample size and phenotypic diversity. This review emphasizes the need for larger, more diverse cohorts and the potential of iPSC-derived models to capture developmental aspects of ASD pathophysiology.

This study employed a multi-faceted approach. First, it involved patient ascertainment, focusing on a cohort of six male ASD patients (mostly high-functioning) and six male controls, with genetic characterization via array-CGH and whole-exome sequencing. iPSC lines were generated from stem cells derived from exfoliated deciduous teeth (SHED), characterized for pluripotency, and differentiated into neuronal progenitor cells (NPCs) and neurons. RNA sequencing was performed on NPCs and neurons, followed by normalization using RUVseq to correct for variations in neuronal proportions. The maturity and regional identity of the cells were predicted using a machine-learning approach comparing the transcriptome profiles with the BrainSpan Atlas. Differential gene expression analysis was conducted using the dream statistical model, and weighted-gene co-expression network analysis (WGCNA) was performed to identify co-expressed gene modules. Functional annotation analysis was performed using DAVID, Ingenuity Pathway Analysis, and cameraPR. Mass spectrometry-based proteomics was performed on NPCs to explore protein level changes. Gene set enrichment analysis was used to assess the overlap between identified genes and known ASD-related genes. Finally, module overlap analysis compared the findings with those from previous studies using fetal/neonatal brain samples or iPSC-derived neurons. Neuronal morphology analysis was also conducted using immunostaining and image analysis.

The study identified two ASD patients with multiple ASD pathogenic variants, supporting an oligogenic model. The iPSC-derived neuronal cells exhibited an expression profile similar to fetal brains during a developmental period relevant to ASD. Analysis revealed dysregulation of co-expressed gene modules in ASD patients, including modules involved in protein synthesis in NPCs and modules related to synapse/neurotransmission and translation in neurons. Proteomic analysis of NPCs showed potential molecular links between dysregulated modules in NPCs and neurons. Crucially, a module related to synaptic molecules was consistently upregulated in iPSC-derived neurons (resembling fetal brain expression) but downregulated in postmortem brain tissue, suggesting developmental differences in gene expression patterns and highlighting the potential of this network as a biomarker for ASD.

The consistent dysregulation of the synapse-related gene module across different studies strengthens its association with ASD. The observed differences between in vitro (fetal-like) and postmortem (adult) expression profiles suggest that the dysregulation of this network may manifest differently across developmental stages in ASD individuals. This finding underscores the importance of studying developmental processes in ASD. The identification of this module as a potential biomarker and therapeutic target warrants further investigation. The study's use of high-functioning individuals alongside the other methodologies contributes to a more comprehensive understanding of ASD, moving beyond the limitations of previous studies focused on limited phenotypes or sample sizes.

This study provides robust evidence for the consistent dysregulation of a synaptic gene module in ASD, demonstrating its potential as a biomarker and therapeutic target. The findings highlight the value of iPSC-derived neuronal models in capturing developmental aspects of ASD, while also emphasizing the need for further studies involving larger, more diverse cohorts and investigations into the developmental trajectory of gene expression dysregulation in ASD.

The relatively small sample size, although larger than in previous studies, may limit the generalizability of the findings. The study focused primarily on male participants, limiting the understanding of potential sex-specific effects. Further research is needed to validate the findings in larger, more diverse cohorts, including female participants, and to explore the functional consequences of the identified gene module dysregulation.