Differential Alternative Splicing in Superior Temporal Gyrus of Autism Spectrum Disorders Brains

Background:  Gene expression studies in postmortem brain in Autism Spectrum Disorders (ASD) have revealed dysregulated immune and neuronal networks. Differential alternative splicing (DAS) in genes targeted by the neural-specific splicing regulator FOX1 (A2BP1) has been reported in ASD brain. Thus, we investigated genome-wide DAS in superior temporal gyrus (STG) of brains of ASD subjects compared to typically developing (TD) controls. STG is an association cortex involved in social perception, joint attention, face perception and speech perception and is implicated in ASD.

Objectives:  To assess DAS in the STG transcriptome in postmortem human brains of ASD subjects compared to TD controls.

Methods:  RNA libraries from 45 subjects (16 ASD, 14/2 M/F and 29 TD, 23/6 M/F) were sequenced to 40M 2×150 bp reads using Illumina sequencing-by-synthesis technology. Alignments to the human genome (hg19) (STAR 2.4.1d aligner) and quantification using the Ensembl 75 transcript database were performed. DAS was assessed using Log-normal with shrinkage models as implemented in Partek Flow, considering Diagnosis (ASD, TD), Sex (F, M) and Age (continuous variable). Gene-Specific Algorithm (GSA) was applied to each transcript separately to identify factors and interactions that affected DAS. After filtering low expressing transcripts, DAS between the groups was considered significant with |Fold-Change| >1.2, and FDR-corrected p<0.3 (nominal p<0.01). Gene Ontology and Ingenuity Pathway Analysis were used to identify relevant biological functions.

Results:  There was no significant age difference between ASD (Average age: 30 years, range 9-56)) and TD (Average age: 25.2, range 5-68) groups. Most transcripts (41.4%) were best described by the model: Dx + Sex + Dx *Sex, followed by the Dx + Sex + Age + Dx*Age +Dx*Sex model (24.9% of the transcripts). 330 genes displayed DAS in ASD vs. TD considering both sexes, there were 795 in ASD-Male vs. TD-Male, and 41 in ASD-Female vs. TD-Female. Genes with DAS were over-represented in immune and neurotransmitter pathways. Some of these have previously been implicated in ASD, such as mTOR Signaling, CREB Signaling in neurons, and Glutamate and GABA Receptor Signaling pathways (Males analysis). There was significant overlap with ASD-implicated genes from the SFARI database with our all ASD vs. TD (33 genes, hypergeometric probability = 1.3E-06) and ASD-Male vs TD-Male (81 genes, p=7.9E-15) differentially alternatively spliced gene lists. Among the genes with DAS in ASD-Male, the RBFOX2-12 transcript was down-regulated. RBFOX2 (RNA Binding Protein, FOX1 Homologue 2) regulates alternative exon splicing in the nervous system and is homologous to the neural-specific splicing regulator FOX1 (A2BP1) which has been reported to be down-regulated in ASD brain (Voineagu et al, 2011). We compared our genes with DAS to the 196 genes with FOX1-dependent DAS identified by Voineagu et al. (2011). There was a significant overlap with our all ASD vs. TD (9 genes, p=0.004) and ASD-Male vs. TD-Male (32 genes, p=4.1E-12) gene lists.

Conclusions:  There is differential transcript isoform expression in STG in ASD which is modulated by age and sex, and which likely contributes to ASD pathophysiology. Future studies will need to confirm these findings.