Evidence for differential alternative splicing in blood of young boys with autism spectrum disorders

Background: Since RNA expression differences have been reported in autism spectrum disorder (ASD) for blood and brain, and differential alternative splicing (DAS) has been reported in ASD brains, we determined if there was DAS in blood mRNA of ASD subjects compared to typically developing (TD) controls, as well as in ASD subgroups related to cerebral volume. Methods. RNA from blood was processed on whole genome exon arrays for 2-4-year-old ASD and TD boys. An ANCOVA with age and batch as covariates was used to predict DAS for ALL ASD (n=30), ASD with normal total cerebral volumes (NTCV), and ASD with large total cerebral volumes (LTCV) compared to TD controls (n=20). Results: A total of 53 genes were predicted to have DAS for ALL ASD versus TD, 169 genes for ASD-NTCV versus TD, 1 gene for ASD-LTCV versus TD, and 27 genes for ASD-LTCV versus ASD-NTCV. These differences were significant at P <0.05 after false discovery rate corrections for multiple comparisons (FDR <5% false positives). A number of the genes predicted to have DAS in ASD are known to regulate DAS (SFPQ, SRPK1, SRSF11, SRSF2IP, FUS, LSM14A). In addition, a number of genes with predicted DAS are involved in pathways implicated in previous ASD studies, such as ROS monocyte/macrophage, Natural Killer Cell, mTOR, and NGF signaling. The only pathways significant after multiple comparison corrections (FDR <0.05) were the Nrf2-mediated reactive oxygen species (ROS) oxidative response (superoxide dismutase 2, catalase, peroxiredoxin 1, PIK3C3, DNAJC17, microsomal glutathione S-transferase 3) and superoxide radical degradation (SOD2, CAT). Conclusions: These data support differences in alternative splicing of mRNA in blood of ASD subjects compared to TD controls that differ related to head size. The findings are preliminary, need to be replicated in independent cohorts, and predicted alternative splicing differences need to be confirmed using direct analytical methods.