Exosomal MicroRNA Expression in Children With Autism Spectrum Disorder

This study will recruit 20 children with idiopathic ASD (defined by DSM-IV criteria) and 20 age- and gender-matched controls with typical neuropsychological developmental patterns. The number of subjects necessary for this study was determined by power analysis. Using a two sample t-test model with 15 ASD and 15 control subjects in each group would have 80% power to detect changes in expression equal to twice the standard deviation of the group mean with Bonferonni corrected level (p<0.0001) based on expectations of measuring 500 miRNA in each sample (Lenth, 2006).

Parental consent and patient assent forms will be distributed to potential subjects at the University Pediatric and Adolescent Center and the Center for Development Behavior and Genetics. To explain the study, children will be told, "The doctors are doing a study about a disease called autism. If you want to be in this study a small amount of blood will be taken from your arm with a needle. The needle will hurt but will go away after a little while. The doctors will ask you to answer some questions about how you feel. You do not have to be in the study. No one will be mad at you if you don't want to do this."

ASD subjects will be assessed using current methods in autism diagnosis: the Autism Diagnostic Observation Schedule (ADOS) and the Checklist for Autism in Toddlers (CHAT). The Vineland Behavior Scales will also be administered. All medical information collected will be coded and de-identified according to standard protocols.

Venipuncture will be employed to collect 2ml of blood from each subject into an EDTA-free vacutainer tube. Samples will be left for 1 hour at 37˚C to allow clotting before storage at 4˚C overnight. The sample will then be centrifuged at 4000 rpm for 20 minutes at 4˚C. Serum will be removed from the clot by pipetting and stored at -80˚C until miRNA extraction is performed.

The decision to measure exosomal miRNA from serum is supported by a study of miRNA in 12 human body fluids. Weber and colleagues (2010) determined that the median total concentration of RNA in plasma (308 ug/L) was greater than that of cerebrospinal fluid (111 ug/L). Furthermore, the authors found the number of detectable miRNAs in plasma (349) to be greater than cerebrospinal fluid (212). The miRNA content of plasma and cerebrospinal fluid has some intriguing similarities. Of the twenty most abundant miRNAs found in cerebrospinal fluid and plasma, 9 overlap. We have previously examined the expression of exosomal miRNA in the serum of human alcoholic subjects. Intriguingly, we characterized the expression levels of brain-specific miRNAs in serum samples from over 30% of those tested. This finding demonstrates that exosomal miRNAs from the central nervous system can be detected and quantified in human blood samples.

Serum miRNA will be extracted using a Qiagen RNeasy Mini Kit according to manufacturer protocol. RNA will be processed with the Genisphere FlashTag HSR RNA labeling kit. RNA Spike Control Oligos, and poly(A) tails will be added to each RNA sample and a biotinylated signaling molecule will be ligated to the RNA. The labeled RNA samples will be added to a hybridization cocktail, incubated, and injected into Affymetrix miRNA 2.0 arrays. After 16 hours of hybridization, the arrays will be washed and stained on an Affymetrix fluidics station 450 using protocol FS450_0003. Arrays will be scanned on an Affymetrix GeneChip Scanner 7G Plus. The cel files will be analyzed using miRNA QC Tool version 1.1.1.0.

A 3-way ANOVA assessing the impact of ASD, age, and gender on miRNA expression will be employed along with a Mann Whitney Test. Targets with a P-value < 0.05 will be selected for further analysis after multiple testing correction with Bonferroni analysis. Significantly altered miRNAs will be queried on Mirbase.org for known mRNA targets. The target list will then be examined using DAVID bioinformatics functional annotation tool for enriched gene ontology categories. Correlations between miRNA expression and medical/neuropsychological characteristics will be calculated using Pearson Correlation statistics, filtering for miRNAs with p-value<0.05 and r>0.5.