Feasibility for Detecting Epigenetic Abnormalities in Autism Brain

Background: There is increasing evidence that de novo and recent mutations at any one of many heterogeneous loci cause a substantial fraction of autism.  These findings are entirely consistent with the high heritability of autism, as evidenced by the high concordance between monozygotic (MZ) twins.  Yet, the etiology of autism remains unknown for a large fraction of individuals, particularly for nondysmorphic, higher functioning males. Epigenetics refers to regulation of gene expression without change in the underlying DNA sequence. Important, analyzable components of epigenetic expression include expression of mRNA and miRNA, DNA methylation, and histone modifications. Epigenetics plays at least some role in the etiology of autism as evidenced by the fact that usually maternal but not paternal interstitial duplications of chromosome 15q11-q13 cause autism. Epigenetic abnormalities could be more widely important in the etiology of autism, and could contribute to the heritability seen in MZ twins if epimutatations arose prior to MZ twinning.

Objectives: To develop and implement methodologies to be used in finding epigenetic aberrations associated with or causing autism focusing on the use of postmortem brain. As a proof of principle, we wished to demonstrate the feasibility of using existing methods to detect the known epigenetic aberrations within 15q11-q13, responsible for Prader-Willi syndrome (PWS) and Angelman syndrome (AS) syndromes.

Methods: To investigate the presence or absence of DNA methylation, a combination of two methods was used.  Methylated DNA immunoprecipitation (MEDIP), utilizing an antibody against methylated DNA, was used to detect the presence of methylated sites.  An enzymatic method that utilized the restriction enzymes McrBC and HpaII and subsequently amplification was used to detect unmethylated sites.  Histone modifications were examined using native chromatin immunoprecipitation (N-ChIP), followed by analysis either using Agilent microarrays or massively parallel Solexa DNA sequencing.  Finally, the expression of 723 human and 76 human viral miRNAs were examined using the Agilent human miRNA (V2) microarray system.

Results: Using methodologies to examine the presence and absence of DNA methylation, we confirmed the ability to detect the known abnormalities in PWS and AS brain.  Using NChIP to detect histone H3 lysine 4 trimethylation, a marker of active transcription, we confirmed its presence exclusively on the paternal SNRPN allele – present in AS but not PWS individuals.  Finally, we found a significant upregulation of three miRNAs in 6 of 20 autism and in 0 of 20 control cerebellum samples.

Conclusions: We have demonstrated the ability to use these methodologies to detect known epigenetic abnormalities in PWS and AS brain and the feasibility of detecting changes in autism brain if they occur.  We plan to apply all of these methodologies to compare autism and control brain tissue. Additionally, we have shown a significant elevation of three miRNAs in a subset of autism cerebellum samples.

ACKNOWLEDGMENTS: Simons Foundation SFARI, March of Dimes, William Stamps Farish Fund, NIH HD-037283, Autism Speaks, NAAR, AGRE, Harvard Brain Bank/Autism Tissue Project.