Finding the origin of complex disease is no easy task but now it seems a seminal breakthrough has occurred. Andrew Feinberg and team at the John Hopkins University Center for Epigenetics have found a DNA methylation pattern in the sperm of fathers with an increased risk of fathering autistic children that is also present in brains of unrelated autistic children.
DNA Methylation Analysis and Offspring Autism Risk
The team performed DNA methylation analysis of CpGs on a cohort of fathers with autism risk in offspring. They were chosen from a cohort of fathers who have autistic children and currently have a young infant of unknown status. The novel analysis also compared the more reliable CHARM array technology (> 7 million CpG sites) to the 450k array (∼485 000 CpG sites). Here’s what they found:
- There were 193 significant differentially methylated regions (DMRs) identified in the CHARM analysis of at risk fathers whose young infants have autistic phenotypes compared to at risk fathers with infants who don’t have these phenotypes.
- The sperm DMRs are large stretches of dozens of CpGs, not single CpG sites, and were consistent with 75 CpG probes from the 450K.
- The DMRs are involved in developmental processes including imprinting disorders.
- SNORD115 regions were represented in 4 of the top 10 DMRs.
- The SNORD 115 region was then compared in an independent sample of post-mortem autism brains (cerebellum) and 18 of the 75 probes from the 450K that overlap with the DMRs from CHARM showed significant differences.
Linking Imprinting Disorders and Autism
The top candidate region with the large-scale methylation change is located in the imprinted SNRPN-UBE3A locus and contains two large and complex ncRNA clusters, SNORD 115 (HBII-52) and SNORD 116 (HBII-85), with the four non-coding exons of IPW sitting between them, in addition to a number of important genes on their sides. This locus is involved in the sister imprinting disorders: Angelman syndrome and Prader-Wili Syndrome. SNORD 115/116 are a region of allele-specific decondensation that regulates neuronal nuclear size and produce a series of complex non-coding RNAs including snoRNAs, lncRNAs, and the ever so strange hybrid sno-lncRNAs.
These findings offer a novel developmental time point for the SNORD 115 ncRNAs in establishing psychiatric disease, particularly those with autistic phenotypes. Ultimately, the research pinpoints a number of unexpected large-scale methylation differences that provide additional evidence linking autism to imprinting disorders with impressive and unprecedented prowess. It also begins to hint at how an alteration in the germ-line may propagate to the brain of an offspring. This leaves us wondering if these alterations are erased during fertilization and are re-instated or are responsible for the transmission across generations in their own right.
Go and check out the powerful findings in this cohort over at the International Journal of Epidemiology, April 2015