Increased copy number for methylated maternal 15q duplications leads to changes in gene and protein expression in human cortical samples

Background: Duplication of chromosome 15q11-q13 (dup15q) accounts for approximately 3% of autism cases. Chromosome 15q11-q13 contains imprinted genes necessary for normal mammalian neurodevelopment controlled by a differentially methylated imprinting center (imprinting center of the Prader-Willi locus, PWS-IC). Maternal dup15q occurs as both interstitial duplications and isodicentric chromosome 15. Overexpression of the maternally expressed gene UBE3A is predicted to be the primary cause of the autistic features associated with dup15q. Previous analysis of two postmortem dup15q frontal cortical samples showed heterogeneity between the two cases, with one showing levels of the GABA(A) receptor genes, UBE3A and SNRPN in a manner not predicted by copy number or parental imprint. Methods: Postmortem human brain tissue (Brodmann area 19, extrastriate visual cortex) was obtained from 8 dup15q, 10 idiopathic autism and 21 typical control tissue samples. Quantitative PCR was used to confirm duplication status. Quantitative RT-PCR and Western blot analyses were performed to measure 15q11-q13 transcript and protein levels, respectively. Methylation-sensitive high-resolution melting-curve analysis was performed on brain genomic DNA to identify the maternal: paternal ratio of methylation at PWS-IC. Results: Dup15q brain samples showed a higher level of PWS-IC methylation than control or autism samples, indicating that dup15q was maternal in origin. UBE3A transcript and protein levels were significantly higher than control and autism in dup15q, as expected, although levels were variable and lower than expected based on copy number in some samples. In contrast, this increase in copy number did not result in consistently increased GABRB3 transcript or protein levels for dup15q samples. Furthermore, SNRPN was expected to be unchanged in expression in dup15q because it is expressed from the single unmethylated paternal allele, yet SNRPN levels were significantly reduced in dup15q samples compared to controls. PWS-IC methylation positively correlated with UBE3A and GABRB3 levels but negatively correlated with SNRPN levels. Idiopathic autism samples exhibited significantly lower GABRB3 and significantly more variable SNRPN levels compared to controls. Conclusions: Although these results show that increased UBE3A/UBE3A is a consistent feature of dup15q syndrome, they also suggest that gene expression within 15q11-q13 is not based entirely on copy number but can be influenced by epigenetic mechanisms in brain.