A specific DNA methylation signature associated with NSD1+/- mutations in Sotos syndrome reveals a significant genome-wide loss of DNA methylation (DNAm) targeting CGs in regulatory regions of key developmental genes

Abstract

Sotos syndrome (SS) is characterized by somatic overgrowth and intellectual disability. Most SS cases (>75%) have mutations in NSD1 (nuclear receptor-binding SET domain protein 1). NSD1 binds near promoter elements and regulates transcription initiation and elongation via interactions with H3-K36Me and RNA polymerase II. To determine if NSD1 mutations impact stable epigenetic marks such as DNA methylation (DNAm), we compared DNAm in peripheral blood DNA from SS cases with NSD1 mutations (NSD1+/-; n=20) to controls (n=30) using the Illumina Infinium450methylation BeadChip. Differential DNAm analysis using non-parametric statistics (with correction for multiple testing) coupled with permutation analyses identified a surprisingly high number (n=2157) of differentially methylated (DM) CG sites (with >20% difference in DNAm) between SS and controls. These sites were distributed across the genome; 95% demonstrated loss of DNAm. Using unsupervised hierarchical clustering of the 2157 DM CG sites, all SS cases with NSD1 +/- clustered as a distinct group separate from controls. Moreover, DNAm at these sites clearly distinguished SS (NSD1+/-) from Weaver syndrome (EZH2+/-, n=5), another overgrowth syndrome which has considerable phenotypic overlap with SS. These results suggest that these DM CG sites constitute a DNAm signature that is specific for NSD1+/-. Also, the DNAm signature was successfully used to reclassify NSD1 variants of unknown significance (VUS) in six cases of SS into functionally damaging (n=1) and non-pathogenic (n=5) variants. The majority of these DM CGs mapped to enhancers and CpG island shores. Analysis of ChIP-seq data showed that NSD1+/- specific CG sites are associated with reduced H3K36me3 marks in both normal blood and embryonic stem cells. Also, Ingenuity analysis showed enrichment in neural and cellular development pathways (p<0.001). We then searched for binding motifs enriched in these NSD1+/- DNAm targets using MEME and JASPAR CORE database; SP1 was the most enriched with binding sites in 41% of the targets (NCOR=0.62). This is the first report of an NSD1+/- specific DNAm signature in SS and that loss-of-function mutations in NSD1 can deregulate the intricate transcriptional balance of key developmental genes. Further elucidation of this signature will significantly impact our understanding of the molecular pathophysiology of SS and identify the specific molecular targets for NSD1 that govern its action in early development.