Adrian Bird, PhD
Professor
Wellcome Trust Centre for Cell Biology
University of Edinburgh
Scotland, United Kingdom
January 22, 2007

DNA Methylation and Disease

The DNA of vertebrate animals is covalently modified by methylation of the cytosine base in the dinucleotide sequence 5’CG3’. In mammals, DNA methylation patterns are established during embryonic development and maintained by a copying mechanism when cells divide. The heritability of DNA methylation patterns allows epigenetic marking of the genome to be stable through multiple cell divisions and therefore constitutes a form of cellular memory. The existence of DNA methylation patterns raises two important questions: 1) How are the patterns formed? 2) How are they read to generate biological outcomes? We have focused on the second question and have identified a set of proteins that recognize and bind to methylated sites in the genome. Most of these proteins are transcriptional repressors that recruit corepressor complexes, which modify chromatin structure to ensure gene silencing.

To understand the biology of these proteins and their role in human disease, we have created mouse gene knockouts for methyl-CpG binding domain (MBD) proteins and identified cases of gene misregulation that can be attributed to the absence of one of these proteins. For example, efficient repression of the Xist gene on the active X chromosome and of exocrine pancreatic enzyme genes in the mouse colon requires MBD2. Interestingly, despite their common DNA binding sites, these proteins apparently do not substitute for one another. Thus, it seems that the bona fide target genes regulated by each methyl-CpG binding protein are distinct. Our work has begun to reveal the molecular basis for this specificity by establishing significant DNA preferences in addition to the requirement for methyl-CpG.

The MBD protein MECP2 is of particular interest as mutations affecting its gene are the primary cause of Rett Syndrome, which is the most common inherited form of mental retardation affecting human females. Delayed-onset symptoms include developmental delay, loss of purposeful limb use, and breathing abnormalities. As there is no obvious neurodegeneration in postmortem brains of RTT patients, the question of reversibility arises and is of obvious relevance for therapeutic approaches to RTT. We earlier created a mouse model for RTT that lacks an intact Mecp2 gene and mimics several features of the disorder, including late onset. Using a mouse with an Mecp2 allele that can be conditionally activated, we asked whether neuronal defects can be rectified if MeCP2 is provided de novo after abnormal neuronal morphology and symptoms have arisen. The results demonstrate that most or all symptoms are in fact reversible. In addition, a deficit in long-term potentiation in the hippocampus is abolished by late reintroduction of MeCP2. These findings have obvious implications for future therapeutic approaches to this disorder and they raise the possibility that MeCP2 is required to maintain gene expression programs in mature neurons.