Advertisement
Letter to the Editor Free access | 10.1172/JCI10912
1Section of Medical and Molecular Genetics, Department of Paediatrics and Child Health, University of Birmingham, Birmingham, United Kingdom.2 Laboratory of Developmental Genetics and Imprinting, The Babraham Institute, Cambridge, United Kingdom.
Find articles by Maher, E. in: JCI | PubMed | Google Scholar
1Section of Medical and Molecular Genetics, Department of Paediatrics and Child Health, University of Birmingham, Birmingham, United Kingdom.2 Laboratory of Developmental Genetics and Imprinting, The Babraham Institute, Cambridge, United Kingdom.
Find articles by Reik, W. in: JCI | PubMed | Google Scholar
Published September 15, 2000 - More info
A.P. Feinberg raises two questions: (a) the origin of the two-domain model, and (b) the organization of enhancers and insulators within chromosome 11p15.5. Our concept that two imprinting control centers exist within chromosome 11p15.5 was developed independently. In a series of reports, we established, first, that loss of imprinting of IGF2 in Beckwith-Wiedemann syndrome (BWS) may be associated with H19 hypermethylation and silencing, consistent with loss of function in a distal imprinting center (1); second, that a BWS-associated maternally inherited inversion with a breakpoint within KCNQ1 was associated with an H19-independent loss of imprinting in IGF2 (2); and, finally, that such H19-independent loss of IGF2 imprinting is frequently found in sporadic cases of BWS that lack chromosomal rearrangements (3). The finding that epigenetic alterations at KvDMR1 and H19 appeared to be mutually exclusive provided us with confirmation of our concept (4). With regard to the organization of imprinting elements within 11p15.5, we agree that it is possible that the CDKN1C (p57KIP2) enhancer could be on the centromeric side, but we favor a telomeric location for several reasons. First, if the enhancer were centromeric, CDKN1C would need its own imprinting mechanism. This is less likely because (a) there is no differential methylation in the human (5); (b) a maternal germline imprint is required for activity of cdkn1c (6); (c) cdkn1c transgenes do not become imprinted (7); and (d) in Dnmt1-deficient mice, cdkn1c is biallelically expressed, but inspection of the gels shows that this could be a low-level expression from both alleles (8), corresponding to the low-level paternal expression in humans. Finally, and importantly, the organization suggested by A.P. Feinberg would require a closed boundary on the maternal chromosome and an open one on the paternal chromosome, but KvDMR1 methylation is maternal (presumably indicating that the boundary is open, as with the H19 upstream region).