We have recently discovered relaxation of genomic imprinting in cancer, a novel type of genomic alteration in man. Genomic imprinting is a differential modification of maternal and paternal alleles in the gamete, leading to expression of a single allele in the offspring, depending on its sex of origin. While genomic imprinting has been demonstrated at the molecular level in insects and mice and suspected to play a role in human disease, there previously has been no direct molecular evidence for imprinted genes in man. We have established that two genes, the insulin- like growth factor II gene (IGF2) and a nearby gene, H19, are reciprocally imprinted in man. Thus, only the paternally derived IGF2 allele and the maternally derived Hl9 allele are expressed in human tissues. Surprisingly, we found that in two childhood tumors, Wilms tumor and rhabdoid tumor, these imprints are relaxed and both copies of the genes are expressed. This alteration occurred in the earliest stages of tumorigenesis. In order to understand the role of relaxation or loss of imprinting (LOI) in human carcinogenesis, we will determine its prevalence in two groups of tumors: childhood cancers that express IGF2 and/or Hl9; and a subset of common adult cancers in which IGF2 has been proposed to play an autocrine growth stimulatory role, such as colorectal and breast cancer. As our preliminary data include the first demonstration of normally imprinted genes in man, we will attempt to identify other imprinted genes, using as a guide regions suggested by homology to known imprinted domains in mouse, and we will determine whether these genes also undergo relaxation of imprinting in cancer. We will also determine whether LOI is related to deregulated gene expression in cancer. We earlier discovered widespread alterations of DNA methylation in malignant and premalignant tumors, and DNA methylation has been tentatively linked to genomic imprinting in other organisms. We will investigate the molecular mechanism of LOI by determining whether it is related to specific changes in DNA methylation, and by attempting to modify the pattern of genomic imprinting in vitro with 5-azacytidine, a drug that causes loss of DNA methylation. Finally, in order to determine whether LOI is potentially reversible, which could have important therapeutic implications, we will attempt to restore the pattern of imprinting by genetic complementation. These studies should provide a foundation for understanding the role of normal genomic imprinting in man and the consequences of altered imprinting in cancer.
Showing the most recent 10 out of 55 publications
