Mammalian Development and the Generation of Mouse Models of Human Genetic Disorders

The long range goals of our work are directed towards an understanding of mammalian development on a genetic and molecular level and the derivation of mouse models for human genetic disorders. Current methods of homologous recombination in embryonic stem cells allow generation of mice with predetermined mutations. Our strategies include the production of mice carrying mutations which are expressed only in specific lineages as well as mice which carry large fragments of DNA derived from yeast artificial chromosomes (YACs). We are using these powerful approaches to investigate the function of specific genes which are important for the following aspects of mouse development, disease, and cancer.

DNA Methylation: Despite a large body of evidence, the importance of DNA methylation in vertebrate gene control has remained a controversial issue because of the indirect and correlative nature of many studies in the field. Over the years, DNA methylation has been hypothesized to be involved in numerous processes including X inactivation, genomic imprinting, virus latency, carcinogenesis, aging, and the regulation of tissue-specific gene expression during development. We have undertaken a direct approach to defining the role of DNA methylation in development and have generated an allelic series with partial loss of function and null alleles of the only known methyltransferase (MTase) gene in mammals. Mutant embryos are unable to maintain normal levels of genomic methylation and die at midgestation, a result which establishes the essential role of DNA methylation in development. We are using this mutant strain to investigate the role of methylation in genomic imprinting, X-inactivation and cancer. A potential problem in our approach is the embryonic lethality caused by the mutation of the MTase gene. This has prompted us to generate a mutant allele which can be expressed in a stage and tissue specific manner which will allow to study the role of DNA methylation in the postnatal animal.

Genomic Imprinting: Imprinted genes are expressed either from the maternal or the paternal allele but not from both. Using the MTase mutant mice we have shown that normal DNA methylation levels are crucial for the maintenance of the imprint and for monoalleleic expression of the genes, leading to the hypothesis that gamete-specific MTase activities install methylation marks on imprinted genes during oogenesis and spermatogenesis, respectively. Our results suggest that at least four different DNA MTase activities are expressed in mammals: (i) a hemi-MTase in all cells; (ii) a de novo MTase in ES cells and in embryos prior to gastrulation; a specific de novo MTase in (iii) spermatogenesis and (iv) oogenesis which recognizes paternally or maternally methylated imprinted genes, respectively. Only the hemi-MTase has been cloned and a major challenge will be to identify the other activities. We are, therefore, using several approaches to define the de novo MTase which establishes the early embryonic methylation pattern and may be involved in the genesis of cancer.

X-Inactivation: The mammalian dosage pathway is activated by a mechanism that counts the number of chromosomes. The "n-1" rule dictates that in diploid cells a single X chromosome remains active regardless of the total number of X chromosomes: XY males have no inactive X, normal females with two Xs have one inactive X and individuals with three Xs have two inactive X chromosomes. X chromosome inactivation in mammals requires expression of the Xist gene which maps to the X chromosome inactivation center (Xic) and encodes an untranslated nuclear RNA. We have used homologous recombination approaches as well as the transfer of large fragments of DNA into ectopic sites to define the genetic and epigenetic control of X inactivation. Our results indicate that (i) Xist RNA is necessary for initiation and propagation of the inactivation process; (ii) DNA methylation is crucial for shielding the active X chromosome against the inactivation process; (iii) that a 450 kb fragment of DNA carrying the Xist gene acts as an inactivation center and is sufficient for initiation, propagation and maintenance of the inactive state. Also, the elements for counting and choosing X chromosomes to execute the "n-1 rule" are part of the transgene.

Cancer: The involvement of DNA methylation in cancer has been controversial: both hypomethylation as well as hypermethylation have been associated with malignant transformation. When the MTase mutation was introduced into mice with a genetic predisposition to colon cancer, a surprising result was obtained: the MTase enzyme level directly correlated with the development of cancer. This represents a novel concept of the role of DNA methylation in carcinogenesis and argues that the MTase enzyme itself acts as an oncogenic determinant. Also, as suggested by work with prokaryotes, DNA methylation may be crucial for DNA repair and for maintaining the stability of the genome. Alternatively, DNA methylation may affect cancer initiation by an epigenetic mechanism, i.e. the alteration of gene expression. We are pursuing experiments to distinguish between these alternatives.

Neurotrophins: Neurotrophins (NGF, BDNF, NT-3, NT-4) are important survival factors for the developing nervous system. It has been postulated that neurotrophins may be useful for the function of postnatal neurons and could be used in the treatment of degenerative diseases such as Alzheimer's or ALS. To directly test the role of NT-3 in postnatal life, we have generated a "floxed" allele which allows the deletion of the gene by stage-specific expression of the Cre recombinase. The Cre recombinase has been targeted to various parts of the brain including the hippocampus and we are assessing whether the lack of NT-3 in the adult may lead to specific neuronal deficits.