Cytosine methylation and human cancer.

Abnormalities of genomic methylation patterns have been attributed a role in carcinogenesis since the early 1980s, when large-scale demethylation of the genome was thought be an early event in multistep colorectal carcinogenesis. In the 1990s, local de novo methylation (with or without global demethylation) at tumor suppressor loci was held to be involved in silencing of tumor suppressor genes. The mechanisms that might mediate methylation and demethylation in carcinogenesis remain obscure, and there are questions as to whether the methylation changes are a cause or consequence of cellular transformation and clonal expansion. It is also important to derive a set of defined criteria by which a tumor suppressor gene can be concluded to have been inactivated by DNA methylation in a manner that contributes to carcinogenesis.

DNA methylation profile of the mouse skeletal alpha-actin promoter during development and differentiation.

Genomic levels of DNA methylation undergo widespread alterations in early embryonic development. However, changes in embryonic methylation have proven difficult to study at the level of single-copy genes due to the small amount of tissue available for assay. This study provides the first detailed analysis of the methylation state of a tissue-specific gene through early development and differentiation. Using bisulfite sequencing, we mapped the methylation profile of the tissue-specific mouse skeletal alpha-actin promoter at all stages of development, from gametes to postimplantation embryos. We show that the alpha-actin promoter, which is fully methylated in the sperm and essentially unmethylated in the oocyte, undergoes a general demethylation from morula to blastocyst stages, although the blastula is not completely demethylated. Remethylation of the alpha-actin promoter occurs after implantation in a stochastic pattern, with some molecules being extensively methylated and others sparsely methylated. Moreover, we demonstrate that tissue-specific expression of the skeletal alpha-actin gene in the adult mouse does not correlate with the methylation state of the promoter, as we find a similar low level of methylation in both expressing and one of the two nonexpressing tissues tested. However, a subset of CpG sites within the skeletal alpha-actin promoter are preferentially methylated in liver, a nonexpressing tissue.

Bisulfite sequencing in preimplantation embryos: DNA methylation profile of the upstream region of the mouse imprinted H19 gene.

In this study we describe a modification of the bisulfite genomic sequencing protocol that enables detection of methylation from as few as five diploid cells from preimplantation mouse embryos. We have used bisulfite genomic sequencing to study the methylation profile of the putative imprinting element upstream of the mouse H19 gene at several stages of embryonic development, including fertilized oocytes and two-cell embryos. The methylation of the H19 imprinting element has recently been described extensively for midgestation embryos, but remains poorly characterized for the preimplantation stages of development, despite widespread changes in genomic DNA methylation occurring at this time. We studied the methylation profile of 35 CpG sites spanning two regions within the H19 imprinting element and found that an overall pattern of allele-specific methylation was maintained at all developmental stages examined, including fertilized oocytes and two-cell embryos. However, allele-specific methylation was not maintained in an absolute fashion subsequent to the first cell division, with a clear flux between partial de novo methylation of the maternal allele and partial demethylation of the paternal allele. Our findings highlight the dynamics of methylation in the early embryo and suggest that it is the overall level of methylation that is responsible for maintenance of the imprinting element and not the methylation of individual CpG sites.

Sequence-specific methylation of the mouse H19 gene in embryonic cells deficient in the Dnmt-1 gene.

We have used Dnmtc/c ES cells that are homozygous for disruption of the DNA methyltransferase gene to address how de novo methylation is propagated and whether it is directed to specific sites in the early embryo. We examined the imprinted H19 gene and the specific-sequence region implicated as an "imprinting mark" to determine whether de novo methylation was occurring at a restricted set of sites. Since the "imprinting mark" was found to be methylated differentially at all stages of development, we reasoned that the sequence may still be a target for the de novo methylation activity found in the Dnmtc/c cells, even though the loss of maintenance the methylase activity renders the H19 promoter active. We used bisulfite genomic sequencing to determine the methylation state of the imprinted region of the H19 gene and found a low level of DNA methylation at specific single CpG sites in the upstream region of the imprinted H19 sequence in the Dnmtc/c mutant ES cells. Moreover, these CpG sites appeared to be favoured targets for further de novo methylation of neighbouring CpG sites in rescued ES cells, which possess apparently normal maintenance activity. Our data provide further evidence for a separate methylating activity in ES cells and indicate that this activity displays sequence specificity.

Detection and measurement of PCR bias in quantitative methylation analysis of bisulphite-treated DNA.

Methylation analysis of individual cytosines in genomic DNA can be determined quantitatively by bisulphite treatment and PCR amplification of the target DNA sequence, followed by restriction enzyme digestion or sequencing. Methylated and unmethylated molecules, however, have different sequences after bisulphite conversion. For some sequences this can result in bias during the PCR amplification leading to an inaccurate estimate of methylation. PCR bias is sequence dependent and often strand-specific. This study presents a simple method for detection and measurement of PCR bias for any set of primers, and investigates parameters for overcoming PCR bias.

Extensive DNA methylation spanning the Rb promoter in retinoblastoma tumors.

The retinoblastoma gene (Rb) is one of the best characterized tumor suppressor genes, and its inactivation is associated with a number of cancers. Previous studies have shown, by restriction enzyme analysis, that the promoter region of the Rb gene is methylated in a significant proportion of primary retinoblastoma tumors. We now report the first detailed methylation sequence analysis of the CpG island spanning the retinoblastoma promoter from hypermethylated retinoblastoma tumors. Our results show methylation is not confined to a specific CpG site, as detected by restriction enzyme studies, but extends to essentially all 27 CpG dinucleotides spanning the retinoblastoma CpG island, including the core promoter. The methylation pattern from each tumor DNA sample is different, ranging from densely to sparsely methylated profiles. Single CpG sites, in particular the E2F transcription factor binding site, as well as blocks of CpGs, were undermethylated in some tumor samples. Possible interference of methylation could be due to the binding of sequence-specific protein factors at these sites in the tumor cells. This study highlights that the dynamics of DNA methylation in cancer cells are clearly different from normal cells and gives an insight into the mechanism of abnormal methylation of CpG islands in cancer cells.