DNA methylation and characteristic histone modifications are established features of heterochromatin. MacroH2A is a histone H2A variant with a large non-histone region of unknown function. MacroH2A was shown to be enriched on the inactive X chromosome. Accumulating data in the literature and data from our work (Barzily-Rokni et.al) establish macroH2A as another component of heterochromatin that contributes to maintenance of the inactive state in addition to DNA methylation. We show strong synergism between macroH2A and DNA methylation in maintenance of gene silencing. We find that macroH2A is functionally involved in abnormal silencing of major tumor suppressor genes in cancer (such as p16 and MLH1): knockdown of macroH2A significantly enhances the effects of demethylating drugs in gene reactivation (Barzily-Rokni et.al). Current work in the lab is addressing the relations between macroH2A1 and H2A.Z in CpG island promoters.

 

 

With the recent developments of direct (iPS) reprogramming, we studied the fate of epigenetic aberrations following reprogramming. Focusing on the silenced TSG p16, our results (Ron-Bigger et.al) show that direct reprogramming can reverse aberrant epigenetic silencing events and that reactivation is robust and persists for many cell divisions both in undifferentiated and differentiated state. Our results provide the first demonstration of reprogramming cancer associated epigenetic defects and demonstrate the potential and the relevance of direct reprogramming approaches for repairing the epigenetic component of cancer. A genomewide analysis of methylation in these cells identified hundreds of sites that like p16 were aberrantly methylated in the donor cell and underwent demethylation following reprogramming. However not all aberrant methylation is reprogrammed and we have identified sites which appear to be resistant to reprogramming. It is currently not known what determines the “reprogramability” of methylation at different loci.

 

 

Genomewide DNA hypomethylation is a frequent feature of tumor cells. Using a mouse model for hypomethylation we demonstrated that hypomethylation can promote chromosomal instability and lead to cancer (Eden et.al; Gaudet et.al). In a later paper (Howard et.al) we identified another part of the molecular mechanism showing that hypomethylation results in T-cell Lymphoma due to activation of transcription from a promoter which is part of a transposable element residing in the genome. The resulting transcript is fused to a truncated version of the Notch1 gene product with oncogenic activity. In recent years we focused on examining the relevance of this mechanism to human cells and human cancer. Weber et.al provides a representative example of the aberrant mechanism we describe. We show that following DNA demethylation, a promoter residing within a Line1 element becomes activated and drives the expression of an aberrant c-Met transcript. Our current work is using experimentally generated data and data on human cancers from the TCGA project to evaluate cases in which alternative promoters (primarily from transposable elements) become activated due to errors in epigenetic maintenance and drive transcription of neighboring genes.

 

 

Recent high throughput genomic profiling of tumors reveals a surprising number of mutations in genes coding for chromatin modifying enzymes. One of the early examples is the case of Snf5. Snf5 is a core subunit of Swi/Snf ATP dependent chromatin remodeling complexes which are involved in transcriptional activation and repression. Snf5 is recognized as a classical tumor suppressor gene since loss of Snf5 results in malignant rhabdoid tumors in humans. We are using a Snf5 +/- mouse model to study the biology of these tumors and address the molecular basis through which loss of Snf5 (and consequent loss of chromatin remodeling activity) leads to transformation (Darr 1 et.al; Darr 2 et.al).

 

 

In collaboration with Prof. Yuval Dor from the faculty of medicine we are developing tools to overcome a long standing obstacle in the study of cell proliferation in-vivo. We have generated a new transgenic mouse strain, expressing a CyclinB1-GFP fusion reporter, which faithfully marks in-situ replicating cells in the S/G2/M phases of the cell cycle. In Klochendler et.al we use flow-cytometry to isolate live replicating and quiescent cells from the liver and study their transcriptome. This analysis reveals gene expression programs associated with cell proliferation in vivo. Interestingly we find that replicating hepatocytes downregulate expression of genes characteristic of liver differentiation and function. Current work in the lab is extending this study to proliferation in other tissues and in other contexts such as regeneration or cancer (Klochendler_2 et.al).