Speaking at the American Association of Cancer Research conference held in Denver recently, Stephen B. Baylin, M.D., professor of oncology and medicine at Johns Hopkins University, declared, “We’re at an exciting era. It’s still all about the genome.” But, rather than focusing on the sequencing, the “era of epigenetics,” as he called it, “focuses on the topology of the genome—how the nucleosomes are arranged in a three-dimensional fashion—to better understand the origins of cancer.”
As Bradley E. Bernstein, director of pathology, Massachusetts General Hospital, later explained, “The term epigenome describes the packaging of the genome. Epigenetic refers to a process that maintains the cell state stably, ideally through mitosis. A classic example is gene repression.” At the biochemical level, they involve contributions from many interacting components, including histome modifications, cancer protein, ncRNAs, and transcriptions factors. “Histome modifications themselves are not sufficient.”
Researchers are finding that “abnormal gene silencing associated with gene promoter DNA hypermethylation is linked to key aspects of chromatin regulation of gene expression, which maintains the state of embryonic stem progenitor cells,” Dr. Baylin said.
“In most of the normal genome, CpG is packaged away, protected from DNA methylation. In the cancer epigenome, however, “many regions ought to be closed but are open and many that should be open have become methylated and, therefore, are closed,” he added. In his work, Dr. Baylin and his lab used a colon cancer cell line to determine genes that appear most frequently. Frequency and DNA hypermethylation were then overlapped.
Individual patients have about 40 mutations, but 300 to 400 instances of DNA hypermethylation, Dr. Baylin reported. “If you only look at the genes that occur with high frequency, you have missed important genes,” he emphasized.
Dr. Baylin’s lab uses Infinium luminescent technology by Illumina for genotyping, querying only a few sets of genes on the genome. “It is a good screen for methylation across the genome,” he pointed out. He added that embryonic stem cells lack methylation. Leukemia and cancer cells, however, have abnormal DNA methylation. “That’s a stark reality in cancer and is a microcosm of the cancer genome.”
DNA hypermethylation seems to play a causative role in the onset of cancer. In fact, he said, “DNA methylation fosters tumorigenesis.” Of 610 candidate genes studied in a colon cancer cell line, 300 were PcG marked in embryonic stem cells or mesenchymal cells. Dr. Baylin plans to study that more broadly using ChIP deep sequencing.
In analyzing tiling arrays for key chromatin marks that are specific to developing cell types, he found that the polycomb (repressive system) mark is really a balance of marks rather than one that is on or off. “It is bivalent chromatic that hold stem cell genes in a poised state.” His working hypothesis is that bivalent chromatin can be remodeled in many cells as they differentiate. “When the system is challenged to renew, the genes are marked by polycomb.”
The epigenetic gatekeepers—like GATA transcription factors and APC—help stabilize stem cells or block cellular renewal. “I propose there are epigenetic steps as well,” Dr. Baylin suggested. “For example, abnormal epigenetic memory could drive cells forward to proliferation.”
In terms of regulation in metazoans, which is critical in regulating behavior in cells, Dr. Bernstein stated, “One thinks of polycombs and trithorax proteins,” which help maintain lineage-specific gene-expression proteins.
In mammalian and mouse stem cells, he explained, “there are many bivalent domains, characterized by the juxtaposition of the repressive polycomb mark and the actived trithorax mark found throughout the genome. These bivalent domains are associated with many developmental regulator genes that are inactive in embryonic stem cells but are rapidly induced along specific differentiation pathways. That helps poise genes for future activation.
Dr. Bernstein’s lab looked at other models, including hematopoietin development, focusing on Pax5 master regulator, a B-cell lineage-specific protein that regulates early-stage B-cell differentiation. Looking at in vivo CD34 cells sorted from cord blood, Pax5 is sitting in a bivalent chromatin state, “whereas, in the CD19 B cells, the polycomb is stripped out,” leaving an active promoter and evidence of p36 elongation. Conversely MyoD, in the hematopoietic lineage, is robustly silenced by polycomb repression.
Additional studies of protein complexes in embryonic stem cells found that PRC2, which catalyzes H3K27 mer3, and PRC1, which blocks RNA polymerase elongation, closely interact in epigenetic repression. All the bivalent domains are contained PRC2, but fewer than half also are contained PRC1. PRC1 and -2 behave differently, with PRC2 triggering a smaller degree of repression than PRC1.
In another in vitro study, all the cancer proteins were lost. “That’s associated with hypermethylation and is not recapitulated in vivo. In vitro is where most of the hypermethylation occurs,” which caused Dr. Bernstein to question whether hypermethylation is, in fact, an artifact. “It may be, but it is very relevant in human cancer.”