The Association for Molecular Pathology Meeting (AMP) was held this year in Salt Lake City, Utah on November 15-18. For me, one of the highlights of this year’s meeting was the lecture given by Dr. Andy Feinberg, who is a professor at Johns Hopkins University School of Medicine Center for Epigenetics and the winner of the AMP award for Excellence in Molecular Genetics.
Dr. Feinberg spoke on “The Epigenetic Basis of Common Human Disease.” He defined epigenetic changes as stable, heritable, modifications of the genome that are not based on actual sequence changes. These epigenetic changes can be modifications of either the DNA, or the DNA associated factors that are maintained through cell division. Examples include DNA methylation, particularly at CpG islands, histone tail modifications, nucleosome remodeling and changes in higher order chromatin structure (such as compaction).
Stochasticity mosaic painting by Andy Feinberg, after a portrait of Conrad Waddington by Ruth Collet,
was featured on the cover of Nature Genetics May 2017 Volume 49 No 5.
In earlier work, Dr. Feinberg’s group developed the CHARM method (Comprehensive High-throughput Arrays for Relative Methylation) to take a comprehensive look at the cancer epigenome. Methylation of CpG islands is associated with transcriptional gene silencing, and removal of methylation can lead to hyper gene expression. The study was an unbiased screen looking for CpG islands and other genomic loci that are differentially methylated, and the outcome of the study was identification of Differentially Methylated Regions (DMRs). What was surprising was that some of the same regions that are differentially methylated in different tissue types, and help make liver tissue different from brain tissue and skin tissue, are the same regions that are differentially methylated between a normal tissue and its cancerous counterpart. Even when comparing the methylation pattern between the same tissue in inbred mice, with homogeneous genotype and identical living conditions, these DNA regions showed high variability between individuals in the level of methylation. Dr. Feinberg’s group hypothesized that what was critical about these regions that were differentially methylated in different tissues, different disease states, and different individuals was their variability. As their studies progressed, they named these regions Variably Methylated Regions (VMRs) and noted that they are not limited to CpG islands, but they are highly enriched around genes important in development and pattern formation.
Dr. Feinberg described how he began to formulate the theory of stochastic epigenetic variation as a driving force of development and evolutionary adaptation. He hypothesized that for complex, multicellular organisms, stochastic variation plays a fundamental role during normal development. While this stochastic variation is mediated epigenetically, the degree of stochastic variation is inherited genetically. Genetic variants that increase epigenetic plasticity may have an advantage, especially in a changing environment.
He went on to say that stochasticity is a defining feature of cancer and aging. Cancer is an environmentally influenced disease, driven in part by factors such as smoking, diet, and sun exposure. Repeated changes in the microenvironment may select for increased plasticity and promote cancerous transformation. Of note is the fact that in cancer, there are large hypo-methylated blocks of DNA sequence that not only result in increased expression levels, but also coincide with VMRs. He showed that repeated sun exposure in elderly skin shows methylation changes in VMRs. These regions of increased plasticity affect genes that give a growth advantage, such as genes in the oxPP pathway that allow cells to be resistant to oxidative stress.
How does all of this fit into a model to explain cancerous cell transformation? Dr. Feinberg painted a picture where development is like a ball rolling down a series of grooves. If the grooves are deep, it is very hard for the ball to jump out of one groove and start rolling down a different groove (i.e., stochastic events are rare). But in cancer, stochasticity is increased and one can imagine the same ball rolling down very shallow grooves, where change to a different fate is more likely. How is stochasticity increased in cancer? First, aging, environmental exposures, and/ or mutation may cause genetic changes (possibly the “variants of unknown significance” so commonly observed in tumors) that affect the epigenetic modification of a gene’s enhancer or promoter for example by disrupting CpG islands. Second, these changes may affect the expression of “epigenetic mediator genes” which can be thought of as “flattening out” the landscape and making stochastic events more likely. Increased expression of mediators such as OCT4, NANOG, and SOX2 can act to increase pluripotency of cancer cells. Finally, the unstable epigenome allows for natural selection to select for changes with a growth advantage, which leads to tumor evolution and metastasis.
Dr. Feinberg stated in his talk that “common variants explain only 1-10% of common diseases;” his model of stochastic epigenetic variation captured my attention as a plausible way to explain at least some of the other 90% of common disease.
In addition to the featured sessions at AMP, we always enjoy the poster talks. This year at AMP, our products and technology were featured in a number of different posters ranging from informatics to technical topics. We've pulled together a few of our favorites for you, click here to download for free. 
