The Epigenetic Landscape

When it comes to investigating the origin of a disease, it’s common to first look into the realm of genetics – but a vital piece that can help us tackle complicated puzzles is often missing: epigenetics. Although still a relatively young field, research proving its significance in understanding disease origins has led to growing interest.

“Whole journals have sprung up on the issue of epigenetics,” says Andrew Feinberg, Professor at Johns Hopkins. “It’s a very young field, so there are many new things coming up, and people are still finding their way around and developing the best ways of investigating.” The development has been facilitated by big changes in technology and mathematics, fields in which Feinberg started his career. He believes that, even though rapid growth has led to teething problems (for example, incomplete information), there’s still a great deal of useful knowledge to be gained.

Early days of cancer epigenetics

Not everyone agrees, however. Despite the strides being made in epigenetics and the increasing research supporting its importance in disease, there are still many scientists who disagree on the field’s usefulness. Feinberg welcomes the criticism and says, “There’s a joke that circulates about epigenetics being an ‘epi-phenomenon.’ It’s healthy for people to push back and make the epigenetics community focus on clarity, what’s uniquely epigenetic, and also how epigenetics and genetic alterations relate to each other.”

This skeptical reaction to epigenetics was very much the case in the early days of cancer genetics, because of the spectacular advances in the purely genetic understanding of oncogenes, for example. But Feinberg and others felt that epigenetics provided a functional context to the understanding of cancer biology because of its strong relationship to gene expression and because cancer, in many ways, represents normal gene expression gone awry.

Feinberg and Vogelstein’s first experiments on altered DNA methylation in human cancer took place in the early 1980s and showed that these epigenetic changes were ubiquitous across tumor types and occurred from the earliest stages of malignancy. From there, Feinberg went on to study Beckwith-Wiedemann syndrome (BWS), a congenital overgrowth disorder that increases the risk of a form of pediatric kidney cancer known as Wilms tumor. With collaborator Michael DeBaun, he showed that the several hundred-fold increase in cancer risk in BWS is caused by an epigenetic abnormality, loss of imprinting, present at birth (1). Wherever the imprinting defect was present, it caused a precancerous condition in the kidney – the smoking gun showing that epigenetic changes can cause cancer.

Painting a landscape

“Conrad Waddington coined the term ‘epigenetic landscape,’ which was the concept of a pluripotent cell committing to different lineages, becoming progressively more differentiated, like a ball rolling in several different potential pathways down a curvilinear landscape,” says Feinberg. “That was a metaphorical point Waddington brought up, but what my colleagues John Goutsias, Garrett Jenkinson, and Elisabet Pujadas, and I have derived is a real epigenetic landscape using principles of statistical physics that embody stochasticity. (2)” In effect, cells take different epigenetic routes to reach their final differentiated states – routes that can be altered by the effects of heredity, external factors, and randomness. This epigenetic variability is leading to new diagnostic and predictive measures that embody variance, not just mean changes – for example, in DNA methylation.

Feinberg and his colleagues have also branched out into studying the epigenomics of common disease more generally. “Dani Fallin, an epidemiologist at Hopkins, and I started to investigate the idea that maybe the epigenome is the funnel through which gene-environment interaction takes place to cause disease,” he says. “People don’t usually get cancer, or most other diseases, based entirely on genetic risk. I think that idea of a purely genetic origin is generally too limiting in genetic testing and in projects looking for risk markers for disease.” To prove his point, they worked on studies exploring the epidemiological aspects of autoimmune and neuropsychiatric disease (3). He says that this line of research is still in its early stages in his and many other laboratories, but there are already important clues emerging, linking prenatal environmental exposure to epigenetic changes in the fetus, and linking hereditary genetic changes to epigenetic alterations affecting gene function in adults.

“It’s important to think about the life cycle in a more integrated way, like prenatal or early-life exposure. We need to see how the epigenome fits into that. More and more data suggest that early-life, prenatal, and even parental exposure is important to what happens in adult life – which is not the same as Lamarckian inheritance, with which I largely disagree,” says Feinberg. “Dani Fallin and I, along with my son Jason, who works with Dani, showed a link between paternal sperm DNA methylation changes and autistic features in children born to mothers who had already delivered a child who developed autism (4).” He notes, “This could be caused by a prenatal exposure, or by a genetic predisposition affecting DNA methylation. Most importantly, the work needs to be independently replicated by other groups to know if it holds up.”

Expanding horizons

There are many elements to consider when it comes to funding disease research, but Feinberg believes that investigators often overlook epigenetics. Given the massive potential benefit the field could offer, he feels that epigenetics could be incorporated more formally into the overall study of human disease risk. We need to think about gene-environment interaction in particular, and how epigenetics stands right at the center of the two in terms of predicting and mitigating human disease.”

Feinberg’s work has continued to show the relevance of epigenetics; an investigation with Christine Iacobuzio-Donahue demonstrated that pancreatic cancer progression is linked to metabolism (3). Feinberg explains, “In that paper, we show that the metastases in pancreatic cancer are driven by genomic regions of change from heterochromatin to euchromatin. These regions show increased plasticity of gene expression, which allows for natural selection of the metastases in the absence of any driver mutation. We’ve proven that metastatic progression is driven by epigenetic changes that arrived within those primary lesions.”

As epigenetics grows and gains more traction in terms of its epidemiological value, there’s still a lot of room for knowledge – but that information won’t help patients unless more researchers are available to the field. Feinberg concludes, “I can’t emphasize enough the importance of cross-disciplinary training of young people going into genetics, epigenetics, and epidemiology. Because epigenetics is a comparatively young area, it is important that its practitioners of the future learn both the strengths and limitations of the field, as well as the statistical tools to distinguish the difference.”