Cracking the codes of human disease

Be they brain or bone, all the cells of our body carry the same DNA code or genome. Each cell type develops differently depending on how the code is read. Until recently, our knowledge of how to decipher those reading instructions has been patchy. Now a consortium of international scientists has delivered an instruction manual for more than 100 cell types. Known as the “Epigenomics Roadmap Program”, the results have been published in more than 20 papers in Nature and affiliated journals. The databank is already providing new insights into human health and disease.

The actual reading instructions employed by the cells are chemicals attached either directly to DNA, or to the proteins that keep it neatly wrapped inside the cell. Because these instructions lie “above” the genome, they are collectively referred to as the epigenome. They change the interpretation of the DNA text “just like annotations on a musical score sheet can lead to different interpretations of the same symphony,” explains Manolis Kellis from the Massachusetts Institute of Technology, senior author of one of the Nature publications.

These annotations are crucial for determining how cells develop and are also deepening our knowledge about the origins of disease. From cancer to brain degeneration, researchers are finding that when errors accumulate in the epigenome, illness can follow. “I am very excited. This is a huge advance in our understanding of how the genome is regulated,” says Sue Clark, who studies the epigenome at the Garvan Institute in Sydney.

Scientists first read the human genome in 2001. The task of decoding the epigenetic markers that tell each cell type how to read the genome is continuing, although our understanding of how the epigenome works has advanced considerably. The chemical marks that annotate DNA guide how the gossamer thin DNA thread is wound and packaged – the result is that some stretches end up more accessible than others. The attachment of methyl groups to DNA, for example, shuts down access, while adding acetyl groups to proteins called histones that package DNA makes these stretches more accessible. The final result is that the same sequence of DNA can be read quite differently in different cell types.

Since the early ’90s, researchers have been trying to trace these annotation marks as cells develop into one tissue or another, or as they change during disease. Tracking them across the more than 200 different cell types in the body and across three billion letters of DNA is a mammoth task. At first, different labs used their own methods, which made it hard for researchers to combine results. That changed with the establishment of the Epigenome Roadmap Consortium, initiated by the American National Institutes of Health common fund in 2008.

By dividing up the work and following common methods, researchers in the consortium have generated a reference map of the epigenome in 111 different tissue types. It’s a significant first step toward the internationally agreed aim of decoding at least 1000 epigenomes within the next decade.

“We now have a much richer picture of how the information in the human genome sequence is used and regulated, and how this varies throughout the human body,” says Ryan Lister, an epigeneticist at The University of Western Australia who is part of the consortium.

The Roadmap – which resembles a collection of 111 musical scores for different cell symphonies – has already been put to work. Importantly, researchers have been able to use this data to trace the origin of cancers. Up to one in 20 cancer patients have a primary cancer site that is unknown. While a tumour might be diagnosed in the liver, for instance, its origins often lie elsewhere and can be difficult to find.

Using the Roadmap datasets, Shamil Sunyaev and his team at the Harvard Medical School were able to use epigenetic annotations to predict the origin of 88% of tumours. That’s important because different cancers respond to different drugs. “The DNA sequence itself remembers the original cell,” he says. Now, analysis of a cancer epigenome will be able to tell doctors where the cancer came from and how to best treat it. Sunyaev is optimistic the technique may become routine in the clinic within two to five years. “This is where the future is,” he says.

Epigenome data is also revealing new insights into Alzheimer’s disease. The disease causes the death of brain cells. Microglia, the brain’s resident immune cells, have long been suspects at the scene of the crime. But are they the perpetrators or merely there to clean up once the damage had been done?

Kellis, Li-Huei Tsai of the Picower Institute for Learning and Memory at MIT, and their colleagues analysed the epigenome of an Alzheimer’s mouse model and found that as the disease progressed, epigenomic markers on the neurons of the mouse and its microglia changed – leading to microglia activation and neural shutdown. The researchers found a similar change in brain tissue from people who had died of Alzheimer’s disease. “We found strong evidence that immune cells might be the driving force behind the common form of Alzheimer’s disease,” says Kellis. “We think it happens early on in Alzheimer’s,” adds Tsai. Researchers may now look into treating Alzheimer’s by targeting immune cells.

Where the roadmap will take us next is unclear, but we know we’re on the right track. “The next 10 years are going to be exciting for the field of epigenomics and human disease,” says Kellis.