A series of fine-tuned maps of DNA packaging in human cells reveal dynamic new views of how the genome’s instructions are carried out to build a person. The maps also offer surprising insights into what goes wrong in diseases such as Alzheimer’s and cancer.
The maps and discoveries made after examining them are being published February 18 in more than 20 scientific papers in Nature and affiliated journals by a large consortium of researchers involved with the Roadmap Epigenomics Project.
Researchers in the project cataloged chemical modifications of DNA and its associated proteins in 111 types of human cells, including embryonic stem cells and cells from several stages of development. Those chemical modifications, called epigenetic marks, include the attachment of molecules, such as methyl groups or acetyl groups, to one of the histone proteins around which DNA winds. Five different types of modifications to histone proteins were mapped in each cell type. Researchers also noted where methyl groups had been attached to the DNA base cytosine, and traced other DNA-packaging processes. These modifications don’t alter genes themselves, but affect how and when genes are used.
More than a decade ago, scientists compiled the complete set of genetic instructions of humans, the human genome sequence. Nearly every cell in the body contains an identical copy of this genetic instruction book. But the genome sequence alone is essentially a list of parts needed for building a person. The new maps supply the directions for snapping those parts together.
Flipping through the maps of each type of cell makes the genome come alive, says Eric Lander, a human geneticist and director of the Broad Institute of MIT and Harvard. “If the Human Genome Project was ‘Human Genome: The Book,’ then this is ‘Human Genome: The Movie,’” he says. Lander was a leader of the genome project, but was not involved in the new work.
Teams of researchers used various techniques to map the epigenetic marks and find active genes — genes from which DNA instructions are copied into RNA and ultimately into proteins — and to reveal which pieces of DNA are important for regulating this gene activity. Inactive areas of the genome were also charted. Some groups also explored the 3-D structure of the genome in different cells.
Researchers also pinpointed the cell types in which genetic variants associated with specific traits, such as cholesterol levels or blood pressure, were active. For instance, genetic variants that influence cholesterol levels were active in the liver. That result was expected because scientists already knew the liver produces cholesterol.
Researchers did not expect to find, though, that genetic variants associated with Alzheimer’s disease were more active in immune cells rather than in nerve cells in the brain. That result was confirmed by studies of mice. Researchers led by computational biologist Manolis Kellis and neuroscientist Li-Huei Tsai, both at MIT and the Broad Institute, mapped active and inactive genome regions in mice prone to developing Alzheimer’s. Many of the same regulatory patterns found in the mice’s brains matched the pattern in human brains, the team reports.
Previous research had suggested that immune cells in the brain called microglia play a role in Alzheimer’s (SN: 1/10/15, p. 12; SN: 11/30/13, p.22), but most researchers thought that neurons were the source of the disease. The new work suggests that genetic factors leading to Alzheimer’s don’t work in neurons at all, but manipulate microglia instead. “It doesn’t just say, ‘microglia are important,’ it’s saying ‘microglia are it,’” Kellis says.
The study also revealed that a master regulator called PU.1, a protein that turns on genes in immune cells, goes awry in Alzheimer’s. That protein may be a good target for new drugs against Alzheimer’s, Kellis says.
Another study examined cancer cells. Scientists thought that DNA-damaging mutations might strike active parts of the genome more often, because those stretches are more loosely packaged and might be more vulnerable. But the opposite was true. Cancer-associated mutations are more likely to occur in tightly packaged, inactive regions of the genome, John Stamatoyannopoulos of the University of Washington in Seattle and colleagues report. Mutations might build up inactive regions of the genome because DNA repair machinery can’t get in to fix the damage, he speculates.
The team compared the mutations patterns of the cancer cells with the epigenetic maps of 106 types of cells. By comparing those profiles, researchers could determine what type of healthy cell a cancer arose from.
For instance, melanoma is a type of skin cancer that arises from pigment-producing cells called melanocytes. The mutations in melanoma cells occurred in the parts of the genome that are normally turned off in melanocytes. The cancer’s mutation pattern didn’t fit the epigenetic profile of other types of skin cells, indicating that the cancer sprang from a pigment-producing cell gone rogue.
Doctors may be able to compile mutation patterns of a patient’s cancer cells and then match that pattern to epigenetic maps of various cell types to figure out the identity of the cell that originally turned cancerous. Such a diagnosis would be helpful for people whose cancer has spread throughout the body and its origin isn’t known.
Data compiled by the project are already available for scientists to use. The information will also be incorporated into a large international epigenome-mapping effort that will examine more cell types from a larger number of people, says Bing Ren, a molecular geneticist at the Ludwig Institute for Cancer Research in San Diego. “This research is crucial for us to understand mechanisms of disease,” he says.