The Womb’s Strange Epigenome

An impossibly long string of DNA letters floats inside each of our trillions of cells. Somehow, incredibly, that one string produces a huge variety of tiny biological machines, from red blood cells that carry oxygen, to brown fat cells that generate heat, to neurons that fire electrical messages. How does one* fixed code manage all that?

The answer is, it’s not fixed. There’s epigenetics: In response to changes in the environment, stuff (chemicals and proteins) interacts with the DNA and affects how the code is interpreted. For example, a methyl group (one carbon and three hydrogen atoms) can latch on to the genome and help switch on or off the expression of the genes below. So in fat cells, say, some suite of genes is methylated, but in brain cells, it’s a different group of genes.

Until fairly recently, research on this so-called DNA methylation focused on places in the DNA code where a G immediately follows a C — CpG sites — which are known hotspots of methylation. When hundreds or thousands of CpG sites appear in a row, it’s called a CpG island, and this intense methylation usually silences the underlying gene. But in the past few years, DNA sequencing technology has allowed researchers to look at methylation patterns not just at the glittering islands, but across the entire genome.

It turns out that at least 70 percent of the genome is covered by methyl groups. And the pattern of methylation across the genome, known as the epigenetic landscape, is more varied than ever expected. But as researchers create more and more elaborate epigenetic maps, they still don’t know much about where those maps might lead us.

This week, a study in the Proceedings of the National Academy of Sciences describes a little-known feature of the epigenetic landscape: a ‘PMD’, or partially methylated domain. PMDs are long stretches of DNA that have markedly little methylation compared with the rest of the genome. Scientists had spotted PMDs before, but only in cancer cells and cells cultured in the lab. The new study looked for PMDs in several human specimens, such as brain, blood and kidney cells, but found them in only one type: placenta.

“Placenta has a completely different landscape in its methylome from other adult tissues,” says lead investigator Janine LaSalle, a professor of immunology at the University of California, Davis. What the placental landscape does, and why it’s different from other tissues, is still a big mystery. But the finding is provocative for researchers (including LaSalle) who are interested in finding biological markers that predict whether a baby has a developmental disorder.

As the interface between a developing baby and its mother, the placenta is crucial for fetal growth and development. Since this nine-inch, one-pound organ is typically discarded at birth, it would be relatively easy (compared with brain tissue, say) to sequence its methylome. An epigenetic glitch in the placenta could be a marker of an environmental exposure or dietary change in the mother. “There is a really growing concern that the placenta is not a protective barrier but more like a sieve for maternal exposures that affect the fetus,” LaSalle says. Alternatively, an abnormal placental methylome might be a sign of a global problem in methylation, which could also appear in other tissues, such as the brain.

LaSalle’s team analyzed the methylomes of 56 placental tissues and found that PMDs cover some 3,800 genes in the placental genome. “Some PMDs appear like ‘footprints’ over a single gene, but some span large clusters of related genes,” she says.

Genes within PMDs are switched off, the study found. Intriguingly, some genes that are in PMDs (and switched off) in placental tissues are actually in hypermethylated regions (and switched on) in other tissues, suggesting that these genes help define the function of the tissue. For example, LaSalle found hundreds of genes that are turned on in placenta but off in brain and lung. Many of these genes are involved in immune responses, which might be important for protecting the fetus during the mixing of fetal and maternal blood.

More than 1,000 genes lie outside of PMDs (and are thus turned on) in both placenta and brain, the study found. “That group suggests that for a large number of genes, placental differences in methylation may be a good surrogate for brain differences,” LaSalle says. Her group has started a study that will compare these epigenetic marks and others from placentas of children who develop autism and from those who don’t. It will take awhile, though, as most children are not diagnosed with the disorder until age 2 or 3.

Before the new study, nobody was sure whether PMDs showed up in normal tissue, or were just an experimental artifact of studying cells in the lab, says Joe Ecker, a professor of plant molecular and cellular biology at the Salk Institute. Ecker’s lab was the first to find PMDs, in cultured lung cells, in 2009. “The interesting part now is, what are these regions?” Ecker says.

This is but one of many questions that will surface as researchers chart the methylome in different kinds of cells. Ecker notes that the NIH’s Roadmap Epigenetics Project is in the process of sequencing complete epigenomes (which include not only the methylome, but patterns in the way DNA wraps itself around proteins, which also affects gene expression) from 100 tissue types, both embryonic and adult. “These landscapes, we’re just beginning to understand what they look like, let alone interpret what they mean,” Ecker says. But that’s what makes the field exciting. “We haven’t been disappointed so far.”

*

Alright, alright, it’s not exactly one fixed code.

(Fun fact from the annals of science: The term ‘epigenetics’ was coined in 1942, long before Watson and Crick figured out the structure of DNA.)