How we’re wired

To uncover the genetic machinery that guides human development, Richard Young is mapping the intricate world of embryonic stem cells

Imagine that you’ve been commissioned to map how the entire North American power grid is connected and controlled, to figure out how a unit of power in, say, Kentucky ultimately hopscotches its way to California. You’ve been given a ream of high-resolution satellite photographs that can help reveal the grid. You’ve also developed a few gadgets that can monitor the flow of electrical traffic throughout the continent. But in order to really understand this entire ensemble of power, you must identify every control room, all electrical generators, and the jungle of power lines that connect every McMansion to every office park.

Richard Young Photo: Sam Ogden

Clearly, the task is insane. There are thousands of control stations, tens of thousands of electrical generators, and millions of miles of transmission lines—all interconnected and overlapping, sometimes looking more like a plate of spaghetti than a feat of engineering. How in the world can you determine the engineering behind this complex network, and how can you be certain that your picture is accurate?

For such a project to succeed, you’d need to create an entirely new technological platform. Now, shrink the entire continental power grid into the nucleus of a cell, and you’ve entered the world of Whitehead Institute Member Richard Young.

Swap the control rooms for regulatory proteins, the generators for the genes that are the engines of the cell, and the grid for the molecular network that guides our genome throughout the entirety of human development, and you’ll start to grasp the scope of the challenge that faces scientists like Young, researchers who are tackling the circuitry that makes us who we are.

The genome is a seemingly endless terrain whose building blocks—chemicals called nucleotides represented by the letters A, C, T, and G—number about three billion pairs per cell. Identifying how all genes and proteins are networked is like mapping the entire grid from a pile of aerial photos.

But over the last few years, Young’s team of biologists and computer scientists has been pulling it off. A project that began with the yeast genome has now scaled up to human embryonic stem cells, and for the first time we are starting to understand the complex wiring that provides these cells such astonishing elasticity.

And while a comprehensive schematic of the power grid helps engineers troubleshoot breakdowns in the system, an embryonic stem cell counterpart of such a schematic might just provide a roadmap for treating diseases that are currently untreatable.

The human challenge

In the winter of 2005, life in the Young lab reached a level of intensity that tested the endurance of some of the most tried-and-true postdocs.

For years, Young’s lab had been working with MIT computer scientist David Gifford to develop a technology platform that could sweep the genome and locate all transcription factors—the proteins responsible for switching genes on and off. While genes contain all the information that defines an organism, transcription factors decide when and where and how much of that information gets put into action. Just a few decades ago it would have taken one scientist 100 laboratory years to understand how one transcription factor functions throughout the mammalian genome. This new platform, however, eventually would shrink a century into less than a month.

Having succeeded spectacularly in mapping every transcription factor throughout the yeast genome, the team was now finalizing the first experiments in which they’d scaled the platform up to human embryonic stem cells, a project that also required the computational expertise of Whitehead’s bioinformatics and IT groups. Not only do human cells add many more levels of complexity, but the team was also rushing to get these results published, suspecting competition from other labs.

“We’d be working through the night, taking turns buying food, eating mostly donuts and Wendy’s burgers,” says postdoctoral scientist Matthew Guenther. “I can remember one frantic 5:00 a.m. search for a Dunkin’ Donuts.”

The deadline pressure made the whole project feel like an episode of the TV show 24, recalls fellow postdoc Richard Jenner. “I’d sometimes get home and watch the show,” he recalls. “Seeing the characters dash between each others’ computers as the time ticked away was eerily similar—minus the hidden nuclear warhead.”

The result of these torrid labors was a map of the three principal transcription factors in the embryonic stem cell genome, an initial circuitry diagram of the wiring that makes a stem cell a stem cell. It’s as if they’d scoured the satellite data, located the three most essential power stations on the grid and determined how these stations communicate with each other. It was not the entire network, but for the first time scientists could see, on a global scale, what makes a stem cell unlike any other kind of cell.

Under the hood

The three transcription factors that the group located are Oct4, Sox2 and Nanog. Of these three proteins, Oct4 stands out as the primary conductor, the one ultimately responsible for an embryonic stem cell’s most tantalizing characteristic: its ability to become any type of cell in the body, a trait known as pluripotency.

“Oct4 is key to stem cells,” says Richard Young. “In fact, it’s the marker that the scientific community uses to demonstrate that is has a mammalian embryonic stem cell. But the problem was, no one knew what it did.”

Collaborating with fellow Whitehead Member Rudolf Jaenisch and MIT’s David Gifford, the team found that Oct4, Sox2 and Nanog preside over a vast network of developmental genes. Since these genes are responsible for guiding a cell’s fate down a particular developmental path, the primary job of these transcription factors is to keep these networks silent. In that sense, these transcription factors endow an embryonic stem cell with pluripotency not through what they make the cell do, but through what they keep it from doing.

Getting to this point required marrying a tried-and-true lab technique with some modern biotech inventions.

For the traditional technique, called “chromatin immunoprecipitation,” researchers take an unsuspecting embryonic stem cell carrying out its normal functions and instantly freeze it with a few drops of formaldehyde. Adding the chemical essentially stops time in the cells; all molecular activities turned suddenly into inanimate statues. The transcription factors are forever locked onto whatever genes they may have been tinkering with at the moment.

Next, the nucleus of the cell is put through a genomic shredder and broken up into millions of DNA fragments, with the transcription factors still fastened onto their gene segment. The researchers then introduce antibodies to this solution. These antibodies are designed to seek out and latch onto Oct4, Sox2 and Nanog. Each antibody is equipped with a magnetic bead, enabling the scientists to then draw the antibody, the transcription factor and its gene fragment out of the solution once they have latched.

Once removed, the antibodies and their magnets are removed from the transcription factors and the gene fragments are labeled with fluorescent tags.

This technique works beautifully for studying one gene, or protein, at a time. But to make the procedure work across the entire genome, Young uses one of biotechnology’s hallmark inventions: the microarray.

Array for the genome

For the last ten years, microarrays have been a laboratory staple for anyone studying genomics. The first scientific paper to report use of the array was published in 1995, and in 1997 researchers described using microarrays to profile an entire eukaryotic genome—in this case, baker’s yeast.

Roughly the size of a stick of gum, a microarray is a small glass slide covered with tens of thousands of DNA fragments, each fragment corresponding to a particular gene. Researchers use the microarrays to measure levels of gene expression. The gene pieces covering the microarray all correspond to specific genes in our cells. It is designed this way so that a scientist can place the microarray into a scanning system with desktop computer software that cross-references the segments to the actual genes. This could tell us, for example, that a particular scrap of DNA on the upper-left-hand corner of the array is taken from gene “A” located on “this particular region” of the X chromosome.

In the Young lab, the fluorescently tagged gene fragments are then released onto the microarray. One characteristic of DNA that makes it so efficient to work with is that it naturally seeks out its complementary sequence. When the gene fragments land on the microarray, they adhere to the matching bits of DNA, and the fluorescent tags allow the researchers to locate where on the array they’ve settled. Once this information is fed into the computer, the scientists can determine where—throughout all 3 billion nucleotide pairs in the genome—each transcription factor was located at the moment formaldehyde froze the cell.

The data is then integrated at David Gifford’s team at MIT’s Laboratory for Computer Science.

“This step is essential given the sheer volume of data that these microarrays produce,” says Young. “We work with thousands of different cells at various time points from a variety of animal models. The Gifford lab is brilliant at integrating all the different data points to build a picture of what the possibilities are. They can assign a probability to any point of the network based on its past performance. This type of information assures that we’re presenting an accurate picture of how the cell is regulated.”

But who regulates the regulators?

The ultimate hope for many embryonic stem cell researchers is to one day create customized therapies for patients without using a human egg cell. For example, imagine that a patient with leukemia visits a clinic. The clinicians scrape off a skin cell from the patient’s arm, then chemically guide the cell to de-evolve back into an embryonic stem-cell state. The cell is then redirected down a particular developmental pathway into a hematopoietic, or blood, stem cell. These new cells are finally placed into the bone marrow and the patient is cured. And because the new cells are derived from the patient, the patient’s immune system doesn’t react.

Before this becomes a reality, we must thoroughly understand two major aspects of cellular biology: First, we need to know what’s happening during somatic cell nuclear transfer, or therapeutic cloning, in which an egg cell reprograms a donated nucleus. Second is understanding how the genome guides cells down specific developmental pathways.

Rudolf Jaenisch and other scientists around the world are pursuing the first task, producing preliminary proof of concept experiments. “We know that when an egg cell reprograms a donated nucleus during nuclear transfer, it doesn’t work some kind of magic,” Jaenisch says. “It’s a biochemical process, one that we can learn, and one that we can most likely reproduce in the lab.”

Young and his team are working on the second aspect, focusing on the mechanisms that guide a stem cell into forming differentiated, mature tissue.

“We want to understand the wiring in an embryonic stem cell so that we can know exactly what transpires in that cell as it develops into a neuron or a blood cell or a pancreatic islet,” he says.

This spring, Young, Jaenisch and colleagues reported in the journal Cell how a network of developmental proteins called polycomb controls the embryonic stem cell genome.

While all cells—including stem cells—share the identical genome, each cell type is only granted access to a select group of genes specific to that cell. All other genes are kept behind a firewall. Polycomb ensures this by tagging the genome with a chemical marker that prevents access to genes. However, polycomb had never been thoroughly studied in embryonic stem cells before.

The researchers reported a significant overlap of polycomb and the key transcription factors Oct4, Sox2 and Nanog in the human embryonic stem cell genome. These parts of the genome where the transcription factors and polycomb converge encode genes involved in the control of most of human development. This suggests that these scientists have found themselves at the core of the regulatory circuitry that controls human development. Because there are close links between the control of development and disease, the discovery places us closer to an understanding of the regulatory dysfunctions involved in many diseases.

Perhaps ten years from now, Young suggest, we’ll pair our new knowledge of regulatory circuits with therapeutics that can act on those circuits when they break down.

“We may not cure Alzheimer’s in the near future,” Young says, “but I’m confident that as we continue to unravel and map human wiring, we’ll be shocked at the range of new possibilities for attacking disease.”

That’s what lies ahead as we continue to map the vast power grid tucked inside of every one of our cells.

****************