New brain map on tap

With the backing of a billionaire, researchers today launched a project that builds on their earlier atlas of the mouse brain and goes after a challenge 2,000 times bigger: a 3-D genetic map of the human brain. And that's not all: They're planning to produce a similar map of the mouse spinal cord, as well as another atlas showing how the mouse brain develops from the fetus to adulthood.

The multimillion-dollar effort could help researchers develop new treatments for maladies ranging from spinal cord injury to autism.

Today's triple play marks a new phase for the Seattle-based Allen Institute for Brain Science, which was founded in 2003 with $100 million in seed money from software billionaire Paul Allen. The first phase of the Allen Brain Atlas focused on the mouse brain - and looked specifically at which genes were active in which areas of the brain.

Genes in the brain serve as chemical switches, providing the instructions for making proteins that have various effects on brain chemistry. For example, a little oxytocin in just the right place gives you a warm, fuzzy feeling. A little less of a protein called p11 could leave you feeling depressed.

Medical researchers can use information about gene expression to figure out the biochemistry behind activity in the various regions of the brain - and the institute says the information in its 180-terabyte online atlas gets a lot of use.

Since the mouse brain atlas was validated in a paper published in Nature little more than a year ago, the research has been cited more than 100 times in other scholarly papers, said Allan Jones, the institute's chief scientific officer. The Web site that houses the brain data records 10,000 unique visits per month, and about 500 of those visitors spend more than an hour each on the site monthly, he said.

Mouse studies can often be applied to human biology as well - and indeed, one of the high-profile spin-offs from the mouse atlas is a Pentagon-funded study on sleep deprivation that may lead to battlefield applications. But there's nothing like having the real thing, particularly when you're studying the brain. That's why the institute is moving ahead with the human brain atlas.

The Allen Institute already has a head start on the human brain, thanks to its studies of gene expression in the human cortex. Today marks the official beginning of a four-year campaign to characterize gene activity in the entire human brain.

Jones said the institute spent about $41 million to create the mouse brain atlas, and about half of that work can be leveraged for the new project. However, he estimated that completing the human brain map would require $55 million more, spread over four years.

"The human brain is 2,000 times as large as the mouse brain," he observed. "The first thing that you're faced with, right out of the gate, is that it's 2,000 times as big."

Breaking down the process

To handle the scientific challenge, the brain-mapping project will be done in two phases.

In the first phase, the human brain will be broken down into about 2,000 smaller structures per hemisphere. Fresh-frozen samples from up to 10 brains, selected from tissue banks around the United States, will be analyzed to produce an inventory of genes specific to each structure. Jones said the process would narrow down the focus from a total of 20,000 genes to between 50 and 500 genes per structure.

Then, researchers will build up a fine-resolution database pinpointing which high-value genes are turned on, right down to the cellular level.

When researchers combine the gene-expression database with insights from other brain-mapping techniques, they should be able to figure out the biochemistry behind typical brain function. And that could help them figure out how to fix glitches in the black box known as the brain.

"These are fairly low-resolution maps that tell you the area ... but you're missing key information about what's going on there at the biochemical level," he told me. "We will be getting to that information with this human brain atlas that we're creating. We're actually going to be able to see what genes are turned on in these particular areas of the brain."

Studying diseases ... and evolution, too

Jones said potential subjects for study include autism, epilepsy, schizophrenia, Parkinson's disease, Alzheimer's disease, multiple sclerosis, spinal cord injuries - and traumatic brain injuries, which affect a significant number of Iraq war veterans.

"There are clearly lots of interesting questions," Jones said.

For example, researchers have spotted subtle differences in gray-matter distribution that may hint at a mechanism behind autism. Using a map of gene expression, researchers might be able to figure out which proteins come into play in the target regions - and come up with new drugs to address the chemical causes of autism.

It will take more than the atlas alone to do such research, said David Anderson, a biology professor and Howard Hughes Medical Institute investigator at the California Institute of Technology.

"The atlas is a starting point," said Anderson, who is a member of the institute's scientific advisory board. "The atlas will identify a set of cell markers that can be used, for example, to identify different regions of the amygdala, which may be implicated in autism. Now what you have to do is take those markers and go out and obtain postmortem samples of autistic brains and age-matched normal brains, and take these markers in parallel and compare them."

The atlas could point to, say, 10 significant markers in the amygdala for comparison. "That would save you the trouble of looking at 4,000 markers," Anderson said.

Anderson said the atlas also could come in handy to assess the evolutionary differences between human brains and the brains of other species, ranging from mice to monkeys. "We could ask whether our brains simply have more of the same kinds of cells, or if we have fundamentally different types of cells in different arrangements," he told me.

More about the mice

In addition to the human brain atlas, the institute plans to delve more deeply into mouse biology. A two-year, $15 million project will produce gene-expression maps for mouse brains at different stages of development, ranging from early formation to adulthood. This would help researchers see how gene expression changes over time.

"Instead of a 3-D atlas, effectively we're going to have a 4-D atlas, where we'll be able to watch movies [showing] data-driven models of the brain as it's growing," Jones said.

The project would focus on 3,000 to 4,000 of the "more interesting" genes thought to be linked to developmental disorders, he said.

Yet another project would target the mouse spinal cord. Jane Roskams, an associate professor at the University of British Columbia's Brain Research Center, likened the spinal cord to an interstate highway, where genes serve as the stop and go signals.

"You've got many roads leading to many different places, and in order to get to those places you have to have an address," said Roskams, who belongs to the International Consortium on Repair Discoveries and served as a scientific adviser to the Allen Institute. "The directions are encoded in the genes that are expressed."

Knowing how gene expression works could help researchers develop new drugs or cell-transplantation procedures to restore the signaling pathways in the spinal cord, Roskams said.

"Being able to restore those signals at each step of the way is the biggest part of the challenge," she told me. "If we lose cells through an injury to just one part of the spinal cord, we've lost the directions that the nerves need. ... This project will tell us those signals."

Work is already under way on the spinal-cord project, with financial assistance coming from a consortium of associations and foundations. The Allen Institute expects the atlas to be complete by the end of this year.

Free and on the Web

Jones emphasized that the data from the three new projects will be freely available over the Web, just as the mouse-brain database is today. Unlike most scientific projects, the Allen Institute doesn't hold back the raw data for its own big publication, but rather puts everything it has into the database as soon as it's available.

"These data sets are so massive that there's no way we can ever be comprehensive about these analyses, and it's better to get the data out there," Jones said.

The open-access nature of the atlas projects reflects the institute's status as a philanthropic organization rather than a business. But the scope of the project is too big for any one philanthropist, even if he's No. 41 on Forbes' list of the world's richest people.

With the ramp-up of the new atlas projects, the institute has set up a series of milestones and budgets so that it can plead its case to additional contributors, chief operating officer Elaine Jones said.

"It's run according to a business model - that's how people see it, so they're far more willing to invest in this type of a project," she said. "In their lifetime, they're going to see tangible results."

But potential donors are told right up front that their "investment" won't bring them future financial advantage. Instead, it's an investment in basic science, and in hope for future generations.

"When you lay that out for people, it's reasonable," Elaine Jones said. "They understand what they're getting in return."

Update for 3 p.m. ET: The three new atlas projects are likely to streamline the process of drug discovery, said Genentech's Marc Tessier-Lavigne, a scientific adviser to the Allen Institute.

Tessier-Lavigne, who is Genentech's senior vice president for research drug discovery, told me in a follow-up phone call that the already-existing atlas of the adult mouse brain has helped researchers search for the genetic clues that could lead to new drug targets. The new efforts will widen the search to include the adult human brain, as well as spinal cord science and developmental issues in mouse models.

"Those are the three clicks of the [computer] mouse that we'd like to be able to make, and we can't make them today. But with this new project, we will," he told me.

So how long will it be before the brain atlases yield new treatments? "Typically, it might take three to four years, starting from a target, to get a drug candidate ready for clinical development," Tessier-Lavigne said. Going through the clinical trials and actually marketing the drug could take another eight to 10 years, he said.

But another track for drug development could take much less time, and that's "what is wonderful about 21st-century biology," he said. A comprehensive atlas of the brain might point up some gene-specific targets that are already familiar to researchers. For example, a drug already being developed for cancer treatment might have an effect on gene expression in the brain as well.

"This type of information may reveal the involvement of various mechanisms in brain disorders that are already being studied in other areas," Tessier-Lavigne said..