Going to the Dogs

Until she discovered dogs. "One day I was standing knee deep in water, videotaping a bunch of dogs that were performing water rescue trials and wondering whether the trait was or was not genetic, when I realized: I'd finally found something that consumed me."

Fueled by that consuming passion, Ostrander went on to build high-density maps of the dog genome, and she led the charge to get the canine sequenced. Ostrander and her colleagues now use this information to identify genes responsible for interesting canine characteristics, such as body size, and for diseases particularly prevalent in pooches, including various types of cancer. "Almost anything big that comes out in dog genomics, Elaine was somehow involved," says Heidi Parker, who did her graduate work with Ostrander at the Fred Hutchinson Cancer Research Center (FHCRC) before following her to her current digs at the National Human Genome Research Institute (NHGRI) in Bethesda, Md.

Humans and dogs are susceptible to many of the same malignancies, including heritable forms of lymphoma, osteosarcoma, and kidney cancer. "And the ability to take advantage of well-documented lineages maintained by constrained breeding makes the dog a powerful system for dissecting out genetic contributions to disease," says NHGRI's Eric Green, who recruited Ostrander to head the NHGRI cancer genetics branch in 2005. "Elaine has carved out a very creative and productive scientific niche with her work on dog genetics and on human cancer, and has emerged as one of the dominant leaders in both fields."

"If Elaine hadn't been around, I think the dog would be a second-rate model that was studied by a handful of people who were only interested in pets," says Robert Wayne of the University of California, Los Angeles, one of Ostrander's many collaborators. "Now the dog has made it on the charts as a unique species that provides special insights into evolutionary biology and also into human disease."

PET PROJECT

Ostrander stumbled onto her work on dogs almost by accident. After finishing her postdoc at Harvard, Ostrander decided she wanted to work on plants. Arabidopsis biologists had recently identified homeotic mutants that had 30 layers of petals instead of one, and Ostrander thought that plants offered a good system in which to address interesting developmental questions. "So I wrote for an NIH fellowship, which got funded, and identified a really outstanding lab at UC-Berkeley, where I was going to do a second postdoc," she says. "But when I got to Berkeley, I discovered there wasn't going to be bench space available for several months. I had already given up my apartment, and more importantly, my parking space. Which, in Boston, forget it! So there was no going back."

Instead, she wandered over to Jasper Rine's lab. He had become interested in making a map of the dog genome. "So I said, 'well, if you'll pay me, I'll come for six months and I can make the libraries and start developing genetic markers. I can get this project off the ground.'" Which is what she did. After going to dog shows, collecting pedigrees, and watching dogs herd, swim, and generally do the things they do, she says, "I gave NIH back their fellowship and I just stayed."

While at Berkeley, Ostrander identified and pinpointed the location of microsatellite sequences - di-, tri-, and tetranucleotide repeats - that could serve as genetic markers. By 1992, she and her colleagues had mapped about 25 such markers. By 1995, after she'd moved her lab and the project to the FHCRC in Seattle, they'd found and published 101. ("Yes, we are a witty people," she laughs.) By 2002, Ostrander thought the dog was ready to compete for the funds to get sequenced. "So I wrote the white paper and it was successful on its first try," she says. With the help of researchers at the Broad Institute, where the sequencing was performed, the dog genome was unleashed in Nature in 2005.

"It took a lot of vision to see the potential of the dog as an interesting model for genetic studies," says Leonid Kruglyak of Princeton University, another of Ostrander's collaborators. "But then it took a huge amount of heavy lifting over many years to actually make it happen, because there weren't any genetic resources. Laying the foundation to getting the dog sequenced, building genetic maps - that required someone with a real stick-to-it-iveness, someone who could keep the faith. I don't think the dog genome would have happened without Elaine, or certainly not as quickly or as effectively as it did."

Wayne agrees. "She single-handedly got the dog on the agenda as one of the first species to get sequenced," he says. In addition, she has organized many meetings on dogs and dog genetics and set up a huge number of collaborative efforts to make the most of the resulting information. Along with Wayne and Carlos Bustamante of Cornell, Ostrander has established a canine "CanMap" project that, similar to the human HapMap project, will map out variations in dogs and attempt to relate those polymorphisms to different phenotypes and behaviors. "That project involved over 1000 samples from 115 or 120 breeds," says Wayne. "Carlos and I would be lost doing that sort of thing on our own. But Elaine has kept us on track and provides the enthusiasm behind the whole thing."

WHEN SIZE MATTERS

In addition to leading large collaborations and developing resources for others in the field, Ostrander and her team have mapped a number of disease genes, including a gene involved in kidney cancer in German Shepherds that causes a very similar syndrome in humans, as well as the border collie equivalent of a human eye disease called retinitis pigmentosa. They're also using the dog system to get a better handle on bladder cancer. In people, Ostrander says, "smoking is a big confounder. So it's been difficult to study the genetics of bladder cancer susceptibility in humans. In dogs the picture is much clearer."

At the same time, she and her colleagues have become experts at understanding the genetic relationships between one dog breed and another. They found that Asian breeds like the Lhasa Apso and the Pekinese share a common ancestry, as do Mastiff-like dogs such as the Bulldog, Boxer, and St. Bernard. That work appeared in Science in 2004.

Then there's the issue of body size. "I think the first slide in every seminar that every dog geneticist shows has a picture of a Great Dane and a Chihuahua or an Irish wolfhound and a toy poodle," says Ostrander. "So the question has been there forever: What makes big dogs big and little dogs little?"

Working with Wayne, Bustamante, and Gordon Lark at the University of Utah, Ostrander and her team set out to find out. Using the DNA samples and careful measurements that Lark had collected from a large collection of Portuguese water dogs, Ostrander and her team first determined that a gene key for small body size lies within a region of a couple million base pairs on chromosome 15. They then analyzed DNA from dozens of large and small breeds, which narrowed that region to about 70,000 base pairs - an interval that coincided exactly with the introns and exons of IGF-1, a gene that had previously been pegged as important for determining body size in mice.

"It was like a finger pointing at IGF-1," says Cornell University's Nathan Sutter, who led the study as a postdoc in Ostrander's lab. "When I saw that signature, I did get up and do a happy dance."

Ostrander, too, was pleased, but she didn't join in the happy dance. "I lost 50 bucks on that result because I was sure it was going to be growth hormone," she laughs. "And growth factor may still turn out to play a role. But when you look at what's been under selection by breeders to make dogs small, it's clear that IGF-1 jumps out as the number one starter out of the gate." Indeed, of the 14 small dog breeds they examined, each had the identical pattern of variation around IGF-1, findings that made the cover of Science in 2007.

Such top-tier publications are well-earned, says Jaime Modiano of the University of Minnesota. "Elaine is tireless and her standards are incredibly high," he says. "Every letter of every word has to be perfect. That's why she's publishing in top journals. It's not just her reputation, it's the quality of her work.

"She's also a leader and a facilitator," he says, and she makes sure that her trainees and collaborators also get things done. Sutter agrees. "Whether you're sick and tired of collecting sequence data or your paper needs to go through another draft, she'll work with you to keep it moving along," he says. The same is true for papers that don't originate in her lab. "We had done a piece of research on Addison's disease in Portuguese water dogs," says Lark. "The journal thought the paper needed more work and I was ready to let it go, but Elaine said, 'No way!' She was prepared to see that that paper got published no matter what. She's absolutely fierce." The paper was published in 2006.

But Ostrander is not motivated by beefing up her CV. She's motivated by a love of science, and of the animal. "She does what she does for very genuine reasons," says Modiano. "She loves dogs and she has been able to take this passion and convert it to an area of research where she's at the forefront. For her it's not about fame and fortune and recognition."

That point was driven home when Ostrander lost her 13-year-old border collie, Tess, in 2006. "I thought that because I knew so much about dogs and I understood everything that was happening to her, I would be prepared," she says. "But when she died, the 150 papers didn't matter, the awards didn't matter, the accolades didn't matter. None of it mattered. Because I'd lost a member of my family. Just like anybody else." That makes Ostrander all the more grateful to the owners, breeders, and kennel clubs who've shared with her their pedigrees, samples, and information about their own beloved family members. "This isn't research we could do in isolation, sitting in our laboratories," she says. "We're in a partnership with the dog-owning community, and we owe them our very best work."