Connecticut’s new partner in science tracks the root causes of disease, mouse by mouse

Bar Harbor, Me. — Every two weeks, the scientists at The Jackson Laboratory get a delivery, the findings from the latest “deviant search.”

The caretakers responsible for the 1.25 million or so mice that populate the lab’s island campus at any one time take note of any critters that seem off, and send them over in boxes, notes attached, for the researchers to examine and divvy up.

Greg Cox

Greg Cox, who is studying muscular dystrophy and other neuromuscular diseases, with lab mice (Photo by Robert F. Bukaty)

“It’s like Christmas,” said Greg Cox, a scientist at “Jax,” who studies neuromuscular diseases. He’s perpetually on the lookout for thin, wasted or paralyzed mice — “sort of the sad mice” — that might have mutations that could lead him to the genes behind diseases like muscular dystrophy and ALS.

These aren’t just any mice. These are the product of a sophisticated, highly controlled and protected mouse-breeding operation.

They live in rooms designed to be impenetrable to the smallest unwelcome microbes, in cages stacked floor-to-ceiling and supplied with filtered air that changes once a minute. They’re cared for by handlers wearing protective suits, who know their inbred charges so well they can spot a potential genetic mutation that even a biologist might not notice.

Mice are central to The Jackson Laboratory, to its history, its financing and its role in understanding human diseases. The fast-reproducing mammals give researchers a model to use to examine and test how genetic distinctions can cause diseases, and how they could be cured.

The lab’s production side raises and sells some 3 million mice a year to researchers around the world. Last fiscal year, the sale of JAX Mice and Services generated $144.4 million in revenue for the nonprofit lab, more than two-thirds of its operating income.

Now Jackson is planning to build an institute in Connecticut, and in doing so, plans to take its work beyond mice, to play a more direct role in translating genetic research into tests and therapies for human diseases.

The vision behind JAX Genomic Medicine — to be located on the UConn Health Center’s Farmington campus and developed with $291 million in state funds — is based on the increasing ability to quickly and inexpensively decode the genes that make up individuals, and the goal of tailoring treatments, even preventive plans, to each person’s particular makeup.

Historically, it’s taken “forever” to move from basic science to therapies for patients, said Mike Hyde, Jackson’s vice president for advancement. “It’s now possible to speed up that cycle from science to treatment,” he said. “That’s sort of the center of what we’d be doing at our facility in Connecticut.”

And the Bar Harbor mice will play a role, offering a way for researchers in Connecticut to test any hypotheses they develop based on human clinical information or computerized data.

Cracking the code

The Roscoe B. Jackson Memorial Laboratory opened in Bar Harbor in 1929, created with money from auto barons — Jackson had been president of the Hudson Motorcar Co. — and led by Clarence Cook Little, a former president of the universities of Maine and Michigan.

Little, a Harvard-trained scientist, wanted to show that the mouse was a good model for human biology. And he wanted to prove that cancer was a genetic disease, not one that spread from person to person like the flu. The remote location would allow the scientists to focus on their research. It also happened to be the area where the Detroit industrialists summered. And, critical in a time before air conditioning, its climate was right for mice, who don’t breed well in warmer weather.

The lab was officially incorporated five months before the stock market crash that ushered in the Great Depression.

Jax mountain

The Jackson Laboratory. Champlain Mountain in Acadia National Park can be seen in the distance (Photo by Robert F. Bukaty)

Today, in another down economy, Jackson is expanding. It has 36 principal investigators leading labs and employs 1,215 people in Maine, 124 at a campus in Sacramento, and 29 elsewhere in the country. The Bar Harbor campus has doubled in the past decade, and there are plans to add more buildings and labs. The Connecticut institute is expected to add another 30 principal investigators and employ 600 scientists and technicians.

Fundamental to Jackson’s expansion plans is the idea that in the coming years, sequencing the human genome will become significantly faster and cheaper, opening the door to a new kind of science.

Knowing the exact code to a person’s genetic makeup — the 3 billion pairs of DNA bases that provide the blueprint for each human being — can give researchers tools to understand the roots of diseases. They can compare genes to identify similar patterns in people with conditions like diabetes or Alzheimer’s. They can look for markers that would indicate a person is susceptible to a disease.

These capabilities have given rise to the concept of “personalized medicine,” the idea of tailoring the treatment a person gets to his or her specific genetic code and condition.

“I believe whether it’s 25 years or 50 years from now, hopefully 25, we’re going to look back and think that medicine today was unbelievably primitive,” said Charles Hewett, Jackson’s executive vice president and chief operating officer.

Patients with anything but very simple conditions commonly get the wrong drug prescribed, Hewett said; often, the next step if the drug isn’t working is to increase the dose, raising the risks of serious side-effects. Cancer treatment can be just as bad, or worse.

“What do we do with cancer? We burn people and we poison people. Radiation and chemotherapy, we burn ’em, we poison ’em,” Hewett said. “And it would be one thing if we had confidence that that was the right drug for that tumor in that person, but way too often it’s just an experiment and it turns out not to be the right drug for that tumor for that person.”

The hope for personalized medicine is that it will make treatment more accurate. Tumors could be tested to see how likely they would be to respond to a particular treatment. (Jackson researchers in California have started putting patients’ tumors into mice, allowing them to test what would most effectively attack the tumors.) A person’s genome could give clues about what drugs might be problematic, or the person’s likelihood of developing a disease, allowing for preventive efforts.

Sequencing the human genome for the first time took 13 years and cost $2.7 billion. Less than a decade later, it can be done in a matter of days for several thousand dollars.

But getting it faster and cheaper isn’t the only challenge. There’s this: The enormous troves of information researchers can now generate about the genetic codes of living creatures have limited use without the tools to make sense of it all.

“That’s actually the big problem for us right now,” said Cox, the researcher who studies neuromuscular diseases. “We can generate the data much faster than we can actually analyze it, and the tools for the analysis … are still developing.”

That’s where people like Matt Hibbs come in.

Cancer and the computer scientist

Hibbs’ background is in data mining and artificial intelligence, not Petri dishes or pipettes. He went to graduate school planning to pursue computer graphics and work at Pixar, the animation studio. Then, in his first year, he took a class in computational biology “and just kind of fell in love.”

“It was the first time it had really come to my attention that a computer scientist could work on cancer and actually make a contribution to that area,” he said in his office at Jackson. A mug on his desk carried the likeness of Mike Wazowski, the eyeball-shaped main character of Pixar’s “Monsters, Inc.”

Matt Hibbs Jax

Computational biologist Matt Hibbs (Photo by Robert F. Bukaty)

With the human genome sequenced, researchers know what the parts — the genes — are. “But for most of these genes, we don’t know what they do,” Hibbs said. “We know that they exist, we know they’re in our DNA, we figure they’re important, but we don’t actually know why.”

Computational biologists like Hibbs are trying to make sense of the data.

On his computer monitor, Hibbs pulled up data shown in what looked like a vertical bar code, with bars in red and green. Humans have 22,000 to 25,000 genes, each of which does one or more things. But they don’t all act all the time, and one way to get a clue about their role is to see when they’re turned on and off.

In the image on Hibbs’ screen, the red and green signified genes that were turned on or off. This can be useful in a number of ways. Researchers can take samples from tumors of 100 patients and measure the activity of every gene in each tumor; perhaps there are some genes that get turned on in all the people who respond well to chemotherapy, but not in people who respond poorly.

In graduate school, Hibbs tried to solve similar puzzles about yeast, predicting what a particular yeast gene did, then testing the hypothesis in the lab. It worked multiple times. But doing the same for mice and humans is far more complex.

Yeasts are single-celled organisms with relatively simple lifespans. Scientists know at least one function of about three-quarters of their genes. Mice and humans, by contrast, have many organs, and lifespans with varied stages. And while one gene in yeast might do exactly one thing, in a mouse or person, it could do different things in different organs, or at different stages of life. In mice, scientists still haven’t identified any functions for close to 60 percent of the genes.

One project Hibbs is working on involves trying to organize the millions of measurements of mouse embryonic stem cells that have been generated around the world to identify genes that could be linked, offering suggestions for other researchers to explore. He has a website that depicts them as networks that look like Tinker Toys, the genes as the nodes, and the connections — documented or potential — the sticks holding them together.

Some of the researchers at Jackson’s Connecticut campus are expected to do work with computers, as Hibbs does. And the work of others will more likely resemble that done by Carol Bult.

From computer to mouse to human, and back

Bult came to Jackson in 1997 after working at a genomic sequencing center. She wanted to work somewhere that could both generate the data and use an organism, like the mouse, to test whether something the data suggests holds true in real life.

The next step, she said, is to connect what researchers at Jackson do more tightly with clinical research involving humans.

Carol Bult Jax

Jackson scientist Carol Bult, who is studying lung diseases (Photo by Robert F. Bukaty)

“For all the strengths that we have in Maine, in mouse genetics and mouse biology, making that bridge between the basic science and the potential clinical application has been a gap,” Bult said.

“Connecticut to me is going to be a real opportunity to build a bridge between what we do and advancing human health, which is what we are really interested in anyway,” she said.

“It’s not just understanding mouse biology. It’s using the mouse to try to understand human biology.”

Bult has made that transition in her research, which focuses on lung development and disease. She’s working with investigators at Massachusetts General Hospital to identify genes that are important for a birth defect known as congenital diaphragmatic hernia, in which infants are born with holes in the structure that separates their abdominal and chest cavities. Many children die, and even those who survive with surgical intervention face lifelong health problems. Understanding what genes are involved in the holes forming could offer a target for treatment.

The research can work in two directions. Families in which the defect has been observed often receive genetic screening, and researchers have found many genes that appear to be potentially mutated. At Jackson, scientists can see if those genes have a known mutation that causes the condition in mice.

Or they can start with the mouse, mine the data to predict genes that could be involved in the defect, and breed mice with those genes turned off. Then, if they find genes that cause the defect in mice, they can screen the people to see if they have mutations in the same genes.

The concept isn’t new, Bult noted, but technology now makes the process far faster. Her lab discovered a gene implicated in the birth defect in mouse models a couple of months ago, she said, and people are already being screened to see if they have the same variations in their genomes.

“Just to see what you’re doing potentially impact human health in such a short time frame is really cool,” she said.

Hyde, Jackson’s vice president for advancement, sees a similar model developing through the Connecticut institute. Researchers like Hibbs, using computers and massive amounts of data, could identify genes that appear abnormal in people with a particular condition. To test whether those genes are related to the condition, they could turn to theirresearchers in Bar Harbor, who could breed mice with those gene variations and see whether the condition appears.

If they find it does, the researchers could then turn to the pharmaceutical industry to see if they have a compound that could target the problematic gene activity. They could test it on the mice. If all that works, it could set the stage for clinical trials using the compound in patients.

“We’re going to create an institute that pulls together doctors, computer scientists, wet lab scientists, patients, pharmaceutical companies, and the Jackson Laboratory will be sort of the nexus where all those people come together and see if we can’t move from discovery to treatment much, much faster than we ever have,” Hyde said.

1.25 million overnight guests

Hewett, the chief operating officer, recalls speaking at an economic development event, following representatives from the tourism-heavy area’s hotel and dining industries.

“I got to talk about the fact that I had way more overnight guests than all of them put together, and that my food was cleaner and more sterile than their food, and my water also,” he said. “And that I had to change the linen, just like they do.”

“The only thing we don’t serve is wine,” he added slyly.

Jackson Laboratory aerial

Jax from above (Photo courtesy of The Jackson Laboratory)

One of the biggest problems Jackson faces is keeping wild mice from the surrounding area out of the production facilities. A local mouse could be forgiven for trying. What’s inside represents the result of years of work to suit mouse tastes: Plenty of the food they like and water that’s been calibrated to their ideal Ph balance. There’s also a lot of, shall we say, reproductive activity going on.

The warehouse at the headquarters smells like the mouse food, strong enough that it permeates to Hyde’s office upstairs.

Unfortunately for any envious local critter, the mouse rooms are designed to keep out just about any foreign organism. Any materials destined to enter are sterilized in an autoclave or with volatile hydrogen peroxide, then taken through a sterile corridor. The handlers who work with the mice enter on a separate level to avoid contaminating the sterile materials on their way in.

Even the bags of grain the mice eat are steam cleaned in an autoclave before making it into the rooms.

Handlers check each mouse daily. They become so familiar with their tiny charges that they can spot a mouse that varies even minutely from the rest of its family.

Which is how Greg Cox gets his subjects.

Tracking the genes

Cox began his career studying Duchenne muscular dystrophy, the most prevalent form of the disorder; he traces his interest in the condition to watching the Jerry Lewis muscular dystrophy telethons every year growing up.

But he didn’t want to be just one more researcher studying it, and he wanted to look at all the causes of neuromuscular diseases. Doing so when you’re a geneticist means looking at spontaneous mutations. Trouble is, those don’t occur frequently, and they can be hard to find.

That is, unless you work at a lab that produces millions of mice a year. In that case, Cox said, “Rare things pop up.”

Each strain of mouse at Jackson is inbred, so every member of the family has the same genetic code, like a colony of identical twins. So when a mutation causes a mouse to walk funny, or have a different coat color, it stands out and will likely land the critter in one of the boxes delivered to Cox and his colleagues.

Once he gets dibs on any thin, weak or paralyzed mice, Cox sets out to figure out the problem.

“Unfortunately, our patients don’t tell us where it hurts, so we have to sort it out,” he said.

The mice go through a range of physiology tests. Some of their problems get traced to birth defects or broken backs, not mutations Cox would study.

Once it has identified the cause of the problem, Cox’s team tries to pinpoint the genetics. Is the mutation one that will get passed on to offspring? If so, is the trait dominant or recessive? They breed mice with the mutation with those from a different strain, allowing them to tease out how the gene shows up. Ultimately, they can pinpoint the gene where the problem occurred and the mutation responsible for it, working backward from the disease to find the culprit gene. The process can take six months to five years.

From there, the researchers look for human patients with the disease to see if they have mutations in the same genes. For one rare disease, spinal muscular atrophy with respiratory distress, Cox’s team started with a mouse delivered from the deviant search and identified the mutation that caused it. They published the results, and scientists in Europe who had been trying to identify the responsible gene in humans used it to find the mutation in their patients.

Of course, mice aren’t humans. They’re a useful model, Cox said, because they’re small, breed relatively quickly and age fast, making it logistically possible to run the tests needed for research at Jackson and labs around the world. A very small, fast-breeding primate — if it existed — would be the best model since it would be the closest animal to humans, Cox said, but work with primates tends to be more expensive and slower.

There are non-mammals used in research that have advantages mice don’t, like fruit flies, which can be tested by the millions and, in weeks, produce results that would take Cox and his mice years.

Zebrafish offer a vertebrate model that, unlike mammals, are easier to track from the start. While mammals develop inside their mothers, fish develop from single cells externally. The developing fish are also clear, so scientists can use fluorescent markers and watch the cells develop into fish under a microscope. Jackson’s institute in Connecticut might include a zebrafish aquarium.

Cox confesses to having some “model organism envy.” But for his work, he said, the mouse is a powerful model.

“You take advantage of your model organism to discover what you can,” he said.