Code Breaker
Geneticist Heidi Rehm ’93 is at the forefront of a genetic revolution in medicine, which may eventually lead to personalized care based on individual DNA.
By Michael Blanding
Photographs by Mark Ostow
This story originally appeared in the Spring 2015 issue.
In 1998, Mark Dunning’s daughter was born deaf. “There were no lullabies, no cooing her to sleep, no baby talk,” he says. “If I wanted to ask, ‘Do you want milk?’ I had to figure out what the sign was for it, then teach it to her again and again.” As Bella began to grow, Dunning and his wife, Julia, realized that Bella’s problems went beyond deafness. Bella took nearly 18 months to walk, and even then she had problems with her balance. She also seemed to have issues seeing in the dark. “I would go into her room at night and hand her something to drink and she would grab at the air,” Dunning says.
When Bella was eight years old, a specialist at Boston Children’s Hospital asked Dunning and his wife if Bella suffered from a series of symptoms, including those of night blindness and difficulty balancing. “Is this something I want to remain blissfully unaware of?” asked
Dunning halfway through.
“Have you ever heard of Usher syndrome?” the specialist asked.
Dunning hadn’t. But that night he looked it up on the Internet to learn it was a rare genetic disease first identified by Scottish ophthalmologist Charles Usher in 1914. Usher syndrome causes hearing loss and balance problems due to inner ear malfunctions. As the disease progresses, it results in deteriorating retinas, as well, which leads first to night blindness and loss of peripheral vision and ultimately to a complete lack of sight. “It described Bella perfectly, but in this horrible clinical way, with these definitive outcomes, including that she would go blind,” says Dunning.
There was no cure. However, they could determine if Bella had the syndrome: a simple genetic test had just been developed that could identify the mutation causing it. Her parents faced a terrible choice: continue to remain unaware of the causes of their daughter’s symptoms or risk learning their worst fears were true.
Since the completion of the Human Genome Project — the massive, international undertaking that sequenced all three billion base pairs of human DNA — these kinds of choices have become more common. In 2012, actress Angelina Jolie revealed she’d had a double mastectomy after testing positive for a mutation of the BRCA1 gene, which can cause the breast cancer that killed her mother and aunt. Following the announcement, referrals for genetic breast cancer testing nearly doubled. And pregnant mothers can now test to see if they’re carriers for the gene causing cystic fibrosis. The test is becoming the norm in prenatal care and leads to an 85 percent abortion rate for those testing positive.
Meanwhile, private companies like 23andMe (named after the 23 pairs of chromosomes that carry human DNA) have for years offered personalized genetic analysis to identify human ancestry. Until recently, the company also returned detailed health information on the risk of genetic disorders, but stopped in February 2015 after running afoul of the FDA. Companies such as Illumina Genome Network still offer genome sequencing through clinics, but the sequencing costs $5,000 to $8,000 and insurance doesn’t cover it.
This rush of genetic information promises to revolutionize medical care, and yet it also raises thorny questions: How much information is too much? How accurately can we know our genetic risks? What actions should we take if we test positive for a genetic mutation? And will these expensive tests create a two-tiered medical system — those who have access to their genetic codes and those who don’t?
For the past two decades, Heidi Rehm ’93 has been on the front lines of these questions. Having created the test for Usher syndrome, among many other genetic tests, Rehm currently directs the Laboratory for Molecular Medicine at Partners Healthcare Personalized Medicine in Cambridge, Massachusetts. There she helps identify genetic disorders for those at risk for disease. She’s also increasingly on the forefront of determining how genetic information is gathered and revealed — whether we have familial histories of genetic disorders.
“If we sequence your genome and find something scary in it that puts you at grave risk, why wouldn’t we tell you what we find?” asks Rehm, sitting in the café at Harvard Medical School, where she is an associate professor of pathology. Petite and dressed in a comfy wool sweater, she wears her wavy brown hair pulled back into a hair clip. Two years ago, in the journal Nature, she was involved in a controversial recommendation that advocated fully disclosing genetic information to patients. “If you go to a dermatologist with an itchy patch on your arm and they find a melanoma, they are not going to keep that information from you. I think in general our society does a pretty good job of evaluating risk and putting information in context.”
At the same time, she says, releasing information the right way is important. That way, patients can best decide what to do with it. After agonizing over whether to have their child undergo genetic testing, Bella Dunning’s parents did decide to do the test. Rehm sequenced the results, finding Bella positive for the mutation. “I couldn’t get off the floor,” says Dunning. “I could carry myself during the day, but as soon as the kids were in bed, I would lay on the floor with the lights off and start to cry. All I could think about was how I was going to watch my daughter go blind.”
However, as he processed the diagnosis his attitude began to shift. A cochlear implant Bella had gotten as an infant had helped her hearing, so now the family got a second implant as a backup in case her vision worsened, and she could no longer communicate through sign language. They also began to protect Bella’s eyes from direct sunlight and changed her diet to include more fatty fish, which had shown to help protect against disease symptoms.
“Whether those things helped or not, they gave us something to do, which made us feel like we were helping our daughter,” says Dunning, who reached out to specialists, including Rehm, to find out more about the disease. “I learned genetics from Heidi,” says Dunning. “She always found time in her busy schedule to meet with me.” During an early conversation, Rehm suggested Dunning start a website to share information with other parents and patients suffering from Usher syndrome. Dunning turned that into the Usher Syndrome Coalition, which shares information on treatment, lobbies Congress for funding, and provides emotional support to those suffering from the disorder.
Now 16 years old, Bella is a straight-A student who is winning blue ribbons in horse-riding competitions, studying for her driver’s license test, and preparing potentially to take part in a clinical trial for a new genetic therapy. “A lot of people when they hear they might have Usher syndrome don’t want to get the genetic test, because they don’t want to know for sure,” says Dunning. But getting Bella’s test results proved very important. “It helped put a name to the problems that Bella had that I had been suspicious of for a long time. Knowing definitively what it was gave me the ability to do something about it.”
Rehm leaves for her office at 6:00 each morning, driving a black Lexus with the vanity plate GENES. (“My first choice was GENOME, but that was already taken,” she quips.) She always knew she’d be a scientist. “For my high school reunion, they showed us what we had written down for graduation for what our career would be, and I said genetic engineer,” she says. “So I was pretty close.” “I” comes out at “ah” — a slight twang in her voice left over, perhaps, from living her first 18 months in Mississippi, where her father attended graduate school for biology. But she spent most of her childhood in Lake George, New York, on “forty acres of land on the side of a mountain,” and an hour and a half drive from Middlebury.
Her ease communicating with patients, however, took time to develop. A math whiz, she was valedictorian of her class, but she was also shy and couldn’t imagine teaching, as her father did. Arriving in Middlebury in 1989, she majored in molecular biology and biochemistry, a new major announced her sophomore year. She spent her senior year working in Bob Cluss’s lab, researching the bacteria that causes Lyme disease. Cluss remembers her as exceptionally devoted to her experiments.
“One day after she had been working the night before, she came into the lab in the morning and went immediately to the bench to start looking at what her results were without even taking her coat off,” he recalls. “You can’t engender that kind of excitement in a student.”
Cluss speculates that going to Middlebury also provided good training for her current career, which involves dividing her time between patients and the lab.
“There’s something about a liberal arts experience that gives you an appreciation for the enormity of the knowledge we have accumulated as a race and allows you to embrace that and take risks but also to be respectful and know your limitations,” says Cluss. “She’s in a unique intersection between basic research, clinical work, teaching, and interacting with patients. There aren’t that many people who are doing all of those things at that level.”
At Middlebury, Rehm also overcame her shyness. While in Sunhee Choi’s chemistry class, she began tutoring a fellow student who was having trouble with the material. Eventually that student invited a friend, who invited another friend, until Rehm was giving repeat lectures to a large chunk of the class.
“It was just an incredible experience where I learned that I loved to teach and communicate my ideas,” she says. “Now I probably give 100 seminars, lectures, and plenaries a year, and I love it.”
Rehm went on to study at Harvard Medical School, where she dove into genetics. “I am a type A personality; I like order,” she says. “There was something about the genetic code that seemed so clear and concrete to me.”
For her PhD, Rehm studied the genetic variants that caused hearing loss, focusing specifically on a genetic malady called Norrie disease that causes babies to be born blind and often, over time, to lose their hearing. Trying to identify a way to treat this hearing loss, Rehm was able to isolate the gene on the X chromosome and to examine its effects on proteins it produced.
In 2000, after receiving her degree, Rehm started a laboratory at Harvard that investigated hereditary hearing loss. Two years later, Partners Healthcare collaborated with the medical school to expand Rehm’s lab to include genetic testing for a wide range of disorders. Rehm helmed the newly created Laboratory for Molecular Medicine, working with other leaders at the Harvard Partners Center for Genetics and Genomics to acquire equipment and hire team members involved in the Human Genome Project. Doing so greatly increased the lab’s capacity to sequence complex genes.
Rehm first developed a genetic test for hearing loss, but others soon followed: for lung cancer; for a heart disorder called cardiomyopathy; and for more targeted disorders like Usher syndrome. In many cases, the tests aimed to give definitive evidence of a malady doctors already suspected. “There is this notion of ending the diagnostic odyssey,” says Rehm. “When patients have syndromes they keep getting more and more tests to find the answer; when you have a diagnosis, you have a much better idea of what the future will hold.”
Since genetic disorders necessarily run in families, tests can also help identify those at risk for disorders before they display symptoms — in some cases saving lives. Consider a heart disorder called hypertrophic cardiomyopathy, which causes defective heart tissue that fails to expand and contract properly. The heart then makes more and more tissue, causing the organ to swell dangerously large and block off blood-vessel flow, leading in many cases to a sudden heart attack. HCM is carried on the dominant gene, meaning that patients only need one gene in a pair to have it and that a patient’s close relatives each have a 50 percent chance of having the disease. Rehm developed a test for it. One of her patients, Lisa Salberg, worried about her daughter, Becca. Salberg’s grandfather, aunt, and sister all died from the disease, and she herself had been diagnosed at 12 and fitted with an implantable cardioverter-defibrillator to prevent heart attack. So Salberg had her daughter annually undergo electrocardiogram tests.
“It was an emotional roller coaster, every time we walked in,” says Salberg. Though her tests routinely came back negative, Salberg’s daughter, from as early as four years old, would wake up complaining about chest pains. When Rehm developed a test for HCM in 2004, Salberg made sure her daughter was among the first to receive it. After the test came back positive, Salberg pushed for a new EKG that confirmed her daughter had the disease and then had an ICD implanted into Becca’s heart when she was 10.
It may have saved her life. One day when Becca was riding a horse that bolted, her heart raced dangerously fast. “It stopped her at 225 bpm and helped get her back to 80,” says Salberg. “Maybe she would have done that on her own, but no one can tell.” Salberg founded the Hypertrophic Cardiomyopathy Association in 1996 to raise awareness for the disease. As director she’s referred many sufferers and potential sufferers of HCM to Rehm. “Heidi has done an extraordinary job of balancing the amazing power of science with the amazing compassion of dealing with people. Heidi, I daresay, is brilliant, and I think she has a very clear and concise picture of what the future of genetics can be.”
That future isn’t clear to the average medical patient who is without obvious history of family genetic disease. The rise in private companies offering genetic information has created confusion about how that information should be used. What does it mean, for example, if you’re told you’re 20 percent more at risk for heart disease? Should you stop eating red meat? Start taking beta blockers? Or just try not to worry?
Four years ago, to help clarify such issues, Rehm joined a Harvard-based study called MedSeq as a coprincipal investigator. Foreseeing that in the near future genome sequencing will be the norm, the study asks how doctors can use that information to help patients rather than to alarm or confuse them. The head of the study, Robert Green, was a student in Rehm’s genetics class at Harvard and admired her clear thinking. (A highly regarded neurologist several years Rehm’s senior, Green studied under Rehm as a fellow in Harvard’s Genetics Training Program.)
“Heidi is very much a leader in terms of genetic sequencing in this country, and someone everyone is drawn to for her intelligence and good sense,” he says.
The study has three parts: to develop a protocol for testing, to determine which genes to test, and to monitor how physicians transmit information to patients. Green hired Rehm to oversee the study’s second part: wading through the genome’s complexity to decide which gene mutations the report should include. There are no easy answers as to what makes the cut. Of the three billion base pairs in the human genome, a full three to five million vary person to person. Some determine physical differences such as hair and eye color; some seem to do nothing at all; and some play major roles in producing organs and enzymes. A mutation in one can lead to a genetic disorder.
The challenge is to determine which of those three to five million variants matter, and by how much. A breast-cancer-causing mutation like the one that affected Angelina Jolie should of course be included, but what about a mild variant for dry skin? Or a late-onset neurological disorder that may not even affect a patient in his lifetime? Even trickier are genes definitely associated with disease but unlikely ever to manifest.
Rehm and her team sifted through journals and genetic databases to offer their best judgments on which variants matter, eventually narrowing the field down to about 4,000. With those genes in hand, the researchers sequenced the genome for 100 patients, returning results for variants. In the end, 95 out of 100 had some genetic mutation that carried risk for disease. Most were carriers for recessive disorders they would never have. However, 20 percent did have the genetic disorders.
Several subjects, for example, had Factor V Leiden thrombophilia, a problem with blood clotting that can be particularly dangerous for women pregnant or on birth control. One had a variant called Long QT syndrome, a heart disorder that can lead to sudden cardiac death, but is treatable with beta blockers. They promptly referred that woman to a cardiologist. With these patients, the team also needed to see what the doctors would do with the information. There are very few geneticists compared to the general population, so it often falls to general practitioners to convey test results. Twenty practitioners participated in the study, and Rehm’s team gave them six hours of training in delivering information accurately.
One patient who had a familial history of breast cancer, for example, was relieved when her test didn’t display mutations for breast cancer genes — but the physician had to explain that even though she may be free of those particular mutations, she may not be free from contracting genetically based breast cancer. Over time, the MedSeq study will trace the decisions doctors and patients make with genetic information — whether they get more or better treatment and if the information affects the outcome of their illnesses.
The same research team has also started a study at Boston’s Brigham & Women’s and Children’s Hospital to sequence the genomes of newborn babies — developing a rapid turnaround of only a few weeks. For this study, Rehm’s team has only considered child-onset diseases, narrowing the number down to about 800 variants that have significant enough probability of disease in childhood. As with MedSeq, the BabySeq study will monitor how treatment of babies diagnosed with genetic disorders differs from those who aren’t diagnosed.
Depending on what these findings reveal, the study could set new standards for patient care and provide new impetus to adopt genetic sequencing, starting at birth, as standard practice. Even so, large-scale genetic sequencing is unlikely to really catch on until costs come down — or until insurance carriers start covering it, which is doubtful in the current environment. “In terms of predictive medicine, I don’t know any circumstance in which genetic testing has been covered,” says Rehm. Even though genetic testing could help catch a problem early, leading to decreased costs, it could just as easily surface a problem the patient didn’t know about — adding costs for care that may not be strictly necessary.
“In some cases you can make those arguments by costs, and in some cases you can’t,” says Rehm. Of course, those arguments are separate from the medical arguments of what will provide the best care and save lives in the long run. As costs inevitably come down and more people take advantage of genetic testing, the question of how it improves medical care will likely become about how people handle information when they receive it. The studies Rehm and her colleagues are conducting will go a long way to determining that — one gene variant at a time.
Michael Blanding is an award-winning writer in Boston, where he is currently a senior fellow at the Schuster Institute for Investigative Journalism at Brandeis University. His most recent story for this magazine was “On the Road,” which appeared in our fall 2014 issue.