An Epi Phenomenon

While exploring the genetics of a rare type of tumor, Stephen Baylin discovered an epigenetic modification that occurs in most every cancer—a finding he’s helping bring to the clinic.

By | December 1, 2012

Professor of Oncology and Medicine
Deputy Director, Sidney Kimmel Comprehensive    
    Cancer Center
Johns Hopkins University School of Medicine
Stephen Baylin always thought he would be a practicing physician. “But I started doing research in high school,” he says. “And over the years it just gradually took over my life.” That early work, which continued to draw Baylin into the lab during his undergraduate years at Duke University, “was a family affair: my father was in on the project, too. So we had lots of talks over the dinner table—and for years after.”

Although Baylin went on to obtain a medical degree from Duke, he never forgot the excitement of doing science. “I enjoyed seeing patients, but it seemed like you had a sort of rote way you had to respond to their problems in terms of the knowledge you could bring to them,” he says. “What was compelling to me was: What else do we need to know? What could we be doing differently?”

As an investigator at the National Institutes of Health (NIH) in the late 1960s and early ’70s, Baylin began to address those questions, studying the genetic underpinnings of an inherited form of thyroid cancer. “My research started out very clinically related, but over the years it became more and more basic,” he says. “Then I stopped seeing patients and put on a lab coat for good.”

“I don’t think young investigators today have that kind of luxury,” says Baylin, who is now deputy director of the Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins. “With funding issues, there’s such pressure to get into things and decide what you’re going to do quickly. But I had the luxury, over many years, to try my hand at research, to go from clinical research to more basic research—and then to come back to studying cell biology and cancer biology in the context of disease. So I’ve come full circle.”

Along the way, Baylin has made fundamental discoveries about the epigenetic modifications that underlie all human cancers—findings that he’s using to develop new treatments for the disease. Here, Baylin talks about his experience as a “Yellow Beret,”  how his early work was declared esoteric, and what it’s like to be a rock star of science.

Baylin from the beginning

War on cancer. Baylin enlisted in his crusade against cancer during the Vietnam war. “They called us the Yellow Berets—MDs who opted to do their public service by engaging in research at NIH. I was interested in an inherited form of thyroid cancer, called medullary thyroid carcinoma.” Baylin found that these tumors produce large amounts of the hormone calcitonin, as well as an enzyme that seemed like a good candidate to serve as a biomarker for tumor detection. “We also found that not all the cells in the tumors were producing these proteins, which introduced me to the concept of tumor heterogeneity”—findings he published in the New England Journal of Medicine in 1970. “So, though medullary thyroid carcinoma is not a very common cancer, it was a terrific one for learning about tumor heterogeneity, hormone production, and biomarkers.”

“My first first-author paper was cited as a piece of arcana that nobody would need to know to do good medicine. I still have a yellowed copy of that piece.”

The emergence of methylation. Thyroid tumors are not the only ones that produce calcitonin—so do lung cancers. “And we wanted to understand why. At around that time—this was the mid-1980s—people knew about DNA methylation and that it could play a part in turning genes off and on. So a graduate student and I started to play around, looking at methylation at the start of the calcitonin gene. We thought that maybe methylation is being lost in these tumors. But we found that it’s actually being gained—at hundreds of chromosomal regions, on hundreds of other genes. That started us on this quest to understand why the start sites of all these genes were piling up DNA methylation.”

Cutting the brakes. It would take another 10 years for Baylin and his colleagues to determine that some of these densely methylated genes encode tumor suppressors whose activities are lost in cancers. “I think the first we found, with my colleague Jim Herman, was the von Hippel-Lindau gene, which contributes to renal cancers. And then with David Sidransky we found that the p16 gene is hypermethylated. So by the mid-1990s we had hit on a couple of bona fide tumor suppressor genes that harbor this change. And I think that’s what helped to legitimize the field.” They also showed that a nucleoside analog called 5-azacytidine disables the enzyme that normally methylates cytosine residues in DNA. By taking the enzyme out of commission, the chemical prevents the excessive methylation and reactivates the genes. Three years later, Baylin and Herman discovered that a DNA repair gene called MLH1 is also silenced by methylation in a sizeable portion of sporadic colon cancers.  “It’s one of the most amazing things,” says Baylin. “I am not aware of any cancer type that’s been studied in any depth that doesn’t show altered methylation. And about half of the classic tumor suppressors involved in hereditary cancers”—genes like p16 and von Hippel-Lindau—are also the targets of runaway methylation.  

Staying stemlike. Baylin and his team are still trying to understand what this collective mis-methylation is doing in various cancers: Which of the modifications drive cancer, and which merely come along for the ride? “People have asked that question about genetic mutations. Now it’s coming into play for epigenetics,” says Baylin. One intriguing possibility is that the suite of modifications locks down genes that are needed to drive differentiation. “That could help keep tumor cells more stemlike,” says Baylin—a troubling hallmark of cancers. The good news: Baylin has found that stemlike cancer cells are just as susceptible to treatment with 5-azacytidine as any other tumor cell. “That could be a big deal, because most cancer therapies hit hard at the non-stemlike cells, allowing the stemlike cells to escape unharmed.”

The epigenome and beyond. Baylin was one of the original PIs—together with Peter Laird of the University of Southern California (USC)—on the epigenetic portion of the Cancer Genome Atlas project, whose goal is to identify all the genomic changes that occur in a wide variety of cancers. Looking at genome-wide DNA methylation “has been very valuable and has added to the story for each of the diseases that have been studied so far.” In glioblastoma, for example, the Cancer Genome Atlas confirmed that patients in whom a DNA repair gene called MGMT is epigenetically silenced are more sensitive to the leading drugs used to treat the disease. But the genome-wide study further revealed that when tumors recur in these patients after therapy, they show an increased burden of genetic alterations—mutations that accumulate in numerous genes. “That observation fleshed out our understanding of the evolution of glioblastoma,” says Baylin—and could lead to more effective approaches to treating the cancer when it recurs.

“I don’t think anything replaces revisiting your data—waking up and coming in and saying, ‘Let me look at that again. What am I missing?’ ”

Expert translation. Along with USC’s Peter Jones, Baylin leads one of eight collaborative “Dream Teams” established by the Stand Up to Cancer initiative. “The mandate was to take what we were learning in the laboratory and, in a 3-year period, make it part of patient management. It’s been terrifically exciting—team science in a way I hadn’t done before.” Their work has focused on 5-azacytidine, the chemical that reverses DNA methylation. “Over the years, the drug has proved to be very toxic. But we’ve found that using it at much lower doses has given us traction in treating common cancers. Treating patients with this agent first sensitizes them to subsequent therapies, so that even patients with advanced lung cancer seem to be responding. We’ve got this in clinical trials now and the drug is very tolerable at these doses, so we think it has great therapeutic potential.”


Baylin believes

All publicity is good publicity. Baylin’s early work was called out in the prestigious New York Review of Books. “It was in a review of some book on medical school training.” The reviewer was railing against the ridiculously esoteric questions that are trotted out during rounds to put med students in their place. “As an example of such an esoteric question, [the review] cited my New England Journal [of Medicine] paper on this enzyme in medullary carcinoma. So my first first-author paper was cited as a piece of arcana that nobody would need to know to do good medicine. I still have a yellowed copy of that piece.”

Seeking informaticists. “There’s currently a desperate need for people with knowledge in bioinformatics. In my laboratory I see individual graduate students or postdoctoral fellows who have a sort of precocious ability to handle the programs you need to do these analyses. I really encourage them to add those skills to their armamentarium because it gives them a breadth and a depth that will make them in demand—and will also accelerate their ability to frame their own questions and mine their own data.”

Mining and mulling. “I don’t think anything replaces revisiting your data—waking up and coming in and saying, ‘Let me look at that again. What am I missing?’ In this age, when everything is moving so fast, you have to say, ‘Wait a minute.’ And really keep doing it. One new question can lead you to re-mine the data and come up with a beautifully important pattern that wasn’t part of your original vision. That’s what’s good about the Cancer Genome Atlas project—not only what’s in the original papers but what’s still sitting there in the database for anyone to revisit.”

Sitting and thinking. “I like the small meetings the best because you can sit in one place and chew on things rather than running around trying to catch the next talk. The big meetings are important for young people, so they can get in and meet a wide swath of the community. But to actually get things you can bring back to the lab, ideas you can grow, I like the small meetings.”

Raising the bar. “The biggest value of journals like Science and Nature is that they hold up a bar that says: ‘Make your work special. Do the extra five experiments that will make your story worth publishing in a high-impact journal.’ Big awards are also like that. You shouldn’t come to work every day and be consciously aware of them. But awards that recognize someone who’s worked hard and done a great job are nothing to sneeze at.”

Redefining cure. “Cancer is such a complex, heterogeneous disease. If you block one signaling pathway, another takes over. So treatment can be tough. But if you can really slow the progression of the disease—slow metastasis or tumor growth—you’ll cut the probability that the next thing will happen to drive the cancer forward. I think that’s a realistic goal. I hope for an absolute eradication of cancer, but I’m less sanguine about that possibility. But reducing cancer to a chronic disease, slowing its progression, detecting it earlier—that’s a given. That’s going to happen.”


Baylin in Baltimore

Indulge. “I’m a well-controlled type 2 diabetic. Bagels are the worst. I love a good bagel, but they have a lot of sugar. So I miss those.” But Baylin is not beyond the occasional well-earned indulgence. In the wake of Hurricane Sandy, Baylin helped bail out the Cancer Center’s mouse colonies when sewage and rainwater flooded the basement. “The faculty were in waders, wandering through the fetid waters, rescuing freezer specimens and animal cages.” The reward? “They had bagels. So I ate half of one. And then I had a slice of pizza. You have to give in to whims and break the rules once in a while, do something crazy—in science and in life.”

Home sweet home. “Baltimore’s really a mystery to people who don’t know it. It’s got a lot of crime, a lot of murders.  But what people don’t realize is that Baltimore is a really great city near the water. The preservation of the waterfront has brought back tourists and conventions. Restaurants keep getting better and better. It’s got a great symphony. So there’s a lot to do here. Baltimore gets some bad press, but it’s a pretty good town to live in.”

Wailin’ Steve Baylin. “I play guitar—badly. And sing. I do it as therapy, for my own amusement. I love the blues. When I still worked in the lab, I’d go to a blues club and get my hand stamped. When the band took a break, I’d run back to the lab to change the fraction collector. Then I’d go back to the club, show my stamp, and catch the next set.”

Double take. In 2010, Baylin was featured in a public-service campaign designed to humanize biomedical research. The promotional event, dubbed “Rock Stars of Science,” paired researchers with celebrity performers for a photo spread that appeared in GQ magazine. “My shoot was with a rapper called B.o.B. I had no idea who that was, but when my grandchildren found out, they were ecstatic. I finally became the grandfather they always wanted.” Baylin didn’t get to see the spread until he picked up the magazine at the 7-Eleven he frequents for his morning coffee. “When I opened it up, the clerk leaned over and said, ‘Is that YOU?’ Then he added, ‘I always knew you were someone special.’”


Greatest Hits
• Characterized the biology and genetics of a rare, heritable form of thyroid cancer.
• Determined that methylation of cytosine bases in the promoter region of genes plays a key role in human cancers. Identified a handful of tumor suppressor genes that are silenced by this DNA methylation and continues to study how these alterations work in concert with mutations to drive the progression of cancer.
• Working with colleague Jim Herman, developed a technique, called methylation-specific PCR, for detecting and quantitating DNA methylation.
• Coordinates efforts to catalog the genome-wide epigenetic changes that contribute to human cancers.
• With research associate Heather O’Hagan and graduate student Wei Wang, discovered that chronic inflammation can rapidly promote the widespread methylation of genes, potentially laying the groundwork for cancer.




Add a Comment

Avatar of: You



Sign In with your LabX Media Group Passport to leave a comment

Not a member? Register Now!

LabX Media Group Passport Logo


Avatar of: nahal


Posts: 1

December 7, 2012

i like this words I don’t think anything replaces revisiting your data—waking up and coming in and saying, ‘Let me look at that again. What am I missing?
Commonwealth Life Perusahaan Asuransi Jiwa Terbaik Indonesia

Avatar of: lisaacs


Posts: 1

December 11, 2012

I like everything about the article and how the scientist comes across, revisit data, do 5 more experiments, reward good work, and give in to a whim.  i especially like how he manages to catch the blues and find time to change the fraction collector.  Baltimore Harbor is a beautiful place with much to do and a beautiful sight. 

BTW, is there such a word was quantitating?  I thought the word was quantifying.

Avatar of: jeenious


Posts: 45

December 11, 2012

The sheer magnetude of information available today is a curse as well as a boon to further progress. No human can learn it all, and no specialist can know all the information that is pouring in outside his research niche. And the generalist -- almost always someone in academia -- can only seek to grasp a large picture, and cannot know the veritable pixels of information emerging at any of the specialists' frontiers.

What is the best hope of pulling all this together? It is the crossovers, those who master two or perhaps three specialties, and can see in the one the data and methodology and focus that can add value to the other.

A lifetime is barely enough, for most of us, to master one specialty, and know its frontiers and its limits, as well as a representative sampling of its informational content. Only the brightest and most energetic and most devoted of us masters two or three specialties.

Where double or triple specialists have but one specialty in common, they can communicate in such a way as to bring, each to specialty they have in common, benefit of their non-mutual specialties, such that the specialty in common is even more likely to benefit.

How often does this happen? Not often enough. An assumption seems to be widely held that merely by making data from one field of specialty available to another, researchers in disparate specialties have all the benefit they need of bouncing what the mono-specialist knows off what everybody else is doing.

The ideal situation would be a chain of information sharers, where each sharer has mastered at least two specialties, one the same and one different, from the next link in the chain.

Where does such a chain exist? It doesn't. At least not formally. Some cross-specialization does occur, fortunately, but perhaps a concerted effort to form networks with the concerted intent to join link to link would be a worthwhile goal.

In no area more than in oncological research, input from multiple specialties is vital to further progress. And even if some giant leaps forward have occurred historically, from contributions made by mavericks, or by such visionaries as Watson and Crick) the flow of new data has increased enormously since 1953, and indications are that flow will continue to increase (cuts in government allocations for research notwithstanding); so the demand for linked communication among specialists increases correspondingly, and the odds of some isolated break-through contribution from someone outside the loop of cross-specialization intercommunication becomes increasingly unlikely.

Avatar of: jeenious


Posts: 45

December 11, 2012

But twere the giftie gie us that the perception of certainty were certainty itself.

Avatar of: mlerman


Posts: 71

December 11, 2012


I know Steve very well. Years ago during the hunt for the VHL gene I suggested to compare the methylation status of chromosome 3p probes in normal and tumor cells (spaced at 50-100kb, which we isolated from a 3p library); that resulted in a paper in PNAS. It amazes me  that epigenetic events are widely beliewed to cause cancer. I know it's a misconception, cancer is caused and driven by mutations in cancer genes only. Silencing of TSG in cancer may results from mutations in insulating genes. Michael Lerman, M.D., Ph.D.  

Popular Now

  1. First In Vivo Function Found for Animal Circular RNA
  2. A Potential Remedy for the Aging Brain
    The Scientist A Potential Remedy for the Aging Brain

    In mice, injected fragments of a naturally occurring protein boost memory in young and old animals and improve cognition and mobility in a model of neurodegenerative disease. 

  3. Nature Index Identifies Top Contributors to Innovation
  4. Your Body Is Teeming with Weed Receptors
    Features Your Body Is Teeming with Weed Receptors

    And the same endocannabinoid system that translates marijuana's buzz-inducing compounds into a high plays crucial roles in health and disease outside the brain.