Harvard Public Health Review Winter 2007
Review home
next story
Previous story- The America Next Door
HSPH home page

A New Twist on Inherited Cancer Risk
dna strand

Cancer casts a particularly dark shadow on some families, while others largely escape the burden of malignancies. Rare gene mutations account for a sliver of familial cancers, while most cancers probably stem from a complex mix of inherited susceptibilities plus environmental and lifestyle factors, from chemicals and UV radiation to hormones and diet.

Only recently have scientists begun to discover what’s behind these inborn susceptibilities. Applying new gene-reading technology to large groups of people with and without cancer, researchers this past year have identified several subtle differences, or “variants,” in the DNA of people with breast or prostate cancer but not in those who are free of cancer.

Professor David Hunter
Gene Hunter David Hunter is searching patient's genomes for gene variants related to disease.

Think of variants as “red flags for cancer risk,” says David Hunter, an epidemiologist and the Vincent L. Gregory Professor of Cancer Prevention at HSPH.

“We’ve known for a couple of centuries that family history is an important factor in disease susceptibility,” says Hunter, who directs the Harvard Program in Molecular and Genetic Epidemiology. “Now we’re starting to draw back the curtain and see not just the principal actors—rare mutant genes with dramatic effects—but a whole cast of characters, that is, minor variations that influence our risk of cancer and other diseases in big ways.”

Recall that a gene is a sequence of chemical DNA units that, like letters of the alphabet, dictate instructions to cells. Some gene variants are akin to alternative spellings that don’t change the meaning of a word: think “color” and “colour.”

Millions of variants dot the human genome—our full set of genes—with no apparent effect on health. Other variants alter the meaning of the words, and thus DNA’s instructions: think “color” and “colon.” (Variants are not to be confused with gene mutations, which are rarer and more dramatic changes in DNA that may cause severe disease, disability, or even death.)

In March and May 2007, Hunter and colleagues identified several gene variants and linked them to heightened risks for breast or prostate cancer. Their work was federally funded through a three-year effort, which Hunter co-directs, called the Cancer Genetic Markers of Susceptibility Initiative, or CGEMS. The CGEMS project’s goal is to search the genomes of thousands of breast and prostate cancer patients for risk-related variants.

For an explanation of these new findings and their importance for public health, the Review turned to Hunter.

Q. How do these new gene variants impact cancer risk?

A. In the study reported in May in Nature Genetics, we identified four variants, any of which, if inherited from both parents, raised a woman’s breast cancer risk by as much as 60 percent. Inheriting just one copy elevated the risk by 30 percent. About 60 percent of women of European ancestry carry at least one of these variants.

In April, we published a report in Cancer Research confirming an earlier discovery in men—namely, that two copies of a variant on the eighth chromosome raised their risk of prostate cancer by 90 percent. A single copy heightened the risk by 30 percent. And in a Nature Genetics study, we reported yet another variant, this one accounting for up to 20 percent of prostate cancers in white men. In this population, new and previously discovered variants taken together help explain one out of four prostate cancers.

Q. Where will the research go from here?

A. The next step is for us to validate in larger populations the connection between these gene variants, also called SNPs (“snips,” short for single nucleotide polymorphisms), and the increased risk of disease. At the same time, we have to keep digging for more variants, because we may find others that are more important.

We also need to learn how the variants act together with non-hereditary factors to raise or lower cancer susceptibility. What’s the impact of lifestyle—for instance, diet, body weight, and alcohol consumption—and of environmental factors? How might cancer risk be mitigated by preventive therapies, such as tamoxifen for breast cancer?

A very likely payoff of our work is that these variants will lead us to new insights about genetic and biological mechanisms that contribute to cancer. Potentially, they’ll point to better ways of preventing and treating the disease.

Q. What about screening large groups to identify people who harbor one or more variants?

A. There’s a mantra that says, “Don’t screen if you can’t intervene.” That is, it’s useless—perhaps harmful—to identify people with elevated cancer risks if there is nothing they can do to keep them from developing breast and prostate cancer. The threshold at which it’s worth identifying high-risk individuals is something the whole public health community, ranging from epidemiology to statistics and health policy, needs to consider very seriously.

Whom would you screen? Everyone? This definitely opens a can of worms. And how would society pay for all that screening, not to mention the public education and supportive counseling that would have to go along with it? Would insurance cover it? Would individuals found to have a risk-related SNP suffer discrimination down the road from insurance companies and employers?

Q. Why are we finding out so much about gene variants now?

A. Several factors have come together just since 2003. That was when the Human Genome Project finished the initial decoding of all the information encoded in our DNA. In 2005, the International HapMap Project published a catalog of millions of SNPs across the genome.

About the same time, in early 2006, biotech companies succeeded in making devices that could examine more than 500,000 SNPs at once, enough to cover the entire genome. The availability of these computer-readable “SNP chips” set the stage for whole-genome scans—that is, for screening DNA samples from large numbers of study subjects, hunting for telltale SNPs that occur in people with common diseases but are less common in healthy people.

Q. What is the CGEMS initiative, and how did you become involved?

A. CGEMS began with an idea I proposed to the director of the National Cancer Institute (NCI) for applying these technical advances to genome-wide searches in large samples of patients—about 8,000 individuals with prostate cancer and 7,000 cases of breast cancer, along with a matched set of healthy subjects, in a large collaborative project called the Breast and Prostate Cancer Cohort Consortium. The NCI allocated $16 million for this three-year project. So far it’s assayed more than 530,000 common genetic variants in almost 5,000 patients and healthy counterparts.

Q. How did you find the breast cancer variants?

A. We used DNA from blood samples from women enrolled in the Nurses’ Health Study, which is based at the Channing Laboratory at Brigham and Women’s Hospital [in Boston]. NCI’s Core Genotyping Facility analyzed DNA from 1,145 postmenopausal women with breast cancer, then compared it with DNA from a same-sized group without the disease.

Of the 530,000 DNA variants, eight had the strongest correlations with breast cancers. By testing those SNPs in additional study subjects, we narrowed the most important SNPs to four. (A British group publishing at the same time found the same group of markers. The Icelandic genetics company Decode identified two breast cancer-related variants, one of them the same as one reported by the British.) The variants were associated with the same gene, FGFR2, which is over-expressed in some breast cancers.

Interestingly, these four variants are located in the crucial sections of the gene that contain instructions for making a protein.

Q. So you’re suggesting that some of the variants lie in long stretches of DNA between genes, the stuff scientists call “nonsense” or “junk” DNA? What does that tell you?

A. It’s likely that “junk” DNA is not really junk: Probably these regions contain bits of DNA that regulate genes nearby. So, the variants we’re finding could be the key to a whole new biological understanding about how our genome works. How do we get by with only about 22,000 genes? It seems we’ve developed the capacity to use this limited amount of information by regulating it in different ways.

Q. All this new information must be gratifying—a validation of the whole idea behind CGEMS.

A. Yes, it is. But the field is moving incredibly fast, and that’s a little daunting. It’s urgent that we address the potential consequences for cancer patients and public health. Our new ability to predict disease risk has huge implications for health policy, risk communication, environmental health, and many other disciplines.

1 | 2 | 3

Richard Saltus has written about science, medicine, and public health for the Boston Globe, the New York Times, the San Francisco Examiner, and the Associated Press.

Review home
next story
Previous story- The America Next Door
HSPH home page
top of page
This page is maintained by Development Communications in the Office of Resource Development.
To contact us with suggestions, comments, and questions, please e-mail: review@hsph.harvard.edu

Copyright 2007, President and Fellows of Harvard College