Brain cancer. Its cause is perhaps the least understood of all cancers. The only known risk factors are heredity and exposure to radiation. The outcomes, however, are clear: survival rates are only 3 to 5 percent three years following diagnosis. Mayo Clinic researchers are collaborating with colleagues in California to chart the road ahead for others in the field. Their maps? The human genome and Mayo’s medical record.

Geneticist Robert Jenkins, M.D., Ph.D., and genetic epidemiologist Ping Yang, M.D., Ph.D., are bringing complementary skills and perspectives to their study of glioblastoma, the most common adult form of brain cancer, which arises out of glial cells, long considered helper cells to neurons. Dr. Jenkins uses genetic techniques at the molecular level to understand the underpinnings of cancer. Dr. Yang studies the Big Picture: populations of patients, survival rates, quality of life, environmental factors and genes involved in carcinogenesis, cancer progression and prognosis. She views their process as a kind of educated angling.

“We’re looking for a special type of fish,” says Dr. Yang. “My role is to find the right pond where the fish can be found. Bob’s role is to find the fish itself. I need a bigger scale to find the high-risk area or chromosomal region where cancer [-related alterations] is most likely to be found. Bob zeroes in on the specific genes that increase the risk of cancer.”

The significance of their efforts is not necessarily an immediate diagnostic test, though that may come eventually. Rather, it’s that this team is able to harness the new knowledge of the human genome and the new tools of bioinformatics, genetic transcription and assays to search for likely culprits – tumor-causing mutations. They can use these tools and the large institutional patient databases to narrow the field and save time and effort. To extend the fishing analogy, modern genomics and its technology serve as the sonar-based mapping devices on their boat. Mayo is uniquely positioned as a leader in this venture, not only because of its technology and experience, but also because it has some of the most extensive patient data sets in the country.

Their six-year collaboration began in 2003 when neuro-oncologist Brian O’Neill, M.D. and Dr. Jenkins applied for a Specialized Project of Research Excellence (SPORE) grant from the National Institutes of Health. These big awards (about two million dollars a year) contain multiple projects. Mayo, through its Cancer Center, has seven cancer SPORE grants in brain, prostate, breast, ovarian and pancreas, myeloma (with Dana-Farber Cancer Institute) and lymphoma (with the University of Iowa) research.

The application required an epidemiology project, so Mayo recruited Dr. Yang to run the epidemiology of glioma (the umbrella term for numerous brain cancers) branch of the SPORE grant.

Collaboration & confirmation

About two years ago, they learned of a large glioblastoma genetics study being performed by the University of California, San Francisco (UCSF), headed by Margaret Wrensch, Ph.D. Dr. Wrensch found a cluster of DNA alterations on chromosome 9 in non-coding regions of two genes in advanced glioblastoma patients.

The alterations are SNPs (pronounced snips), single nucleotide polymorphisms, lone mutations of just one nucleotide in the four chemicals that comprise DNA. The vast combinations of DNA nucleotides give humans and other species their traits such as hair and eye color, an inclination to be talented in music or sports, and predisposition to certain diseases. Just five SNPs (three in a locus called 9p21on chromosome 9 and two in locus RTEL1 on chromosome 20) were repeatedly found in patients with advanced glioblastoma in both studies.

“By happenstance we had been collecting glioblastoma data, and we thought we might do a glioblastoma study in the future,” Dr. Jenkins says. “They were finding some suspicious mutations. We then did an analysis of our glioblastoma patients’ data, and lo and behold, it replicated what Dr. Wrensch found.”

Results were published in Nature Genetics (2009). A separate paper appeared in that issue detailing another replication of the UCSF/Mayo Clinic results by researchers elsewhere. The significance in the Mayo Clinic findings is that the same outcome was found in an entirely different population of patients from the UCSF genome wide association study, confirming the UCSF results and the study’s reliability.

Revealing the risk

Using Dr. Yang’s analogy, chromosomes 9 and 20 comprise the pond, the five SNPs the fish.

If a person should have one or more of these SNPs, their risk of developing glioblastoma tumors is increased, but it doesn’t mean they will develop a brain tumor unless environmental factors trigger its development.

“In the United States, the risk of having a brain tumor is one in 10,000,” Dr. Jenkins says. “If you happen to carry one of these SNPs, it rises to one in 7,000.”

Dr. Yang says environmental factors are important, but results have been inconclusive – and still an unsolved mystery. She says the literature includes full-mouth dental X-rays, cell phones and even allergies (proposed by UCSF) among a handful of possible environmental factors related to glioma, but there have been no replicated studies, including her own ongoing work, not yet published.

Drs. Jenkins and Yang had analyzed in total 610,000 SNPs, but to replicate UCSF’s results, they focused on 13 SNPs across five chromosomes that UCSF found suspicious. “The three common regions identified in the two studies, 9p21, RTEL1 and TERT, provide a solid foundation for future research to begin to discover the origins of glioma,” says Dr. Wrensch.

Tumor suppressor genes

A couple of SNPs are within the coding region of other genes called tumor suppressor genes, which allow the cell to undergo mitosis or duplication. Three other papers on melanoma in the same issue of Nature Genetics highlight mutations in these genes in melanoma patients.

“In both melanoma and gliomas, tumors themselves will delete these genes,” Dr. Jenkins says. “Tumors undergo a genetic rearrangement, and one of the processes in that is they delete these tumor suppressor genes.”

SNPs found in TERT and RTEL1 are of interest because they are regulators of telemorase, an enzyme involved in making telomeres, the long ends of chromosomes. A tumor needs to activate telemorase to become cancer.

“Telemorase is active in all of our germ cells and stem cells, and they’re basically immortal — if we all had telemorase, we’d be immortal,” Dr. Jenkins says. “If the chromosome keeps getting shorter and shorter, pretty soon it chews into the genes on the chromosome and that kills the cell. So, basically in all other cells in your body telemorase is turned off – that’s what everyone thinks of as the aging process.”

In cancer cells, telemorase is activated and the result is runaway cell division. “These SNPs being present in telemorase activators has real promise as a potential mechanism to understand why an individual might be predisposed to cancer,” Dr. Jenkins says.

The deletion of 9p21 in high grade glioma tumors has been known for 20 years so the gene has been thoroughly investigated.

“Even if you think it’s an old story, there might be something new in there after all,” says. Dr. Jenkins. “I think we’re going to find some novel genetics in an old region.”

Either way, Drs. Jenkins and Yang have more fishing expeditions.

“In the next phase, we will characterize the alterations and mechanisms in top candidate genes,” says Dr. Yang. “My job is to look for the right chromosomes in the right individuals.”

“The important thing to understand next is the interactions between the association we’ve found and the environment and other exposures,” says Dr. Jenkins. “We need a card-carrying epidemiologist for that kind of analysis. Ping is just that, and already has added a lot to what we’re doing.”

— Tony Fitzpatrick, September 2009