Tilling for cancer's 'needle in a haystack'

Just as no two snowflakes are alike, neither are two humans the same. Even identical twins may possess some differences hidden in their DNA. But to find these differences, one would need a roomful of DNA sequencers, a small army of technicians, and millions of dollars.

Those hurdles are daunting enough to scare off not just curious twins, but also medical researchers puzzling out the mechanism of disease. It turns out that lining up two genetic sequences and finding the differences is much, much easier said than done.

But a study published this summer in Nucleic Acids Research shows that Ecotilling — technique developed in Dr. Steven Henikoff's lab in the Basic Sciences Division — is a whiz at honing in on isolated mutations. These genetic blips, although rare, may lead to some of our most common diseases, including cancer. Ecotilling was first developed to look at plant DNA but is now shown to work on humans. The technology promises to help tease out the increasingly complex relationship between genes and disease.

Don't call me 'mutant'

Most of us might bristle at being labeled a mutant, but on a genetic level, it's the truth. "We're all accumulating rare mutations, naturally," explained Dr. Bradley Till, lead author of the paper. "Sometimes it's from going out in the sun, or exposing yourself to cigarette smoke. There are a variety of things that can cause mutations to occur in our body. And they can occur naturally, as well, just as errors in DNA repair and DNA replication."

Such random genetic changes are termed "single-nucleotide polymorphisms" (SNPs), pronounced "snips." Often a random change makes no difference to the body's function. Sometimes, it offers an advantage or a disadvantage. And in some cases, a single change confers both an advantage and a disadvantage. This is the type of profile that we normally associate with a common genetic disease.

The textbook example is a fairly common variant that causes sickle-cell anemia but also guards against malaria. Other variants are similarly double-edged. For instance, a common genetic variation affects synthesis of folic acid, a B-vitamin, and increases risk of colon cancer. Researchers suspect this disease gene may have conferred some advantage in the ancient past, when human diets were different, and thus it spread through the population.

But the type of variants Henikoff's lab seeks to find are more fleeting. The change might just have happened in that one person. Or it might pass down through a few generations, but not become so common that it's an official "type." Any variant that's present in less than 5 percent of the population is officially considered to be rare.

Finding rare mutations is especially important, Henikoff said, for what's known as the "common disease — rare variant hypothesis." This debated theory holds that many different rare genetic changes could lead to the same disease — that is, if the changes affect the same biochemical pathway. For instance, a recent study in Dallas showed that people with very low levels of "good cholesterol" (and thus at high risk for heart disease) often had a rare genetic mutation somewhere in a gene involved in cholesterol metabolism.

"In this Dallas study they found about two dozen potentially damaging mutations, and they were all different," Henikoff said. This means indirect methods would fail to catch the genetic culprit. "Family studies, association mapping — none of that would work here."

The most common approach for actually tracking down genetic mutations, known as the Sanger method, sequences every person's DNA and compares the results to pinpoint the difference. For a change that shows up in only one in 1,000 people, that's tedious work. And since the method looks for common changes, a rare genetic mutation could be overlooked.

"It's a problem that's been out there," Henikoff said. "How do you detect rare SNPs?"

Funded by the National Science Foundation, Henikoff's group originally developed their mutation-detection technology to look for new mutations in plants. Scientists who want to study gene function speed up the genetic clock by exposing living organisms to chemicals that can cause a mutation; then they look for the mutation and see what effects it has wreaked. TILLING — an acronym for "targeting induced local lesions in genomes" — works by pre-scanning the samples to find the mutation and then sequencing only that portion of the DNA to pinpoint the difference. Ecotilling uses TILLING to find mutations in different populations, sometimes called "ecotypes."

The recent study applies the Ecotilling technique for the first time to human DNA. As a practice run, the team analyzed 90 human samples under intense scrutiny by geneticists at the University of Washington, part of a national project to describe the diversity of the human genome. Henikoff and colleagues recognized a chance to check the performance of their technique on humans.

Results showed that Ecotilling turned up 24 of the 25 changes found by the UW researchers. The new method also discovered seven "rare" mutations that each appeared only once in the 90 DNA samples.

Now that the technique has passed this initial test, the researchers hope to apply the method to cancer cells. "Cancer mutations are like a needle in a haystack," Henikoff said. Those changes occur at frequencies of literally one in a million. In addition to finding genetic mutations that increase the risk of cancer, they hope to look for the genetic changes that arise in tumor cells and cause them to lose control of cellular division. Detecting changes in tumors is especially challenging, Till said, because "you have the further problem that some of the cells from a single sample will have the mutation, and some of them won't. And this becomes a very big problem with standard sequencing."

Fast, cost-effective technology

By increasing efficiency, Ecotilling promises to stretch cancer research dollars. A study last year sequenced breast-cancer tumors to look for rare mutations. The Center team calculated they could undertake a comparable study for roughly $1,000 per sample, rather than $50,000. Such savings could benefit upcoming projects such as the Cancer Genome Atlas, which proposes to sequence and study several hundred tumors starting next year.

Till is quick to point out that he's a plant geneticist, not a cancer expert. But he thinks the technology can benefit medical research.

"We hope to be able to create a technology that's relatively cheap, fast and easy to use, that will enable cancer researchers to study tumor samples, and learn more about the mechanisms and the progression of cancer."