Searching through the scientific literature, Scott then located Ervin Epstein, a dermatologist at the University of California, San Francisco, just 40 miles away, and an authority on the genetics of basal cell nevus syndrome. Scott called Epstein, introduced himself, and said, "I think we've got your gene." He launched into an explanation, mentioning patched genes, Drosophila, and so on.

Epstein now likes to say that his first thought was, What's a Drosophila? although when pressed he'll admit he knew. "But I sure as hell didn't know what a patched gene was, or anything about developmental biology." Epstein told Scott it was "nice" that he thought they had his gene, and the two agreed to work together to see if it was true.

Epstein, who describes himself as "a dermatologist who sees patients as his day job," had been searching for the genetic basis of BCNS since 1987, when he read a series of papers in the journal Nature on a flight from Cleveland to San Francisco. The articles announced the discovery of the gene that causes retinoblastoma, a rare and deadly cancer of the eye that afflicts children. Like BCNS, retinoblastoma comes in two versions. One affects newborn infants and is characterized by multiple tumors. The other hits children when they're older and is usually characterized by only a single tumor.

Back in 1971, Alfred Knudsen, a pediatrician working at the M.D. Anderson Hospital and Tumor Institute in Houston, had proposed an explanation. In the early-onset version of retinoblastoma, he suggested, children inherit a defective gene from one parent. These children are partway to getting the disease the moment they are born. Then, something as simple as an error in DNA replication in a single eye cell, causing a defect in the normal gene that was inherited from the other parent, would send that cell on its way to becoming a tumor. In contrast, Knudsen speculated, children who develop retinoblastoma later in childhood are probably born with two good copies of the gene but are unlucky enough to have both copies in a cell rendered defective. This would take longer, causing the cancer to show up at a later age.

"Think of it as analogous to losing the brakes on a car," suggests Epstein. "If you have a factory that's producing rear brakes that fail 1 in 1,000 times, and front brakes that fail 1 in 1,000 times, only 1 in 1 million cars will lose their brakes entirely. That's not very many. But if the factory is producing all its cars with defective rear brakes, then 1 out of 1,000 will eventually lose control, and that's a lot of cars."

In 1987, when researchers at the Massachusetts Eye and Ear Infirmary in Boston and the Whitehead Institute for Biomedical Research in Cambridge, Massachusetts, announced that they had identified and cloned the gene that causes retinoblastoma, they also confirmed Knudsen's hypothesis. The gene—known as RB, for retinoblastoma—was the first "tumor suppressor" gene ever discovered, so called because cancer develops only when the gene is defective or missing.

The similarity between retinoblastoma and BCNS caught Epstein's attention. "The idea that retinoblastoma comes in two versions—a rare hereditary type, in which there are multiple tumors with an earlier age of onset, and a sporadic type, with single tumors and a slightly later age of onset—threw new light on what happens in basal cell carcinoma," he says. "Basal cell carcinomas more routinely come in people who are not in the first flush of youth, like me, and have been in the sun too much, and they get one or two basal cell carcinomas. But I also knew there was a rare, heritable form called basal cell nevus syndrome, which happens early and leads to multiple carcinomas."

Epstein and his colleagues set about trying to find and clone the gene responsible for BCNS by repeating the steps that the Boston and Cambridge groups had taken in studying retinoblastoma.

First, they obtained blood and tumor samples from as many people with BCNS as they could find by putting notices in medical journals and talking to researchers at meetings, hoping physicians would get the word and send samples from their patients. Then the scientists started painstakingly going through the 23 pairs of chromosomes of each patient, one by one, looking for defects.

"Tumor suppressor genes are often inactivated by deletions in one copy of the gene," explains Epstein. "Frequently this deletion is not just of a small part of the gene but of a large part of that arm of the chromosome. If you have some way of detecting the two copies of the gene in a normal cell's DNA, you can see if both copies are present in the blood. Then you can look at the tumor and see if one copy of the gene has been lost during its growth."

Epstein and his colleagues started on chromosome 1 because they had seen some hints in the journals that this was the problem area. "Sure enough," says Epstein, "we found some deletions in 25 percent of basal cell carcinomas in area 1q and we said, 'Dynamite, we've put in our thumb and pulled out a plum.' And then we got a very large family from New England with BCNS and looked to see if it mapped there—and it didn't at all! It was clearly not the right site."

Over the next seven years, Epstein and his colleagues made little progress. Luckily, they weren't alone in the search. In 1992, Allen Bale of Yale University, a cancer geneticist, announced that his team had traced the gene responsible for BCNS to an area of chromosome 9.

"Then everyone else found that families with BCNS all linked to the same chromosome," says Epstein. The gene's position was soon narrowed down to a region of chromosome 9q that was a mere 5.5 million base pairs (5.5 megabases) long. As Epstein describes it, it was as though the human genome were the equivalent of the entire Encyclopaedia Britannica and they had managed to narrow down the location of the gene first to one volume, and then to one entry. But that entry was still 50 pages long, and the defect in the gene they were looking for could have been the equivalent of a typo in a single word. "And you can't even tell what the words are," he adds, "because they're all run together and there's no punctuation."

This is where Epstein stood in 1995 when he received the phone call from Scott, which he now refers to as his deus ex machina. Soon Scott's and Epstein's groups, working together, proved that the human patched gene actually lived in the same 5.5 megabase region of chromosome 9q that had been linked to BCNS and basal cell carcinoma.

Even with that information, it looked as if they might be on the wrong path. When they searched for gross defects in patched in the DNA from the 86 families with BCNS that Epstein had collected, "we could find nothing," he recalled. "It was extraordinarily discouraging."

The only thing left to do was look for point mutations in the gene—which, being the equivalent of single typos in the Encyclopaedia Britannica, could be time-consuming and maybe futile. Using every trick in the book, the two teams finally did identify some patched gene mutations in a handful of patients with BCNS, as well as in a few of the sporadic basal cell tumors. They still needed something that would nail it down, however: a single case, a sample of DNA, or several samples from a family that would be unambiguous. That's where Jenica Chekouras came in.

— Gary A. Taubes

< Previous  |  Top of page  |  Next >

Basal cell carcinomas occur most frequently on exposed areas of the body, such as the face. They may be open sores, reddish patches, shiny bumps, or pink growths with a slightly elevated border and a crusted indentation in the center, such as this one on a patient's nose.

Photo: From "Basal Cell Carcinoma, the Most Common Cancer," ©1986 by the Skin Cancer Foundation

© 2008 Howard Hughes Medical Institute. A philanthropy serving society through biomedical research and science education.
4000 Jones Bridge Road Chevy Chase, MD 20815-6789 | (301) 215-8500 | e-mail: webmaster@hhmi.org