Cancer Risk Exceeds Outlook in Gene Therapy, Studies Find

ew studies suggest that gene therapy might have a greater chance of causing cancer than previously thought, adding to safety concerns that have troubled the fledgling field.

Gene therapy often uses partly disabled viruses to carry genes into human cells to correct gene defects. The new studies, one of which is being published today, show that these viruses tend to land on or near genes in the human cells. When that happens, the gene hit by the virus can be inadvertently switched on or off, contributing to unexpected effects like cancer.

"It's sort of potentially rather bad news for gene therapy because you're hitting just what you'd rather not be hitting," said Dr. Frederic Bushman of the Salk Institute for Biological Studies, who has done some of the recent work.

The gene therapy field was jolted several months ago when two children in France who had been essentially cured of a rare immune disorder by gene therapy developed leukemia, a blood cancer. Scientists believe the virus used in the gene therapy inserted its genetic material into or near a cancer-promoting gene and turned it on.

Scientists had long known that this could theoretically occur but had thought the chances were small. That is because much of human DNA is not part of a gene and has no known function. Some scientists presumed that the viruses would land at random locations, and therefore would probably not hit a gene, let alone a cancer-promoting one.

But the new studies show that the viruses land on genes more often than if the process were random.

In the study being published today, in the journal Science, Dr. Shawn M. Burgess and colleagues at the National Human Genome Research Institute looked at 903 cases in which murine leukemia viruses, very similar to the type used in the French gene therapy, infected human cells in culture.

They found the virus hit a gene 34 percent of the time, more than the 22 percent that would have been expected if the process had been random. Moreover, the virus had a striking preference for the beginning of the genes, a site that is key to turning genes on or off.

That mean the leukemia cases in the French children were not so surprising after all, they said.

"This is definitely showing that it can integrate into the most important regions of gene," said Dr. Xiaolin Wu, a postdoctoral researcher and first author of the paper. "The risk is definitely much higher than what we thought before."

The scientists also studied 379 cases of infection with H.I.V., the virus that causes AIDS, and one that, in a disabled form, is also being considered for use in gene therapy. That virus landed inside genes 58 percent of the time, though not it did not favor the beginning of the genes. That was similar to findings reported last year by Dr. Bushman.

Both the murine leukemia virus and the H.I.V. are retroviruses, which integrate their genetic content into the genome of the cells they infect. That makes them desirable for gene therapy because the corrective gene can become a permanent part of the human cells and the therapy might not have to be repeated periodically.

Some other viruses do not integrate their genetic material into the human genome and might thus be safer. But another new study shows that one of the more commonly used ones, known as adeno-associated virus, is not free of risk either.

That virus occasionally does integrate into the genome of the cells it infects. And when it does so, the integration occurs inside genes 72 percent of the time, according to the study, published on-line by Nature Genetics this month.

Still, some scientists say the risk of cancer should not be exaggerated. "We can't rule out that it would never cause cancer, but there's lots of data suggesting that it hasn't caused cancer in animals," Dr. Mark A. Kay of Stanford, lead author of the Nature Genetics study, said of adeno-associated viruses.

He and others said that the findings about the retroviruses were also not that big a cause for concern because it usually takes turning on more than one gene to cause cancer and no cancer cases had been found in the scores of American gene therapy trials. They said the cancers in the French children might be unique to that circumstance because the corrective genes were put into cells that multiplied rapidly, giving an advantage to fast-growing cells like those of cancer.

"In most gene therapy this doesn't happen," Dr. Kay said. "You are putting the gene into quiescent tissue that is not rapidly dividing."

Scientists say they are not sure why the viruses seem to land on genes.

Some scientists said that rather than doom gene therapy, the studies could provide clues to finding or designing safer viruses.