Just as gene therapy finally seems to be living up to its promise, a study has revived a lingering worry about the viral vessel that many efforts rely on to ferry therapeutic genes into patients. This “vector,” a stripped-down version of adeno-associated virus (AAV), was thought to be safe because it rarely knits the human DNA it carries into a cell’s chromosomes, where it might activate cancer-causing genes. But a study of dogs with hemophilia, treated with AAV up to 10 years earlier, has shown that the vector can readily insert its payload into the host’s DNA near genes that control cell growth.

The new data, described in a conference talk last month by a Philadelphia-based research team, are “good news and bad news,” says gene therapy researcher Charles Venditti of the National Human Genome Research Institute. By slipping into the chromosomes rather than floating free, the therapeutic DNA might have longer lasting benefits. Integration “happens and it may actually be essential for long-term expression” of a needed protein, says physician-scientist David Lillicrap of Queen’s University, who attended the 9 December 2019 talk at the annual meeting of the American Society of Hematology (ASH) in Orlando, Florida. But the findings also fuel a debate about whether AAV vectors could pose an unacceptable cancer risk. “We don’t know enough yet” to say, according to Lillicrap.

Another viral vector, used in some early gene therapy trials, caused cancer in a few children after it integrated its cargo into the chromosomes. AAV seemed to be a safer alternative because genes introduced by the modified virus generally become a free-floating loop, known as an episome, in the cell’s nucleus. AAV vectors have helped drive the recent surge of successful gene therapies. These include one approved by the U.S. Food and Drug Administration last year for spinal muscular atrophy, a fatal childhood neurological disease, and a treatment for the blood-clotting disorder hemophilia A that’s expected to receive FDA approval this year. In the hemophilia treatment, AAV infects liver cells and turns the organ into a factory for making the clotting protein that patients depend on.

Yet doubts about AAV’s safety have simmered for nearly 20 years, since a study found that in newborn mice given high doses of the virus, it could integrate its genetic cargo into the animals’ DNA and cause liver cancer. Many gene therapists argued that findings in newborn mice are not relevant to human adults. But the new warning comes from larger, older animals: adult dogs with hemophilia A, in which a clotting protein called factor VIII is missing.

In seven of nine such dogs, AAV vectors successfully supplied a replacement copy of the gene for factor VIII and restored stable production of the molecule. In two of those dogs, however, blood levels rose further after about 3 years, reaching about four times the original level by 7 to 8 years, gene therapy researcher Denise Sabatino of the Children’s Hospital of Philadelphia reported at the ASH meeting.

After ending the experiment and studying the livers of six dogs, her team found that in every one of them, AAV-ferried DNA—the factor VIII gene or, more often, fragments of regulatory sequences—had integrated in many spots across the genome in the dogs’ liver cells, sometimes near genes affecting cell growth. Some of those cells had divided more than other cells, forming pockets of multiple cells or “clones” in some animals. Sabatino’s team suspects—although it can’t prove—that the insertions activated growth genes, explaining both the clones and the rise in blood levels of factor VIII in the two dogs. (Sabatino declined to discuss the results because they have not yet been published.)

To some researchers, the results are encouraging: The integration levels were relatively low, the dogs had seemingly healthy livers, and their factor VIII levels held steady. “I don’t think there was anything too unexpected,” says Andrew Davidoff of St. Jude Children’s Research Hospital, which sponsored the first successful gene therapy trial in people with hemophilia B, which also used an AAV vector. Indeed, integration of AAV’s therapeutic DNA could explain why levels of clotting protein appear stable in that trial’s patients, after 9 years for some. DNA carried in episomes, in contrast, could be lost over time as the cells divide, because only one daughter cell would inherit the replacement gene.

But others in the gene therapy field worry that it’s only a matter of time before such clones acquire another growth-driving mutation and become tumors: “What if the dog lived another 5 years?” Venditti asks. Such a risk could arise not only in the liver, but in other tissues targeted with AAV treatments, such as neurons and muscle cells, some researchers say.

Sabatino’s data give new urgency to plans to look for integration of AAV-carried genes in other long-term dog studies and in planned liver biopsies from the St. Jude hemophilia patients, says Lillicrap, who is now studying this in his own colony of nine dogs. Meanwhile, researchers emphasize that people receiving AAV-delivered gene therapy should be monitored for signs of liver cancer for longer than the 5 years of follow-up now required by FDA. A spokesperson for the agency said only that it is “aware of” the Sabatino findings.

*Correction, 19 June, 2:50 p.m.: An earlier version of this story misstated the type of hemophilia for which a gene therapy treatment is expected to receive FDA approval in 2020.