Washington

Blood cancer: leukaemia has been detected in two patients who have received gene therapy. Credit: A. & H.-F. MICHLER/SPL

The world of gene therapy was shaken last year when a child treated in a French trial developed leukaemia. Researchers had pinned their hopes on this being an unfortunate one-off. Now those hopes have been dashed with the emergence of a second, almost identical case that could jeopardize the future of gene therapy.

The latest case centres on a three-year-old boy treated in a gene-therapy trial led by Alain Fischer at the Necker Hospital for Sick Children in Paris. Just under three years ago, the child was cured of severe combined immunodeficiency disease (SCID), a condition that disrupts the development of the immune system. But researchers revealed last week that the boy was diagnosed with leukaemia just days before Christmas.

The news last August that a child in Fischer's trial had developed cancer left nations divided over their regulatory response (see Nature 419, 545–546; 200210.1038/419545a). This time there was greater accord. Britain has stopped the world's only active SCID trial, and the US Food and Drug Administration (FDA) has suspended more than two dozen similar gene-therapy trials in a variety of diseases — trials it had allowed to continue after August's incident.

The setback is all the more serious because it arises in a trial that was widely viewed as gene therapy's only true success. Fisher has so far cured nine boys — including the two who now have leukaemia — out of 11 patients. And when the first child was diagnosed with cancer, some argued that it was an isolated event (see Nature 420, 595; 200210.1038/420595b).

But Christof von Kalle of the Cincinnati Children's Hospital says analysis of the two boys' cells shows that the same molecular events probably caused the cancers. In both boys, the retroviral vector used to deliver the corrective gene has integrated itself into a stretch of DNA in or near a gene called LMO2, which can cause childhood leukaemias. Scientists are now satisfied that the disease behaves enough like leukaemia to describe it as such.

Kalle estimates that the vector is likely to insert itself near to LMO2 in between 1-in-50,000 and 1-in-100,000 cells. And because each child in the trial receives a dose of about a million corrected cells, each patient could carry at least one cell with the vector near LMO2. Kalle is analysing samples from all of the children — about 15 in all — who have been treated in SCID gene-therapy trials to find out whether they are carrying such cells.

In the meantime, an FDA advisory committee on gene therapy will meet on 28 February to decide what the adverse event means for other studies. The American Society of Gene Therapy will also convene a committee to examine data from human and animal trials of retroviral vectors. And the National Institutes of Health's Recombinant DNA Advisory Committee (RAC) postponed its planned meeting on 17 January, but will convene shortly to discuss these issues.

At their last meetings, both the FDA committee and the RAC recommended that gene therapy in SCID could proceed with appropriate monitoring and informed-consent programmes. RAC chair Ted Friedmann of the University of California, San Diego, says that this approach is “still valid”, although it could be revised in light of the new case.

The case is an enormous setback for the field — and for patients with SCID. “The treatment works very well,” Fischer says, “but the risk is not acceptable.”