91av

“GREAT medical hope”, “lethal experiment”, “up-and-coming treatment”… In turns, gene therapy has been called each of these. So when French researchers announced last month that a young boy given a therapeutic gene had developed leukaemia, there was a sense of resignation among researchers.

Resignation but not despair. After the knocks that gene therapy has received over the past 12 years, researchers in the field are taking the news in their stride. Far from triggering a bout of hand-wringing and self-flagellation, the child’s illness has confirmed gene therapists’ understanding of the dangers of the treatment, and reinforced the recognised need to find better ways to deliver genes. It’s too early to tell precisely how this case will affect the future of gene therapy but there is a feeling that the field will emerge stronger than ever.

Gene therapy hasn’t seen this kind of optimism since the first trial in 1990 for an immune deficiency disorder. The experts thought that for diseases caused by single inherited mutations, such as cystic fibrosis, muscular dystrophy and thalassaemia, all they needed to do was add a healthy version of that faulty gene. And they would deliver the genes by co-opting the natural talents of viruses, which entwine their own DNAwith ours. “The strategy was so simple and beautiful that we all got so excited,” remembers Savio Woo of the Mount Sinai School of Medicine in New York, a former president of the American Society of Gene Therapy. The man who first proposed the idea of gene therapy, W. French Anderson of the University of Southern California School of Medicine, was equally beguiled. “We thought it would immediately translate into cures,” he says.

But it didn’t. Hundreds of unsuccessful trials followed, and experts began to doubt they’d ever solve the huge technical problems gene therapy was throwing at them. The field was harshly criticised by government agencies. Then disaster struck. When a young volunteer, Jesse Gelsinger, died inexplicably after a massive immune reaction in a trial in 1999, many thought the field would fold entirely.

Instead, Gelsinger’s death led to a wholesale reappraisal of the way gene therapy was conducted – its procedures, ethics and goals. A new sense of realism emerged. And in September 1999, as this catharsis took place in the US, a group of researchers across the Atlantic at the Necker Hospital for Sick Children in Paris, led by Alain Fischer and Marina Cavazzano-Calvo, were notching up gene therapy’s first cure, for another immune deficiency called X-SCID. At last somebody had shown that gene therapy could work.

Which is why it came as such a blow in October when the same team announced that one of its young patients had developed leukaemia, almost certainly triggered by the virus used to insert the gene into the boy. Immediately, questions were asked about whether all gene therapy trials should be stopped until safety could be assured. It certainly looked possible as regulatory agencies around the world halted all similar trials. The US froze recruitment for a forthcoming trial, while Germany and Italy suspended trials that used viruses similar to the one employed in France. Only Britain decided to continue testing the therapy, but with extra warnings and monitoring.

But a couple of weeks later, having had a chance to discuss the case, both the US advisory committee to the Federal Drug Administration and the European Society of Gene Therapy recommended lifting these bans. Instead, researchers were urged to consider carefully the risks and potential gains for each disease being treated.

The reasoning was that alternatives to using viruses to shuttle genes into cells are years away from the clinic and the children involved in Fischer’s trial almost certainly had no more than a year to live. So with no alternative treatment, leukaemia could be a risk worth taking. “It’s not so much a setback for gene therapy,” says Richard Mulligan of Harvard University, who is a member of the FDA advisory committee. “In my mind, it has eroded the risk-benefit ratio.”

Although observers were quick to draw parallels between the French boy and Gelsinger, there are crucial differences. Gelsinger’s death was unpredicted, unexpected and to this day unexplained. Cancer, on the other hand, has always been the most obvious risk in the back of gene therapists’ minds. Insert a gene into human cells and it could disrupt their normal functioning. Nobody knows how big this risk might be. But the child’s case, tragic as it is, simply confirms a suspected problem. The challenge now is to work out how great the risk of cancer actually is, and how to deal with it.

Fischer’s group has already discovered an impressive amount about the boy’s leukaemia, which has helped to contain the damage. The boy was born in 1999 with a severe combined immune deficiency known as X-SCID. This is caused by mutation of an X-chromosome gene known as γc, and it prevents two types of white blood cells, T cells and natural killer cells, from developing. With no defence against infection, sufferers usually die in their first year unless a bone marrow donor can be found.

Fischer and his colleagues collected stem cells from the boy’s bone marrow and infected them in the lab with a retrovirus engineered to carry a healthy copy of the γc gene. Although Fischer thinks that only about 50 of the treated cells picked up a working copy of the new gene, the fact that they were stem cells – precursors that can divide over and over to produce new blood cells – meant that when they were returned to the infant, they generated all the immune cells he needed.

By May this year, Fischer’s team had treated 11 children. All were doing well, living normal lives. Similar success was reported at London’s Great Ormond Street Hospital (91av, 12 October, p 4). But in August, the boy – Fischer’s fourth patient – caught chicken pox. His T cell count stayed high after the infection and within a month it rocketed. It wasn’t long before Fischer and his team concluded that the boy had a form of leukaemia.

They immediately suspected there might be a problem with the location of the therapeutic gene. Retroviruses, including the one used by the French researchers, integrate themselves into their host’s DNA. The advantages are that the added gene can stay active for the lifetime of the cell, and when the cell divides, each daughter cell inherits the gene too. The downside is that it’s not easy to control where the virus will end up.

And this seems to have been the problem. Fischer’s analysis shows that in at least one bone marrow cell, the retrovirus inserted itself and the therapeutic gene into a regulatory region of a gene on chromosome 11 called Lmo2. It looks as though when the new gene became active, so did Lmo2. That might not be so bad except that Lmo2 is a cancer-causing “oncogene”. Lmo2 could have triggered rapid cell division, producing a host of identical T cells – leukaemia.

The big question is how often will this type of thing happen? There are about 3 billion places the virus could have landed, and only around 300 oncogenes, so statistically it’s unlikely that the virus would hit a danger spot. Some viruses, however, are known to prefer certain spots to lodge in, and if one of these is in or near an oncogene it would increase the odds of cancer.

Yet until now, cancer has only been seen in one animal test of gene therapy and never in human trials, leading researchers to suspect it poses only a minuscule risk. But Mulligan speculates that it could crop up more and more as we move into bigger trials. Early-stage clinical trials, designed for assessing safety rather than effectiveness, may have looked satisfactory only because the gene transfer was not very efficient, he points out. Quite simply, they seemed safe because genes weren’t infecting many cells. The real risks will only appear during later-stage trials designed to test the efficiency of treatments.

One factor that affects the level of risk may well be genetic differences between people which make some patients more susceptible to disease than others. Environmental factors, too, could make a difference. In the French case, chicken pox and a family history of cancer are just two variables to take into account. “We want to pass a clear message to the public that the treatment was to blame,” says Mulligan. But there could be other factors behind the boy’s leukaemia which will make it harder to determine the precise risk to others.

The treatment procedure itself may have avoided some potential hazards. Removing the bone marrow stem cells from the body before infecting them with the retrovirus eliminated the danger of an acute reaction to the virus. It also solved the problem of inserting the virus into the right cells, because they were the only ones in the dish. However, the flip side is that treating only a few stem cells and using them to repopulate the whole immune system amplifies a small risk into a big problem, says Mark Kay of Stanford University in California, who chaired the European Gene Therapy Society debate on the case. Precisely because the treatment is selecting for proliferating cells, it may be uniquely risky.

Theoretically, any virus that integrates its own genetic material into human DNA carries a similar risk, and many other therapies rely on such vectors. Indeed there is no point in giving up on integrating viruses in favour of those that remain loose, warns Alan Kingsman, CEO of British firm Oxford Biomedica, which researches and produces viral vectors. They don’t stick around long enough to help in most conditions and almost any virus chosen to deliver DNA into cells will leave its traces. But at least the only unknown with integrating viruses is where those genes will land, not the order or number of them.

Kingsman also says that Fischer’s virus is now fairly old technology. Vectors can now be engineered to inactivate the signals that might, in rare cases, switch on an oncogene. In any case, Kingsman argues, viruses are still the only realistic choice at present, since other methods are just “horribly inefficient”.

Other researchers think that viruses will never be part of optimal therapy because they are too complicated and costly. Researchers have tried alternatives such as encapsulating therapeutic genes in fatty globules called liposomes, or in the circular chromosomes called plasmids, which are found in bacteria (91av, 11 May, p 17). It’s also possible to inject “naked” DNA straight into a target tissue. But all these processes are fairly hit-and-miss, and still incredibly inefficient.

At Stanford University, however, Michele Calos and her colleagues improved the efficiency of one of these techniques and developed a way to control exactly where the gene lands. They have teamed up plasmids containing a therapeutic gene with other plasmids containing a gene for a bacterial “integrase” enzyme. Integrases tend to cut the host DNA only at specific points and chaperone the gene into those spots.

At the moment, Calos and her colleagues have tested this approach only in animals and cultured human skin cells. They have a good idea where the integrase places the genes in mouse chromosomes – just two locations – and they are now searching for the sites favoured in human DNA. Each tissue will be different, so it won’t be a quick job, but if the insertion points favoured in a particular tissue don’t look safe and suitable, it is possible to evolve new integrases in the lab to do a better job.

The researchers have used their idea to insert the gene for factor IX – a clotting agent that is missing in one type of haemophilia – into mouse liver cells, and with reasonable efficiency. While it will be at least a year before the group has enough data to start human trials, Calos is very optimistic about the future.

For now, though, cancer remains a risk. This will inevitably mean that the conditions tackled first will be life-threatening ones, where patients have little other hope. But researchers are trying to put the leukaemia case in perspective. Drug treatments carry risks, as do bone marrow transplants. And the list of gene therapies that are showing early signs of success is too impressive to give up (see “The next wave”). To this day, SCIDis the only disease that’s been cured by gene therapy. Now there is hope for conditions including heart disease, cancer, Alzheimer’s, Parkinson’s, AIDS and even chronic pain.

Only as trials go on and researchers get a better idea of the cancer risk will we decide which treatments become routine. If the risks are high, we’ll have to wait for alternative technologies, rather than giving up. By the time you or I suffer one of these diseases, we may well be offered a gene therapy.