Matt Ridley

Bacteria are happy to absorb little rings of D N A called plasmids and adopt them as their own. Moreover, each bacterium is a single cell. Human beings have 100 trillion cells. If your goal is to genetically manipulate a human being, you need to insert a gene into every relevant cell, or start with a single-celled embryo.

The discovery in 1970 that retroviruses could make D N A copies from R N A suddenly made 'gene therapy' seem, nonetheless, a feas-ible goal. A retrovirus contains a message written in R N A which reads, in essence: 'Make a copy of me and stitch it into your chromosome.' All a gene therapist need do is take a retrovirus, cut out a few of its genes (especially those that make it infectious after the first insertion), put in a human gene, and infect the patient with it.

The virus goes to work inserting the gene into the cells of the body and, lo, you have a genetically modified person.

Throughout the early 1980s, scientists worried about the safety of such a procedure. The retrovirus might work too well and infect not just the ordinary cells of the body, but the reproductive cells, too. The retrovirus might reacquire its missing genes somehow and turn virulent; or it might destabilise the body's own genes and trigger cancer. Anything might happen. Fears about gene therapy were inflamed in 1980 when Martin Cline, a scientist studying blood disorders, broke a promise not to try inserting a harmless recombinant gene into an Israeli suffering from the genetic blood disorder thalassaemia (though not by retrovirus). Cline lost his job and his reputation; the result of his experiment was never published. Everybody agreed that human experiments were premature, to say the least.

But mouse experiments were proving both reassuring and disappointing. Far from being unsafe, gene therapy seemed more likely to be unworkable. Each retrovirus can only infect one kind of tissue; it needs careful packaging to get the genes into its envelope; it lands at random anywhere among the chromosomes and often fails to get switched on; and the body's immune system, primed by the crack troops of infectious disease, does not miss a clumsy, home-made 248 G E N O M E

retrovirus. Moreover, by the early 1980s so few human genes had been cloned that there was no obvious candidate gene to put in a retrovirus even if it could be got to work.

None the less, by 1989 several milestones had been passed. Retroviruses had carried rabbit genes into monkey cells; they had put cloned human genes into human cells; and they had put cloned human genes into mice. Three bold, ambitious men - French Anderson, Michael Blaese and Steven Rosenberg - decided the time was ripe for a human experiment. In a long and sometimes bitter battle with the American federal government's Recombinant D N A Advisory Committee, they sought permission for an experiment on terminal cancer patients. The argument brought out the different priorities of scientists and doctors. To the pure scientists, the experiment seemed hasty and premature. To the doctors, used to watching patients die of cancer, haste comes naturally. 'What's the rush?'

asked Anderson at one session. 'A patient dies of cancer every minute in this country. Since we began this discussion 146 minutes ago, 146 patients have died of cancer.' Eventually, on 20 May 1989, the committee granted permission and two days later Maurice Kuntz, a truck driver dying from melanoma, received the first deliberately introduced (and approved) new gene. It was not designed to cure him, nor even to remain in his body permanently. It was simply an adjunct to a new form of cancer treatment. A special kind of white blood cell, good at infiltrating tumours and eating them, had been cultivated outside his body. Before injecting them back in, the doctors infected them with retroviruses carrying a little bacterial gene, the only purpose of which was to enable them to track the cells inside his body and find out where they went. Kuntz died, and nothing very surprising emerged from the experiment. But gene therapy had begun.

By 1990, Anderson and Blaese were back before the committee with a more ambitious scheme. This time the gene would actually be a cure, rather than just an identification tag. The target was an extremely rare inherited disease called severe combined immune deficiency ( S C I D ) , which rendered children incapable of mounting C U R E S 2 4 9

an immune defence against infection; the cause was the rapid death of all white blood cells. Unless kept in a sterile bubble or given a complete bone marrow transplant from a fortuitously matched relative, such a child faces a short life of repeated infection and illness.

The disease is caused by a 'spelling' change in a single gene on chromosome 20, called the ADA gene.

Anderson and Blaese proposed to take some white blood cells from the blood of a S C I D child, infect them with a retrovirus armed with a new ADA gene, and transfuse them back into the child's body. Once again, the proposal ran into trouble, but this time the opposition came from a different direction. By 1990, there was a treatment for S C I D , called P E G - A D A , and it consisted of ingeniously delivering into the blood not the ADA gene, but A D A itself, the protein made by the equivalent gene in cattle. Like the cure for diabetes (injected insulin) or for haemophilia (injected clotting agents), S C I D had been all but cured by protein therapy (injected P E G - A D A ) . What need was there of gene therapy?

At their birth, new technologies often seem hopelessly uncompetitive. The first railways were far more expensive than the existing canals and far less reliable. Only gradually and with time does the new invention bring down its own costs or raise its efficacy to the point where it can match the old. So it was with gene therapy.

Protein therapy had won the race to cure S C I D , but it required painful monthly injections into the hip, it was expensive and it needed to continue for life. If gene therapy could be made to work, it would replace all that with a single treatment that re-equipped the body with the gene it should have had in the first place.

In September 1990, Anderson and Blaese got the go-ahead and treated Ashanthi DeSilva, a three-year-old girl, with genetically engineered ADA. It was an immediate success. Her white-cell count trebled, her immunoglobulin counts soared and she began making almost a quarter of the A D A that an average person makes. The gene therapy could not be said to have cured her, because she was already receiving and continued to receive P E G - A D A . But gene therapy had worked. Today more than one in four of all known 2 5 0 G E N O M E

S C I D children in the world have had gene therapy. None are definitively cured enough to be weaned off P E G - A D A , but the side-effects have been few.

Other conditions will soon join S C I D on the list of disorders that have been tackled by retroviral gene therapy, including familial hypercholesterolaemia, haemophilia and cystic fibrosis. But it is cancer that is undoubtedly the main target. In 1992 Kenneth Culver tried an audacious experiment that involved the first direct injection of gene-equipped retroviruses into the human body (as opposed to infection of cultured cells outside the body and transfusion of those cells back in). He injected retroviruses directly into brain tumours of twenty people. Injecting anything into the brain sounds horrifying enough, let alone a retrovirus. But wait till you hear what was in the retrovirus. Each one was equipped with a gene taken from a herpes virus. The tumour cells took up the retrovirus and expressed the herpes gene. But by then the cunning Dr Culver was treating the patient with drugs for herpes; the drugs attacked the tumours.

It seemed to work on the first patient, but on four of the next five it failed.