91av

A WOMAN finds a lump in her breast. Her doctor injects a few cells from a
biopsy into a machine. Within minutes, a computer reveals the tumour’s type,
where it originated and what state it has reached. Best of all, it details the
drugs most likely to halt the tumour. Michael Bishop from the University of
California at San Francisco thinks we might be only two decades away from
creating this marvellous machine. It will be powered, in a sense, by the gene
revolution.

It was Bishop and his UCSF colleague Harold Varmus who made the Nobel
prizewinning discovery that cancer could be triggered by damaged genes. That was
a quarter of a century ago. Since then, about a hundred genes have been found to
contribute to tumour formation, and hundreds more may yet be found. If every
tumour contains a dozen or more such mutations, the potential number of tumour
types is astronomical. But handling such complexity will soon be routine, Bishop
says. In the growing field of genomics, researchers are learning to cope with
tens of thousands of genes at once.

Bishop sees three important steps in making cancer more manageable. First,
find every “oncogene” capable of triggering cancer. Second, learn how to
classify every tumour according to the genetic changes it contains. And finally,
devise therapies based on the tumour’s genetic make-up.

The sequencing of the human genome will let scientists complete the first
step quickly. Chromosomes in many cancer cells are visibly broken, or have
sections that are inverted or moved around. Since researchers have been
cataloguing these break points for years, they already know where the damaged
genes are, but not what they are. Soon they will be able to read the critical
sequences directly from the genome database.

Other candidate cancer genes will pop out of the database because they are
similar to the genes that control cell growth in other organisms, such as yeast
and roundworms. Bishop says biologists will be able to use DNA
chips—DNA-coded wafers that latch onto specific DNA sequences—to
hunt down any remaining cancer genes.

Researchers will compile a dossier of genetic defects in every type of
cancer. Then, one glance at a tumour’s genetic fingerprint will reveal where it
originated in the body. After comparing tissue samples from cancers at different
stages, it will also be possible to tell what stage a tumour is at—whether
it is about to spread.

Once cancer is diagnosed, says Bishop, therapy can in theory be tailored to
each patient or even to each tumour, based on its specific genetic changes. That
might include gene therapy to replace damaged genes. But it might be more
practical to get round the differences between tumours by exploiting what they
have in common.

One strategy would be to repair a group of broken genes that show up in many
tumours. Lesions in a gene called APC, for instance, are present in
nearly all colon cancers, while many different types of tumours have defects in
a gene called p53. By looking at damage to other genes in a particular
tumour, researchers would know whether repairing p53 or APC,
or using drugs to counteract the effects of the damaged genes, would be enough
to stop the growth.