The possibility of such a pill follows successful research on mice and monkeys at the Institute for Human Gene Therapy at the University of Pennsylvania Medical Centre, USA.

Its scientists introduced genes into the animals that they were able to control with a drug that can be taken orally. As the dosage of the drug was increased, so the effects of the genes were raised as well.

If this can be made to work in humans, it would represent a major step forward in the development of gene therapy - a novel technique that aims to cure disease by fixing or replacing faulty genes.

"We're excited about these results because they create many new opportunities for experimental and then clinical applications of gene therapy that we couldn't consider before," said Dr James Wilson, the Institute's director. "This system will allow us to modulate the expression of a delivered gene in response to changes in a patient's disease."

Harmless virus

The Penn researchers used a virus, stripped of its disease-causing genes, to carry the desired, therapeutic genes into the animals' cells.

Two viruses were used. One was reloaded with a gene that makes erythropoietin, or Epo, a protein that stimulates red blood cell production and is used to treat anaemia.

The other was given the genes for a protein complex that regulates the production of Epo by affecting the way DNA is read.

This complex, known as a transcription factor, was then switched on by a drug called rapamycin. Significantly, the amount of rapamycin administered controlled the level of Epo produced by the cells exposed to the virus.

"It's a very precise gene switch, a kind of molecular rheostat," explained Dr Wilson. A common example of a rheostat is the so-called dimmer switch used in some household lights.

Better techniques

The research, published in the journal Science, is another step forward in the search for a more effective gene therapy. Although scientists have developed a number of ways of introducing genes into the body, getting the genes to express the desired proteins in the right place at the right time has proved problematic.

A technique such as that proposed by the Penn team would be a major advance on conventional treatments that rely, for example, on a simple injection of a desired protein to treat the effects of a faulty gene.

Patients can experience peaks and valleys in the amount of drug in the blood, which rises sharply after administration and then declines steadily until the next dose.