Swimming mice suggest memory pill is possible

Geoff Watts looks at the prospects for a pill that could prevent memory loss with age.

Like dreams of a drug to make you clever or witty or good at maths, talk of a pill to improve memory sounds suspiciously like an advanced case of wish fulfilment. But Karl Peter Giese of University College, London, as realistic a man as you could hope to meet, is serious about the idea of a memory pill. And, like all good brain scientists, he has a mouse model to prove it.

Giese is head of learning and memory at the Wolfson Institute for Biomedical Research, now housed in the old but splendidly refurbished building once occupied by University College Hospital.

Most biologists, he says, have believed that learning involves changes in the synapses, the junctions by which adjacent nerve cells or neurones communicate with one another. He and his collaborators wanted to test the possibility that changes in the electrical excitability of nerve cells might also be playing a part. It seems they do. It was while pursuing this line of purely basic research that he stumbled on the potential practical application that now dominates his work.

To operate as the body's message carriers, whether in the brain or elsewhere, nerve cells generate a small voltage that keeps their interiors electrically negative with respect to their surroundings. They create this difference in electrical potential by transporting certain ions (electrically charged atoms) across their outer membranes.

A nerve impulse is simply a brief loss of this voltage: a localised depolarisation of the membrane that propagates itself, wavelike, along the length of the nerve. In the wake of the impulse, the membrane rapidly repolarises itself: recreates its lost electrical potential, in other words.

Giese's principal interest is in ion channels: pores in the membrane that open at critical moments to allow specific ions to pass through them. "For example," he says, "the potassium channel is selective for potassium ions, and regulates their outflow into the surrounding space. Potassium channels are responsible for the repolarisaton that prepares a neurone for the passage of the next nerve impulse."

The channels themselves are actually built out of various proteins. The recipes for making them are stored in the genes. One particular protein sub-unit found in certain potassium channels affects a process called "slow after hyperpolarisation" or AHP.

The effect of AHP is to limit the speed of the repolarisation process - presumably to prevent impulses following one another too closely along the nerve and so overexciting the system.

Inside the hippocampus - the part of the brain associated with learning and memory - AHP has been found to increase with advancing age. Could this be why the capacity to learn and memorise diminishes as we get older?

Giese is working with mutant mice that have lost the gene that codes for the protein regulating AHP. In these mice, AHP does not increase with age. If AHP really is important, these animals should stay smart, even in their dotage.

"We tested the animals using tasks in which performance is known to be affected by age. One is the water maze. This is a tank of opaque water with a platform submerged just below the surface. You put the tank in a room with walls marked in different ways. When the mouse is placed in the pool, it swims around until, by chance, it finds the platform. It can then stop swimming. After several tries, it can use the cues on the walls of the room to navigate its way straight to the platform, even though it can't see it."

At the age of two years (old for a mouse) half of all normal mice are no longer able to learn well. But not so the mutants, says Giese. "When the mutant mice grow old they have the same ability to learn that you find in young animals." In other words, the age-related loss of spatial memory does seem to reflect an increase in the extent of after hyperpolarisation. Prevent that increase, and memory can be maintained.

But does this have any bearing on humans? Giese thinks so. "We know that the same gene and the same protein exist in humans. So we believe a similar mechanism should apply. Of course, it's not possible to genetically manipulate humans. So what we want to do is mimic the effect of the mutation using a drug."

He is already testing one candidate substance. And a colleague, David Selwood, is hoping to develop others. The art of devising new drugs to disable a protein relies on knowing something of that protein's three-dimensional molecular structure. One way of finding out is to use X-ray crystallography.

Selwood, who is head of biological chemistry at the Wolfson, says that proteins that are bound to cell membranes are difficult to crystallise. "So instead we have to use molecular modelling techniques to work out the structure." But he reckons he should have a drug suitable for testing in no more than six to 12 months.

In the longer term, Giese hopes, his work could have some bearing on the prevention of Alzheimer's disease. "We hope to study other mouse models of the earliest stages of the disease to find out."