Cannabis plant extracts could potentially form the basic ingredients for a market-leading diabetes drug, the scientist who developed a former world-beating treatment for the condition believes.

Professor Mike Cawthorne led the team that developed GlaxoSmithKline's Avandia, which became the company's second-biggest selling drug until sales plunged in 2007 after a study linked it to a higher risk of heart attacks.

"I sincerely believe it is possible to improve on it (Avandia), and plant-based medicines could be one way to do that," he told Reuters in a telephone interview.

Cawthorne is collaborating with GW Pharma, a specialist developer of cannabis-based medicines, at a new laboratory dedicated to looking for plant-based treatments for diabetes.

The GW Metabolic Research Laboratory will look the different cannabinoid molecules that have been found within the cannabis plant, as well as range of other plants extracts.

There are 60-70 cannabinoid extracts, though only one of those — THC — has the psychoactive properties traditionally associated with the plant.

The reseachers will conduct preclinical studies to evaluate them all as possible treatments for diabetes, with a view to getting licensing deals if they strike it lucky.

Cawthorne said that the cannabinoid CBD, used along with THC in GW Pharma's Sativex drug, has been seen to raise levels of 'good' cholesterol in animals.

While 'bad' cholesterol can build up in the blood vessels and cause strokes or heart attacks, 'good' cholesterol is thought to protect against heart attacks.

Cannabis-related diabetes treatments have a chequered history, with Sanofi-Aventis discontinuing development of its Acomplia obesity drug after European authorities requested that it was withdrawn from sale over fears of psychiatric side effects.

It had been seen as Sanofi's biggest new drug hope, and the withdrawal dealt a blow to CB1 receptor antagonists — the class of medicines to which Acomplia belonged — in general.

Cawthorne says he is working with actual cannabis extract rather than its synthetic equivalent, giving the basic ingredients of his potential treatments very different pharmaceutical properties.

The centre would look to treat specific symptoms of diabetes, such as non-alcoholic fatty liver or increased energy expenditure, rather than focusing on specific molecules, which is the route the pharmaceutical industry has taken to date.

"One needs to... not worry too much about the individual targets, but look and see what individual plant-based materials can do to (treat) the whole disease," he said.

"There really have been relatively few developments in finding new diabetes drug treatments... This new approach might be more productive in answering the unmet clinical need."