Fresh insights into changes in the brain linked to depression could pave the way for new treatments, according to scientists.

Their study also sheds light on why a particular category of antidepressant drugs stops working in some people, researchers believe.

Experts at the University of Edinburgh have pinpointed a key molecule that they think may protect the brain from depression.

The team studied mice bred to have defects in their ability to activate the molecule, known as eIF4E.

These animals showed signs of depression, including reduced levels of the hormone serotonin - a hallmark of the condition.

The mice also demonstrated behavioural changes linked to depression, such as lacking interest in food.

Treatment with a commonly-prescribed antidepressant called fluoxetine also failed to produce a response in the mice, the university said.

Scientists said this suggests that activation of the molecule is required to experience the beneficial antidepressant effects of fluoxetine, in a category of medicines called selective serotonin re-uptake inhibitors (SSRIs).

This could help explain why some patients stop responding to SSRIs, researchers say.

Other studies have shown that eIF4E is involved in in regulating protein synthesis in the brain, with defects in the molecule linked with other neurological conditions, such as autism and Fragile X syndrome.

The latest study marks the first time the molecule has been implicated in depression, the research team said.

They believe the findings could lead to the development of new medication for depression, which affects around one in four people in the UK each year.

Dr Christos Gkogkas, of the university's Centre for Discovery Brain Sciences, said: "Our study reveals that altered protein synthesis through eIF4E is a key cellular process in the brain that can go awry in depression.

"Importantly it may explain why some people with depression become resistant to treatment with SSRIs. This knowledge can help us design a new generation of antidepressants."

The research, published in the Journal of Neuroscience, was funded by the Wellcome Trust, the Royal Society, the RS Macdonald Charitable Trust and the Patrick Wild Centre.