Scientists have developed a treatment which may be effective against the most common and deadly form of brain cancer.
Tumours can be difficult to treat
Glioblastoma multiforme (GBM) usually grows so quickly that it kills within a year of diagnosis, and neither surgery, drugs or radiotherapy can stop it.
But a team from Cedars-Sinai Medical Center in Los Angeles has boosted survival of lab rats with the tumour by injecting them with a protein.
Details are published in the journal Molecular Therapy.
The researchers used a genetically modified virus to deliver a small protein called hsFlt3L into the brains of lab rats who had developed GBM.
They found that the protein increased the number of immune cells in the brain, and significantly slowed tumour growth.
Seven out of 10 rats given a high dose of the protein survived for over a year. There were no signs of adverse side effects.
In contrast, rats treated with a dummy injection died from their tumours within one week.
Among rats treated with hsFlt3L, 33% were completely tumour free at three months, while all those who survived for six months or longer had no tumours at all.
Researcher Dr Maria Castro said: "Importantly, our study is the first to show that GBM tumours shrank or were completely eliminated in lab rats, which is likely due to the ability of the protein, hsFlt3L to stimulate the production of fully mature immune cells within the brain.
"Since gene therapy has given us the tool to deliver this protein, our hope is to translate these laboratory studies into clinical trials in patients with GBM."
GBM tumours develop in the supportive tissue of the brain and grow quickly, often becoming very large before a person experiences symptoms and is diagnosed.
Surgery is typically performed to remove as much of the tumour as possible and followed with radiation and/or chemotherapy to slow progression of the disease.
But despite aggressive treatment, the tumour recurs and patients usually die within a year.
Dr Castro said it was hoped to start clinical trials of the new treatment within the next three years.
Dr Jeremy Rees, an expert in neuro-oncology at University College London, described the research as "exciting".
But he said many other studies had produced promising findings in animal experiments without translating to human disease.
Previous attempts to harness the patient's own immune system to destroy the tumour had also been tried many times before without success.
"At this point it is too premature to say this will be a major breakthrough."
However, Professor Lawrence Young, of the charity Cancer Research UK, called the research "very exciting".
"Similar immunotherapeutic approaches are currently being trialled in humans with different cancers with promising early results."