Deciphering the signature of prostate cancer

Prostate cancer - the answer's in the genes

Many of us know someone who’s been affected by prostate cancer – the disease is the most common cancer in men in the UK. Thanks to the efforts of doctors and scientists, more than three quarters of men diagnosed with prostate cancer now survive beyond five years – in the 1970s it was less than a third.

But although huge progress has been made, there are still big challenges to be overcome.

Perhaps the biggest of these is finding a way to tell the difference between prostate cancers that are slow growing and are unlikely to cause any problems, and those that are aggressive and need urgent treatment.

New research by our scientists, published this week in Lancet Oncology, is offering a tantalising glimpse of how this could be possible in the future.

You might think it would be sensible to treat all prostate cancers straight away. But the side effects of treatment can be quite serious – most men won’t want to risk incontinence or impotence unless they really do have a dangerous illness.

And we know that some prostate cancers grow so slowly that they won’t cause any problems in a man’s lifetime – so it’s vital to be able to separate these harmless cancers from the life-threatening ones.

Cell cycle progression genes

Professor Jack Cuzick, one of our researchers based at Queen Mary, University of London, has found a way to spot this difference by measuring the genetic ‘signature’ of prostate cancer. The technique measures levels of a number of genes involved in controlling how healthy cells grow, collectively called ‘cell cycle progression genes’. And in fact two of our scientists – Sir Paul Nurse and Sir Tim Hunt – won a Nobel prize for discovering one of these genes back in the 1980s.

Cancer is caused by cells growing out of control, so cell cycle progression genes really are at the heart of the disease. The theory is that by measuring how active these growth genes are, it’s possible to estimate how fast the cancer will grow, and how aggressive it will be.

Recent research has suggested that this type of genetic signature can predict how long patients with other cancers such as breast, brain and lung cancers survive.  So building on this work, Professor Cuzick and his team were keen to find out if it could be used to solve one of the biggest riddles in prostate cancer research.

Testing the new approach

The researchers tested this approach using around 700 tumour samples from two groups of men – some from the UK, others from the US – who had already been diagnosed with prostate cancer. The men were diagnosed at least nine years ago, meaning that the researchers could compare the real outcomes of their disease with the outcome predicted by the genetic test.

The team collaborated with scientists in America to analyse the samples. One group of men had been treated with surgery to remove the prostate. The other group had cancer that hadn’t spread outside the prostate and didn’t need to be treated immediately. This second group had been given the ‘watch and wait’ option, avoiding treatment for cancers that are unlikely to cause a problem. This set-up allowed the researchers to look at two groups of men who had been managed differently by their doctors.

Reading the signature

To read the genetic signature of each tumour, the researchers used samples of tissue that had been taken when the men were diagnosed or treated. They measured the levels of RNA – a messenger molecule produced by genes – to give them an idea of how active the cell cycle progression genes had been in each tumour. In all, their test measured the activity of 31 cell cycle progression genes. These were chosen because they give a good average measure of the many other genes that control cell growth.

For the group of men who had had their prostates removed, the researchers tested if the genetic signature could predict if the men’s cancers had returned. The results were exciting – they found that tumours with higher levels of RNA from the chosen genes were more likely to have come back after surgery, and the signature could also predict if men were likely to have died after the disease returned.

For the ‘watch and wait’ group who hadn’t been immediately treated, the researchers tested whether the genetic signature could predict if the patient had gone on to die from prostate cancer. Again the signature gave good results – higher scores meant the patient was more likely to have died from the disease.

Importantly, the gene signature test worked best when it was combined with other information about the cancer. The researchers gave each tumour a ‘risk score’ based on the stage the cancer was at, the patient’s PSA score (the levels of a protein produced by the prostate in the patient’s blood), and other factors such as the amount of healthy tissue removed around the prostate if the patient was given surgery. Adding the gene signature score gave more accurate predictions, showing that the test was giving valuable extra information about the cancer.

These are really encouraging results – it looks like the gene signature test could help doctors to predict much more accurately which men can safely ‘wait and see’, and which need immediate treatment. The test also helps to identify men whose cancer appears less harmful, but who in fact have a higher risk of dying from their disease.

So what’s the next step?

The test isn’t yet ready to be part of normal diagnosis and treatment for prostate cancer, but the researchers are working towards this goal. And the results need to be repeated in larger numbers of men, with different backgrounds, to find out how widely applicable they are (for example, there’s evidence that men from certain ethnic minorities are more at risk of prostate cancers – whether this affects their genetic signature needs to be pinned down).

And as the researchers point out in their paper, the gene signature doesn’t measure every gene that might help to predict the outcome of the disease. More research could help to improve the test so that its predictions are even more accurate. Other research in this area is also giving intriguing clues to the genes that could predict if the cancer is likely to spread.

Towards personalised medicine

Finally, and even further in the future, researchers would like to know whether  the gene signature could help doctors decide which treatment will work best for each patient. We won’t know the answer until the signature has been tested as part of clinical trials comparing prostate cancer treatments.

This study has brought us one step closer to more personalised treatment for prostate cancer – it’s a great example of how research into the genes at the heart of cancer could help to improve outcomes for patients. Although it’s still early days, we’ll be watching this space.

Nell

• Watch Nell Barrie talking about the new research on our website

Reference:

Cuzick J et al (2011). Prognostic value of an RNA expression signature derived from cell cycle proliferation genes in patients with prostate cancer: a retrospective study. The Lancet Oncology PMID: 21310658