European Journal of Cancer
Issue 4, 2009

Dr Patricia Steeg is Chief of Women’s Cancers at the Laboratory of Molecular Pharmacology, Center for Cancer Research, National Cancer Institute (Bethesda, Maryland, USA). Her research has focused on molecular and translational aspects of tumour metastasis and she gave a well-received Brinker award presentation at the recent San Antonio Breast Cancer Symposium.

You have shown that metastatic cells are quite different from those in the primary tumour?

This is not just my work; it has now been supplemented by multiple lines of research. I discovered the first metastasis suppressor gene, nm23, in melanoma and breast cells 20 years ago. Unlike a tumour suppressor gene, a metastasis suppressor gene has no effect – or a very limited effect – on the primary tumour. Instead, it reduces the number of metastases. Tumour suppressor and metastasis suppressor genes work in very different ways.

There are now 20 known metastasis suppressor genes and they all have the same function: to inhibit the ability of the tumour cell to make a new home in a distant organ, to prevent its interaction with local micro environment and subsequent proliferation.

Further, others have shown clearly that some compounds have qualitatively different effects on primary tumours than they do on metastases.

Third, while gene expression profiling studies show that primary tumours and metastases are closely related, they are not identical. Studies of matched primary tumours and metastases have been conducted at the RNA, DNA and protein levels and show important differences.

What changes are needed in drug development?

 Most studies have not used the right pre-clinical models. In pre-clinical studies in mice, researchers inject tumour cells under the skin, and examine whether the drug slows growth of the tumour over a few weeks. A few studies have done a better job and looked not only at the effect on the primary tumour but also on metastases; some drugs have a dramatic effect on one but not the other. There are good preclinical models of metastatic cancer – mostly xenografts. We need to use these to get data we can believe in.

We also need to develop drugs that target metastatic colonization, the outgrowth of tumour cells in a distant organ. The question is how to test these drugs in the clinic? Typically, in phase I trials we test the drug on patients with advanced metastatic disease. But drugs which suppress the formation of metastases wouldn’t melt metastases that have already formed.

We need fundamental changes and it will take time. Cancer researchers have done a good job in targeting proliferation and tumour growth, but we haven’t targeted the metastatic process. There is more to tumour progression than growth. We need to find the important molecular changes in metastasis.

It’s been argued that drugs can’t alter the metastatic process?

That’s true to some extent – the horse is already out of the barn – but not completely. A woman with lymph node positive breast cancer is at risk of metastatic disease. The first step of metastasis has been done; the cells have already gone through the circulatory system and are sitting as micrometastases in distant organs. If left, there likely will be metastatic colonization and progressive growth of cells in a distant organ.

But these cells are not autonomous; they have to survive in a new environment. Metastatic colonisation is not just about growth and it should be perfectly druggable if we can figure out how it works. After surgery, 94% of breast cancer patients have no distant metastases, so maybe we can attack metastatic colonisation in the vast majority. It gives us a whole different set of targets.

When we find agents we will have to do clinical trials differently. We usually only start the big adjuvant trials when we have responses and improvements in patient survival in the metastatic setting. But if we’re targeting metastatic colonisation, we won’t have this data in an already metastatic population. Maybe we can do smaller adjuvant trials, or find new ways to determine which agents to take further. Could early clinical trials use biopsies to determine that the drug hit its intended target?

Why has it taken so long to get this far? 

20 years ago, the hallmark of metastases was thought to be instability. My discovery of nm23 suggested that underneath the instability a series of consistent genetic changes were occurring, and this went against the prevailing view.

Metastasis work is also difficult. There is no in vitro assay; we have to use complex mouse models. But clinical trials based on the discovery of nm23 and the other metastasis suppressors are starting; in fact data on metastatic colonization is coming in from all angles: metastasis suppression, drug development, expression profiling. Different groups of people are coming to the same conclusion – that this is a good therapeutic target – and it’s very helpful.

At San Antonio, the audience was mostly clinicians who are using drugs and don’t know why they’re not working. They understood the relevance of this work.