4th EUTROC
Biology of molecular subtypes in ovarian cancer
Professor David Bowtell – Peter MacCallum Cancer Centre, Melbourne, Australia
We’re at a stage where we’re essentially cataloguing all the different mutations that occur in different types of ovarian cancer and the challenge before us, I think, is to take that knowledge and convert it into an integrated picture of the way that cancer survives. So in an original Hanahan and Weinberg conceptual framework that they provided they suggested that there were six things, essentially, that cancer cells needed to do to become cancer cells. So what I was proposing is that the challenge before us is to take this catalogue of mutations and convert it into that kind of conceptual framework. Then I stepped through the sorts of hallmarks that we recognise at the moment for high grade serous cancers; we had to narrow the question because ovarian cancers are so different. For high grade serous cancers I was talking about the molecular subtypes which are a feature of them and the need to understand the biology behind the different subtypes. I was also talking about the fact that p53 mutations occur in essentially 100% of high grade serous cancers and so are really fundamental to the biology of those tumours. There are still aspects of that that we don’t understand very well.
We talked about the hallmark of BRCAness, the frequency of BRCA mutations, BRCA1 and BRCA2, in high grade serous cancers and pointed out that although they’re very common and dysfunction of that pathway occurs in nearly 50% or maybe a bit more than 50% of high grade serous tumours, there’s a significant fraction where there are no obvious mutation and where they have apparently intact homologous recombination repair; so one of the groups that falls into that is the cyclin amplified group which occurs in patients without germline mutations in either BRCA1 or BRCA2. So then essentially I was talking about how that leads to one of the other really key features of these tumours which is very widespread copy number change; how, at least in one instance, we think that that drives one of the molecular subtypes but also how that copy number change is heterogeneous. So if we look in tumour samples that are collected either in different samples from within a patient that are collected either spatially, so across different parts of the abdomen, or temporarily, across different times from pre-treatment, primary surgery, through to recurrence, so after the patient has been through one or several lines of chemotherapy. Those copy number changes actually vary very substantially between the samples and that reflects the evolution of the tumour and its ability to evade therapies, for example.
What trials and drugs are being developed to target the p53 mutation?
There is actually a EUTROC trial that is being put together to look at a small molecule that refolds inactivated p53. So that might be actually a very interesting tumour type to direct that kind of therapy towards because it’s really fundamental to the biology of these tumours.
Are we moving to a more personalised approach for diagnosis?
Yes, I think it’s a graded process. We’ve gone from a situation where we think of ovarian cancer as a single disease. I don’t think we actually ever really did think of it as a single disease because there are clearly different histotypes that people have been aware of, but it’s been essentially treated as a single disease in terms of its clinical management. Now we’re gradually translating from that situation to what you’re describing which is a highly individualised, personalised approach. So in between, the first way station has been the recognition that the different histotypes deserve different therapies, so there are a number of clinical trials now that are directed towards the different histologies of ovarian cancer. The next step along that journey is to take that stratification further, so stratification on germline mutations in BRCA1 and 2 is obviously a target group. The next step again is to, for example, think about clinical trials in molecular subtypes, if there are obviously different therapeutic targets in different molecular subtypes you can imagine the design of trials in that way. Then progressively, I guess, we go on down trying to target it more and more. One of the most interesting areas will be the targeting of therapies, the design of individualised therapeutic approaches in the relapsed setting. So the majority of epithelial ovarian cancer is high grade serous cancers, the majority of those patients have an extremely good response to platinum; it’s going to be very hard to really change that first line setting therapeutic approach but where things essentially fall apart is in the relapsed setting. So there are some guidelines in terms of first relapse based on the length of time between the end of treatment and relapse in terms of guiding how the patient should be treated. But after patients have failed second line treatment it then essentially becomes a shotgun approach in terms of the way that they’re managed. So I think the opportunities for individualising therapy are probably greatest in the recurrent disease setting. But to do that we need to get samples so that we can identify the biomarkers, understand the mechanism of drug resistance that’s occurring in that acquired resistance setting and then start to individualise therapy based on that.
Is there a need for more biobanks to expand this?
I think it’s moderately high. I think it’s moderately high; I think there are now some quite large collections that are well annotated for ovarian cancer. I think the greater need is actually a collection of samples from patients on treatment, particularly with molecularly targeted therapies and a collection of those paired samples from patients who were originally sensitive to treatment and now have become resistant to whatever the treatment is. So I think the challenge is there to change what is standard of care. At the moment we’ve got this crazy situation where we’ve put a huge amount of effort into identifying prognostic and predictive markers in a primary treatment setting and it’s essentially established now as part of the way we think about modern cancer management. When we come to the relapsed setting, the conversation goes with a clinician, “As a researcher I’d be really interested in getting hold of some samples in a relapsed setting,” and the clinician will say, “Well it’s pretty hard to get biopsies because it’s not standard of care.” And you reply, “Well why isn’t it standard of care?” “Well, because there are no biomarkers to guide therapeutic decision making.” “Well, why aren’t there any biomarkers?” “Well because people haven’t collected any samples to discover the biomarkers.” So we’re in this sort of crazy loop, in a sense, that needs to change. We need to be able to get samples in a recurrent setting, we need to be able to identify mechanisms, from that will come the biomarkers and then it will become standard of care to individualise treatment. I think that’s where the big uplift is likely to be, at least in high grade serous cancers.
What were some of the highlights from this meeting?
I’ve greatly enjoyed the meeting; I think this has been a wonderful meeting. It’s quite small in terms of the number of people here, a little under a hundred people, I think, but the advantage of having these very focussed meetings is that they’re small but really dense and rich in terms of the quality of the science and its appropriateness to the things we’re studying.
