IGCS 2012 - Vancouver, BC, Canada
The Cancer Genome Atlas and high grade serous ovarian cancer
Dr Andrew Berchuck – Duke University Medical Center, Durham, USA
http://ecancer.org/tv/conference/157/1685
Dr Andrew Berchuck, we’re very pleased to see you here because you’ve been looking at something that’s actually very complex but could be extremely useful - the Cancer Genome Atlas and how it impacts on a disease you’ve been talking about her, high grade serous ovarian cancer. Can you tell me first of all what you’ve been doing here?
Well, high grade serous ovarian cancer is the type of ovarian cancer that causes the most deaths from this disease. There are several other types but they tend not to be as lethal. So the Cancer Genome Atlas Project was funded by the National Cancer Institute of the United States and the goal was to perform a comprehensive genomic characterisation of most of the common forms of cancer. What I mean by that is looking across the whole genome at mutations in specific genes; looking at changes in what we call amplification or deletion of genes; changes in copy number, really any type of way in which you can study the genome is being done. Rather than on a gene by gene, very specific, way, just looking at certain genes that we think might be interesting, it’s looking comprehensively across the whole genome.
Now, to the lay person it sounds impenetrable because there’s a lot of complexity in the genome, isn’t there?
Absolutely but if you look back over the past decade or so there has been a stepwise progression of knowledge from the Human Genome Project that told us there were about 18-20,000 genes and really was the first map of the human genome to then the HapMap project which has mapped human variation between individuals and so this is something that has been building, it didn’t just come about out of nowhere.
And then there are interactions so it might not be just one gene.
Yes, with the genomic characterisation, in this case of high grade serous ovarian cancer, we’ve found that in fact most of these tumours have a huge number of genetic alterations, that’s what’s kind of scary about this type of cancer is that for some types of cancer it appears to be driven by just a very few events and you can wrap your arms around that and think about targeting it with drugs. With ovarian cancer there’s just a myriad of genetic alterations and it really is intimidating.
What have you, in fact, been able to distil from the data that you’ve gleaned on ovarian cancer?
We’ve found that the p53 gene, it’s a tumour suppressor, is the most frequently mutated gene in this disease and essentially every single one of these cancers has a mutation in p53. We also now know that the BRCA1/2 genes are the next most frequently mutated and, again, those are not in every cancer. So in what we thought of as a disease that’s maybe one disease, we’re now starting to develop subsets based on whether or not a tumour might have this or that mutation and then starting to begin to think about maybe we could treat those as different diseases, not just all lumped together as one disease, one size fits all treatment.
So is some clarity coming from it which would give you some clinical directions in which to steer?
Yes, again getting back to the BRCA1 and 2 genes, those genes are in a pathway that’s involved in homologous DNA repair. There are a number of other genes in that pathway that also can be mutated or amplified, altered in some way that’s abnormal, and overall about 50% of these cancers have alterations to some gene in that DNA repair pathway. So there’s actually a class of drugs under development called PARP inhibitors that seem to work best in cancers that have an alteration in some gene in that pathway. Right now there are clinical trials going on in ovarian cancer to see if those drugs will be a valuable new weapon in our armamentarium in fighting this disease.
Of course, it is difficult to know which ones of these genetic processes are going to yield a harvest, doesn’t it? There’s something else called methylation that comes into this as well, isn’t there?
Yes, methylation is another way in which genes can be regulated so, in this case, you’re not directly damaging the gene itself but there’s a switch next to every gene that’s an on/off switch and when the gene is methylated the gene is turned off and when it’s not methylated it’s turned on. So that’s another approach to potentially modulating the behaviour of a cancer.
But you’ve found some things in ovarian cancer?
Well there are definitely a number of changes, alterations. Right now we’re not sure that any of those are immediately approachable in terms of translating into therapy.
Some signalling mechanisms as well, I see Notch and pOX come into this, don’t they?
Again, because the TCGA project involved a comprehensive genomic characterisation we’ve been able to look pretty much under every stone, so to speak, and so there are certain pathways by which cells regulate their growth, like you mentioned the Notch pathway, PI3K, Ras pathway. There are drugs under development that target these pathways and, again, hopefully in the future we’ll be able to look at a patient’s cancer and target more specific therapies that match their cancer.
So from this genomic characterisation what’s in it for the hard worked cancer doctor so far?
I think we all feel that knowledge is power. For the longest time we knew very little about the molecular pathogenesis of cancer in general. Right now the TCGA project is providing a detailed anatomy of the molecular pathogenesis of cancer and I think with a lot of research and discovery it remains a bit unpredictable how this will develop in terms of whether it’s improved early detection, prevention, treatment. I think all that is a possibility and is open.
But you are reasonably optimistic that this blue skies thinking is going to go somewhere practically?
Yes, again I think the clinically relevant applications are hard to predict with any specificity, but one of the really exciting things is that all this data is being made publically available on the internet and so it really is fuelling the entire research community around the world to advance knowledge.
Let me ask you briefly, if I may, about another small paper you’re doing here on the outcome of metastatic Lynch syndrome associated gynaecological cancers. What’s happening here, what did you find?
It’s been known for some time that in the breast-ovarian hereditary cancer syndrome, which is due to inherited mutations in BRCA1 and 2, there’s an increased risk of both ovarian and breast cancer but, interestingly, in the women who develop cancer their response to chemotherapy and survival is actually better. So it’s a bit of a paradox there – you have an increased risk of cancer but your response to treatment is better.
Because you can find a switch to turn off?
Yes. There is just a difference there in that because BRCA1 and 2 mutations reduce DNA repair capability, chemotherapy works by causing DNA damage and if that DNA damage isn’t repaired a cancer cell will die, which is a good thing. So chemotherapy can cause DNA damage and in women who don’t have a BRCA1 or 2 mutation that might be repaired and the cancer cell might survive whereas with a BRCA1 or 2 mutation it’s less likely. So what we found in this current paper is a similar sort of paradigm, potentially, in Lynch syndrome. Lynch syndrome is another cancer syndrome, it’s also sometimes referred to as hereditary non-polyposis colorectal cancer syndrome, or HNPCC. That’s because colorectal cancer is the most common type of malignancy in individuals who inherit these mutations but it turns out that endometrial cancer is the second most common malignancy and there’s up to about a 40% lifetime risk of endometrial cancer in women with Lynch syndrome. What we found in this particular study is that in patients with Lynch syndrome associated cancers, primarily endometrial cancers, when they had metastatic disease and were treated with chemotherapy or radiation, they actually had a very, very favourable survival, more than we would predict or suspect if we were talking about our global experience in this disease.
That’s really interesting. Now, because of all these genetic factors you’ve got something called the Ovarian Cancer Association Risk Consortium and you’ve got BRCA1 and 2, which are strong risk factors. You’re looking for others and there are real hopes of intervention to reduce the risk of people affected, isn’t there? What’s happening here?
Correct. As I’ve alluded to earlier in our conversation, the BRCA1 and 2 genes, when they’re mutated a woman’s risk of ovarian cancer is much increased, so up to in the range of 40-60%, which is quite high. So there has been fairly clear consensus around the paradigm of letting these women have their family and then around age 40, when they’re approaching menopause, to do prophylactic removal of the fallopian tubes and ovaries and you can prevent most of the cancers in the women who carry BRCA1 and 2 mutations. Again, that’s been well accepted because the risk of cancer is so high. But we think that there are probably other genetic factors that also increase risk of ovarian cancer but that these increase risk much less so you don’t see the striking clusters of cancer in families. With these smaller increases in risk you have to do population studies with thousands of women to see a slight increased risk or decreased risk associated with a certain genetic variant.
There is, however, a really tough prospect for some of these women. They face bilateral salpingo-oophorectomy, also they might go on to face bilateral mastectomy; it’s pretty tough for them, what are the psychological issues that they face?
Prophylactic surgery is not the only option. Again, with our studies in the Ovarian Cancer Association Consortium where we’re trying to identify these other genetic risk factors, we have about fifty groups working together on this research over the past five years and we now have about a dozen genetic variants that affect risk. So our idea here is that by combining the genetic risk factors with epidemiological risk factors, for example oral contraceptive pill use reduces risk of ovarian cancer, having children, breast feeding, other risk factors like endometriosis increase risk, the idea would be to work towards a personalised risk stratification where at one extreme we tell women with a BRCA1 or 2 mutation, “You have an extraordinarily high risk, we really do for sure recommend prophylactic surgery,” on the other extreme, women with very low risk, probably no intervention. But with the new genetic risk factors and other epidemiological risk factors we may be able to find that moderate risk group that maybe have a 3-5% lifetime risk and there not only prophylactic surgery but potentially screening, which is not recommended currently in the general population, typically ultrasound type of radiology scans or the CA 125 blood test are recommended… well they’re not recommended for screening but those are the two that are most studied as potential screening tools in ovarian cancer.
They may become…
The problem is ovarian cancer is fairly rare in the population so most screens with those modalities that are positive are false alarms, false positives. But if you start to focus screening on a population with a higher risk it might perform better.
It all sounds like a really exciting area. What’s your advice to cancer doctors right now?
Just keep your finger on the pulse of what’s happening. As I said, it’s very exciting that we’re gaining the knowledge of the molecular anatomy of cancer and I think that the clinical implications will vary between different types of cancer. In breast cancer, in colon and some others there are already widely used targeted therapies that have come out of that knowledge; we don’t have that in ovarian yet and we may never but in ovarian it could be better prevention that is the fruit of this labour.
And the role of the doctor, clinician, in making all of this happen, whether it’s a cancer clinician or perhaps another clinician, what should that be?
One of the really interesting things is that you’d assume when we have exciting new knowledge that it’s immediately disseminated and it becomes integrated into practice. But in fact we know from hard experience that there’s a process that takes place once a discovery is shown to be probably useful there have to be guidelines developed and physicians have to be educated and there really is a whole process of getting physicians to accept this, and patients, and to understand the pros and cons and strengths and weaknesses. So the implementation process is something that the people doing the research often take for granted and forget about but without it there’s a great lag in the implementation of these exciting discoveries.
Well let’s hope that the implementation will take place a bit quicker thanks to your visit here to ecancer.tv. Andrew, thanks a lot.
It’s a pleasure to be here.
