IGCS 2010, 23-26 October, Prague
The Cancer Genome Atlas
Dr Douglas Levine – Memorial Sloan-Kettering Cancer Center
Here at the IGCS you’ve been speaking about the Cancer Genome Atlas and the Cancer Genome Project, can you speak a little bit about this, give us some background?
I’ve been very fortunate to be part of the project for a number of years now. The Cancer Genome Atlas Project is a collaborative project between the National Cancer Institute and the National Human Genome Research Institute. It started as a pilot project to see if they really could comprehensively catalogue all of the genomic aberrations in a number of cancers. They picked three cancers to do as a pilot project which included glioblastoma, ovarian cancer and lung squamous carcinoma. The glioblastoma project was completed approximately a year ago and the ovarian cancer project is just finishing up now and the lung cancer project is still on-going.
So what did they map in this project?
The idea is that they want to perform a comprehensive and integrative catalogue of all the genomic changes and so they are mapping alterations in the tumours that occur at the DNA and the RNA level; the project does not include any protein studies at this time.
So I imagine there are going to be millions of changes, some of them are going to be relevant and some of them are just going to be spontaneous or acquired. So how do you decide which are the important changes?
With current technology, which is always improving, we can try to survey all the changes that exist in the genome, the question then is which ones are the important ones. So you can take a couple of approaches, you can look at the genomic changes that are common across cancers and you can find things that just are relevant to cancer in general, and then you can look at the genomic changes that are unique to ovarian cancer and try to figure out what makes ovarian cancer different from other cancers.
So in ovarian cancer what has come out, or is it too early to say?
No, it’s not too early to say. The first analysis has been completed and it has been quite successful. I would say that some of the excitement is around the number of copy number changes that were found in ovarian cancer to the exclusion of frequent recurrent mutations. So in many other cancer types you often find a small number of genes, maybe half a dozen to ten genes that are frequently mutated at a significant event rate. For ovary cancer we found that the TP53 gene is mutated and altered in almost all of the tumours and then the genes that code for BRCA1 and BRCA2, which is unique to ovary cancer and also to breast cancer, are commonly mutated. The other aspect of ovarian cancer that was somewhat of a new finding is just the extent of the copy number changes that are present across all the tumours and these are focal amplifications and focal deletions.
So what does this mean for therapy?
Once you have a comprehensive survey of the changes you can then use that information to do things to advance the mission of improving the outcomes and treatments for your patients which, in this case, is women with ovarian cancer. So looking at the copy number changes, you can group them into pathways that may be targeted by new molecular agents, so one way to use the data is to develop better therapy. You can look for mutations, of which case we have common mutations in BRCA1 and 2, and there is a whole class of drugs now that people are familiar with, called PARP inhibitors, which target defects in the BRCA1 and 2 pathway. You can also look at the genomic changes to identify better diagnostic predictive signatures to figure out which patients are going to do better, which patients might do worse and which patients might be more likely to respond to a particular therapy.
Probably one of the main things that has come out for ovarian cancer, because BRCA1 and p53 is quite well-known, is the copy number.
That’s one of the new things but I wouldn’t say that BRCA1 and BRCA2 were already known. So we already knew that germline mutations are very common in BRCA1 and 2 but there’s a whole bunch of sematic mutations that we identified as being mutated in 6% of cases. So when you consider that we knew that BRCA1 and 2 were mutated in the germline in 15% of cases, but now you add 6% on top of that, you’ve increased the number of mutational defects by 30-40%. That’s a significant contribution when you consider the attention that is already being focussed on BRCA1 and 2. The second thing we found out that was also a novel finding is the extent of BRCA1 methylation which also occurred in 10% of cases. So now if you look at the defects in BRCA1 and 2, you’re up to 30% which you’ve doubled the knowledge that we had before this project was undertaken. So this may be unique to ovary cancer and may have corollaries in breast cancer but it’s probably unique from other solid tumours.
What are the percentage of women, for instance in the States, now that are actually screened for BRCA mutation?
It’s somewhat centre specific because it’s not just the ability to do that screening. So we obviously can screen people with today’s technologies but they patients have to get to the genetic counselling individuals, they have to have reimbursement to have the testing done. At our centre we’re trying to screen every patient with ovarian cancer and we’re quite successful, certainly we don’t get every patient but we’re pretty successful. But there are lots of logistics that are involved so I don’t think it’s being done widely across the country but it certainly is recommended by many national organisations.
Lots of people don’t appreciate that we knew about BRCA1 and 2, but what we’ve learned more about it is a lot more than just what we knew before. So the copy number changes are new, the lack of recurrent mutations is also interesting. So we thought we’d find a couple of genes that were going to be frequently mutated and we found mostly the genes that we already knew. There are a couple of new candidate genes, there’s a tumour suppressor called FAT3 that has been mutated in a number of cancers that may be useful. But it was somewhat hoped that we would find other genes with recurrent mutations but the negative finding is equally important as a positive finding, it will prevent others from doing potentially unnecessary research.
That’s important to publish negative results, which doesn’t happen.