78 Papers on p53

"So now I had moved to the UK. I decided I wanted to do research. I started my clinical training but I wanted to start a PhD. So I stopped my clinical training and I said, ‘Okay, the time has come. I’m going to do a PhD no matter what.’ It’s a big decision when you’ve reached a fairly advanced level of your clinical training, but it was something that I’d always wanted to do. So I started my PhD in a lab in Cambridge with Dr. James Brenton, who is currently a principal investigator at the Cambridge Institute . My PhD was on a technique called ‘expression profiling’ of ovarian cancer. Rather than looking at genes one by one, you get a read-out of the RNA level all at once from the same tumour. At the time that was very new and exciting. Now it’s the norm and who would do anything else? But then I drifted a bit and I said to James, ‘I want to take a subject that is believed to have been very well studied, and I want to read everything about it to try to understand what people do in scientific research.’ And I said I wanted to do what we call ‘a systematic review’ on p53. p53 is one of those genes that get mutated in many cancers. We call it the ‘guardian of the genome.’ It belongs to a family called the tumour suppressor genes, and its function is to detect errors and to make a decision to either correct the error or kill the cell. So if the errors in the genes are so vast that they cannot be repaired, p53 signals for self-destruction of the cell, a process we call ‘apoptosis.’ But if p53 is not functioning because it’s mutated, or deleted, or altered, then it gives a licence for the cell to transform and to become cancerous. So this was a very clear concept and I wanted to see how it related to ovarian cancer. I looked at every paper that had the keywords ‘p53’ and ‘ovarian cancer’ and I narrowed the search to ones that looked at what we call the prognostic, or predictive significance of p53. There were 78 of them. “We’re already seeing the huge power of this sort of data collection. For example, the link between taking the oral contraceptive pill and the reduction of the risk of ovarian cancer by 50% or so. ” The prevailing view, back then, was that in cancers that have intact p53, if you gave chemotherapy, you induce DNA damage. p53 plays a role in self destruction, so then p53 will kill the cells. So those cancers would respond well to chemotherapy. In cancers where p53 was not working—because it was mutated or deleted—there’s nothing to induce self-destruction, and therefore when you give chemotherapy you get resistance. So I did a systematic review with two of my colleagues. We studied exactly what the scientists had done from the very beginning. How did they make a call whether p53 was mutated or not? What was the number of samples of ovarian cancer patients that they looked at, the statistical power of their study? If they had used an antibody to detect the expression of P53, which antibody was it? Which company? All the details. If they used sequencing, what were they sequencing? Were they sequencing the entire gene, or what we call ‘hotspots’ in the gene? It was so important that I did this for my own scientific career, because as soon as I started doing it, I quickly realised that there was a huge variation in the methodology used. What one researcher defined as a non-functioning p53 might not be the same definition used by another researcher. So you couldn’t really marry up all the papers, and reach a final conclusion. More than 8,000 ovarian cancer samples from about 8,000 women were included in those 78 studies from all around the world. And the conclusion was that really, there was no conclusion to be made, because of the variation in the methodology. That was very interesting and very good to find out very early on in my PhD. It was the motivation to then say, ‘Okay now what have we learned from all of these studies?’—I sound very arrogant here—’Can we design the study that can answer this question?’ If we’re going to do that we’ve got to get samples that are fresh, frozen samples, not paraffin-embedded, old samples. We have to make sure that most of the sample has tumour, and not a lot of normal cell contamination, because that can bias your sequencing result. Also, if you want to sequence, you have to sequence the whole gene and not just part of the gene, because otherwise you’d be missing mutations. We were applying quite basic principles, if you think about it. It wasn’t rocket science. But it was quite difficult to implement, because there weren’t a lot of fresh frozen samples around that you could do the test on. It was also very difficult, at the time, to sequence the whole gene. Cost also came into it. “Ovarian cancer is full of examples of areas that we just don’t know about.” So we had the design, but we didn’t have the cohort of samples until we met David Bowtell , who’s an Australian researcher on ovarian cancer. I talked with him and David said, ‘This is a great idea. Let’s do it.’ He had a big collection of samples from Australia where they had been collecting fresh samples and then freezing them at minus 80 to preserve the quality of DNA. He sent me the first 45 samples, and I did the first round of processing in James’s lab. 44 out of the 45 samples had the p53 mutation. I said, ‘Well I clearly have done something wrong, because previously everybody had reported that it’s only in about 50%.’ “When you start reading some books, you quickly realize that yes, we know a bit about how things work, but a lot of the details are lacking.” I remember I emailed David, who had then returned back to Australia, and I asked, ‘What’s special about the 45th patient?’ He said, ‘Nothing.’ Then, about a week later, a colleague of his, who was on cc, responded to the email and said, ‘Actually, there was something different, because this patient, when reviewing her slides, didn’t have high grade ovarian cancer. She had low grade ovarian cancer.’ All the others had high grade ovarian cancer. So then we said, ‘Okay, the immediate interpretation is that, in fact, the p53 mutation is present in all high grade cases of ovarian cancer. To confirm that, we wanted to look at the next hundred cases. But now we had to apply to a committee. The committee looked at our application and said, ‘You’re wasting your time. p53 has been studied for years and years and years. You must have made errors. This is a valuable resource; you cannot use it for this purpose.’ It was very frustrating. I decided to appeal and explained that these were not errors, as we had sent the samples blindly to a reference laboratory and got back the same results. These were true results. Anyway, finally the committee agreed, and we got the next 100 samples. It turned out that 96.7% of high grade serous ovarian cancers had p53 mutations. That changed the idea that the p53 mutation is a decisive factor in whether or not chemotherapy will work. It is not. But the exciting part was also finding out that the p53 mutation is a very important—if not essential—component of the start of high grade serous ovarian cancer. Well yes. For a start, you could, for example, look for circulating tumour DNA for p53 mutations as a screening tool, which I know people around the world are testing. You could look for it in the vaginal fluid at the same time as having a Pap smear—because if you know that the mutations are essential, then you can detect those mutations early, and then you’ve got a good chance. But also, in terms of a mechanistic understanding, you can now try to understand what’s so important about the mutation of p53 for ovarian cancer cells. Why is that a requirement? So these are areas of active research. We started this work in 2001, just before I started my PhD, and the paper was published in 2010. We don’t get the samples, in fact, what we got was the DNA. They’re shipped by FedEx or DHL, and kept cold in dry ice, which is solidified carbon dioxide and keeps the temperature at about -20. David and his colleagues really had the vision when he started that collection and by the time I met him, he already had a large number of samples. Now there is awareness of the need, so when I started my lab, the first thing I did was to start a collection of frozen samples, because I knew how important it is. I started it here in 2010. That is when I moved to Oxford. They are all scientific journals, such as the New England Journal of Medicine . Essentially you go onto the PubMed website , and you do a search using the terms ‘p53’ and ‘ovarian cancer.’ You get the big list, and then you start to make a shorter listing. I think you’re very right in that. Patients can really help themselves a lot by looking and reading, and then discussing it with their doctor and going for a second opinion—and more opinions, if needed."