Drugging the cancer genome: opportunities and challenges
Professor Paul Workman – Institute of cancer Research, UK
You’re really filling the gap that a lot of the pharmaceutical industries don’t want to risk, is this correct?
Yes, what we see is an incredible opportunity to do innovative new drug development. Masses of data coming out of cancer genome sequencing, functional screens, we have lots of molecular targets to work on. On the other hand, it’s a very difficult time economically for the industry – lay-offs, risk aversion and so what has developed is this problem that’s described as the Valley of Death. On the one hand we’ve got exciting biology going on in academic research and we’ve got very powerful pharmaceutical companies that know how to develop a drug but there’s this gap in between, a chasm really, in between and our job is to take on these difficult targets. These targets might be difficult either because they’re risky biologically, in other words although there’s been some proof of concept in maybe an animal model, it hasn’t really been shown that you can develop a drug and that it will work if you develop the drug. So there’s biological risk. On the other hand, you can have targets where there’s very strong technical risk, there’s a potential druggability gap. So you might have a gene that’s clearly important in cancer and you can definitely show that if you modulated that gene, you can show by genetics that it would cause a therapeutic effect, but you don’t actually know that you can make a drug against it. Does it have a pocket, does it have a mechanism that you could design a small molecule or an antibody against? Particularly a small molecule is a problem. So we take on these two types of targets and we take them to the point where we’ve done enough to show, for example, that a biomarker will change in a tumour cell; probably that if you gave a prototype drug to an animal tumour you would get a response. Usually that’s the point where the pharmaceutical industry or a biotechnology company would come in with us and say OK, we’ll co-invest now on this project and then move it as fast as possible to get it to the clinic.
You’ve actually been quite successful, you had sixteen potential drugs, I think, six of which you took to phase I.
So over the last five or six years we’ve identified sixteen drug development candidates, these are ready to move towards the clinic. We’ve taken six of our drugs into phase I clinical trial in the Royal Marsden Hospital, which is our partner organisation. And of course the great success is that our drug, abiraterone, that was designed and synthesised in our institute and initially trialled in our institute and the Royal Marsden, has been approved in America and Europe and Canada. Of course we’re excited about that, now we want to see our next drugs move forward equally successfully.
How long did it take for abiraterone to get to be approved?
Abiraterone took a long time, it was actually synthesised and published around 1993 and its development moved relatively slowly. It was perceived as a very risky approach for a couple of reasons. One is that it wasn’t at all clear at that time that hormone refractory prostate cancer, which is the big clinical problem, it was not at all clear that this disease remained hormone dependent. In fact, work with abiraterone in the clinic, it was that work that showed, really confirmed, that patients with this hormone refractory late-stage metastatic prostate cancer are still dependent on androgen drive, it’s simply that they don’t respond to the particular drug. So there was some resistance about the biology and there were also concerns about the side effects, as to whether if you inhibited steroid synthesis you might have severe side effects. So it took some time for the trials to be done in a non-profit setting and then the biotech company Cougar came in and resourced the trials and, of course, Janssen then eventually came in and took it through to approval.
I think it exemplifies, that story exemplifies, something that we’re trying to really improve upon. It demonstrates the risk aversion problem but it also demonstrates the value of a non-profit academic group taking on that target, championing it through to showing proof of concept, not just pre-clinically but actually also clinically, in clinical trials, but then also shows the value of a biotech and a pharma company coming in. Now what we need to do is to do it much faster in the future. So if you take the case of the HSB90 inhibitor drug that we’ve developed, and I guess that’s the next example that’s coming through, it’s now in phase II trials. There’s another example where I suggested this target, generally it was seen as very risky, probably adverse side effects, never been done before, some concern. We took it through to, first of all, showing with an actual product inhibitor, working with the NCI and Cancer Research UK, we took an example drug into the clinic and showed biomarker changes which is important, we also showed some clinical activity. In the end that drug didn’t go forward because of liver effects, which we predicted, actually, probably would be a problem. So in parallel with that early clinical work we started a small molecule project. We ran a high throughput screen in our academic group, we discovered early prototype inhibitors, we then partnered with a small biotech company in the UK who became Vernalis and within two years from starting the high throughput screen we had a development candidate. So only in a two year period and within a couple of years of that we had the drug candidate in the clinic. We licensed the whole project to Novartis, Novartis then took this on and now Novartis trials at ASCO, just a couple of weeks ago, revealed 20-30% response rates in trastuzumab refractory breast cancer and in non-small cell lung cancer EGFR mutant patients and ALK translocation patients. And so that shows that you can maybe try to emulate the success of abiraterone but do it in, say, five or seven years rather than the twenty years or more that it took.
We’re here at WIN and there’s a lot of talk about biomarkers. Biomarkers are changing the way people think about drug development, can you just say two words on this?
Biomarkers, I think, are critical in a number of ways. First of all we’ve moved from the one size fits all, all patients get the same drug, to defined subgroups, often genetically defined subgroups with a particular mutation – a translocation, an amplification. And those patients, based on their genetics, should be responsive and therefore you need a companion biomarker to predict which patients will be sensitive and that’s working well and we have good examples of the success of that and that’s the way forward. Because if you use a biomarker, you’ll enrich the trials for patients who will be responsive so you don’t need to treat the patients who won’t respond, it’s better for patients. You’ll have a higher success rate because you’re only treating the patients who have a good chance of responding and the trial will be quicker and it will be cheaper than if you treat all comers. So it’s good for everyone – it’s good for patients, it’s good for the costs and therefore the pharmaceutical industry and the healthcare providers and so on. So I think that’s really important.
Then there’s another group of biomarkers which we call target engagement or proof of mechanism biomarkers, pharmacodynamic biomarkers. These allow you to show that the target is being inhibited by your drug, that the pathway is being modulated and can quantitate that. So you can say, “I can show 95% inhibition of my target pathway for 24 hours,” and the pre-clinical models will give you a target to aim for and in the clinical trial you can shoot for that exposure. We apply this in something that we call the pharmacologic audit trail. This audit trail is around, first of all, using the predictive biomarker to select the right patient and then using the target inhibition biomarker to tell you if you’re getting the right dose, the right schedule, the right frequency of administration. If you couple that together with the tumour effectiveness and the side effects, if you put all that information together, you can come out with a much more rational and sensible schedule for having the best chance of getting a response and therefore a successful outcome from the whole drug discovery programme. So one of the things that’s been discussed at this WIN meeting is the problem of how do we translate these great ideas from basic research into patient benefit; how do we do it at a time when we have a massive amount of biological information, genetics, and yet we have a financial climate that’s adverse, we have an attitude to risk that’s become rather conservative. If you couple all these things together, particularly the use of biomarkers, it should make the cost equation add up, it should make it more feasible for industry to develop drugs for smaller groups of patients, so no longer blockbuster large populations but very important genetically defined subpopulations. So the biomarkers are critical to all of this. To be successful you have to start with the biomarker at the same time as the drug, you actually have to think about the clinical trial right at the very beginning – what’s the target population? So that the biomarker is ready when the drug comes to the trial.
In fact do they call them the companion biomarkers?
Companion biomarker and drug, absolutely, that’s right.