Steve Rees: Homing in on the right targets
Image Credit: Steve Rees, Photo courtesy of Steve Rees
Click here to listen to the full podcast episode
Kat: Steve Rees leads the Discovery Biology group at AstraZeneca, where he and his colleagues are focusing on this very first step of the drug discovery journey: coming up with the new ideas and targets that could lead to the new medicines of tomorrow.
Steve: So the choice of the target in a drug discovery project is the most important decision that we make. The target is the protein, It's normally an enzyme or receptor, that we choose to take into a drug discovery program to find a molecule that will either activate or inhibit it, to result in treatment in disease. So taking an example that most people would be familiar with. So everybody who's had hay fever takes a medicine that's called an antihistamine. The medicine in there is a small molecule that binds to something called the histamine receptor. That receptor is found on immune cells in the body, nd in essence, it switches off that receptor to stop the inflammation response. That receptor, the histamine receptor, is what we call the drug target the target of the medicine that results in the treatment for the disease.
Kat: So why do we actually need to find more targets? Have you not got enough?
Steve: So our primary drive isn't necessarily to find more targets, it's to find better targets. It's to find targets, which if we can take them through the drug discovery process, we'll be able to find a medicine that works in disease. So a fundamental challenge facing the entire drug discovery industry is that many of our medicines fail and those medicines fail when they reach a phase two clinical study. And that's the first time that we test a medicine in patients to ask the question, does it work in disease? Most of the time we get that wrong, 60, 70, 80% of the time we get that wrong. And the reason we get that wrong is because we don't fundamentally understand disease. And perhaps more importantly, we don't understand the differences within the same disease. And taking lung cancer as an example, 20 years ago we believe there were three types of lung cancer. Today we know there are 40 different types of lung cancer and each of those different types of lung cancer may require a different medicine.
Kat: And so how have we found these targets in the past?
Steve: So for most of the last 25 years, and certainly since the sequencing of the human genome, the approach that we've taken is we have a receptor on an enzyme that we've identified and we ask the question, which disease is this receptor or is this enzyme involved in? And that involves things like understanding which tissue is it expressed in. So if it's expressed in the heart, maybe it's a target for heart disease. If expressed in the pancreas, maybe a target for pancreatic disease and on and on. And we then do research science to try and build evidence that says that yes, this target is expressed in disease. The change that we're now moving towards is to turn this paradigm on its head. And we start with disease and we work backwards from the patient to ask the question, which targets do we see that have changed in the patient that we could potentially take forward to find medicines for?
Kat: So what's changing now, how are you trying to find these better targets?
Steve: So what's changed over the last 20 years is the technologies that have been available to us to really understand disease better. And this starts with the patient, understanding what disease looks like in the patient through being able to identify patient tissue. And the biggest transformation has been the advances in sequencing, the ability to sequence all the genes in the genome. And it's possible today to take skin cells from any one of us, tumor cells from a cancer patient, and sequence the genes of that tumor and we can do it for less than about $800. 20 years ago that cost a hundred million dollars. We can now do it for $800. And what that allows us to do is to understand which genes are changed, in which disease state to begin to understand two things. Firstly, how disease varies across different people, but also to understand which genes are changed, so which drug targets are changed in that disease, that we could try to find medicines for.
Kat: So let's talk about how you're trying to use this information to find better targets. I mean, how does that actually work? How do you go from all this information about the genes that you have in patient tissues to actually then finding better targets?
Steve: So the major advance in recent years has been our treatment of cancer. So cancer is largely a genetically driven disease, it's caused by changes in genes within cells that cause that cancer tissue to develop. So by taking cancer tissue from patients, we can sequence that cancer tissue, or we can ask the question for all of these people who have lung cancer, what are the genes? What are the genetic changes that we're seeing in those patients and how do they relate to disease? And that allows us to say, well, in this particular group of patients, we know that this particular protein has changed. It's mutated, it's become more active or it's become inactivated. And that leads to the hypothesis that if we found a drug that either switched that protein on or switch that protein off, it would have activity in disease. And this is one of the reasons, or the primary reason, why the field has now been so successful at delivering cancer medicines into the clinic, because we now understand what cancer looks like at the genetic level, at the patient level. We're able to develop drugs against those targets that are changed in the patient. And most importantly, through the advances in what's the so-called precision medicine or diagnostics, we're able to identify the right patient to give the medicine to where it works.
Kat: So let's drill into this process even more. So you identify a molecule, a protein, the gene that you think is really important in this disease, it's changed in the disease. You're like, right, we're going to develop a drug against it. How can you be sure that that is actually the thing that you think you've got, that it is actually involved in it? How do you know that the gene that you found is really doing what you think it is in the disease?
Steve: Well, as always, you can never be completely sure, but the methods that we would use again, taking cancer tissue as the example, is we would do experiments in the laboratory where we would introduce that mutation into cells, in tissue cultures, we call it in the laboratory, and we ask the question, what does that change cause those cells to do? If that change causes the cells to grow faster, or maybe that's some evidence that actually this is a change that leads to cell growth in cancer. We then have methods in the laboratory where we could take that same gene and we could delete it. We can turn it off. We can remove it from cells and ask the question. What happens when we do that? And we tend to work through a series of experiments, both in cell models, as we call it in the laboratory and culture, then moving potentially into disease models in other organisms may be including the likes of transgenic mice, whereby we would ask the same question to try and understand the role of that change in those so-called model systems to just build up a body of evidence that tells us that actually, yes, if we are able to find a medicine that acts at this target, it's likely to work in disease.
Kat: And this seems very sensible and in some ways, well, this seems a very obvious thing and why hasn't this really been done before? Because we've had things like knockout mice, where you can take out genes and things like this. So what's really changed.
Steve: So it hasn't been done before, but for me there are three or four major technology advances in the last seven or eight years that have transformed our ability to validate targets. The first we've already discussed, the ability to sequence a genome for just a few dollars, completely transformative in terms of our ability to identify genes associated with disease.
Steve: The second is the development of CRISPR. So many of us have heard of CRISPR on the news. CRISPR was identified eight years ago. What CRISPR allows us to do is to specifically turn on or turn off any gene we choose in the genome of any cell. And that allows us to ask the question, what happens to this gene If it's turned on? What happens to this gene if it's turned off? And how does that relate to disease? And those are experiments that A-level scientists can now do today.
Steve: The third of course is big advances in our ability to manage and interpret data. We're now able to use artificial intelligence, knowledge graphs to take all of this data, to take the world's data and ask the question. If I surveyed the entire world's data, if I surveyed all of AstraZeneca's data, what does that tell me about which genes could be involved in disease?
Steve: When you bring those three technologies together, we have an incredibly powerful engine that allows us to generate new ideas, new hypotheses that we simply couldn't do before. And then of course, we now have the experimental methods in the laboratory that allow us to test those hypotheses and test them very quickly again, in ways that we couldn't do even three or four years ago.
Kat: One of the things that amazes me about biological research in the past few years is the scale of automation. And during my PhD, I tried to knock out one gene, very unsuccessfully over several years, and now you're able to do thousands and there's robots and all kinds of things. What sort of scale are we talking about now of being able to try different ideas and check out different targets?
Steve: Again, at a scale that was unimaginable a few years ago. I mean, in my early years in the industry, I cloned a gene called the 5-HT5A receptor. That took 12 months. In the lab today we could clone that receptor and it would take about four hours and that's due to the advances in the technologies that we have available to us.
Steve: But, picking up on your question. We have incredible interest in the science of functional genomics and what we do there is we have this technology called CRISPR and rather than just deleting a single gene, we have 20,000 CRISPRs and each one of those will delete a different gene in the genome. And those CRISPRs are aided by something called a microtiter plate, which is about the size of your iPhone, and in a single microtiter plate we're able to do 384 separate experiments to ask the question, what happens to the biology I'm interested in if I delete 384 genes.
Steve: So you scale that up to 20,000, take an example of the sort of experiment we do in cancer. We're very interested in what leads to resistance to our medicines. So if we have a cancer medicine, it's been in the clinic to treat lung cancer, the way that we do that is that we take cells in the laboratory that we believe are good models of lung cancer, we take our CRISPR technology and we'd run what we call a whole genome screen that allows us to delete each and every gene in the genome of those cells on a one by one basis. And the question we ask is which genes that, when we delete them using that CRISPR technology, allow those cells to grow in the presence of that medicine. Those genes then represent potential new targets in lung cancer, which allow us to run a new drug discovery project, to identify a molecule that inhibits that new target, which we believe we could then use to treat patients who are resistant to the original medicine.
Kat: So that's one example of something that you're trying to do. What have you done so far? Are there any success stories that have come out of this kind of approach?
Steve: Well, the science is very early. It's only in the last year or two that we've been able to run these functional genomic screens using libraries of CRISPR reagents. So in terms of success stories, our successes today, our new projects entering AstraZeneca's drug discovery portfolio that have been discovered using this approach. And we're now in a situation where around about a quarter of our new projects originate from ideas that we've generated internally at AstraZeneca using these various different methods to identify new targets, to move into the portfolio.
Kat: You've talked quite a bit about your approaches in cancer, but obviously other diseases are available. So does this kind of approach work in other diseases? What sorts of conditions are you looking into?
Steve: So this is the real power of these approaches in that they work in any disease and every disease. Taking functional genomics as the example, all we need to run a functional genomic screen is a cellular model that we believe is predictive of the disease state. So in our laboratory today, we are running functional genomic screens to identify new targets in chronic kidney disease, in heart failure, in a variety of different lung diseases and also in CNS diseases as well. So potentially these methods allow us to better understand any disease and to bring forward new targets across the full range of diseases that are affecting each and every one of us.
Kat: So bringing all of this together, what are the benefits of finding more targets, better targets? How is this going to change drug discovery? And then ultimately what's this going to do for patients?
Steve: So in terms of patient benefit, this will lead to more medicines, this will lead to better medicines and, combining these technologies together, it will lead to a situation where we only ever give a medicine to a patient where we have high confidence that that medicine will work because in essence, we tailored that medicine to the specific disease that that patient has had. From a pharmaceutical industry perspective, drug discovery perspective, if we're successful, we will significantly reduce the failure in phase two clinical studies, it should also reduce the timelines to take new medicines to patients. So, taking these various innovations together, our aspiration is that we can make more medicines, we can deliver those medicines faster and we can do that at a cost which is significantly lower than the cost that we do today.
Kat: Steve Rees from AstraZeneca. And if you want to find out more about CRISPR and how it works, just scroll back to our previous episode on the history of genome editing.
