Q&A: Personalized medicine	February 2009

David Sidransky, M.D., chairman of the board of directors at Champions Biotechnology Inc., is one of nine renowned scientists profiled in this month’s R&D Directions. Below, the 20-year oncologist, known for his work in early cancer detection (he was named by Time as one of the top physicians and scientists in 2001), expands on the topic of personalized medicine and the tough challenges researchers still face in developing effective targeted therapies.

 
Q: What are the major advantages of personalized medicine?

A: My work has been aligned in molecular markers. We’ve always dreamt about the ability to put together markers in patients and being able to benefit them. The big advantage of course is that you’re able to select a therapy that’s really going to help patients. The first thing you do is you void the toxicity of therapies that don’t work. One way is knowing that it doesn’t work, the other one is seeing the toxicities in the body – the hair loss, and everything else. The other rising advantage of personalized medicine is that, frankly, even to talk about it 20 years ago was almost ridiculous because what else were you going to give somebody if not the drug that was available. Would you really not give them that drug? Now, luckily we’re in a situation where the number of drugs that have been approved is growing. Frankly, not even the pharmaceutical companies have the ability to go through every different type of cancer in every drug. So you’re left with drugs that have a long track record, that have been around for a long time. Some new drugs that have very specific indications and may have or are known to have potentially some activity in other cancers, you just don’t know what percentage of patients are responding or how to give it to them. The notion that you can finally personalize medicine and pick those drugs is obviously very attractive for all those reasons.

Q: Do you see the pharmaceutical industry – big pharma, in particular – beginning to shift more of their research toward personalized medicine, perhaps realizing the benefits of that approach?

A: That’s a complicated questions for big pharma. I’ve been an adviser and insight to these companies for a long time. I would say that the general answer up until recently was no. There is not a lot of incentive because the thing they wanted to do was put a drug through and get it approved. Then the real goal is just to expand its use. I think that there are two things now that are pushing the pharmaceutical companies into this area as well. The first one is the FDA has announced that if you have a targeted therapy over the next few years, they really want to see these companies coming forward with a diagnostic test that identifies that target. If you’re going to hit receptor A, they would like you to figure out, or at least measure receptor A so we know if it really works in people with receptor A. That’s one thing that’s getting them a little bit more anxious about making sure they do that. The other one is now that there are many more drugs, it’s not clear sometimes how they are going to fit into clinical practice. So the biomarker or the personalization allows them to select a group of patients, both for their trials and then for the clinical indication. Increasingly, as it’s not possible to just say I’m going to approve a drug for all lung cancer first line, they’re interested in getting a piece of the pie that they know is at least a piece of the pie that they can define and define the market. They’re more and more interested in doing that, to be able to get a clinical regulatory strategy beyond the biologic underpinnings that drive the development of the drug.

Q: With so many clinical trials that ultimately miss their target or end up exposing patients to unnecessary side effects, do companies perhaps see personalized medicine as potential solution to the high costs of drug development?

A: That’s a little bit too visionary. I think it does maximize their return on investment because just the ability to get a drug into the market is all you need to justify your return on investment. Anything short of that is zero, that’s the bottom line in the pharmaceutical industry. If it gets them over the hump and gets them approved, even if it’s a small piece of the pie or anything else, that’s still a big attraction for them. Again, it’s not that visionary, but I think if it gets them over hump, then obviously it’s working in their favor.

Q: How have advances in biomarker development played a role in personalized medicine?

A: I think you have to compare the dream with the reality. There are a few good biomarkers and that’s it. You can probably count them in one or two hands. There’s a couple of mutation markers, there’s a methylation marker, there’s one expression marker, there’s few that really have those characteristics where they either with certainty or near certainty will tell you that a patient will or will not respond to a specific therapy. That’s the tough part. We’ve been at this for a long time. Realistically, the targeted agents have been coming in at the back end. A lot of the difficulty was in trying to get markers for very complex therapies that both were themselves complex and hitting and damaging cells in so many different ways. And then also in the multiple combinations that were the mainstay of chemotherapy for most of the last few decades. The good part about that is now with more targeted therapies, there’s a lot more impetus to do that and also more rationale. Again, if you’re going to hit receptor A, let’s start at looking at receptor A as a biomarker. It’s not so complex even though sometimes it does end up being complex. It’s at least getting out of the gate that makes a lot of sense. I think that itself is also changing, but the number of biomarkers that really predict accurately with great, great sensitivity is only a few.

Q: Why don’t we have more personalized medicine?

A: There are a lot of biomarkers, but there’s probably only a few winners and, just like with drugs, we’re trying to figure out how to make that happen. There are a few good examples, but they all came about by different ways and methods and approaches, and it’s a challenge. I think that’s the other big part of it. Why don’t we have more personalized medicine? We have a lot of drugs, why don’t they don’t all have a specific marker to tell us if it’s going to work? Well, that’s not so easy. Picking the right biomarker can be as complex as finding a new drug sometimes. That’s the other part of it. There are a few good examples and a lot of reason to pursue them, but we still don’t have the complete answers how to get them.

Part of what we do at Champions is actually figure out how to get these markers at the same time as we get the drugs to work. That’s another way that potentially our models can help bring those markers to fruition.                                   

Q: What inspired you to get involved in the area of personalized medicine?

A: That’s pretty straightforward. I became an oncologist in 1988; 1989 I got certified. For 20 years, we’ve all been talking about one way of being able to do this. Why? Because we see all the toxicities and the failures that we have. We really believe from the other point of view, because I do a lot of laboratory research, we really believe that understanding the biologic pathways has to translate into us being a lot smarter in how to give these drugs. Personalized medicine really brings that together, it brings the science and the clinical medicine together to say, ‘hey, how can we really benefit our patients the most?’ Really, I see it in three ways, one is diminished toxicity. No. 2 – get better predictability, get drugs out there alone or in combination that really work. And the third one is can we find drug combinations or models that will help us actually cure patients. For most patients with advanced disease, once the cancer comes back, most cancers with few exceptions are almost incurable. The only chance you have is right after surgery, either with surgery or addition of chemo radiation right after surgery. We really believe is can we finally find the Achilles’ heels in these tumors where maybe one drug alone can’t do it, but maybe if you find in these models two or three can, that you finally move those combinations forward and cure patients. And that’s what it’s all about.

To me, it was a very natural evolution of what I was doing, which is trying to translate our findings into the clinic. That’s why we are clinician scientists. We want to make those changes.     

Q: Can you explain how Biomerk Tumorgrafts – the test you developed and brought to Champions to evaluate anti-cancer drugs before they are given to patients – works?

A: There’s really two main indications that we have now. The simple concept of course is to take a tumor from a patient – it has to be fresh and you have to implant it in a mouse that has an immune deficiency. It needs to be done within two hours. That’s the real limitation. If you do it [in] over two hours, the take rate is very low. Once that happens, though, and the tumor takes over the next few weeks, it’s then propagated only from mouse to mouse.  So the tumor gets big, the mouse is sacrificed, the tumor is cut into small pieces and re-implanted again. Then after three or four times, you have hundreds and thousands of these tumors available to test directly in the mice. Now that’s never been done before. Up until a couple years [ago] when we started to do this, the only thing that was available were standard xenografts, and those were cell lines. They would take a tumor and grow it in a petri dish or in plastic and propagate only as cells. People did it because it was just a lot more feasible. You could grow it, you can transport them, you can do whatever you wanted with them. The problem is the cells change; they got good at growing in plastic and got good at dying from a lot of different drugs. People very early on started to realize that just because things worked on them in the laboratory – that you could kill these cells – it wouldn’t actually translate into any benefit for patients. The pharmaceutical industry finally began to use them only as a checkoff box, and said, ‘Ok, well, these are my xenografts I’m testing, one them works, fine, lets move on.’ The realized there was just no prediction to the model.

The advantage of the way we do it is that the predictability is as close as we could tell, close to 100%. The patients that we’ve [tested] and we’ve seen what they’ve gotten, we can see 100% prediction. If predicts drug A, you can see the surrogate in the patient. In the patients.

That, of course, drives the whole process. Even though it’s tedious, and this is the down side – it’s a tedious, time-consuming, expensive process – it’s something that has so much value, it’s worth the effort. There’s two main uses for that. One of them is for patients that can afford it, they can actually get the tumor implanted when they have surgery, it can be grown. Then, whether they need chemotherapy now or later as an insurance policy, you can test the various drugs and ensure them that the predictability is nearly certain that whatever the model predicts will be what happens to them. So that obviously is very attractive, but very expensive. Patients may have to dish out tens of thousands to a couple of hundred thousand dollars to get that done over time.

Q: How are pharmaceutical companies benefiting from your tumorgraft technology?

A: In addition to individualized patients, we’re collecting and harvesting these tumors on our own, and we now have a rapidly growing library of these tumors that we’re offering to the pharmaceutical company as the most predictive model that they can ever put their hands on. That’s really changing the whole equation because now it’s going from, let me just do them because I need to do a checkbox, to let me do them because I can actually predict what’s going to happen to me when I go into the clinic. What that provides for the pharmaceutical companies is a tremendous acceleration in their time development and much higher predictability for getting the drug into the clinic, which all translates into a lot of savings.We’re offering the service to the pharmaceutical companies. We’ve signed deals with ImClone, with Centocor from J&J, and the list is growing, as well as smaller biotechnology companies. It really for the first time I think changes the equation, where instead of just going into the clinic and having no idea what’s going to happen and no way to really inform you of which patients to try, now you can really rely. If I do 10 lung cancers and three tumors respond, it’s like three people responded. You can go into a clinical trial in lung cancer and expect to have three out of 10, or 30% of patients, respond. It completely changes the equations of how you begin to develop your trials.