Only a fraction of possible cancer medicines gain final approval

[00:00:00] Host Amber Smith: Upstate Medical University in Syracuse, New York invites you to be "The Informed Patient" with the podcast that features experts from Central New York's only academic medical center. I'm your host, Amber Smith. The majority of the promising new cancer drug candidates never make it to market. Only a small fraction of drugs that make it through clinical trials will end up with FDA approval. For help understanding some of the reasons for this, I'm talking with Dr. Jason Horton. He's an assistant professor of cell and developmental biology, orthopedic surgery and radiation oncology at Upstate. Welcome to "The Informed Patient," Dr. Horton.

[00:00:40] Jason Horton, PhD: Hi Amber. Thank you for inviting me.

[00:00:42] Host Amber Smith: There are different types of research, and I think it's important to explain that and where the idea of translational research fits in. Can you please explain the difference between basic, clinical and epidemiological research?

[00:00:55] Jason Horton, PhD: Sure. So basic research, which is where I spend a lot of my time, really deals with the fundamental aspects of, in the case of medicine, biology, how different types of cells talk to each other, how different molecules interact, things at a very fundamental level where we're just really trying to understand how nature drives some of these processes.

Now, clinical research is at the other end of the spectrum, and that's really devoted to understanding how patients might react to drugs, how different drugs are used to treat different diseases, and ultimately whether these drugs work and have the desired benefit, as well as demonstrating that they're safe for use.

So in between we have these other concepts of applied research, which is where we take that basic research that I talked about initially, and how we use that information to begin to move things toward a very practical application.

And then in between applied research and clinical research is what we call translational research. And this is where we move different drugs or different devices, things of that nature, between the applied phase and into the clinic. And so this is where we start doing preclinical models, some animal research, some more complicated cellular research, things like that where we can screen out a lot of things that may be toxic to humans or, also, ineffective.

Unfortunately, and kind of, I think, what brings us to the topic today is that's the spot where a lot of things don't ultimately carry over into clinical research and making real impact.

[00:02:33] Host Amber Smith: From, from day to day, do basic scientists collaborate much with clinical research? Because that would make it translational, right?

[00:02:41] Jason Horton, PhD: In a certain sense, yes. I think at the basic science level, people are really trying to focus on discovery and understanding the real fundamentals, but it is really important for clinicians in medical research and the basic scientist to sort of work hand in hand and bridge that gap between the two through the applied and translational disciplines.

[00:03:02] Host Amber Smith: Well, let me, before we get into drug development, epidemiological research, that would be more like population studies?

[00:03:09] Jason Horton, PhD: That's correct. And that's an important part of this too, because you have to be able to study these things in the right populations of people, the right populations of animals. Certainly there's a lot of drugs and diseases which have genetic basis to them. And if you're not studying the right population of people that has, say, a particular trait or particular gene, that carries a disease, or that codes for a disease, you may not be testing the right people and ultimately kind of doing some fruitless research.

[00:03:40] Host Amber Smith: Well, let's talk about how long it takes for a drug or a device to be developed.

[00:03:45] Jason Horton, PhD: So that's a good question, and it varies quite a bit. I think the average sort of description is that it can take anywhere from five to 30 years for a drug to go from the the bench at the basic level to the bedside.

And it varies a lot, both in terms of how long it takes to progress through those different phases, but also how the partners that are involved in whether you're able to, in an early stage get sponsorship and interest from industry who ultimately see some potential for investment in a particular technology and wants to try to capitalize on that potential in the future.

[00:04:23] Host Amber Smith: So how does it go? You have to start with an idea first, right?

[00:04:27] Jason Horton, PhD: Absolutely. It all starts with the idea and again, a lot of that starts at the basic science level where we can find things like different targets, say, a certain molecule that we know is associated with a particular disease, and then find different molecules or chemicals or even other proteins or different biomolecules interact with our target, and try to understand how those things work together, and if we can interrupt the function of the protein or of the cells that are diseased and kind of go from there.

[00:05:01] Host Amber Smith: Now, I know publishing or sharing your results at conferences and things, you have to let other scientists know what you're doing, right?

[00:05:11] Jason Horton, PhD: Well, that's absolutely a critical part of it is, is getting the information out there, and a big part of it, you want to be careful when you're making these discoveries, that you're sharing it with the world and the scientific process. But then also you want to be able to protect intellectual property, which is how through the patent process where we can develop something that has potential, potential to make money obviously, but also the potential to deliver cures or interventions that are useful.

And you want to be able to protect your investment in that process. And so, really very early on, we get involved when we're doing a drug development type thing, we get involved with, at Upstate, SUNY has an intellectual property office.

[00:05:53] Host Amber Smith: Do you have any examples of a drug that moved through this system or the process quickly?

[00:06:00] Jason Horton, PhD: I can give you an example of one. I don't know how quickly it went through the process. But there is a particular leukemia, chronic myelogenous leukemia, that has a very consistent chromosomal translocations. That's where two parts of the genome in our cells, which are supposed to stay in a particular order, in a particular place, become lost, and then they join up to each other.

And when that gene is expressed as a protein, we have a protein that has parts of two different genes stuck together. And so in this case, for chronic myelogenous leukemia, it is termed, or it's what was called the Philadelphia chromosome because it was discovered in Philadelphia. And what people found is that there is a small molecule that fits into the protein that's produced from this gene. Initially it was called imatinib, but then sold as the trade name Gleevec. And that was really an important development in medicine overall, as an example of a precision targeted medicine.

So they were able to take the structure of the protein that was formed by the fusion of these two genes, look at the structure of that protein and find a little pocket where they could stick an another molecule into it. And that molecule disrupted the function of the whole protein and essentially, wiped out the cancer in those patients.

[00:07:24] Host Amber Smith: Wow.

Well, let's talk about some of the reasons that a drug might drop out of development along the way.

[00:07:32] Jason Horton, PhD: That's really the what's sort of called the valley of death. So we can test and develop a lot of these small drug compounds or small molecules in the lab. And they do great things in a dish. But we start to lose them to attrition when we test them in the sort of translational phase, say in small animal models, mice, etcetera. And we find that they may provoke some sort of toxicity or that they're metabolized in a strange way that makes them no longer active or makes them overactive and again, leads to toxicity, or they just don't ever get absorbed. So those are kind of the main things that would come out in the translational phase. But then ultimately, there's quite a bit of difference between mouse and man when we go to test of these molecules.

And ultimately, certainly the different drugs have different effects in different species. And so a lot of drugs fail at each of these phases along the way. And so we start with these sort of lead compounds, and by the end of the drug development process, we're kind of left with one.

I can speak to a particular case in a drug that I'm working on developing now, where it was picked out of a screen of approximately 50,000 small molecules, and was found to have very specific activity against one particular pediatric cancer that we're studying. And over the course of several years, the team that discovered that, went through... so they discovered this and published it. In about 2011, they finished their first clinical trial with the drug, and this was done at the NIH (National Institutes of Health). They published their first clinical trial on the drug in approximately 2017. And while they found that it had fantastic outcomes for some patients, about 25% of these patients experienced dose limiting toxicity. And so they decided that it wasn't worth pursuing that drug anymore. So that's an accelerated process where the drug had been around for years. It was sort of an orphan drug. They went and revisited it as part of this screening library. And then ultimately after testing it both in cells, then in animals, and then in humans, in some cancer patients, that they found that while it was reasonably effective at each of those stages, once they got it to humans, they found that it was really toxic in some of them. And so they decided to stop the trial there.

[00:09:49] Host Amber Smith: This is Upstate's "The Informed Patient" podcast. I'm your host, Amber Smith. I'm talking with Dr. Jason Horton. He's an assistant professor of cell and developmental biology, orthopedic surgery and radiation oncology at Upstate, and we've been talking about the process for new medications to get approved in America.

What are some of the existing barriers for a health provider who thinks of a drug or a device that they believe might help patients? How do they typically proceed with that idea, or can they?

[00:10:22] Jason Horton, PhD: Well, that's a very important question, and I think that's one of the particular challenges out there is being able to find and take those ideas and pass them along to the right people in a collaborative nature, or in a collaborative endeavor to really see where it can go.

Certainly it's a challenge for physicians just to find time to do basic and translational and ultimately clinical research. And that's one of the great things of being at an academic medical center like SUNY Upstate, is that it definitely facilitates the interaction of those clinicians with the basic scientists and really is able to make those connections at those very early stages and then be able to sort of build on that initial idea and keep chugging it forward, and hopefully leading to a point where you can get to a trial.

[00:11:11] Host Amber Smith: Do you think it's better to have doctors who care for patients and doctors who do research, separately? Or can there be physician scientists who excel at both?

[00:11:21] Jason Horton, PhD: Oh, there are absolutely physician scientists that excel at both. Certainly physician scientists are indispensable in this, but really it's a team effort. And so you need specialists in a lot of different areas that can look at some of these questions and some of these ideas from a lot of different angles.

There's sort of a saying that the more training you get, the less you know, and that's because we become also specialized. And so, really, interdisciplinary research, at least in my opinion, is really the way that we can make these important developments in science become clinical realities.

[00:11:54] Host Amber Smith:

What is the "bench to bedside and back" development pipeline?

[00:11:59] Jason Horton, PhD: Yes. That's really a philosophy of where something might be, and I would even say take it one step further back, and say it starts at the bedside, goes to the bench, and then back to the bedside. Again, sort of flipping that arrangement around.

And so a lot of this stuff, in my experience, a clinician will come to me with a problem that they've had or that they experience in their practice, and maybe they don't have a great way to address it. I'll sort of think about it in a lot of different ways.

I may talk to a few colleagues and sort of collect a bunch of different ideas of how we might be able to help. What can we do in the lab? How can we test this idea? What do we know about it? And certainly for all the things we know about medical science, we're really just on the early edge of understanding how a lot of these systems and cellular biology and physiology work together. You know, in the basic science end, we become so focused on one discrete aspect that we kind of have our blinders on to some of these parallel things. And then, as we make our way through the translational process toward clinical trials, we begin to narrow some of those things down.

And so that also kind of runs in parallel to a process where we fail drugs. We want to fail drugs faster. And what I mean by that is we want to get those drugs, which are probably never going to make it to a clinical trial for various reasons, like I said earlier-- toxicity, ineffectiveness, metabolic complications. We just want to get those out of the way faster so we can really get from a very broad range of possibilities down to a few discreet promising ones as quickly as possible, both as a time, but also as a resource saving endeavor.

[00:13:43] Host Amber Smith: Like the example you gave before, though, that the drug that went all the way through rodent trials, but then it was toxic in humans. If rodents aren't the best model for testing things that are going to be used in humans, what is the alternative to that?

[00:13:59] Jason Horton, PhD: And that's one of the really hot topics in translational and biomedical research at large, and certainly something I'm beginning to be involved in, again, in the context of the drug that I talked about earlier. There's been a recent emphasis on things like microfluidic or micro physiologic systems, where we're able to use human cells and tissues, grow them outside the body, from a variety of different donors. We can sort of sample a reasonably sized population in a relatively small footprint using some complicated tissue engineering approaches with these patient derived or human derived cells and really get testing in human tissues, in human cells as early as possible.

[00:14:46] Host Amber Smith: Can you test a drug on, say for example, you've got cancer cells that you saved from a tumor, or you've got normal cells that you scraped off someone's arm. Can you test a drug and know what it's going to do with simple cells that have not yet turned into an organ, let alone an organ system? I mean, how do you, they might react with the cell one way, but once the cell is attached to a body and the body is functioning, will they react the same way?

[00:15:16] Jason Horton, PhD: Well that's a complicated question because they may and they may not. So, you know, one of the big things that's going on in my lab right now is understanding how this drug that we're testing in the lab, we want to kind of know how... It's primary toxicity is to the liver, and this didn't really come up when we were testing just tumor cells. I mean, we knew about it in the background because that other group, like I said in the clinical trial, had found this liver toxicity. And we had tested it in animals, on sort of recovering a lot of their experiments, or redoing a lot of their experiments in our hands, to try to understand toxicity.

We made a slightly different formulation of that drug that we hope, and at least our preclinical data seems to suggest that we can prevent the toxicity by causing the drug to accumulate specifically in the tumor rather than in the liver where it's doing damage. We developed a nanoparticle encapsulation technology with some scientists in the department of pharmacology here.

[00:16:18] Host Amber Smith: So you don't need the full liver?

[00:16:21] Jason Horton, PhD: You don't necessarily need the full liver. I guess I was focusing on that end on, on how we can show efficacy. But then on the toxicity side, you don't necessarily need the whole liver. The early indications seem to be that it is a specific type of cell in the liver that we can model. And so we're developing a micro physiologic system now, essentially a liver on a chip model in my lab. And so that was with the grant that I just got from the cancer center is going to develop this liver on a chip system where we can integrate cells from a number of different donors and kind of get a broad representation of different genetic variability.

We have some notion that this drug in particular affects the process of bile transport. And in that initial population that was published by the people at NIH, where they said about 25% of their patients showed toxicity. In a follow-up paper, they were able to attribute that to single nucleotides, so single A T Cs or Gs, variants between different individuals, and show that those single nucleotide polymorphisms are single letter changes in how a gene is spelled, predicts whether a person is going to have this toxicity syndrome or not. But ultimately the process that all of these different genes play a part in is how bile exits a hepatocyte and then enters the digestive system.

If they can't put the bile outside of the cell. The hepatocyte makes bile. If it can't export that bile and put it into the bile ducts, the hepatocyte dies. That's the sort of mechanism of toxicity for this particular drug. And so we're trying to now curate liver cell samples from a variety of different donors and understand what that level of toxicity is, and if there are different ways that we can augment that, either by changing the molecule that we're using, or as I said earlier, in this nanoparticle encapsulation direction, enhance its ability to go to the tumor and stay in the tumor so that it doesn't cause the injury to the liver.

[00:18:27] Host Amber Smith: So what you're doing, then, essentially means we won't need cages full of lab rats anymore.

[00:18:33] Jason Horton, PhD: I mean, I think that's one of the, definitely one of the things that we're trying to alleviate.

There's certainly ethical concerns with doing animal experiments. There's costs associated with doing those animal experiments. And then ultimately, a lot of in the case of this drug in particular that I was talking about a moment ago, the mice don't seem to have very much trouble until you get at very, very, very high sort of not really physiologically achievable doses in humans before they start to show a lot of toxicity.

And so that's one where, probably the toxicity didn't show up in the animal model, and so it went to that clinical trial and then they found the toxicity there -- in some patients.

[00:19:17] Host Amber Smith: This is like revolutionary stuff.

[00:19:19] Jason Horton, PhD: I'd like to think so. It's certainly interesting to do and gratifying and it's certainly not anything I thought it would have been doing when I started here going on eight years ago now. So it's been a really exciting time to be here and to be doing this sort of research.

[00:19:36] Host Amber Smith: Well, before we wrap up, I wanted to talk about what the ideal translational research team would be like. Who would be the members of that team, and what would their roles be?

[00:19:47] Jason Horton, PhD: Well, sure. At the one end of the spectrum, you have somebody like me who's a bench scientist, and those folks also include graduate students, postdoctoral fellows, and other people that work in the lab, technicians, et cetera, that really do a lot of the very early stage hands-on research and really try to weed out some of the either ineffective things or really just screening things at the very earliest stages to get rid of drugs that just don't have any particular viability.

Moving through there you have people who are engaged in the translational phase with that level. Again, it's a mix of basic scientists, some clinicians, oftentimes veterinary researchers will also participate. And then at the clinical end you also have the physicians who are managing the patients, who are overseeing the trials. Running parallel to the basic scientists and the clinical scientists we also have epidemiologists and statisticians who are really the gatekeepers to say whether something is good, is effective, is safe based on the numbers and the data that come out of it.

And so we have what are called different levels of trial blinding. So in a double-blind trial, which is considered the gold standard, where neither the patient nor the physician who's overseeing the trial knows exactly which patient is getting the intervention or the experimental intervention or the standard of care or control intervention, to study that efficacy.

And so really it's those statisticians that sort of are able to say definitively whether a drug works or not based on the unbiased data that we hope will come out of the studies. And then finally we have our partners both in administration at a place like Upstate, the office of tech transfer, who helps us manage intellectual property and patents, but then also facilitates interactions with industry where these things are able to be scaled up and large investments are able to be made for a device that's very promising.

[00:21:44] Host Amber Smith: How would you go about improving quality control in research? How would you weed out... you talk about failing drugs faster, but what would you do to avoid just bad ideas from the outset?

[00:21:58] Jason Horton, PhD: Well, I think a big part of ruling out bad ideas is getting a lot of divergent opinions, and I think at a certain level you have to weigh them each with their own merit. But at the end of the day, the more people you get involved and the more perspective that you bring to a particular question, probably the better insight you're going to have in whether there's viability, say, in testing a drug, if there's a reason to keep going with it. If this is a reasonable target, what are some of the sort of known unknowns, at least to the person who's asking the questions and try to identify as many different facets of these questions as possible.

[00:22:36] Host Amber Smith: Does AI (artificial intelligence) have a role in any of this, do you think, now or in the future?

[00:22:40] Jason Horton, PhD: Well, that's a fantastic question and very timely.

So that is something that is definitely actively being used. But there are also certain barriers that we have to think about there. So artificial intelligence just for the audience, is essentially a computational algorithm that is used to automate a lot of decision making processes. And so in this context it may be understanding how different molecules physically fit together, or how a drug interacts with a particular protein. It's something that we make the use of the computers to iterate through these different interactions that might be present. And I'm just using this as a particular example. And I think in that context, artificial intelligence and machine learning is extremely powerful.

But on the flip side of that, we also have to be careful to protect patient privacy and protected health information, because right now at the pace that the artificial intelligence technology is growing and accelerating and really kind of becoming a part of everyday life, we have to be careful that there are the appropriate protections there because these things are so new that it's a little bit of a cat and mouse game on the safe artificial intelligence, or the benevolent use of artificial intelligence, in different levels, both of biological research, but I think where it really becomes concerning is when we get into clinical research.

[00:24:01] Host Amber Smith: Well, Dr. Horton, thank you so much for taking time to explain this to us. I appreciate it.

[00:24:06] Jason Horton, PhD: Thank you very much. It was a pleasure being with you.

[00:24:08] Host Amber Smith: My guest has been Dr. Jason Horton. He's an assistant professor of cell and developmental biology, orthopedic surgery and radiation oncology at Upstate. "The Informed Patient" is a podcast covering health, science and medicine, brought to you by Upstate Medical University in Syracuse, New York, and produced by Jim Howe. Find our archive of previous episodes at upstate.edu/informed. This is your host, Amber Smith, thanking you for listening.