Video Episode 64 - Professor Justin Stebbing

[00:00:00] Speaker A: Foreign. [00:00:05] Speaker B: Welcome to a special episode of the Bio Rossi Few and Far between podcast. I'm your host, Chris o'. Brien. In today's fast moving and often noisy biotech landscape, it's easy to miss the exciting new discoveries just below the surface. That's why we're launching a new segment dedicated to spotlighting the trends, technologies and discoveries that deserve your attention. And we will be enlisting the insight of former guest and friend of the show, Professor Justin Stebbing, Editor in Chief of Oncogene, consulting oncologist and key figure across biotech and oncology. In this episode, Justin and I will zoom in on solving the unsolvable with AI, the promise of the GLP1 revolution, and the discovery of T regulatory T cells as announced in this year's Nobel Prize for Medicine. We'll also explore emerging innovations in today's biotech and some developments that may not be worth the hype. We're beyond excited to share this episode and we're already lining up future conversations with Justin, so please let us know what you think. Send us your feedback, comments, and suggestions for upcoming episodes. Okay, let's start the podcast. Professor Justin Stebbing, welcome back to Few and Far Between. I have been really looking forward to this conversation. [00:01:15] Speaker C: Me too, Chris. [00:01:17] Speaker B: So in this world that we live in right now, where there's an enormous amount of noise and it's hard to separate out the signal, I thought that that would be the purpose of this call would be to talk about things that you're observing that you're seeing that you think are exciting or interesting and have some kind of an impact on the way we should be thinking about where life sciences is going, where the biotech industry is going, et cetera. So we'll just kind of dive in and we're just going to go through some topics. Can we start with AI? I know that there's been some. Well, how to not start with AI almost is the question. It feels like it's everywhere right now. And I know in June we saw De Novo AI publishing the Alpha design showing functional synthetic proteins that are entirely de novo in living systems. That seemed like an interesting place maybe to start, you know, do you feel, well, how would you like to weigh in on this, on the AI drug discovery? [00:02:26] Speaker C: I mean, without meaning to sound horribly immodest, at the beginning of the COVID pandemic in January 2020, I used AI and published a couple of papers in the Lancet about an existing drug that we wanted to find that could treat patients with COVID pneumonia. And we came up with Lilly's baricitinib. And the reason we wanted an existing drug is so we'd all know its safety profile. Yeah, that was FDA approved to treat Covid pneumonia. To treat Covid pneumonia. Nine months later and it's still the drug with the greatest mortality benefit in Covid pneumonia. [00:03:10] Speaker B: Now. [00:03:14] Speaker C: Obviously we didn't have the large language models then and we have them now. And what we're seeing is truly staggering, such that AI looks like it's superior in every metric we may want to measure compared to AI plus human. So it won't surprise anyone learn that if we're designing synthetic protein or a new protein or looking for a new drug target or a drug receptor interaction, that AI is superior to any human involvement. And human involvement with all the psychology and motivations that interlays with that actually decreases the output. On the other hand, Chris, in the clinic, apart from one or two notable exceptions such as the COVID story, we're not seeing thousands or even tens of drugs developed using AI. [00:04:13] Speaker B: And I think the first couple we've seen, we've seen some that have failed in the clinic early. [00:04:17] Speaker A: Right. [00:04:17] Speaker C: And we've seen some that have failed. So people are arguing with AlphaFold Generation 3 and the newer versions like the one you just mentioned, we will see a new wave of drugs which will help people live better and longer lives, which is all we're trying to do, improve people's quality and quantity of life. Treating the right person at the right time with the right disease, with the right drugs. So the drug probably treats fewer people than the big blockbusters, but in those people it does treat, it treats them super effectively with very little toxicity. But the reality is the proof is going to be in the pudding. And I'd like to see those drugs, if you look, for example, that the biggest drugs that everyone's talking about now, the GLP1s, obviously in obesity, that they weren't developed using artificial intelligence. Now, people often talk about their neurologic effects and we're trying to understand AI as a brain and we're trying to understand the use of GLP1 drugs on the central nervous system. We're going to see Alzheimer's data this year from GLP1 drugs and phase 3 clinical trials, just by way of one example. Again, that's not been developed using AI, but often AI is used as a companion in clinical trials, in the process of development and so on. So it seems to be a good add on. It helps obviously with productivity gains. It's useful. But in terms of AI, truly Discovering new things and creating new things as opposed to creating plausible text. I really don't think we're there yet. That's just my view. [00:05:58] Speaker B: So I want to double click on a couple things you said. You said you used AI in the COVID example, but this obviously predates large language models. So what was that then? Was that a kind of a. [00:06:11] Speaker C: No. [00:06:11] Speaker B: Neural network thing that you did to kind of look for. Yeah, you explain it. Let me guessing. [00:06:16] Speaker C: Yeah, it was not a neural network. It was an NLP based program. It included tons of documents, well about 2 billion documents that it used to make higher order inferences using natural language processing to connect disparate data points that with all the humans in the world you could never connect A to D because a human would need to go A to B, B to C, C to D. And it creates what we call higher order functions that humans just aren't capable of. [00:06:50] Speaker B: So in that example. [00:06:52] Speaker C: Right. [00:06:54] Speaker B: A human was still essential because you, you got, you know, you got sort of hey, the, the AI says this looks promising, I suppose. And so then you have to figure out how to test it and, and, and whether or not you, you buy the sort of hypothesis. What do you think is different now because you're saying you said it as if it' but I don't think it is yet. That AI systems without human involvement are better than the combinations. Mostly what I've seen. Certainly there's a control question that people have about if we allow the AI to be unmediated by humans, who knows what it'll do. That sort of thinking. But also I've just generally, you know, this is an analogy from a couple of years ago, but people were saying that um, in chess this is, this might be from three years ago. So, so it just may not be true anymore in chess that the combination of an AI and a human could beat an AI. And so. Yeah, so what, what do you think? [00:07:51] Speaker C: I mean we saw quite a few years ago AlphaGo. [00:07:55] Speaker B: Yes. [00:07:56] Speaker C: And then beat everybody, beat everybody and go. Which is I think a harder game than chess. And then we saw different. Then we saw the next version of AlphaGo as a mystery guest meet. Beat the original program of AlphaGo. [00:08:11] Speaker B: Okay. [00:08:13] Speaker C: But when it comes to large language models playing chess, a bit like quantum computers doing simple arithmetic, they're not necessarily the best systems for that. We know that they're very good at evaluating everyone. Everything that's ever existed on the Internet. Yes. Know that they can ingest huge amounts of data and they create very plausible text. And the newer versions Hallucinate much less thankfully. But when you evaluate them using certain metrics, it's really clear that they're still not very good at some things. Discovery to your point about protein synthesis being a very, very good example, and that they can be flawed. With all the foibles about encouraging suicide or making people fall in love. Could AI murder someone? Having said that, the general direction of the conversation is about AGI, artificial General intelligence, where AI is more capable than humans. And the other, the other point is that even if it doesn't cause productivity gains, because everyone says it's associated with productivity gains, we're acting like it's associated with productivity gains. People are hiring fewer people just by way of one example, because we have AI. And even if the AI isn't that useful, people are still hiring fewer people, in my view, because of the anticipation of the wave that's coming. [00:09:47] Speaker B: I think that's right. You know, the. I have this favorite cartoon that I saw recently, which was. It's a bunch of guys with CEO on their chest and they're all, they're all chanting and they say, what do we want? AI? When do we want it? Now. Why do we want it? We don't know. And there's an element of this, right, that people are saying, well, hey, if this is a reason, if this is a logic for cost reduction, I'm interested in it, says the cfo, maybe says the CEO as well, and lots of work organizations. And so what we're seeing is a friend of mine who works in education describes this as pulling up the ladders. You know, that the entry level positions in law, in consulting, in a whole bunch of fields are drying up right now. Will they come back? Who knows, right? [00:10:31] Speaker C: I mean, to that point, to that point, just very, very specifically, if you develop a new drug and you put it through an FDA process and you perform clinical trials, there's no AI program that can speed up those trials, the efficacy of the drug or the toxicity. And that's a massive expense. We're now talking $200,000 per patient on many trials. And that's a really huge expense. [00:11:04] Speaker B: Yeah, yeah. It's funny, I listened to a podcast recently, but there's a whole bunch of folks, technology people, not. Not bioscience people, basically saying, you know, the curve, the curve is so, so fast in AI improvement that, you know, we're going to cure all major diseases in the next 15 years or something, or most major diseases, and we're going to massively extend human lifespan and, you know, blah, blah, blah. You've you've heard these, you've heard these arguments. They're notably absent from a lot of those arguments is people who understand biology and, and, or understand the regulatory process. So when you sort of imagine what 10 years from now looks like, if I can ask you to put your speculative hat on, do you think it will look significantly different from the way it looks now in terms of like cancer? If we talk about oncology, sort of a deep area of expertise of yours, do you think that it'll be dramatically different or do you think we'll continue to make the kind of improvement that we've seen over the last few years, 15 or 20 years, which has been very significant? [00:12:09] Speaker C: I think advances in knowledge, contrary to what newspapers or magazines say are made in small steps, not giant increments. Generally the anticipation is with AI just like it was perhaps with the immunotherapies a decade ago. Yep. That we're going to have another giant leap. But we're still not seeing those molecules developed using AI that can treat people. And until we see those, for me it's still an execution story. It's still a show me story. [00:12:40] Speaker B: So maybe the right posture then to have is. I think it's also, it's very, very dangerous to be skeptical. Skeptical to the point of. So maybe skeptical is not right. But to be a pessimist about an AI pessimist. I think that's a dangerous posture because you know, I know plenty of people who said I played around with ChatGPT, it's not, not that great. You know, it's like, well, have you played around with it today? Because it is actually. [00:13:03] Speaker C: No, I agree with you. I think issues like the ethical issues or I mentioned encouraging people to fall in love or cheat or commit bad behaviors or the way AI hallucinates, all of those are relatively minor issues that can be overcome. And all new technologies have their foibles if you like. And that's just one of those. Over time others will come up, others will be overcome. In terms of using as a force for good. Clearly it's a really helpful thing to have. But using it to really save lives, sure it will improve diagnostic precision. But even with that, even when you know these days a lot of children are going to be using scissors or pens less because of AI and computers. [00:13:52] Speaker B: We all know that's a great point. [00:13:54] Speaker C: But even in high end medicine and a recent trial showed that endoscopists, people performing colonoscopies. [00:14:04] Speaker B: Yeah. [00:14:05] Speaker C: Using artificial intelligence to improve adenoma detection, detection of pre cancerous polyps or Cancerous polyps. That was found to be really helpful. But when you withdrew the AI, they'd lost some of their skills. [00:14:19] Speaker B: Oh, interesting. [00:14:20] Speaker C: And when you ask students to write essays with and without AI and you monitor using functional MRI their brain waves, it's very clear that when they use AI, no surprise, their thinking is considerably less. They are cognitively lazy. [00:14:39] Speaker B: Yes. [00:14:40] Speaker C: So although it's very useful, the fact is quite simply making us dumber. Sorry to be so candid. [00:14:49] Speaker B: At scale. [00:14:50] Speaker A: Right. [00:14:50] Speaker B: That's the planetary scale. [00:14:53] Speaker C: At a planetary scale hasn't escaped the attention of a lot of people because it's not perfect and it's not very good. To the very first question you asked me at discovering new things. Yeah. [00:15:08] Speaker B: So I want to just double click one more, go one level deeper on the question of AI medicines, what would have, what would be a signal that if you saw it, you would say, okay, this, this changes my. This changes my conclusion here. What would have to be true for you to believe. Okay. No, there is the possibility of. Of course, we're still going to have a clinical trial process, maybe a slightly abbreviated one in some cases, but maybe that's possible. But that would bring a whole bunch of promising stuff to the clinic and bring. Bring this golden age of the future of blah, blah, blah, human health that some folks are talking about. [00:15:45] Speaker C: I would say that there's quite a lot of unsolvable problems in medicine. Okay. [00:15:52] Speaker B: Yeah. [00:15:54] Speaker C: In my field, treatments for glioblastoma and pancreatic cancer haven't changed for a decade or two. [00:16:01] Speaker B: Yeah. [00:16:03] Speaker C: In neurology, treatment for motor neurone disease and related conditions is very important. I can go through every specialty. Although there's a new drug approved by the FDA this week for idiopathic pulmonary fibrosis. [00:16:17] Speaker B: Yes. [00:16:18] Speaker C: It's a, It's a shocking condition with a survival of less than five years. It would be to see AI making a difference, not in the easier to treat things. [00:16:30] Speaker B: Yeah. [00:16:30] Speaker C: Like diabetes or high blood pressure or not saying that they're easy to treat. But we've got a lot of drugs available and a lot of drugs that work. It would be to see AI making a difference in diseases that traditionally we have, as humans, found it really difficult to treat. Now I understand why we find that difficult to treat. Pancreatic cancer cells are, you know, derived from pancreas cells which live in an acidic environment with no oxygen. So when you treat them with noxious chemicals, it just says wing it on. [00:17:06] Speaker B: Yeah. Right. [00:17:06] Speaker C: But it would be really to see artificial intelligence making a dent in diseases that even as a hardened oncologist that I find scary to treat. [00:17:18] Speaker B: Yeah, that's a great. I like that a lot as a takeaway. So, you know, keep your eyes peeled for breakthroughs in tricky, very difficult conditions. [00:17:28] Speaker C: It's not a breakthrough though, Chris. It could be an incremental improvement that has to be statistically significant in a randomized trial, obviously. No kidding. But also clinically significant. And different people mean different things by clinical significance. Some people say, well, making someone with pancreatic cancer live from a year to 15 months. So some people say different things with regards to clinical significance and that's probably a separate conversation. But if you can make people live meaningfully longer, that's a good place to start and that's objective. [00:18:05] Speaker B: But so look for significant enhancements in human life for people who have difficult conditions that we don't have good answers on. I like that a lot. Okay. If you think about oncology and the path forward in oncology, it's been an incredibly productive like your working life has seen massive, massive changes in our ability to treat lots of cancers. And we have a whole bunch of targeted therapies now. Doesn't cover the waterfront, but we have a bunch of them that are, that are, have a profound impact on the way we treat those cancers. When you look forward now, do you think the opportunity like 10 years from now, do you think we're more likely to say that oncology remained? I'd submit maybe it's been the most fertile ground over the last 10 or 15 years. Do you think that's likely still true. And we just had this Huntington's result announced in the uk which is really exciting. How do you think about the path forward? [00:19:08] Speaker C: Look, 50% of all research and development funding is spent on cancer. You can ask you whether that's right or wrong. We all know one in two people are going to develop cancer. The biggest risk factor for cancer, bigger than obesity, smoking and so on, is age. We're living longer, people are developing cancers beyond the age of 50. Our genomes no longer protect us from developing cancer. Maybe because we've, evolutionary speaking, done what we've needed to do, we don't want too many old people in society, etc. Etc. [00:19:41] Speaker B: That's a depressing fact that I did not have stored in my 58 year old brain, Justin. [00:19:47] Speaker C: But I think the reality, Chris, is that although oncology will still be hugely popular and for me, the marriage of research and patient care, the link from the laboratory to the clinic and the clinic to the laboratory is fascinating. I actually Think that an organ we're barely scratching the surface of, that we don't really understand will come to the forest. Many people can argue we don't really need another breast cancer drug. If I see a young woman with a HER2 positive breast cancer, no matter, and it hasn't spread, her chances of cure in the high 90s percent. [00:20:27] Speaker B: Yeah, extraordinary. [00:20:29] Speaker C: We don't really understand in the same way a lot of people don't understand what goes on inside the large language model, although we're now seeing publications on that, such as one in Nature about how deep sea seat works. The Chinese model. We don't really understand, nor do we have drugs for brain conditions. [00:20:46] Speaker B: Yeah. [00:20:48] Speaker C: Alzheimer's, whether the amyloid hypothesis is correct or not. But we are just scratching the surface of neurology. And to me, the next Genentech will be neurology based because of that. [00:21:01] Speaker B: Yeah, that's. It's been a dark. In the same way that it's been encouraging for cancer, it's been a dark, you know, decade plus in conditions, in a lot of neurology conditions. We have a podcast episode coming up shortly with a company that is looking at biomarkers in psychiatric conditions, schizophrenia specifically. [00:21:26] Speaker C: But to that point, there is not a single FDA approved psychiatric biomarker. [00:21:32] Speaker A: Right. [00:21:32] Speaker C: There is not a single psychiatric condition. Chris. Yeah. You can measure a biomarker. Not one in schizophrenia, depression, you name it. [00:21:42] Speaker B: Yeah. [00:21:42] Speaker C: So we're very lucky in oncology that I rarely have a patient that I treat without a. But without a companion biomarker. May not be a companion diagnostic. But whether it's something as simple as PD L1 expressed on the surface of cancer cells. To turn off the immune system. [00:22:00] Speaker B: Yeah. [00:22:01] Speaker C: So we block it to inhibit the inhibition of the PD L1 turning off the immune system. A bit like removing the cloak in Star Trek. [00:22:10] Speaker B: De Cloaking. Yes. [00:22:11] Speaker C: Right. Whether it's that or some precise molecular rearrangement that we're looking at that we can target with the drug. We have none of that in psychiatry. Yeah. And so I think what you're saying is correct, that on top of neurology, I would also say psychiatry as well, but I think we know so little about it. But I actually think also that thinking of the brain, you know, within a cancer, there's probably 10,000 signaling pathways going on at any one time. And within the brain there's probably about a million. Now, these numbers may not be accurate, but that's probably accurate in terms of scale. [00:22:52] Speaker B: Directional. Yeah. [00:22:53] Speaker C: Right. Sometimes I think it's actually quite useful to think of cancer as a little brain. So we can actually model it. It's autonomous, within us, but connected to us. Yeah. And sometimes I try to think of it like that when we are thinking of the hardest to treat cancers, but unfortunately that's not led to improved patient outcomes. What's the most frustrating thing for me compared to oncology and where I think we're going in neurology is there's not a single neurological development in the laboratory that I can think of in the last few years that's helped a patient in the clinic. Yes, we've got wonderful functional mri, optogenetics. We understand the electrical conductivity in the brain. We now are starting to understand consciousness. On the subject of LLMs. But what developments have there been in the laboratory that have led to precise developments in the clinic? Well, I'm not really sure. If you take multiple sclerosis, which is a neurological condition, every single drug just works by suppressing the immune system. Yeah, I don't really think of that as neurological. That's why people with multiple sclerosis get so many horrible infections, for example, on these drugs. I don't really think of that as immunological. All they're doing is calming the T cells down. [00:24:17] Speaker B: So treating the symptoms rather than. Yeah, yeah, that makes a lot of sense. [00:24:22] Speaker C: Okay. [00:24:22] Speaker B: Within oncology, are there particular areas that you think are exciting right now, where, you know, what are the places where you think the most interesting stuff is happening? [00:24:34] Speaker C: I think that I'm going to answer that in a slightly unusual way. I'm going to tell you what I think is less interesting. [00:24:41] Speaker B: Fine. [00:24:42] Speaker C: So I think antibody drug conjugates, where there's a huge amount of hype about, is less interesting. I think it's glorified chemotherapy. [00:24:53] Speaker B: That's interesting because a lot of money is being spent in that space right. [00:24:56] Speaker C: Now and AstraZeneca's biggest drug is going to be in HER2 for years to come. And it's a fabulous and effective drug. I just think we've hit the ceiling of them and they're hugely toxic. Yep. I think in oncology, it's very exciting for me to have AI diagnostics where AI looking at mammography is amazing compared to humans. But in terms of the developments, I still think that I'm excited potentially that AI discovered drugs will make a difference. But to me, it's still an execution story and I'm a bit anti hype as opposed to overexcited. I'm getting old and cynical. Chris. [00:25:40] Speaker B: Hey, aren't we all on some Level. Yeah. Well, this is the challenge, isn't it? Is you want to find this middle ground of, you know, being open to being persuaded by new evidence and willing to consider new pathways, new tools, et cetera. But you don't want to sort of buy into, you know, hey, guys, we're all going to live forever. You know, that feels a little overblown. [00:26:03] Speaker C: Yeah. And would you want to live forever? And if you were going to live forever, what age would you choose to stop aging at? [00:26:09] Speaker B: Yeah, right. [00:26:10] Speaker C: I think because living forever and aging forever may not be a lot, that. [00:26:14] Speaker B: Would be less appealing. Correct, Correct. Okay, that makes sense. So kind of broadly, in cancer, lots of. I think you're saying lots of places show promise. [00:26:26] Speaker C: Lots of places show promise. But we've also had a lot of disappointments. Car t in solid malignancies. [00:26:32] Speaker B: Yes. [00:26:32] Speaker C: Appointment. The immunotherapy add ons beyond the original CTLA4PD1PD L1 mechanisms. Okay. We've got lag 3, but that's just approved in melanoma. Okay. The next generation of therapeutics, sure, they're coming, but they look very, very incremental and none of them are targeting the really tough oncology questions. [00:27:00] Speaker B: Yeah, to your earlier point. [00:27:01] Speaker C: To my earlier point. So. [00:27:05] Speaker B: Okay, well, we'll see what comes out of the labs over these next couple of years. But overall, it's not that you're saying, I think that oncology doesn't show lots of promise. It's that neurology feels like it's on the edge of some really exciting breakthroughs and then oncology will have its continue to. There'll be frustrations and there'll be wins. Is that right? [00:27:26] Speaker C: Completely. I mean, we've seen the exponential improvement in oncology. We haven't seen it yet in neurology. And maybe now with the critical mass from data collection, from biobanks, from use of LLMs, we will be able to help patients with Alzheimer's, motor neuron disease, and so on in the clinic. [00:27:47] Speaker B: So. Okay. I want to give you a chance to just pick a random innovation that's worth talking about. I've said this to you offline, but I will say this publicly. I love the daily updates from you which cast such a wide glance or gaze across the scientific universe. So you can pick whatever you'd like to talk about. But first, do you think that the, you know, the upcoming move to oral GLP1s, how big of an impact do you think that's going to have? You know, it certainly seems in the US like there are a Lot of formerly fad people in Hollywood or formerly normal looking people who've become skinny. Do you think this is another step, function or just the people that are already doing it, stop giving them some shots. [00:28:37] Speaker C: I have a pretty strong view that it's not particularly innovative at a number of levels. Number one, the weight loss with the injectables, with the weekly subcutaneous injectables where people barely feel the needle is in the high teens. Percent. [00:28:50] Speaker B: Yeah. Okay. [00:28:52] Speaker C: The weight loss with the oils is in the low teens or 11 12%. [00:28:57] Speaker B: Gotcha. [00:28:59] Speaker C: Number two, the oils are hard to make. Novos one, you need tons of protein. Lily's one, which we've just seen data for. You don't need tons of protein. But the weight loss was a bit disappointing. Okay. Okay. And I don't think they've got food restrictions. You need to sit up, drink water, yada, yada yada. I actually don't think having a weekly injection is that big of a deal. [00:29:26] Speaker B: Yeah. Okay. That's an interesting point. Especially as you say, if it's when. [00:29:29] Speaker C: You don't feel it and you know you've taken it, you don't need to remind yourself and so on. But you know, even something as simple as GLP1s, why is. No one can answer this simple question. Why is the weight loss in non diabetics so much greater than in diabetic patients? [00:29:48] Speaker A: Okay. [00:29:48] Speaker B: I didn't know that. You know, that's fantastic. [00:29:51] Speaker C: It's a scale order of magnitude greater than in non diabetic patients. It's. It's just. It's just incredible to. [00:30:02] Speaker B: Yep. [00:30:03] Speaker C: Well. [00:30:04] Speaker B: And it seems to have this impact on all kinds of addictive behaviors. Right. We've got, I think a bunch of data coming out on that. We don't understand that either. [00:30:11] Speaker C: Right. Gambling. Yeah. [00:30:13] Speaker B: Why would it, why would it impact gambling? [00:30:16] Speaker C: Because it changes people's motivations. [00:30:18] Speaker B: Yeah. Okay. [00:30:20] Speaker C: Some people think it blunts people's motivations. And you know, in this era of limbic capitalism that we live in, yes. GLP1s will eclipse caffeine and alcohol as our favorite drugs will be zombified. [00:30:36] Speaker B: Greg. Nope. No exciting breakthroughs. [00:30:38] Speaker A: Nope. [00:30:38] Speaker B: No. No billionaires. [00:30:41] Speaker C: Exactly. That's funny going into space, right? [00:30:45] Speaker B: Yes. Yeah. Yeah. Well, we'll just have. We'll just have to stick with the ones that we have trying to get off the planet. Okay. Gene therapy and gene editing. We saw this extraordinary result, it seemed to me, in Huntington's that was announced. I know, it's. I know. [00:31:00] Speaker C: It's one by unicure Say it again. A small trial. [00:31:04] Speaker B: Small trial, yes. [00:31:05] Speaker C: Yeah. But really exciting to the neurology point. Huntington's. If you've ever seen a patient with Huntington's career, awful. It is one of the most upsetting things you see and these people, because in generations, it affects younger people. Each generation, they know what's coming and they're in diseases like Huntington's and als, where you keep your cognition, unlike in Parkinson's or Alzheimer's. Obviously, to me, I just find it shocking. So it's great that they can focus on the Huntington gene. [00:31:45] Speaker B: Yeah. [00:31:46] Speaker C: But that's the point. With gene therapy, it most the only diseases it can target are monogenetic. Yeah. Cancer, diabetes, high blood pressure. They're polygenic diseases. Yeah, yeah. [00:31:59] Speaker B: It won't fix everything. [00:32:00] Speaker C: It won't fix everything. It's difficult to think of a world where gene therapy might be democratized because it's so expensive. [00:32:10] Speaker B: It's just too expensive. Yeah. [00:32:12] Speaker C: But it would be nice to think of that world. [00:32:14] Speaker B: It would be nice to think of that world. It's maybe two orders of magnitude cost reduction to get there though. Right. It's not even a 50% reduction or something like that. Okay. I want to give you a chance to share whatever you'd like to talk about. So what, what are you, what do you think that we've not talked about that is exciting or noteworthy or scary? That's been happening in the, in the broader bioscience world. [00:32:45] Speaker C: So this is, you know, papers in the bioscience world. You know, we're being deluged with data. Yeah, yeah. The ability to interpret hues, huge swathes of data every day is just, you know, it's just incredible. It's just incredible to me. I mean, only this week, if I think of major discoveries going back to the neurology point, even though I'm a card carrying oncologist. [00:33:18] Speaker B: Yes. [00:33:20] Speaker C: Researchers have discovered a very small group of brain cells that play a particular part in long lasting pain. And those neurons in a brain area called the parabrachial nucleus. [00:33:32] Speaker B: Okay. [00:33:32] Speaker C: Switch on in response to a painful stimulus, but they remain active afterwards. So chronic pain as an oncologist is often, you know, is a massive problem, obviously. And the drugs have their own issues, as if we need to tell anyone that. But when the researchers block the activity of those neurons in mice, the chronic pain went away, although the short term pain stayed. So it's super exciting to know if there's. [00:34:01] Speaker B: Which would sort of be a best case scenario, wouldn't it? Because we don't Want people to lose their pain receptors, But. [00:34:06] Speaker C: Exactly. But as a general rule, you know, this week's been the week of the Nobel Prizes, right? [00:34:11] Speaker B: Yes. [00:34:12] Speaker C: So the Nobel Prize in chemistry went for finding structural MOFs, as they're called. The Nobel Prize in Medicine and physiology went for the discovery of a specialized subset of T cells that helps control the immune system. [00:34:29] Speaker B: Explain that if you exist. That's a fundamental breakthrough in our understanding, as I understand it. So will you give a little more on that? [00:34:36] Speaker C: Sure. So a group of three immunologists, Mary Brunko, Fred Ramsdell and Simon Sakaguchi, won the Nobel Prize for discovering what we call T regulatory T cells, which are a class of immune cells that helps prevent the body from attacking its own tissues and understanding how it works. So it's really useful in diseases like infections, cancer or graft versus host disease. So they described them over 30 years ago. And it's super important in diabetes and autoimmune disease. But these are rare cells. Interestingly, in clinical trials I've been involved in, in cancer, what we try to show is that when cancers are responding, we're getting rid of the T regulatory cells because they damp down the immune system and we're replacing them with cytotoxic T cells. So they're not always a good thing. They're not necessarily a good thing in cancer because what they do is the cancer actually utilizes them as part of an immunization strategy to calm down the immune response in cancers. But people are very optimistic that T regulatory cell based therapies could reach the clinic. But for me, they need to reach the clinic for me to be super excited. [00:35:55] Speaker B: That's the validator. [00:35:56] Speaker C: Yeah, that is the validator. The history of drug development is not one of necessarily huge excitement. It's been one of many. [00:36:05] Speaker B: Lots of great ideas. Founder of the clinic. Yeah, sure, yeah. But it is incredible. Just understanding these kind of core mechanisms of action. That's really, really exciting. So then I'll close with a question about sort of basic science. So again, you read everything. Well, you read a lot. Do you read more than a normal, highly focused person can achieve? So when you see breakthroughs coming, where are they coming from? What's the global distribution on interesting stuff that you're intrigued by? Obviously there's stuff coming out of the United States, there's stuff coming out of the uk. Where else is interesting or interesting? Sort of core science, things that enhance our knowledge. [00:36:55] Speaker C: Last year, more than a third of clinical trial starts were in China. [00:37:00] Speaker B: Yeah. [00:37:01] Speaker C: And previously China was thought about as copying People making drugs not respecting patents. We have seen a massive step shift where you can look at the number of deals between pharmaceutical companies in America and Europe with Chinese biotechs. You can see the value of those deals. You can see that they're not me too drugs. They're totally new molecules focused on new pathways where it looks hugely exciting. So we're actually seeing China as a new originator. [00:37:46] Speaker B: Yeah, I've definitely heard some wringing of hands in places like Cambridge, Massachusetts about, you know, what does this mean for the. Because, you know, I think typically or historically we were able to say with some degree of self satisfaction that most of the cool stuff was coming out of the sort of, you know, the northeast corner in the U.S. san Francisco and that area. The sort of. There's still cool stuff coming out of that, of course. Yeah. [00:38:14] Speaker C: We're seeing for the first time we're seeing real innovation coming from China. Now I'm entering. I'm editor of Oncogene, which is Springer Nature's cancer journal. We see issues with a lot of papers from China as well. I have to double click on issues that are flagged to us and that's frequently from China. So as long as it's handled with a healthy degree of skepticism, I don't really see the problems here. [00:38:43] Speaker B: I heard an interesting description, this from a corp. A corp dev guy in pharma who said the signal is still interesting. In other words, the data may not be good enough for a trial that ran in China alone to then advance through the process with FDA or EMA or whomever. But it's still less risky than starting from scratch. And that there's a lot of enthusiasm for trying to understand what was accomplished. And then maybe you need to run a new ARM or a complete new trial in to order order to get there, but you have a higher confidence in the end result completely. [00:39:20] Speaker C: You certainly will need a new ARM or a new child because of ethnic differences between. [00:39:24] Speaker B: Yeah, of course, at a minimum. [00:39:26] Speaker C: And other issues. Yeah, basically. [00:39:34] Speaker A: Yeah. [00:39:35] Speaker C: All right. [00:39:36] Speaker B: I think that's a great one for us to close out on. Professor Justin Stebbing. It is always a pleasure to chat with you and see what the view looks like from your vantage. So thanks again. [00:39:47] Speaker C: Thanks, Chris. [00:39:48] Speaker B: Welcome, producer Adam. [00:39:50] Speaker A: Hey, Chris. It's great to have Justin Stebbing back. Right? [00:39:54] Speaker B: Yeah. [00:39:56] Speaker A: So let's dive right in into his insights on AI. His comments seem to place AI in the center of a spectrum which we've heard many times on the podcast before. It's either the future star or it's something that's making us just a little bit dumber. [00:40:10] Speaker B: Yeah, I think I'd say he is appropriately skeptical. So he's pushing us to wait for evidence before we conclude that the era of precision medicine is here. And some hyperbolic podcasters are saying we're going to live forever or which, that's just around the corner, stuff like that. That Justin is nowhere near there. On the other hand, he does think there's real potential for breakthrough treatments and we'll all be watching that carefully. And then on the other hand, I think he's arguing that it is going to make a lot of us dumber. And so the question, in a way is which bucket are you going to be in? AI enabled getting smarter or AI crippled or something like that and getting dumber. And I think both buckets seem plausible right now. [00:40:54] Speaker A: Agreed. So while we're talking about that, let's go back to drug development before AI. Justin's comments on GLP1 reminded me of our article that we did with Dr. Andy Dwyer on how we are in the golden age of GLP1. Is it set, like Justin said, to eclipse caffeine and alcohol as our favorite drug? [00:41:16] Speaker B: Yeah, that's a fascinating question, right? I mean, I think we're all seeing extraordinary impact. Probably many of us know someone who's lost a significant amount, a significant amount of work weight on a GLP1. And that's. That's seems magical in many ways. The thing that's exciting now is, is to see the potential impact it's having on lots of other addictive behaviors. And I think we're, you know, we're all kind of crossing our fingers that that will, that will prove out, because, boy, that would be positive. On the other hand, the idea of a nation or a world where Everybody's taking a GLP1 is vaguely dystopian. So I think I get sort of both sides of where people come at this. But on balance, I think, I think he's saying this is an exciting change and more, certainly more positive than not. [00:42:04] Speaker A: I agree. I do agree with that. So let's talk about the comments Justin made about the recent discovery of T regulatory T cells, how they could be a breakthrough for chronic pain. But he's waiting for the clinical trials first. How does that make, does it make you feel good to be in the clinical trials industry at this point? [00:42:26] Speaker B: It does. I think that there's hope that we can make trials faster and more efficient. And there's lots of stuff, AI and regulatory and different regulatory regimes, et cetera. That are all looking at this right now and trying to solve it. It remains the biggest obstacle, biggest challenge I guess, that we're facing in getting treatments to market safely and effectively. But yeah, I think he's saying, you know, an understanding of something as fundamental as the T regulatory T cells and kind of what that means for medicine is still really, really exciting, even if we're not, even if we haven't proven it out in the clinic yet in the form of treatments. And so, you know, a worthy, I think he was arguing, a worthy recipient of the prize. [00:43:16] Speaker A: Very much so. I think, I think as long as clinical trials remain the true standard of validation when it comes to drug development, I think that we have a better chance of a safer, more and having also these pieces become more diverse in which groups they're available to. [00:43:36] Speaker B: Yeah, I think AI can help us a lot in creating more efficient clinical trials. I don't think it's going to eliminate the need for clinical trials in humans. I can't imagine a version of that where we'd all be comfortable with that, at least not in the near term. [00:43:52] Speaker A: So let's talk about China. I know we spoke about it earlier this year when we talked about patents. Are you surprised by what Justin said about this massive shift to China as the starting point for one third of clinical trials into in 2024? [00:44:11] Speaker B: Yeah, it's extraordinary. And so I think we're seeing a few things in China that I didn't see coming. I don't think many people did. You know, the movement, the economists had a great piece on this about innovation in China and the movement from a place that did me too, drugs, you know, they would, they would read the papers and, and you know, kind of read the posters and go to the conferences and then develop similar, you know, biologics to things that looked promising, that sort of stuff, that's fine. There's a place for that in the world. But now we're seeing really exciting innovation coming out of China and we're seeing pharma in the US and biotechs in the US saying hey, there's a way to de risk the process of getting the very risky process of getting a medicine safely to patients. If we can get interesting data that's being developed in China, at the end of the day, I think still a lot of those trials, maybe most of them are going to have to end in the United States because I think regulators are still going to want to see and in Europe, regulators are still going to want, want to see success in clinical research in local populations. I think that's likely to remain the case. [00:45:17] Speaker A: Very good. So before we go, we should talk about this new episode format with just. I know we're trying to break it into a new podcast for Bayarossi that's a little bit different. Different than our, than our current podcast. What can some of our audience do to make this a must listen? [00:45:36] Speaker B: Yeah, well, first tell us, tell us if you like it, because we, we think that the idea of having somebody help us separate the wheat from the chaff, find the signal amidst the noise, that. That's really the concept here. Justin is the person that I go to when I don't understand something. I want a point of view on it. And he's really, really. I think he's a really cool, clear communicator and he reads everything. So the thought was, let's see what he's finding exciting and let's see if that can be material for a different version of Few and Far Between. We're going to Keep interviewing biotech CEOs. I love those conversations and they're really exciting and interesting, and I think we fill a spot in the communications ecosystem that is not otherwise very well covered. So this is an addition to, rather than a change for us. And yeah, if you're listening and you like that idea, tell us why. Tell us what you'd like to hear more of, if there are topics you'd like us to cover, if there are things you think could have been stronger, or of course, if you liked it, like follow, share, and tell us that we are not immune to positive feedback. [00:46:40] Speaker A: That's a great point. And when it comes to comments, you know, you could leave them on LinkedIn. You're more than welcome to send them to our webpage. Contact us sectioniorofsky.com and of course, as you said, Chris, you know, there's also the opportunities for our streaming services such as Apple podcasts and things like that. So, yeah, I'm really looking forward to the next one. And I bet you and Justin are, too. [00:47:05] Speaker B: I for sure am. I've already got a couple of questions to teed up and I'll look for it. Hopefully we'll get some suggestions from folks who are listening. Lots more to come. [00:47:13] Speaker A: All right, that's great. [00:47:14] Speaker B: Thanks, Adam.