Plain English with Derek Thompson - The Biggest Breakthroughs in Science Happening Right Now

Episode Date: December 27, 2023

If you're looking for a hopeful and mind-expanding conversation to round out the year, this one is for you. It's our breakthroughs of the year episode, covering 2023's biggest achievements in science and tech, including space technology, life extension, fusion, gene editing, vaccines, and, of course, GLP-1s. It has become a 'Plain English' tradition—after weeks of stories that often take us into sad areas, like anxiety, depression, and war—to close the year with a nerdy conversation about the most important developments at the frontier of science and technology. Today's frontier guide is Dr. Eric Topol. He is the founder and director of the Scripps Research Translational Institute and a bestselling author on the future of medicine. If you have questions, observations, or ideas for future episodes, email us at [e-mail].  Host: Derek Thompson Guest: Eric Topol Producer: Devon Baroldi Learn more about your ad choices. Visit podcastchoices.com/adchoices

Galaxy Lights, Coachella, Lightning Bolt Necklaces. 20203 was the year of Scandival. On March 3rd, one cheating scandal launched a reality TV investigation that generated hundreds of conspiracy theories, thousands of podcast episodes, and millions of dollars in revenue. I'm Jody Walker, host of an American Scandival. One retrospective story told in three salacious parts. Listen, December 26th, on the Ringer Reality Feudel. feed.

Today is one of my favorite episodes of the year. It is a plain English tradition after weeks and months of unpacking new stories that take us into sad areas like anxiety and depression and war and death and loneliness and tech nonsense. This episode is an oasis of optimism in a news environment that loves its doom and gloom. It is our breakthroughs show, where we run down the most interesting and profound to scientific and technological achievements in the last year. In my capacities as an Atlantic writer, I reported on this piece by emailing several of my

favorite thinkers in tech and science to tell me what they considered the most important achievements in 2023. And first, before we go into the bulk of this episode, which really does focus on the incredible frontier of biotech, I want to talk about two nominees that are in the world of space and energy. So first, in space, some people I talked to said the coolest thing that happened last year. And maybe the most important discovery in terms of understanding the origins of human life was the retrieval of material from a seven-year NASA mission.

I talked about this a bit last week, but just a quick reminder, NASA launched a spacecraft to collect bits of a nearby asteroid called Benu, B-E-N-N-N-E-E. This spacecraft visited the asteroid and returned to Earth this year, landing in Utah. When officials got a closer look at the specimen that it gathered, they saw water molecules inside clay materials. Now, this is important because one of the major theories for why the Earth has water, and therefore why it's been conducive to life, why you and I and everyone that you know exists, is that water came to this planet via an asteroid delivery system.

Dante Loretta, the NASA mission's principal investigator, told the New York Times that the reason that the Earth is a habitable world, the reason we have oceans and lakes and rivers and rain, is because asteroids, like this one, landed on Earth four billion years ago and brought us water from space. What this made me think of is in the opening scene of one of my favorite movies, Prometheus, an alien comes down to Earth and drinks a potion, which dissolves into a trillion pieces of DNA or primordial genetic goo, and that seeds the beginning of humanity. But this, the Benu discovery, is among the most significant confirmations that, no, actually, we came not from aliens,

but from asteroids, and from the long, long chemical processes that follow their crashing into Earth and leaving behind small pools of water. The second nominee is in the field of energy. And speaking of space objects, the sun heats our solar system through an energy process called fusion. This is an energy process we've never been able to efficiently recreate on Earth.

Inside the sun and other stars, heat and light are thrown off as atoms crash into each other and merge in a process called fusion. This is the opposite, by the way, of a nuclear reactor, what you and I know is a nuclear power plant, which uses fission, right, the splitting of atoms to release heat. Fusion, on the other hand, has always been a space stream. A total impossibility on this planet. But in the last 13 months, we've had two historic breakthroughs. And because this next part is a little bit complicated, I'm just going to read you what I

wrote for the Atlantic. Quote, last December at the Lawrence Livermore National Laboratory in California, December 2022, 192 lasers blasted a diamond encasing a small amount of frozen hydrogen and created, for less than 100 trillionths of a second, a reaction that produced about three megajoules of energy. In that moment, scientists said, they achieved the first lab-made fusion reaction

to ever create more energy than it took to produce. Seven months later, they did it again. In July 2023, researchers at the same ignition facility nearly doubled the net amount of energy ever created by a fusion reaction. Startups around the world are racing to keep up with science labs. New fusion companies like Commonwealth fusion systems and Helion are trying to scale this technology.

Will fusion heat your home next year? Fat chance. Next decade, cross your fingers. Within the lifetime of anyone listening to this podcast, conceivably. So, space, asteroids, Earth, fusion, those are two of the coolest breakthroughs of the last year. But my top honors in breakthrough of the year were almost entirely in biotech, as I said. In the last 12 months, we've seen America's first FDA-approved CRISPR therapy, vaccines for malaria and RSV, an amazing experiment that I think of as what-if face paint cured cancer. And of course,

We return again to the GLP-1 revolution in diabetes and weight loss drugs. Today's guest is Eric Topal. He is the founder and director of the Scripps Research Translational Institute and a best-selling author on the future of medicine. He is our trusty guide today through the frontier of the biotech revolution. I'm Derek Thompson. This is plain English. Dr. Eric Topal, welcome to the show.

Great to be with you, Derek. So my goal for our next 40-ish minutes together is to give people an appropriately optimistic safari guide to the most important, the most interesting, the most wondrous breakthroughs happening right now in biotech. And I should begin by saying, I am so very far from being an expert in these subjects, but for better or worse, this is the part of the world, this is the genre of news that I've become the most interested in in the last two years, probably since maybe my conversation's with you about the MRNA vaccine breakthroughs. So I gave you a list of today's topics, just so you had a rough roadmap of the journey.

But before we set off, I want you to tell me, of all the breakthroughs, the science reports, the AI research, the published papers, the cover stories in science and cell and nature, I want to know what is Dr. Topol's nomination as the most important or interesting breakthrough in science in the last year? Well, there are many ones that, of course, you can imagine, are worthy to get kind of special recognition. But the one that I found the most intriguing of all was the work from Stanford, Tony Weiss, Coray, and his colleagues on internal clocks. It was the cover of nature just a couple of weeks ago. And what it was the first time to be able to tell the age of 11 different organs of the body by different plasma programs, proteins such that this will really advance the science of aging and the ability to influence aging at an organ-specific level. And 20% of us have advanced accelerated aging of one organ.

So this was a breakthrough. This has not really been shown before. And it took, you know, really a monstrous effort of looking at thousands and thousands of proteins and figuring out if they were specific to an organ. and then showing that these proteins were linked to outcomes like heart failure or Alzheimer's or all the other organ-specific type adverse things. So it was a big contribution among many that I highly regard. I had no idea you were going to say this.

I know nothing about internal clocks, but I'll tell you the first question that occurred to my mind when you said internal clocks and you kept talking was it's interesting to think that when someone asks me, how old are you? I have one number I give them, 37. But I have been alive for 37 years. That is my external clock, so to speak. My liver, you know, I drink maybe a little bit too much whiskey and scotch than I should. Maybe my liver is 45.

Maybe my heart is 23. Maybe my brain is 34. Maybe my pancreas is 55. Tell me how in some, hopefully, not science fiction, but science fact future, someone listening to this podcast could imagine getting a birth certificate for all of their organs. What kind of a test would it take to learn, well, you're a 37-year-old man, but your heart is this age, your pancreas is this age, and your liver is this age. How would someone begin to get that information in the future? Yeah, well, it's available today for your total body, you know, the so-called epigenetic clock of Steve Horvath,

where it looks at methylation markers, and it can tell you very accurately, what's your biological age? So you could say, well, Derek, you're 37, but your biological age is 42. But that's a total body, and that's what the best that we have right now for a clock. Now, to be able to drill down into 11 most important organs is a new thing. And so this is going to very likely become available, you know, widely in a way to do checkups and people much better than that we can today. We don't want to have to always resort to MRI scans or CT scans or that sort of thing.

And also, you know, doing these liquid biopsies that are starting to get some traction and cancer, you know, it isn't clear that they're going to be useful to prevent not only cancer, but other conditions as well. So that's why this was a finding that really stands out. It's not yet ready for daily, you know, to give you a readout, like you said, is that whiskey hurting you hurting your liver to some degree. But it will. This is where we're headed. It's really cool. Do we know one more follow-up question on this before we get into the meat of today's episode. Do we know if these full body biological clock tests are truly predictive, by which I mean if there's a 60-year-old who's told that their body is the biological age of a 30-year-old, and there's a 30-year-old who's told they have a biological age of 60,

can we really expect that 60-year-old to live like 30 years longer than the 30-year-old? I don't know that my full math is right there, but is it predictive on that kind of lifespan level? Yeah, actually, it turns out they're pretty darn accurate, you know, looked at him just, you know, you know, tens of thousands, even not hundreds of thousands of people. The tricky part is, you know, I kind of look at it if you want to give it a very simplistic reductionist description, it's like the rusting of your body. You know, some people don't rust very much and others do. But it's very generalized and rudimentary.

It doesn't tell you where the problems lie. But the reason why this is a big deal also is because if we're ever going to start to find ways to promote aging, you know, the reduction of aging, this, you know, is decelerating our aging process. It's important to have these kind of metrics because it's unlikely that, you know, any particular intervention is going to have a quick whole body effect, but it might on a particular organ. So that's why this is especially promising. It's a path towards a regulatory, towards approval someday of agents that, promote healthy agent.

So when I wrote my piece for the Atlantic about the breakthroughs of the year in 2023, my award for the most significant breakthrough in science and tech went to the breakthrough in CRISPR. So in December, the FDA approved the world's first medicine based on CRISPR technology. This has been made by pharmaceutical companies based in Boston and Switzerland. It is a treatment for sickle cell disease, which is a chronic blood disorder that affects about, 100,000 people in the U.S., millions of people around the world, most of them black. Before we dig into a little bit about the frontier of CRISPR, can you explain for a lay audience? What is CRISPR and how does it work?

Right. Well, this is the biggest life science breakthrough of our time. This is the ability to edit the genome. The problem is that CRISPR, which was basically started, you know, 10 years ago, was a blunt tool. You know, basically it cut across both strands of DNA. And so it was very highly disruptive, which is what was approved for sickle cell. And since then, we have much more precise ways to edit, namely basis. editing and prime editing.

So while I really like the CRISPR sickle cell, obviously it was great to see a first FDA approval and approvals in other countries like the UK. The problem with it is it isn't direct. It's working on, you know, alpha fetal hemoglobin. And so it's not correcting the sickle cell. Right. Right. Sickle cell is caused by a genetic mutation that affects the production of hemoglobin,

which is the protein that carries oxygen in red blood cells. And tell me if I'm wrong, this treatment doesn't correct the genetic mutation. It essentially works on another gene that has stopped the production of fetal hemoglobin, which does not sickle, doesn't produce these sickle-shaped cells which clog and create extreme pain in anemia. it turns that gene from red to green, and suddenly the body starts producing fetal hemoglobin, which does not sickle. And so these people still have the genetic mutation, uncorrected,

that causes sickle cell disease or sickle cell anemia, but a separate gene has been edited in order to produce this fetal hemoglobin that essentially overrides their disease by filling their body with fetal hemoglobin, which most adults do not produce. Is that the general picture here? Well, you did it really well. That is an excellent explanation, but it also cuts to the chase here, which is we didn't fix the genetic defect. It was a bypass or in a workaround path, knowing that we had this only rudimentary way to do genome editing, the original genome editing. It's only in recent years that we have these really incredible tools that you could fix the sickle cell gene. Or basically, you know, get close to any gene now can be fixed without.

having to do disruption of double-stranded DNA. So that's what's exciting is what's in store. In fact, you know, we'll go back to sickle cell with actually correcting sickle cell. But look, this is great because theoretically it can help a lot of the people with sickle-cell disease. The problem, as you know, Derek, is that it's incredibly complex. I mean, my goodness. You know, that's why I like the precise ways to do genome editing where you just get one shot and you're done and you're cured and it's specific to the gene. This one is hardly that. This is, you know, basically having it involving, you know, bone marrow wipe out a month in the hospital, incredible, you know, expense. You know, a lot of people will never be able to access this very complex form of treatment of. So on the one hand, it's a momentous advance that we have approval, but it's just the beginning

of how exciting a genome editing field is going to get in the years ahead. I'm glad that you mentioned some of the problems, because in this next question, I'm going to both, well, first pour cold water on this discovery and then hopefully get us hyped up again. So there are really fair concerns that I found in my reporting on price and access. As you alluded to, these treatments are. are incredibly complicated. They involve bone marrow, blood transfusions. They require weeks or months long stays in the hospital. They also cost millions, not one, several millions of dollars,

at least in the sticker price. Now, 70 to 80 percent of sickle cell anemia patients in the world live in sub-Saharan Africa. The average GDP per capita in sub-Saharan Africa is $2,000 a year, which means when you multiply it out very simply, you're talking about a treatment that would require a thousand people's annual salary to afford one treatment in the U.S. that'd be the equivalent of us talking about a $70 million treatment. So an incredibly expensive treatment, at least in terms of the sticker price, even as it is an amazing scientific breakthrough. So that's the cold water, the complications of, and the price of this treatment mean that

it's not maybe going to immediately have this enormous effect on the number of people suffering from sickle cell disease in the world. Okay, get us hyped up again. you on your substack talk to David Luz, a molecular biologist and a chemist at the Broad Institute, who is working on this next generation of CRISPR that you've alluded to, base editing and prime editors. If I asked you to like peel back the curtain just a bit to preview, you know, the most exciting CRISPR research in either early clinical trials or pre-clinical research, like what is just over the horizon that people should be so excited about? Well, David really is incredibly creative and has come up with, the inventor of these two much more refined advanced forms of editing.

And so one of the ways it's being done today is in people with very high cholesterol, familial hypercholestrelemia, where 10 people were recently reported that have this F8 condition, where their bad cholesterol, LDL is, you know, several hundred despite the fact that they're on medications for that. And they get one shot of a base editor goes to the liver, fixes the specific PCSK9 that's responsible, and then their LDL is down at good levels and should be a lifelong treatment. So it's very different than the sickle cell that you just reviewed. To one shot intravenous fixes the gene specifically and could have a lifelong, cure of a very serious condition. But to get you even further, if this works, it's safe.

Then we talk about instead of everybody having to take cholesterol medicines for all their life, what about you just go get a shot, you know, when you're young and you don't have to worry about forever, you know. So the expense is absurd today. It's intolerable, you know, millions of dollars. But as the price comes down and the volume of people potentially to get genome editing increases, it's possible someday. We're not talking about in the near term, but you know, a decade plus from now, that these could get to very reasonable costs at scale. And that's just an example. I mean, we're talking about, you name the genetic condition and then the extension of that kinetic condition.

So there are a lot of people who don't have familial hypercholosteremia, but they have still what's known as polygenic, high cholesterol. They could benefit from such an approach. And that's kind of, you know, way most chronic conditions people are. So there's almost unlimited potential if the price can come down and if it's one-off and if it's a cure, if it doesn't have the so-called off-target effects or on-target, issues, if we don't have problems that are unforeseen in the longer haul, if we can surmount

these obstacles germ genome editing is just an amazing therapeutic in the years ahead. I have one more follow-up question on CRISPR before we move on to the next arena of breakthrough. I'm curious to know what you see as the rate limiting step for these, for base editing and prime editing to be delivered and reached the kind of phase three clinical trials that the sickle cell disease CRISPR therapy reached. Is it we haven't yet perfected the therapy itself, the technique of editing at the atomic level, the A's into G's and putting in the missing CCTs for cystic fibrosis, or is it that we don't yet know where many of the single gene mutations are? And so we're still having a treasure hunt to figure out what those gene mutations are

and maybe even where the polygenic mutations live that are responsible for more complex diseases like Alzheimer's or dementia. So is it the technique that we're working on? Is it that we're still looking for the right genes to target? Or is it something entirely different? What do you see is the most significant rate limiting step here? Well, it's really not any problem as far as knowing the genes because we've got that down in 7,000 conditions. And so the list is long that can be genome edited for treatment.

The problem is the delivery. So if it's a liver, okay. If it's in the blood, like sickle cell, okay. But, you know, when you start getting to other organs, then we got a problem. The eye, you could do local delivery, but then to the heart, the brain, other parts of the body, it's a problem.

So delivery has to improve. That's a rate limiting step right now. There are a lot of ideas of how that's going to go forward, but that's the one that puts the list at a relatively limited number of conditions. Just to round this out, I mean, talk about the extended use of genome editing.

Take 63 genes in a person to block the ability to have a pig transplant organ. 63 genes so that you don't have to take immunosuppressin drugs. You could just genome added all of them at once. It's like, whoa. So, you know, we're talking about, as you well know, there's a tremendous shortage of organs for heart and lung and liver and the ability to use animals with CRISPR by, you know, taking their organs and putting them through genome editing. That just gives you that.

And, of course, all the cancers that we could, you know, leukemia is and others that we can do genome editing outside the body. So it isn't just things in the body. fact, outside the body is not a delivery issue. So that gives you another edge. And so people should think of germ line. We don't want to go there. Genome editing as just a tool that is just, you know, diverse. And we're just getting out of the starting block here. That's so interesting. And just thinking about all the different diseases that you're talking about, right? Like you've alluded to congenital blindness, to heart disease, to maybe diabetes or cancer,

to high cholesterol, maybe coronary artery disease, to the rejection of a pig organ transplant. I mean, we're just talking about so many different diseases and that really, I think, establishes CRISPR as really this platform technology where it's not about any one disease. It's about the ability to make all sorts of things possible that are currently impossible, which is, you know, in the most optimistic vision, what technology is all about. One last thing I just want to point out is, you know, Alzheimer's, which we don't still have anything really to help significantly in Alzheimer's. David and I spoke about the idea of changing our APOE4 gene to APOE2 and the twofer you get to take out your risk of Alzheimer's and get a longer

life. So that's out there. Jennifer Dowden is also. commented on that. So we're talking about getting to, you know, conditions that were never even conceivable through better delivery and improvements that are to come in gene editing. Let's move on to vaccines. Most of our conversations, I would say, have been about vaccines. And this has been another good spell on the vaccine front. So 15 months ago, the first malaria vaccine developed by University of Oxford Scientist was endorsed by the World Health Organization. it has already been administered to millions of children, which is incredible because malaria is one of the leading causes of death for children worldwide. But demand is still outstripping supply, and that's why it's so fantastic that a second malaria vaccine called R21 was recommended this year, 2023 by the WHO.

It seems to be cheaper, seems to be easier to manufacture. And then in addition to a second malaria vaccine, the other vaccine that really caught my eye because it was administered to my four-month-old daughter is the FDA-approved. several vaccines against RSV, which is so common that an estimated 97% of children catch it before they turn two in addition to it being a serious risk for older Americans as well. I want you to, so you got malaria, we got RSV, you might have others half of mind. I'd like you to actually begin by scoping out. Do you have a big picture explanation for why we seem to be in a golden age for vaccine research? Well, it isn't just MRNA, the package with nanoparticles.

That's one part. But, I mean, some of the, like the RSV vaccines, which are pretty striking, you know, a single shot and not necessarily relying on RNA, the malaria veins of vaccines, not MRNA vaccine. So what we're seeing is vaccinology has had decades to warm up. And at the same time, these pathogens have had. had many decades without a vaccine. I mean, we're talking about malaria and RSV, and there's so many other pathogens where we had nothing for, you know, 30, 40, 50 years. And we still have nothing for many important diseases that are killers like tuberculosis and so many others.

But what is happening is, you know, you're watching all these dead triumphs. You know, it could be Zika, it could be Ebola. You know, you name the condition of pathogen. and vaccines are in the works. And we're seeing triple vaccines. So in the next couple of years, there'll be an RSV COVID flu vaccine and, you know, things like that. So this is a typical thing in science that most people don't realize is when you,

the mRNA nanoparticle breakthrough, which was incubating for three plus decades, is emblematic of all what's been going on in vaccinology. You know, some of this is that we've been able to sequence the virus, understand the structural biology of the proteins involved. That's what helped us a lot with RSV. But some of it is just that, you know, science takes a long time to get these remarkable triumphs and we're seeing them.

And, you know, at some point, we're not going to have pathogens that we can't make really good vaccines to guard against. And that's exciting. It's extraordinary. There's nothing more potent in vaccines in our armamentary. to prevent diseases. And by the way, we're talking about cancer, vaccines, prevent a vaccine someday, vaccines perhaps to prevent, you know,

Alzheimer's, coronary disease, you know, the work. So it isn't just against infectious diseases that we're talking about. I, as a general sort of, as somebody who's interested in science and tech more broadly and some of the themes in the history of progress, have always been interested in this concept of twin ideas or sometimes called simultaneous invention. I mean, the theory of evolution was essentially discovered by Darwin and other European scientists in the same year. You have, what is it? Charles Wheatstone and Samuel Morse invented the telegraph in the same year, 1837. Elijah Gray and Alexander Graham Bell filed patents for the telephone on the exact same day in 1876.

I think we kind of saw this in artificial intelligence. It was two years ago, toward the end of the year in the fall of 2021, 2022, excuse me, where you didn't just have Chachibout Chachibit and Dali and Mid-Journey. All of these tools seem to have this sudden simultaneous Cambrian explosion, and this might be taking the metaphor a little bit too far, but it seems like we are having this reunion party for vaccine success stories that's happening around the same time. And I'm just wondering whether maybe it's luck, right?

I mean, the theory of simultaneous invention and the theory of twin ideas just sort of says, there are ideas that are in the ether that lots of people just sort of pick up around the same time based on technology just generally being at the same place in Boston and New York and L.A. and London. And so people just invent the telethone at the same day. That's just how it happens. Do you have a more sophisticated theory

for why the last few years have been this sort of pull back the curtain, aha, everything is ready moment for vaccines being approved? Because as you said, of course, the research goes back decades, but the date of approval is a date of approval. And we seem to be getting a cluster of them. Any possible reason why we seem to be sort of accelerating in terms of our success in vaccinology recently? Well, I love the analogies you've made parallels to, you know, the AI world with transformer models and, you know, the various old inventions that, you know, with simultaneous, you know, this is a really tough question. I wonder about it myself. What really is the explanation? I mean, I think there's parallel efforts that go on. I mean,

we've seen this up for, for example, RSV vaccines. They were worked on, you know, at NIH, at GSK, at Pfizer, and then the time it takes to get it done is typically you can't change it that much. So sometimes it's just the fact that, you know, there's a target, there's a goal, and, you know, it's kind of the usual time that's required. I think that also accounts for what's happened in the GPT world where, you know, there was a transformer, a preprint in 27, and it took, you know, that many years. And all of a sudden, like you said, there was a party of all these different large language models, or, you know, really basically multimodal models coming out at once because

they got, they went to that incubation phase. So I think that's what it really is. Because when you talk about vaccines, you're talking about understanding the pathogen, sequencing the pathogen, which has only become, you know, much more common, not just sequencing the pathogen, but lots of people who have that pathogen. So that, along with the structural biology, the idea of delivery, which sometimes is through an MRNA, but of course it's a diverse way to get it into cells. So these all things were happening together and a similar timeline. And I think that's what's enabled so many new ways. I mean, the RSV was being,

pursued before COVID and it enabled COVID but it didn't hit until after COVID but you know if it hadn't been for the work that was being done with RSV to understand the so-called pre-fusion protein it accelerated our ability to interfere and get high neutralizing antibodies to SARS-CoV-2. Let's talk about AI. In the last year I've read stories about you didn't using artificial intelligence to predict protein shapes to read radiology, to read radiology reports, to interpret radiology, reports to assist in diagnoses for complex diseases. This is a pretty bewildering landscape for me, like the intersection of AI and medicine.

But fortunately, you literally wrote the book on AI and medicine. What stands out to you as the most significant recent work at that intersection of AI and healthcare? Well, I think the biggest contribution thus far is the so-called Alpha II, which is as you know, the ability to predict from a linear sequence of amino acids, you know, 1D, or dimension, 3D at the atomic level, structure of basically the entire world's proteome, every protein, 200 million proteins in the world, with variable levels of confidence, but nonetheless here it is. And just to put that in context, you know,

where I work at Scurice Research, I had colleagues that would take three years to crystallize a protein, and now they can do it in, you know, three minutes. So, I mean, this is just extraordinary, and it's enabled by, you know, the work of DeepMey and a transformer model, which has now been the birth of many other transformer models in life science. And the same thing that you just reviewed is happening in the medical space as well, which basically is we're not now, you know, large language models as a term is about to become obsolete. It's not just language, it's images, it's speech. And, you know, that is getting us to a point where making medical diagnoses much more accurately is going to be in the near future. I mean, we're starting to see it now,

good validation of that. But the biggest thing so far to show that you could change the world of life science or biomedicine clearly is this alpha-fold to an inch derivatives. There's many different transformer models now. You can even invent

proteins that were never known to nature. Don't exist in nature. And, you know, the biggest thing this week was the discovery of a whole new structural class of antibiotics. The last one took 38 years. And James Collins and his colleagues at Weiss Institute in Harvard did this, and it's a

stunner. So that's basically, and that's not even transformer. That was just the old deep learning at all, I say, because it's, you know, less than a decade now. But look what's going to happen here. It's just mind-blowing. Connect this to how the technology could actually affect people's lives. So let's say I'm a scientist. I read a sequence of amino acids. I build an accurate, perfectly accurate, to the nano-adam, accurate, 3D model of the protein. What do I do with that? How do I use a perfect 3-D model of a protein to develop a drug or help a class of patient?

Yeah, well, that's the template to build on an antibody to that protein. You know exactly where to bind or to build a... small molecule, like a pill as you could take to know exactly where the business ends of that molecule are and the pockets and how you get in and these so-called cryptic, which is, you know, the hidden parts of the protein. So it basically unlocks, you know, it's a treasure chest for making drugs, and that's why you're seeing the acceleration of drug efforts now like we've never seen before. So, you know, whereas it might have taken 10 or 20 years to come up with a new molecule that has big impact, we're going to see that down to very short periods of time. And we're

going to see a lot of AI helping to invent the drugs. And the only question I phrase is, shouldn't the AI get some credit? Because the humans are, you know, kind of pushing the buttons. But the AI is doing a lot here that we don't fully understand. So it's sort of like if someone, we're developing an antibody to some kind of malady, right, some kind of disease. And historically, we've been sort of reaching into our bag and trying to fit random keys into a lock, and we don't know the shape of the lock. And so it's just like, trying key number one, doesn't work. Try and key number two, doesn't work. But if we could somehow x-ray the lock and we knew exactly where all of its little points up and points down and how thick this part is

and how fat that part is, we could say, oh, this is exactly the kind of key that we need in order to open this closed door and then we just go off and 3D print that key or just have some key maker make that key, it slips and perfectly opens up. That's obviously incredibly simplified. I'm sure this process takes many years, but it goes from blind to visible.

I love the metaphor, Derek. That's perfectly right. That, you know, structural biology, a lot of people don't get, but an x-ray of the lock and the key, yeah, there it is. Very good. Cool.

All right. Well, that sounds great. I mean, where are we close to, are there, the same way that the first target up for CRISPR turned out to be sickle cell disease, do you know if there are certain diseases or certain classes of diseases where this kind of protein intelligence is making it easier for us to pick that lock? Like, is there some disease that you've read reports where it's like, oh, we're making promising progress on this kind of thing thanks to these 3D models of protein that we're able to develop.

Well, I mean, there again, the list is long. And it's, you know, you only will say that it clicked when you actually have a drug that doesn't have toxicity and really does the job. I mean, since we started to see AI apply to new drugs, I mean, one of the things that we've never had a good drug for is to prevent scarring of organs in the body, whether that's the liver or the heart or you name it, lungs. And, you know, that's one of the first drugs that came out of AI. It was a drug, you know, that was developed through AI.

But one of the things that a lot of people don't realize, Derek, is that there was a drug that was found by AI mining during the pandemic, baricitinib, which is life-saving and proved by the FDA fully for, you know, saving lives with severe COVID. And if it wasn't for AI, we wouldn't have known that this, drug, which is used for rheumatoid arthritis and alopecia areata. So that's another thing, is that using AI to look at these atomic structures and then going through all the known drugs, the 20,000 drugs and how they work and say, oh, well, here's a good one. So repurposing drugs also will be accelerated. Two more categories I want to ask you about. The first is, so in last

year's breakthroughs essay, one of the weirdest examples of a science breakthrough that I found that was recommended to me was a liquid solution that revived the organs of dead pigs. And this year's category of, wait, what is the news that some scientists figured out a way to engineer a common skin bacteria. They engineered it to carry bits of tumor material, of tumor information. And when they rubbed a concoction of this engineered skin bacteria

on the head of mice in a lab, the animals produced T cells inside the body that sought out that tumor, the attending tumor, and attacked it.

So, as I joked in the article that I published for the Atlantic, the jokey way to summarize this is face paint that cures cancer, or skin cream that cures cancer, that fights cancer, right? Like you rub this topical chemotherapy or topical, you know, carty cell on your forehead, and it somehow goes to fight the distal cancer. First, please do your best to correct anyway in which I've utterly bastardized the summary

of this particular piece of research. And second, maybe make us a little bit smarter about what exactly was done here and why it might be so important. Because at least one scientist I really, really respect in the, in the, Bay Area said that this was the most interesting breakthrough that he saw in the last 12 months. Yeah, well, you know, when you asked me about the top breakthroughs, you know, overall, you know, I'd say that our ability to either enhance the immune response or take it away, suppress it, is going to new levels that we've never seen before. And this is just one example of that.

So while it's in mice and it's with mice that have melanoma tumors, the ability to rev up our immune system, of course, we know that. I mean, we need to, in fact, that's where, again, gene therapy, gene editing, genome editing is being used. But we know that we can squash a lot of tumors by revving up our immune system, particularly our T cells, our cytotoxiciller T cells. So this made a lot of sense, you know, that is it's just another way through the skin, through a skin tumor melanoma, that you should be able to do that. Now we have to see whether it's going to hold up in humans, but it's encouraging. The point here is that it's just one of so many different ways we're going to be attacking cancer because probably what we're starting to realize now overall is that the basic reason why

cancer spreads, which is a killer. It isn't the cancer. It's the spread to metastasis is because our immune system can't squash it. And especially as we get older, our immune system, as you know, we have senescence, immunosinessence. So this is a way to rev up our immune system. And of course, it could be widely applicable. And using the skin, using bacteria as a carrier, fine. You know, whatever way you've got to get it in. You know, get it into the body to make the person, ideally, to recognize that tumor specifically, so it doesn't kill other cells, just gets the cancer. And so, you know, this is where the whole field is headed. Yeah. I recently read Sadratham Mukherjee's book on cells. I think it's called The Song of

the Cell. And he has this really poetic chapter on T-cells, where he says, T-cells are the body's way to distinguish self from non-self. Because T-cells attack the disease within us, but there's research that suggests that if you take my T-cells outside of my body and they put them into my friend's body, they don't do the job of my friend's body. They can only tell the difference between me and the disease inside of me. So they know the difference between self and non-self. And so the sort of metaphorical way that I've come to understand the power of T-cells is we need to find ways. You mentioned that what is cancer? Cancer is a failure in a way to recognize non-self, right? It's the immune system's inability to recognize the non-self that is metastasizing, growing inside of

the self. And we're finding these little ways to make the T-cells smarter at recognizing non-self. Is that essentially it? With CART therapy and maybe with, you know, this, this bacterial engineering in mice that we're talking about now, we're finding little ways to raise the IQ of our T-cells. Absolutely. I mean, we got two things going on here. One is the cancer hijacks the cells. It figures out a way to evade our immune system. And then the other is that our impaired ability as a person in general to get our immune system to recognize the cancers, the cell. So yeah, I mean, this is a really nice, you're very good out of analogies, Derek. And I think that's a real, real, talent. So I got to give you a lot of credit for that. Well, unfortunately, something else is very good analogies, and it's chat GPT. So

before it puts you out of work in cardiology, it's going to put me out of work in analogy making. Let's close on GLP-1s. We did several episodes in GLP-1s last week, or maybe it was two weeks ago. I think these things are absolutely fascinating. I think they're fascinating in terms of what they do for people with diabetes, fascinating for

their effects on weight loss, not just the gLP ones, but also the dual agonist, the triple agonist that are coming out, we're attacking more and more, or mimicking, I should say, not attacking, mimicking more and more hormones and having a larger and larger effect on weight loss.

But I know that you, in addition to all these hats you wear as a cardiologist, you're also really interested in the effects we're seeing in terms of it reducing cardiac events, like heart failure, stroke. Maybe just talk a little bit about these unexpected surprisingly, maybe even miraculous side effects that we're seeing with this GLP1 plus class of medications. What are you seeing here? Yeah, I mean, I wouldn't call them side effects. What I'd

call them is that the problem with the obesity or underlying diabetes, metabolic syndrome, these conditions are pro-inflammatory throughout the body. And we haven't really had good agents to block inflammation. And even before you lose the weight, for example. So what these GLP1 drugs, and, you know, as we got the dual receptor and the triple receptor, we're increasing the potency, increasing the anti-inflammatory effect, it looks like as well. And that's what's putting up a, you know, the reducibility to reduce what's called preserved

ejection heart failure, which is half of heart failure, which largely is from obesity. The ability to prevent heart attacks, strokes, and heart cardiovascular deaths has been seen in trials that were completed in this year. But that's just the beginning. This is headed, the problem, of course, these drugs right now, they're injectable, they're very expensive, But pill forms are coming. Expense will inevitably get down. It has to go much, much lower, of course. But the availability or access to the meds.

But we're going to see very likely this becoming not just the breakthrough for obesity, as was first heroin, but an across-the-board benefit for many, many, whether it's liver disease, kidney disease is hard, you know, possibly Alzheimer's is a very big trial that's going on right now that'll read out in 2025. We're talking about, you know, most chronic diseases of man that this could be a potential remedy for if it becomes, you know, very inexpensive and doesn't have to require, you know, frequent injections. So that is exciting. There's so many things that are going right now that are, extraordinary, but this is one that we're just seeing the beginning of it right now.

You know, we basically have seen most of the work has been with semi-glutide, which is, you know, relatively weak hitter, and there's many more potent ones coming along. And, you know, we don't know about the long-term side effects of these drugs over many years. We don't know how to wean people off these drugs, so they don't have to take it for life. There are many unanswered issues, but the effects that the biologically, level are exciting. And we haven't really had a drug that does this. The only other drug that decreases inflammation safely that we rely upon is in the heart world is statins. It's the only other one that's big. But this has the potential for being much bigger because of its pronounced

effects across, you know, not just obesity, but all the other things I've mentioned. It's interesting to me because, and maybe this is wrong, and it's just me being a Johnny come lately to the biotech space. But it's my sense that typically there's a discovery process in science followed by FDA approval. And with GLP-1s, it's like we had FDA approval followed by discovery process. Like, we're like, oh, my God, this thing that we already knew to be effective at treating type 2 diabetes and also type 1 diabetes, it turns out it's incredible for weight loss. Oh, my God, it turns out is doing this thing for, you know, reducing people's appetite for candy. Oh, it's also good for cardiovascular health. Oh, it also might reduce rates of Alzheimer's.

It's like it almost sounds made up, right? I said in one of my earlier episodes, like, you know, Derek, you can't just name a bunch of good things and list them one after another and say, GLP1 does that as well. But like we are like in that honeymoon phase with his drug where it does seem like the number of positive, I won't call it side effects, I'll call it consequences of taking the drug. Yeah, it's really. Yeah, benefits. It's really sensational.

You know, you could get rid of cravings and gambling and alcohol abuse. I mean, the list is just, it's absurd how long it is. is. But, you know, one thing I'd say here is how dumb we were, okay? Because the first of these drugs got approved in 2004, okay? And the reason we're so dumb is that we didn't jump on this 20 years ago. If we had GPT4 20 years ago and we'd say, hmm, what do you think we could do with this drug? and say, you know, maybe not just think about it for glucose control. Maybe you could do a lot more. And, you know, with GPT4 would say, yeah, go for it.

Make a long half-life and try obesity because it has a lot of promise. And by the way, once you hit the hypostomans and the limbic system, you know, there's a lot you can do here. But it took 20 years there to figure that out. Okay. So there's the intersection between GPT and GLP here, which is really interesting. And I actually think this could be likely the biggest drug class in medical history, but it will be a segue to others as well.

I want to close even, well, I'll make it with this penultimate question because it's a little bit more of a downer. There are some safety questions about the GLP 1 class of drugs. I know you talk to Peter Atia, who is concerned about, or I should say looking at two side effects that he's a little bit worried of. Number one is it seems to raise people's heart rate while they're sleeping. You're the cardiac experts, so I'll have you comment on that. And then second, there's a muscle loss component that you put people in dexas scanners, and they tend to lose roughly equal amounts of fat and muscle. I talked to Robert Lustig on this show, again, two weeks ago, and when he brought this up, my response was, that's not good, but it's my hope that we can find some way to raise muscle mass, whether it's medically

or just by the endocrinologist or internist physician telling their patient on the GOP-1s, hey, you need to strongly consider having a heavy lifting regimen because the muscle loss, especially as you get older, can be difficult and not entirely healthy. How do you feel about some of these side effects that we're seeing? There's nausea, which is more common, but then there's also these fears about elevated heart-rated sleep and muscle loss. Yeah, I looked into the elevated heart rate. It is common, but it's usually just a few beats per minute. The problem is some people get up to even 20 beats per minute of the resting heart rate heightened by taking these drugs. So that's something to keep an eye, and we don't want drugs that are going to increase heart rate 10 to 20. It isn't common at all to see that.

But when it does occur, that should be at least a yellow flag. Now, and we don't know why either. That happens. And we don't fully understand the mechanism of these drugs, which also needs work. That gets us to your question, which I think is fundamental, is about loss of muscle mass. There's a recent study with MRI. Of course, it's from the company. But it's better than a Dexas scan that basically is saying,

there's really not that much muscle mass loss. And then, of course, as you are getting at, there are ways to counter the potential muscle loss with more protein intake or weightlifting and doing particular exercises to counter that because we don't want this so-called sarcopenic obesity with people who are falling and frail because they've lost skeletal muscle and even bone density.

So this is an unknown yet. There's mixed data. It's a bit concerning for sure. There's also the kind of known unknowns that we could see that we, you know, so haven't given at high doses for years, what's going to happen to people since, you know, it's really, it's, it's horrible that these companies are not doing anything to try to get the drugs off of people. They're just committing them, oh, this is like you're insulin and you're a diabetic. Take it for life. No. Why? So we should be seeing work like that because there well could be some side effects that are concerning we haven't seen yet. Because, you know, it's kind of, I think you got to, it's too good to be true. And there's going to be, besides nausea and GI side effects and besides the ones that we just discussed, who knows what else we might see over time. So we need more work on the muscle mass prevention, loss of that. And we also should be open-minded about seeing things we haven't yet,

haven't yet surfaced. The longest follow-up of people at high dose of a glyp one is 40 months. That's not very long in the big picture. I want to close by braiding my first question and my last question, and that is, if we understand with this new class of drugs that medication that mimics, you know, glugon like peptide one and gip and glucagon, I mean, those are the three hormones that are hit by Reda Trutide, which is the latest class of the GLP1 Plus drugs, if we recognize that mimicking those kind of hormones is what this is all about and we don't want people to have to stab themselves in the leg once a week for the rest of their lives, is there a crisper solution here? I'm like, I mean, is there some way that we could figure out a genetic, monogenetic or polygenic

place where if we do some kind of base or prime editing, we could change the genome to have a different kind of production of GLP, GIP, GIP, Gugugan. Have you ever thought about there being some way that, like, obviously, like, GLP is, like, shown us the way forward in terms of, you know, the weight comes down, there's all sorts of benefits. But maybe we could do this not at the end of a needle, but rather through CRISPR. Is there any way that's plausible, or am I way off base there? No, you're not off base. I think you're off time, think it might take quite a long time to get to that.

You know, I think it intersects more with what I started with about the internal clocks, that if you can work on these three peptides, the chance of you changing organ aging is enhanced. But to try to simulate the triple receptor agonist, you know, with a base or prime editing, it's potentially doable. The question is, you know, what tissue are you going after? And, you know, a lot of these effects, particularly the inflammation, has appeared to be meted through the brain. That's not one of the delivery places right now that we can get to. So if you really want to start to change the triple peptide story, we got to work on delivery much better.

What I think you're more likely to see, Derek, is that peptides are even smaller that are in pills are going to cross the blood-brain barrier easier so that we could take advantage of that property of lowering the inflammation of throughout the body through the brain with small molecule peptide agonis. So that's kind of where I see things are going. But who knows, you know, 20 years from now, your forecast could come true. Well, God willing, Eric Topol, thank you very much. This is fun. Thank you. Thank you for listening. Plain English is produced by Devin Baraldi. Our holiday schedule will be a little bit different than typical.

We'll be coming at you once a week on Wednesdays. Happy holidays, and we will see you soon.