Ground Truths

“A few years ago, I might have chuckled at the naiveté of this question, but now it's not so crazy to think that we will be able to take some sort of medicine to extend our healthy lifespans in the foreseeable future.”—Coleen Murphy

Transcript with external links

Eric Topol (00:06):

Hello, this is Eric Topol from Ground Truths, and I'm just so delighted to have with me Professor Coleen Murphy, who has written this exceptional book, How We Age: The Science of Longevity. It is a phenomenal book and I'm very eager to discuss it with you, Coleen.

Coleen Murphy (00:25):

Thanks for having me on.

Eric Topol (00:27):

Oh yeah. Well, just so everyone who doesn't know Professor Murphy, she's at Princeton. She's the Richard Fisher Preceptor in Integrative Genomics, the Lewis-Sigler Institute for Integrative Genomics at Princeton, and director of the Paul Glenn Laboratories for Aging Research. Well, obviously you've been in this field for decades now, even though you're still very young. The classic paper that I can go back to would be in Nature 2003 with the DAF-16 and doubling the lifespan of C. elegans or better known as a roundworm. Would that be the first major entry you had?

Coleen Murphy (01:17):

Yeah, that was my postdoctoral work with Cynthia Kenyon.

Eric Topol (01:20):

Right, and you haven't stopped since you've been on a tear and you’ve put together a book which has a hundred pages of references in a small font. I don't know what the total number is, but it must be a thousand or something.

Coleen Murphy (01:35):

Actually, it's just under a thousand. That's right.

Eric Topol (01:37):

That's a good guess.

Coleen Murphy (01:38):

Good guess. Yeah.

Eric Topol (01:39):

So, because I too have a great interest in this area, I found just the resource that you've put together as extraordinary in terms of the science and all the work you've put together. What I was hoping to do today is to kind of take us through some of the real exciting pathways because there's a sentence in your book, which I thought was really kind of nailed it, and it actually is aligned with my sense. Obviously don't have the expertise by any means that you do here but it says, “A few years ago, I might have chuckled at the naivety of this question, but now it's not so crazy to think that we will be able to take some sort of medicine to extend our healthy lifespans in the foreseeable future.” That's a pretty strong statement for a person who's deep into the science. First I thought we'd explore healthy aging health span versus lifespan. Can you differentiate that as to your expectations?

Coleen Murphy (02:54):

So, I think most people would agree that they don't want to live necessary super long. What they really want to do is live a healthy life as long as they can. I think that a lot of people also have this fear that when we talk about extending lifespan, that we're ignoring that part. And I do want to assure everyone that the people in the researchers in the aging field are very much aware of this issue and have, especially in the past decade, I think put a real emphasis on this idea of quality of life and health span. What's reassuring is actually that many of the mechanisms that extend lifespan in all these model organisms also extend health span as well and so I don't think we're going to, they're not diametrically opposed, like we'll get to a healthier quality of life, I think in these efforts to extend lifespan as well.

Eric Topol (03:50):

Yeah, I think that's important that you're bringing that up, which is there's this overlap, like a Venn diagram where things that do help with longevity should help with health span, and we don't necessarily have to follow as you call them the immoralists, as far as living to 190 or whatever year. Now, one of the pathways that's been of course a big one for years and studied in multiple species has been caloric restriction. I wonder if you could talk to that and obviously there's now mimetics that could simulate that so you wouldn't have to go through some major dietary starvation, if you will. What are your thoughts on that pathway?

Coleen Murphy (04:41):

Yeah, actually I'm really glad you brought up mimetics because often the conversation starts and ends with you should eat less. I think that is a really hard thing for a lot of people to do. So just for the background, so dietary restriction or caloric restriction, the idea is that you would have to take in up to 30% less than your normal intake in order to start seeing results. When we've done this with laboratory animals of all kinds, this works from yeast all the way up through mice, actually primates, in fact, it does extend lifespan and in most metrics of health span the quality of life, it does improve that as well. On the other hand, I think psychologically it's really tough to not eat enough and I think that's a part that we kind of blindly ignore when we talk about this pathway.

Coleen Murphy (05:30):

And of course, if we gave any of those animals the choice of whether they want to start eating more, they would. So, it's like that's not the experiment we ever hear about. And so, the idea for studying this pathway isn't just to say, okay, this works and now we know how it works, but as you pointed out, mimetics, so can we target the molecules in the pathway so that we can help people achieve the benefits of caloric restriction without necessarily having to do the kind of awful part of restriction? I think that's really cool, and especially it might be very good for people who are undergoing certain, have certain diseases or have certain impairments that it might make it difficult ever to do dietary restrictions, so I think that's a really great thing that the field is kind of getting towards now.

Eric Topol (06:15):

And I think in fact, just today, it's every day there's something published now. Just today there was a University of Southern California study, a randomized study report comparing plant-based fasting-mimicking diet versus controlled diet, and showed that many metabolic features were improved quite substantially and projected that if you stayed on that diet, you'd gain two and a half years of healthy aging or that you would have, that's a bit of an extrapolation, but quite a bit of benefit. Now, what candidates would simulate caloric restriction? I mean, what kind of molecules would help us do that? And by the way, in the book you mentioned that the price to pay is that the brain slows down with caloric restrictions.

Coleen Murphy (07:10):

There's at least one study that shows that.

Coleen Murphy (07:13):

Yeah, so it's good to keep in mind. One of the big things that is being looked at as rapamycin, looking at that TOR pathway. So that's being explored as one of these really good mimetics. And of course, you have things that are analogs of that, so rapalogs, and so people are trying to develop drugs that mimic that, do the same kind of thing without probably some of the side effects that you might see with rapamycin. Metformin is another one, although it's interesting when you talk to people about metformin who work on it, it's argued about what is exactly the target of metformin. There's thought maybe also acts in the TOR pathway could affect complex one of mitochondria. Some of the things we know that they work, and we don't necessarily know how they work. And then of course there's new drugs all the time where people are trying to develop to other target, other molecules. So, we'll see, but I think that the idea of mimetics is actually really good, and that part of the field is moving forward pretty quickly. This diet that you did just mention, it is really encouraging that they don't have to take a drug if you don't want to. If you eat the right kind of diet, it could be very beneficial.

Eric Topol (08:20):

Yeah, no, it was interesting. I was looking at the methods in that USC paper and they sent them a box of stuff that they would eat for three cycles, multiple weeks per cycle. It was a very interesting report, we'll link to that. Before we leave the caloric restriction and these mTOR pathway, you noted in the book that there some ongoing trials like PEARL, I looked that up and they finished the trial, but they haven't reported it and it's not that large. And then there's the FAME trial with metformin. I guess we'll get a readout on these trials in the not-too-distant future. Right?

Coleen Murphy (08:57):

Yeah, that's the hope that especially with the Metformin trial, which I think is going to be really large the FAME trial, that just to give the listeners a little background, one of the efforts in the field is not just to show that something works, but also to convince the FDA that aging could be a pharmaceutical, a disease that we might want to have interventions for. And to do that, we need to figure out the right way to do it. We can't do 30-year studies of safety and things to make sure that something's good, but maybe there are reasonable biomarkers that would tell us whether people are going to live a long time. And so, if we can use some of those things or targeting age-related diseases where we can get a faster readout as well. Those are reasonable things that companies could do that would help us to really confirm or maybe rule out some of these pharmaceuticals as effective interventions. I think that would be really great for consumers to know, is this thing really going to do good or not? And we just don't have that right now in the field. We have a lot of people saying something will work and it might and the studies in the lab, but when we get to humans, we really need more clinical studies to really tell us that things are going to be effect

Transcript with audio and relevant external links, recorded on 6 Feb 2024

Eric Topol (00:05):

Hello, this is Eric Topol with Ground Truths, and I have a remarkable guest with me today, Professor Michelle Monje, who is from Stanford, a physician-scientist there and is really a leader in neuro-oncology, the big field of cancer neuroscience, neuroinflammation, and she has just been rocking it recently with major papers on these fields, no less her work that's been on a particular cancer, brain cancer in kids that we'll talk about. I just want to give you a bit of background about Michelle. She is a National Academy of Medicine member, no less actually a National Academy of Medicine awardee with the French Academy for the Richard Lounsbery Award, which is incredibly prestigious. She received a Genius grant from the MacArthur Foundation and is a Howard Hughes Medical Institute (HHMI) scholar, so she is just an amazing person who I'm meeting for the first time. Michelle, welcome.

Michelle Monje (01:16):

Thank you. So nice to join you.

Long Covid and the Brain

Eric Topol (01:18):

Well, I just am blown away by the work that you and your colleagues have been doing and it transcends many different areas that are of utmost importance. Maybe we can start with Long Covid because that's obviously such a big area. Not only have you done work on that, but you published an amazing review with Akiko Iwasaki, a friend of mine, that really went through all the features of Long Covid. Can you summarize your thoughts about that?

Michelle Monje (01:49):

Yeah, and specifically we focused on the neurobiology of Long Covid focusing on the really common syndrome of cognitive impairment so-called brain fog after Covid even after relatively mild Covid. There has been this, I think really important and exciting, really explosion of work in the last few years internationally trying to understand this in ways that I am hopeful will be beneficial to many other diseases of cognition that occur in the context of other kinds of infections and other kinds of immune challenges. But what is emerging from our work and from others is that inflammation, even if it doesn't directly initially involve the nervous system, can very profoundly affect the nervous system and the mechanisms by which that can happen are diverse. One common mechanism appears to be immune challenge induced reactivity of an innate immune cell in the nervous system called microglia. These microglia, they populate the nervous system very early in embryonic development.

(02:58):

And their job is to protect the nervous system from infection, but also to respond to other kinds of toxic and infectious and immune challenges. They also play in healthy conditions, really important roles in neurodevelopment and in neuroplasticity and so they're multifaceted cells and this is some population of those cells, particularly in the white matter in the axon tracks that are exquisitely sensitive it seems to various kinds of immune challenges. So even if there's not a direct nervous system insult, they can react and when they react, they stop doing their normal helpful jobs and can dysregulate really important interactions between other kinds of cells in the brain like neurons and support cells for those neurons like oligodendrocytes and astrocytes. One common emerging principle is that microglial reactivity triggered by even relatively mild Covid occurring in the respiratory system, not directly infecting the brain or other kinds of immune challenges can trigger this reactivity of microglia and consequently dysregulate the normal interactions between cells and the brain.

(04:13):

So important for well-tuned and optimal nervous system function. The end product of that is dysfunction and cognition and kind of a brain fog impairment, attention, memory, ability to multitask, impaired speed of information processing, but there are other ways that Covid can influence the nervous system. Of course there can be direct infection. We don't think that that happens in every case. It may not happen even commonly, but it certainly can happen. There is a clear dysregulation of the vasculature, the immune response, and the reaction to the spike protein of Covid in particular can have very important effects on the vessels in the nervous system and that can trigger a cascade of effects that can cause nervous system dysregulation and may feed directly into that reactivity of the microglia. There also can be reactivation of other infections previous, for example, herpes virus infections. EBV for example, can be reactivated and trigger a new immune challenge in the context of the immune dysregulation that Covid can induce.

(05:21):

There also can be autoimmunity. There are many, we're learning all the different ways Covid can affect the nervous system, but autoimmunity, there can be mimicry of some of the antigens that Covid presents and unfortunate autoimmunity against nervous system targets. Then finally in severe Covid where there is cardiopulmonary compromise, where there is hypoxia and multi-organ damage, there can be multifaceted effects on the nervous system in severe disease. So many different ways, and probably that is not a comprehensive list. It is certainly not a mutually exclusive list. Many of these interactions can happen at the same time in the same individual and in different combinations but we're beginning to wrap our arms around all the different ways that Covid can influence the nervous system and cause this fairly consistent syndrome of impaired attention, memory, multitasking, and executive functions.

Homology with Chemo Brain

Eric Topol (06:23):

Yeah, well there's a lot there that you just summarized and particularly you highlighted the type of glia, the microglia that appear to be potentially central at least a part of the story. You also made analogy to what you've seen with chemotherapy, chemo brain. Maybe you could elaborate on that.

Michelle Monje (06:42):

Yeah, absolutely. So I've been studying the cognitive impairment that can happen after cancer therapies including chemotherapy, but also radiation and immunotherapy. Each time we develop a new model and dig in to understand what's going on and how these cancer therapies influence the nervous system, microglia emerge as sort of the unifying principle, microglial reactivity, and the consequences of that reactivity on other cell types within the nervous system. And so, understanding that microglia and their reactive state to toxic or immune challenges was central to chemotherapy induced cognitive impairment, at least in preclinical models in the laboratory and confirm by human tissue studies. I worried at the very beginning of the pandemic that we might begin to see something that looks a lot like chemotherapy induced cognitive impairment, this syndrome that is characterized by impaired attention, memory, executive function, speed of information processing and multitasking. When just a few months into the pandemic, people began to flood neurologists’ office complaining of exactly this syndrome. I felt that we needed to study it and so that was the beginning of what has become a really wonderful collaboration with Akiko Iwasaki. I reached out to her, kind of cold called her in the midst of the deep Covid shutdown and in 2020 and said, hey, I have this idea, would you like to work with me? She's as you know, just a thought leader in Covid biology and she's been an incredibly wonderful and valuable collaborator along the way in this.

Eric Topol (08:19):

Well, the two of you pairing up is kind of, wow, that's a powerful combination, no question. Now, I guess the other thing I wanted to get at is there've been many other studies that have been looking at Long Covid, how it affects the brain. The one that's frequently cited of course is the UK Biobank where they had CT or MRI scans before in people fortunately, and then once they had Covid or didn't get Covid and it had a lot of worrisome findings including atrophy and then there are others that in terms of this niche of where immune cells can be in the meninges, in the bone marrow or the skull of the brain. Could you comment on both those issues because they've been kind of coming back to haunt us in terms of the more serious potential effects of Covid on the brain?

Michelle Monje (09:20):

Yeah, absolutely and I will say that I think all of the studies are actually quite parsimonious. They all really kind of point towards the same biology, examining it at different levels. And so that UK Biobank study was so powerful because in what other context would someone have MRI scans across the population and cognitive testing prior to the Covid pandemic and then have paired same individual tests after a range of severity of Covid infection so it was just an incredibly important data set with control individuals in the same cohort of people. This longitudinal study has continued to inform us in such important ways and that study found that there were multiple findings. One is that there appears to be a small but significant atrophy in the neocortex. Two that there are also abnormalities in major white matter tracts, and three, that there is particular pathology within the olfactory system.

(10:30):

And we know that Covid induces as a very common early symptom, this loss of smell. Then together with those structural findings on MRI scans that individuals even with relatively mild acute disease, exhibited long-term deficits in cognitive function. That fits with some beautiful epidemiological studies that have been done across many thousands of individuals in multiple different geographic populations. Underscoring this consistent finding that Covid can induce lasting cognitive changes and as we begin to understand that biology

Jim Collins is one of the leading biomedical engineers in the world. He’s been elected to all 3 National Academies (Engineering, Science, and Medicine) and is one of the founders of the field of synthetic biology. In this conversation, we reviewed the seminal discoveries that he and his colleagues are making at the Antibiotics-AI Project at MIT.

Recorded 5 February 2024, transcript below with audio links and external links to recent publications

Eric Topol (00:05):

Hello, it's Eric Topol with Ground Truths, and I have got an extraordinary guest with me today, Jim Collins, who's the Termeer Professor of Medical Engineering at MIT. He also holds appointments at the Wyss Institute and the Broad Institute. He is a biomedical engineer who's been making exceptional contributions and has been on a tear lately, especially in the work of discovery of very promising, exciting developments in antibiotics. So welcome, Jim.

Jim Collins (00:42):

Eric, thanks for having me on the podcast.

Eric Topol (00:44):

Well, this was a shock when I saw your paper in Nature in December about a new structure class of antibiotics, the one from 1962 to 2000. It took 38 years, and then there was another one that took 24 years yours, the structural antibiotics. Before I get to that though, I want to go back just a few years to the work you did published in Cell with halicin, and can you tell us about this? Because when I started to realize what you've been doing, what you've been chipping away here, this was a drug you found, halicin, as I can try to understand, it works against tuberculosis, c. difficile, enterobacter that are resistant, acinetobacter that are resistant. I mean, this is, and this is of course in mice models. Can you tell us how did you make that discovery before we get into I guess what's called the Audacious Project?

Jim Collins (01:48):

Yeah, sure. It's actually a fun story, so it is origins go broadly to institute wide event at MIT, so MIT in 2018 launched a major campus-wide effort focused on artificial intelligence. The institute, which had played a major role in the first wave of AI in the 1950s, 1960s, and a major wave in the second wave in the 1980s found itself kind of at the wheel in this third wave involving big data and deep learning and looked to correct that and to correct it the institute had a symposium and I had the opportunity to sit next to Regina Barzilay, one of our faculty here at MIT who specializes in AI and particularly AI applied to biomedicine and we really hit it off and realized we had interest in applying AI to drug discovery. My lab had focused on antibiotics to then close to 15 years, but primarily we're using machine learning and network biology to understand the mechanism of action of antibiotics and how resistance arise with the goal of boosting what we already had, with Regina we saw there was an opportunity to see if we could use deep learning to get after discovery.

(02:55):

And notably, as you kind of alluded in your introduction, there's really been a discovery void and the golden age of discovery antibiotics was in the forties, fifties and sixties before I was born and before you had the genomic revolution, the biotech revolution, AI revolution. Anyways, we got together with our two groups, and it was an unfunded project and we kind of cobbled together very small training set of 2,500 compounds that included 1,700 FDA approved drugs and 800 natural compounds. In 2018, 2019, when you started this, if you asked any AI expert should you initiate that study, they would say absolutely not, there's going to be two big data. The idea of these models are very data hungry. You need a million pictures of a dog, a million pictures of a cat to train a model to differentiate between the cat and the dog, but we ignored the naysayers and said, okay, let's see what we can do.

(03:41):

And we apply these to E. coli, so a model pathogen that's used in labs but is also underlies urinary tract infections. So it’s a look to see which of the molecules inhibited growth of the bacteria as evidence for antibacterial activity and we could have measured and we quantified each of their effects, but because we had so few compounds, we just discretized instead, if you inhibited at least 80% of the growth you were antibacterial, and if you didn't achieve that, you weren't antibacterial zero in ones. We then took the structure of each molecule and trained a deep learning model, specifically a graphical neural net that could look at those structures, bond by bond, substructure by substructure associated with whatever features you look to train with. In our case, making for good antibiotic, not for good antibiotic. We then took the train model and applied it to a drug repurposing hub as part of the Broad Institute that consists of 6,100 molecules in various stages of development as a new drug.

(04:40):

And we asked the model to identify molecules that can make for a good antibiotic but didn't look like existing antibiotics. So part of the discovery void has been linked to this rediscovery issue we have where we just keep discovering quinolones like Cipro or beta-lactams like penicillin. Well, anyways, from those criteria as well as a small tox model, only one molecule came out of that, and that was this molecule we called halicin, which was named after HAL, the killing AI computer system from 2001 Space Odyssey. In this case, we don't want it to kill humans, we want it to kill bacteria and as you alluded, it turned out to be a remarkably potent novel antibiotic that killed off multi-drug resistant extensively drugs, a pan-resistant bacteria went after to infections. It was affected against TB, it was affected against C. diff and acinetobacter baumannii and acted to a completely new mechanism of action.

(05:33):

And so we were very excited to see how AI could open up possibilities and enable one to explore chemical spaces in new and different ways. We took them model, then applied it to a very large chemical library of 1.5 billion molecules, looked at a subset of about 110 million that would be impossible for any grad student, any lab really to look at that experimentally but we looked at it in a model computer system and in three days could screen those 110 million molecules and identified several new additional candidates, one which we call salicin, which is the cousin of halicin that similes broad spectrum and acts to a novel mechanism of action.

Eric Topol (06:07):

So before we go further with this initial burst of discovery, for those who are not used to deep neural networks, I think most now are used to the convolutional neural network for images, but what you use specifically here as you alluded to, were graph neural networks that you could actually study the binding properties. Can you just elaborate a little bit more about these GNN so that people know this is one of the tools that you used?

Jim Collins (06:40):

Yeah, so in this case, the underlying structure of the model can actually represent and capture a graphical structure of a molecule or it might be of a network so that the underlying structure itself of the model will also look at things like a carbon atom connects to an oxygen atom. The oxygen atom connects to a nitrogen atom and so when you think back to the chemical structures we learned in high school, maybe we learned in college, if we took chemistry class in college, it was actually a model that can capture the chemical structure representation and begin to look at sub aspects of it, associating different properties of it. In this case, again, ours was antibacterial, but it could be toxic, whether it's toxic against a human cell and the model, the train model, the graph neural model can now look at new structures that you input them and then make calculations on those bonds so a bond would be a connection between two atoms or substructures, be multiple bonds, interconnecting multiple atoms and assign it a score. Does it make, for example, in our case, for a good antibiotic.

Eric Topol (07:48):

Right. Now, what's also striking as you set up this collaboration that's interdisciplinary with Regina, who I know of her work through breast cancer AI and not through drug discovery and so this was, I think that new effort and this discovery led to this, I love the name of it, Audacious Project, right?

Jim Collins (08:13):

Right. Yeah, so a few points on the collaboration then I'll speak to Audacious Project. In addition to Regina, we also brought in Tommi Jaakkola, another AI faculty member and marvelous colleague here at MIT and really we've benefited from having outstanding young folks who were multilingual. We had very rich, deep trained grad students from ML on Regina and Tommi's side who appreciated the biology and we had very richly, deeply trained postdocs, Jon Stokes in particular from the microbiology side on my side, who could appreciate the machine learning and so they could speak across the divide. And so, as I look out in the next few decades in this exciting time of AI coming into biomedicine, I think the groups will make a difference of those that have these multilingual young trainees and two who are well set up to also inject human intelligence with machine intelligence.

(09:04):

Brings the Audacious Project. Now, prior to our publication of halicin, I was invited by the Audacious Project to submit a proposal, the Audacious Project is a new philanthropic effort run by TED, so the group that does the TED Talks that's run by Chris Anderson, so Chris had the idea that there was a need to bring together philanthropists around the world to go for a larger scale in a collective manner toward audacious projects. I pitched them on the idea that we could use AI to address the antibiotic resistance crisis. As

“The history of science, it turns out, is filled with stories of very smart people laughing at good ideas.”—Katalin Karikó

Ground Truths podcasts are now available on Apple and Spotify!

The list of obstacles that Kati Karikó faced to become a scientist, to make any meaningful discovery, to prevail over certain scientists and administrators who oppressed her, unable to obtain grants, her seminal paper rejected by all of the top-tier journals, demoted and dismissed, but ultimately to be awarded the 2023 Nobel Prize with Drew Weissman, is a story for the ages. We covered them in this conversation, which for me will be unforgettable, and hopefully for you an inspiration.

Recorded 30 January 2023, unedited transcript to follow soon

Jonathan Howard is a neurologist and psychiatrist who practices at NYU-Bellevue and posts frequently on Science Based Medicine.

Transcript, unedited, with links to audio

Eric Topol (00:05):

Well, hello, Eric Topol with Ground Truths and I'm really pleased to have the chance to talk with Jonathan Howard today, who is a neurologist and psychiatrist at NYU at Bellevue and has written quite an amazing book published a few months months ago called We Want Them Infected, so welcome Jonathan.

Jonathan Howard (00:27):

Hey, thanks so much for having me. I really appreciate it.

Eric Topol (00:30):

Yeah, I mean, there's so much to talk about because we're still in the throes of the pandemic with this current wave at least by wastewater levels and no reason to think it isn't by infections at least the second largest in the pandemic course. I guess I want to start off first with you being into the neuropsychiatric world. How did you become, obviously caring for patients with Covid, but how did you decide to become a Covidologist?

Jonathan Howard (00:59):

Well, I developed a strong interest in the anti-vaccine movement of all things about a decade ago when a doctor who I trained with here at NYU in Bellevue morphed into one of the country's biggest anti-vaccine doctors a woman by the name of Dr. Kelly Brogan. I knew her well and we were friends; She was smart and after she left NYU in Bellevue, she became one of the country's most outspoken anti-vaccine doctors and really started leaving off the wall things that germ theory didn't exist, that HIV doesn't cause AIDS. When Covid struck, she felt that SARS-CoV-2 was not killing people because she doesn't believe any virus kills people and so I became very fascinated about how smart people can believe strange, incorrect things and I dedicated myself to learning everything that I could about the anti-vaccine movement. In 2018, I wrote a book chapter on the anti-vaccine movement with law professor Dorit Reiss.

(02:01):

And so when the pandemic came around, I was really prepared for all of their arguments, but I got two very important things wrong. I thought the anti-vaccine movement would shrink. I was wrong about that and I was also really caught off guard by the fact that a lot of mainstream physicians started to parrot pandemic anti-vaccine talking points. So all of the stuff that I'd heard about measles and the HPV vaccine, these are benign viruses, the vaccines weren't tested, blah, blah, blah. I started hearing from professors at Stanford, Harvard, UCSF, Johns Hopkins, all about Covid and the Covid vaccine.

Eric Topol (02:40):

Yeah, we're going to get to some of the leading institutions and individuals within them and how they were part of this, and surprisingly too, of course. Before we do that in the title of your book, We Want Them Infected, it seems to bring in particularly the Great Barrington Declaration that is just protect the vulnerable elderly and don't worry about the rest. Can you restate that declaration and whether that's a core part of what you were writing about?

Jonathan Howard (03:21):

Yeah, the title of the book is to be taken literally. It comes from a quote by Dr. Paul Alexander, who was an epidemiologist in the Trump administration and he said in July 4th, 2020, before anyone had been vaccinated, infants, kids, teens, young people, young adults, middle age with no conditions, et cetera, have zero to little risk so we want to use them to develop herd, we want them infected. This was formalized in the Great Barrington Declaration, which was written by three doctors, our epidemiologist, none of whom cared for Covid patients, Jay Bhattacharya at Stanford, Martin Kulldorf who at the time was at Harvard, and Sunetra Gupta who is at Oxford. If I could state their plan in the most generous terms, it would be the following that Covid is much more dangerous for certain people, but we can relatively easily identify older people and people with underlying conditions.

(04:19):

It's much more benign for a healthy 10-year-old, for example and their idea was that you could separate these two groups, the vulnerable and the not vulnerable. If the not vulnerable people were allowed to catch the virus develop natural immunity that would create herd immunity. They said that this would occur in three to six months and in that time, once herd immunity had been achieved, the vulnerable people who have been in theory sheltering at home are in otherwise safe places could reenter society. So it was really the best of both worlds because lives would be saved and schools would be open, the economy would be open. It sounded very good on paper, kind of like my idea of stopping crime by locking up all the bad guys. What could go wrong? It was a very short document. It took about maybe an hour to write.

(05:17):

I imagine there were some nefarious forces behind it. One of the main instigators of it was a man by the name of Jeffrey Tucker, who sounds like a cartoon villain and he worked at the, I forget, is the American Enterprise Research Institute. It was some right-wing think tank and he is a literally pro child labor. He wrote an article in 2016 called Let the Kids Work, which suggested that children drop out of school to work at Walmart and Chick-fil-A I'm not making that up and he's overtly pro child smoking. He feels that children, teenagers should smoke because it's cool and then they can quit in their twenties before there are any bad harms from it. Needless to say, the Great Barrington's premises that one infection led to permanent immunity didn't really work out so well, but they were very influential. They had already met with President Trump in August of 2020 and the day after their Great Barrington Declaration was signed, they were invited to the White House. This was October 5th, 2020 to meet with Secretary Human Health and Secretary Services, Alex Azar, and they are advisors to Ron DeSantis. They became mini celebrities over the course of the pandemic and it was a very pro infectious movement. As I said, the title of the book, We Want Them Infected, and they did.

Eric Topol (06:42):

Right. In fact, I debated Martin Kulldorf, one of the three principals of the Great Barrington Declaration. It was interesting because if you go back to that debate we brought out, at least I tried to highlight the many flaws in this. You've mentioned at least one major flaw, which was to this virus. There's not a long-term immunity built by infections. It's just, as we say with vaccines the immunity for neutralizing antibody production and protection from infections and severe Covid is limited duration for four to six months, and at least for the antibodies and maybe the T-cell immunity is longer, but that doesn't necessarily kick in and quickly. So that was one major flaw, but there are many others, so maybe you could just take that apart further. For example, I like your analogy to lock up all the bad guys, but compartmentalizing people is not so easy in life and I think this is a significant concern of the idea that is, while you indicated there may be some merits if things went as planned, but what else was a flaw of that argument or proposition?

Jonathan Howard (08:11):

So yeah, this could be a 10-hour conversation and I think importantly, we don't have to speak hypothetically here. A lot of defenders of the Great Barrington Declaration will say, oh, we never tried it, but they promised that herd immunity would arrive in three to six months after lockdowns ended. So we don't have to speak theoretically about what would've happened had we done it. Lockdowns ended a while ago and we don't have herd immunity. They were very clear on this. Dr. Kulldorf tweeted in December 2020 that if we use focus protection, the pandemic will be over in three to six months. So, what could have gone wrong if about 250 million unvaccinated Americans contracted Covid simultaneously in October and November of 2020? A lot of things, as we said, they dichotomized people into vulnerable and not vulnerable, but of course it exists on this. The only bad outcome they recognized was death.

(09:11):

They felt that either you died or you had the sniffle for a few days and you emerged unscathed. Separating vulnerable people from not vulnerable people is a lot easier than it sounds and I think by way of comparison, look at the mRNA vaccine trials. You can read their protocols and the protocols for these trials were 300-400 pages of dense policies and procedures. The Great Barrington Declaration, if you go to their frequently asked questions section, they made some suggestions, which sound great, like older people should have food delivered at home during times of high transmission, but setting up a national or even statewide food delivery program, that's a lot harder than it sounds. When asked about that later, Dr. Bhattacharya has said they could have used DoorDash, for example. So it was just very clear that no serious thought went into this because it was really an unactionable thing.

(10:21):

It's not as if public health officials had billions of dollars at their disposal and they weren't many dictators. They couldn't set up home food delivery programs overnight like they suggested and two months after the Great Barrington Declaration was published, vaccines became available so it became obsolete. Not that vaccines have turned out to be the perfect panacea that we had hoped for, unfortunately, but the idea that young people should continue to try to get natural low immunity in favor instead of vaccination became at that point obscene, but they still are anti-vaccine for young people and for children, which I find despicable at this point.

Eric Topol (11:07):

Right, the data is unequivocal that there's benefit acro

is an award-wining entrepreneur and innovator in technology, especially A.I., a member of the editorial board of Harvard Business Review, and an outstanding communicator which makes him a frequent media guest and often featured in The Economist, WSJ, and Financial Times. is chock full of interesting analyses and podcasts on tech and A.I.

Here’s his summary of our extended and fun discussion

I hope you find our conversation interesting and informative.

A snippet of our conversation below

Transcript of our conversation 8 January 2023, edited for accuracy, with external links

Eric Topol

It’s a pleasure for me to have Liv Boeree as our Ground Truths podcast guest today. I met her at the TED meeting in October dedicated to AI. I think she's one of the most interesting people I’ve met in years and the first time I've ever interviewed a professional poker player who has won world championships and we're going to go through that whole story, so welcome Liv.

Liv Boeree

Thanks for having me, Eric.

Eric Topol

You have an amazing background having been at the University of Manchester in physics and astrophysics. Back around in 2005 you landed into the poker world. Maybe you could help us understand how you went from physics to poker.

From Physics to Poker

Liv Boeree

Ah, yeah. It's a strange story, I graduated as you said in 2005 and I had student debt and needed to get a job I had plans to continue in academia. I wanted to do a masters and then a PhD to work in astrophysics in some way, but I needed to make some money, so I started applying for TV game shows and it was on one of these game shows that I first learned how to play poker. They were looking for beginners and the loose premise of the show was which personality type is best suited for learning the game and even though I didn't win that particular show we were playing for a winner take all prize of £100,000 which was a life changing amount of money had I won it at the time. It was like a light bulb moment just the game and I’ve always been a very competitive person, but poker in particular really spoke to my soul. I always wanted to play in games where it was often considered a boy’s game and I could be a girl beating the boys at their own game. I hadn't played that much cards in particular, but I just loved any game that was very cutthroat which poker certainly is. From that point onwards I was like you know what I'm going to put physics on hold and see if I can make it in this poker world instead and then never really looked back.

Eric Topol

Well, you sure made it in that world. I know you retired back in about 2019, but that was after you won all sorts of world and European championships and beat a lot of men. No less. What were some of the things that that made you such a phenomenal player?

Liv Boeree

The main thing with poker is well the most important ingredient if you really want to make it as a professional is you have to be extremely competitive. I have not met any top pros who don't have that degree of killer instinct when it comes to the game that doesn't mean it means you're competitive in everything else in life, but you have to have a passion for looking someone in the eye, mentally modeling them, thinking how to outwit them and put them into difficult situations within the game and then take pleasure in that. So, there’s a certain personality type that tends to enjoy that. The other key facet is you have to be comfortable with thinking in terms of probability. The cards are shuffled between every hand so there's this inherent degree of randomness. On the scale of pure roulette which is all luck no skill to a game like chess which has almost no luck (close to 100% skill as you can get) poker lies somewhere in the middle and of course the more you play the bigger the skill edge and the smaller the luck factor. That's why professionals can exist. It's a game of both luck and skill which I think is what makes it so interesting because that's what life is really, right? We're trying to get our business off the ground, we're trying to compete in the dating market. Whatever it is. We're doing our strategy, the role of luck life can throw your curved balls that you can do everything right and still things don't go the way you intended them to or vice versa, but there's also strategies we can employ to improve our chances of success. Those are the sort of skills that poker players particularly this idea of gray scale probabilistic thinking that you really have to hone. I've always wondered whether having a background in science or at least you know studying having ah a scientific degree helped in that regard because of course the scientific method is about understanding variables and minimizing uncertainty as much as possible and understanding what cofounding factors can bias the outcome of your results. Again, that's always going on in a poker player's mind, you'll have concurrent hypotheses. Oh, this guy just made a huge bet into me when that ace came out, is it because he actually has an ace or is it because he's pretending to have an ace and so you've got to weigh all the bits of information up as unbiased as possible in an unbiased way as possible to come to a correct conclusion. Even then you can never be certain, so this idea of understanding biases understanding probabilities I think that’s why a lot of top poker players have backgrounds in scientific degrees a very good friend of mine he had a PhD in in physics. Especially over time poker has become a much more sort of scientific pursuit. When I first allowed to play it was very much a game of street smarts and intuition in part because we didn't have the technological tools to understand really the mechanics of the game as well. You couldn't record all your playing data if you were playing just in a casino unless you were writing down your hands. Otherwise, this information wasn't getting stored anywhere, but then online poker came along which meant that you could store all this data on your laptop and then build tools to analyze that data and so the game became a much more technical scientific pursuit.

Eric Topol

That actually gets to kind of the human side of poker. Not the online version —especially since we're going to be mainly talking about AI the term “poker face” the ability to bluff is that a big part of this?

Liv Boeree

Oh, absolutely. You can't be a good poker player if you don't ever bluff because your opponents will start to notice that so that means you're only ever putting your money on the line when you have a good hand so why would they ever pay you off. The point of poker is to maximize the deception to your opponents so you have to use strategies where some of the time you might be having a strong hand and some of the time you might be bluffing where you might have a weak hand. The key is this is getting into the technical sort of game theory side of it, but you want to be doing these bluffs versus what we call value bets as in betting with a good hand with the right to sort of frequency. You need these right ratios between them, so bluffing is a very core part of the game and yes having a poker face obviously helps because you want to be as inscrutable to your opponents as possible. At the same time online poker is an enormously popular game where you can't see your opponent's faces.

Eric Topol

Right, right.

Liv Boeree

Yet you can still bluff which could actually lead us into this topic of AI because now the best players in the world are actually AIs.

Eric Topol

Well, it's interesting because it takes out that human component of being able to bluff and it may be good for people who don't have a poker face. They can play online poker and be good at it because they don't have that disguise if you will.

Liv Boeree

Right.

Game Theory and Moloch Traps

Eric Topol

That gets me to game theory and a big part of the talk you gave at the TED conference about something that I think a lot of the folks listening aren't familiar with— Moloch traps. Could you enlighten us about that because what the talk which of course we’ll link to is so illuminating and apropos to the AI landscape that we face today?

Liv Boeree

Yeah, I’ll leave it for people to go and watch the TED talk because that's going to be much more succinct than me to explain the backstory of how it came to be called a Moloch trap because Moloch is a sort of biblical figure a demon and it seems strange that you would be applying such a concept to what's basically a collection of game theoretic incentives, but essentially what a Moloch trap is the more formal name for it is a multipolar trap which some of the listeners may be familiar with. Essentially a Moloch trap or a multipolar trap is one of those situations where you have a lot of competing different people all competing for 1 particular thing that say who can collect the most fish out of a lake. The trap occurs when everyone is incentivized to get as much of that thing as possible so to go for a specific objective, but if everyone ends up doing it then the overall environment ends up being worse off than before. What we're seeing with plastic pollution – It’s not like packaging companies want to fill the oceans with plastic. They don't want this outcome. It doesn't make them look good. They're all caught on the trap of needing to maximize profits and external and one of the most efficient ways of doing that is to externalize costs outside of their P&L by using cheap packaging that perhaps ends up in the lakes or the oceans and if everyone ends up doing this but well basically you're a CEO in a decision of I could do the more expensive selfless action, but if I don't do that then I know that my competitors are going to do the selfish thing. I might as well do it anyway because the world's going to end up in roughly the same outcome whether I do it or not because everyone ends up adopting this mindset they end up being trapped in

The science to advance our understanding of the aging process—and to potentially slow it down—has made important strides. One of the leading scientists responsible for this work is Professor Tony Wyss-Coray, whose work has particularly focused on brain aging but has implications for all organs. I believe his December 2023 Nature paper on blood proteins that can track aging for 11 of our organs is one of the most important aging reports yet.

Here is the audio and transcript of our conversation, recorded 20 December 2023, with a few relevant external links.

This is the last Ground Truths post for 2023 and I hope you’ll find it informative. I look forward to sharing many more exciting, cutting-edge biomedical advances with you in 2024!

00:10.38

Eric Topol

Hello this is Eric Topol and for this edition of Ground Truths. I'm so delighted to have with me Professor Tony Wyss-Coray of Stanford, a Distinguished Professor at Stanford and who directs the Knight Initiative for Brain Resilience. So welcome Tony.

00:30.19

Tony Wyss-Coray

Thank you, thank you for having me, Eric.

00:32.84

Eric Topol

Well, I've been following your career and your work for decades I have to say and what you just published a couple weeks ago in Nature. The cover paper about internal organ clocks. It blew me away. I mean it's a built on a foundation of extraordinary work. I thought we could start with that because to me that's really a breakthrough in that when we think of aging and how to gauge a person aging with things like the Horvath clock of methylation markers or telomeres or —not at all specific to any part of the body, just overall, l but you published an extraordinary work about plasma proteins for 11 organs that predicted the outcomes things like heart failure and Alzheimer's so maybe you could tell us about this. Seems to be a big deal to me.

01:28.41

Tony Wyss-Coray

Thank you so much I'm honored. Really, you know I think if you work on this stuff, especially for several years it feels sort of obvious to do it? But I think you know it is in a way. It is. Pretty simple. So what we argued is that the thousands of proteins that you know are present in our blood. They must originate from somewhere now a lot of proteins are you know, produced by cells throughout the body. But some proteins are very specifically produced. For example, only in the brain or only in the liver or only in the heart because they have specialized functions and we have you know being taking advantage of that in clinical medicine where you measure. Often you know one of these proteins to sort of diagnose pathology in a tissue, but we took this It's just a level further and said, well, let's just find out of thousands of proteins that we can measure assign them to specific organs and tissues. And then see whether they change with age and many of them turn out to change. We found you know about 1500 proteins or so in the study that we did although that number can grow dramatically if we you know keep.

03:01.11

Tony Wyss-Coray

Improving our technologies or techniques to measure them and many of them come from the brain or from other tissues and because they change with age. They tell us something about the aging of that organ. And as others have shown in the field including Steve Horvath is that that prediction of the age if it doesn't really match exactly your actual age contains information about the state the physiological state or the risk to develop. Organ-specific disease.

03:37.75

Eric Topol

Right. And you found that about 1 in 5 people had evidence of accelerated aging of 1 organ which of course is really starting to nail down ability to detect aging you know to localize it and um. What strikes me Tony is that now because we're seeing at the cusp of advancing in the science of aging a field that you have done so much to propel forward and one of the issues has been well, how are we going to prove it. We can't wait for 20 years to show that. Whatever intervention led to promotion of healthy aging. But when you have a marker like this of organ specificity, it seems like the chances of being able to show that intervention makes a difference is enhanced would you say so?

04:29.28

Tony Wyss-Coray

Yeah, absolutely I think that's one of the most exciting aspects of this that we can now start looking at interventions whether they are you know a specific intervention that tries to target the aging process, or you know just that. Let's say a cholesterol lowering drug or blood pressure lowering drug does that have a beneficial effect on the heart. For example, on the kidney or you can also start thinking of lifestyle interventions where they actually have an effect right? If you started exercising you collect your blood before and then a year after you have an exercise regimen does that actually change the age that we can measure with these different clocks.

05:22.55

Eric Topol

Right? Well I mean it's really a striking advance and by a marker of aging so that gets me to your other work. You've done well over 10 years which is that you could identify that given young blood. First of course in mice and then later verified in people could improve cognitive function in older whether it's experimental models or in people. So what are your thoughts about that is that if that's something you've been ruminating on for many years and I’m sure there are places around the world that are trying to do this sort of thing. What do you think of that potential?

06:11.40

Tony Wyss-Coray

Yeah, so there really this recent observation or study really came out of you know that finding that young blood can change the age of different organs and you know we. We were not the first to show this. We showed it for the brain but Tom Rando who studied muscle stem cell aging showed this you know a few years earlier in the muscle and we worked with Tom to explore this for the brain, but it shows sort of that this you know the composition of the blood. It is really not just reflecting the age of organs and tissues. But it actually also affects them. It directs them in a way and so you can speculate that you know if you had an organ that shows accelerated aging. Because some of the factors end up in the blood. They might actually induce aging in other tissues and so promote the aging process and people in the field have also shown that this is true for specific cells. We call them senescence cells. So these are a specific type of cell that seem to somehow stop dividing and assume the state that releases inflammatory factors these cells too. They seem to almost infect the neighborhood where they live in with an age promoting sort of.

07:41.95

Tony Wyss-Coray

The secretome , as we call it, so they release factors that seem to promote aging locally but potentially across the organism and interfering in that could potentially have rejuvenating effects and so that brings us back to this observation that.

08:01.23

Tony Wyss-Coray

Young blood could potentially rejuvenate organs We know old blood can accelerate it at least in mice. So could we neutralize the age promoting factors in people and could we deliver sort of the rejuvenating factors. Now what's been frustrating for me is that it has been incredibly challenging to identify the key factors.

08:33.30

Tony Wyss-Coray

I think we became to realize as a field that there is not 1 factor. There's not 1 magic factor that will keep us young or keep our organs young but rather different cells and different organs in our body seem to respond in different ways actually to this young blood. Can show this with molecular tools. We can show that every cell actually responds. So if you take a mouse an old mouse and you give it young blood every cell in that mouse shows a transcription of the response to the young blood.

09:10.80

Tony Wyss-Coray

Some of them may regenerate mitochondria and others activate other pathways. We see that stem cells respond particularly well the stem cells of the Immune system hematopoietic stem cells um while other cells show less of a response. And that to me suggests that they respond to different factors in the young blood and that you know they have very specific um receptors Probably that recognize some of these beneficial factors and then respond in a specific way. So that’s what we need to.

09:33.16

Eric Topol

Right.

09:48.63

Tony Wyss-Coray

Figure out I think as a field to translate this really to the clinic is what are the key factors and will it be possible to make a cocktail that sort of mimics Nature's you know elixir

10:06.13

Tony Wyss-Coray

I Said this before it's almost like the fountain of youth is within us, but it just dries out as we get older and if we could figure out what are the key factors that that make up this fountain. We could potentially you know either, as a treatment, deliver it again or reactivate that found and so that the body produces these factors again.

10:34.73

Eric Topol

Well, you know that's something that years ago I was very skeptical about and because of your work and others in the field. I've come a long way thinking that we're on the cusp of really identifying ways to truly promote healthy aging. And so this is a really you know extraordinary time in our lives I wonder you of course mentioned 2 critical paths that have been identified the senescent cells—removing them— or the infusion of young plasma. Would you say it's too simplistic to reduce this to decreasing inflammation or is that really the theme here, or is it much more involved than that.

11:28.48

Tony Wyss-Coray

I think inflammation ha

David Liu is an gifted molecular biologist and chemist who has pioneered major refinements in how we are and will be doing genome editing in the future, validating the methods in multiple experimental models, and establishing multiple companies to accelerate their progress.

The interview that follows here highlights why those refinements beyond the CRISPR Cas9 nuclease (used for sickle cell disease) are vital, how we can achieve better delivery of editing packages into cells, ethical dilemmas, and a future of somatic (body) cell genome editing that is in some ways is up to our imagination, because of its breadth, over the many years ahead.

Recorded 29 November 2023 (knowing the FDA approval for sickle cell disease was imminent)

Annotated with figures, external links to promote understanding, highlights in bold or italics, along with audio links (underlined)

Eric Topol (00:11):

Hello, this is Eric Topol with Ground Truths and I'm so thrilled to have David Liu with me today from the Broad Institute, Harvard, and an HHMI Investigator. David was here visiting at Scripps Research in the spring, gave an incredible talk which I'll put a link to. We're not going to try to go over all that stuff today, but what a time to be able to get to talk with you about what's happening, David. So welcome.

David Liu (00:36):

Thank you, and I'm honored to be here.

Eric Topol (00:39):

Well, the recent UK approval (November 16, 2023) of the first genome editing after all the years that you put into this, along with many other colleagues around the world, is pretty extraordinary. Maybe you can just give us a sense of that threshold that's crossed with the sickle cell and beta thalassemia also imminently [FDA approval granted for sickle-cell on 8 December 2023] likely to be getting that same approval here in the U.S.

David Liu (01:05):

Right? I mean, it is a huge moment for the field, for science, for medicine. And just to be clear and to give credit where credit is due, I had nothing to do with the discovery or development of CRISPR Cas9 as a therapeutic, which is what this initial gene editing CRISPR drug is. But of course, the field has built on the work of many scientists with respect to CRISPR Cas9, including Emmanuel Charpentier and Jennifer Doudna and George Church and Feng Zhang and many, many others. But it is, I think surprisingly rapid milestone in a long decade’s old effort to begin to take some control over our genetic features by changing DNA sequences of our choosing into sequences that we believe will offer some therapeutic benefit. So this initial drug is the CRISPR Therapeutics /Vertex drug.

Now we can say it's actually a drug approved drug, which is a Crispr Cas9 nuclease programmed to cut a DNA sequence that is involved in silencing fetal hemoglobin genes. And as you know, when you cut DNA, you primarily disrupt the sequence that you cut. And so if you disrupt the DNA sequence that is required for silencing your backup fetal hemoglobin genes, then they can reawaken and serve as a way to compensate for adult hemoglobin genes like the defective sickle cell alleles that sickle cell anemia patients have. And so that's the scientific basis of this initial drug.

Eric Topol (03:12):

So as you aptly put— frame this—this is an outgrowth of about a decade's work and it was using a somewhat constrained, rudimentary form of editing. And your work has taken this field considerably further with base and prime editing whereby you're not just making a double strand cut, you're doing nicks, and maybe you can help us understand this next phase where you have more ways you can intervene in the genome than was possible through the original Cas9 nucleases.

David Liu (03:53):

Right? So gene editing is actually a several decades old field. It just didn't quite become as popular as it is now until the discovery of CRISPR nucleases, which are just much easier to reprogram than the previous programmable zinc finger or tail nucleases, for example. So the first class of gene editing agents are all nuclease enzymes, meaning enzymes that take a piece of DNA chromosome and literally cut it breaking the DNA double helix and cutting the chromosome into two pieces. So when the cell sees that double strand DNA break, it responds by trying to get the broken ends of the chromosome back together. And we think that most of the time, maybe 90% of the time that end joining is perfect, it just regenerates the starting sequence. But if it regenerates the starting sequence perfectly and the nuclease is still around, then it can just cut the rejoin sequence again.

(04:56):

So this cycle of cutting and rejoining and cutting and rejoining continues over and over until the rejoining makes the mistake that changes the DNA sequence at the cut site because when those mistakes accumulate to a point that the nuclease no longer recognizes the altered sequence, then it's a dead end product. That's how you end up with these disrupted genes that result from cutting a target DNA sequence with a nuclease like Crispr Cas9. So Crispr Cas9 and other nucleases are very useful for disrupting genes, but one of their biggest downsides is in the cells that are most relevant to medicine, to human therapy like the cells that are in your body right now, you can't really control the sequence of DNA that comes out of this process when you cut a DNA double helix inside of a human cell and allow this cutting and rejoining process to take place over and over again until you get these mistakes.

(06:03):

Those mistakes are generally mixtures of insertions and deletions that we can't control. They are usually disruptive to a gene. So that can be very useful when you're trying to disrupt the function of a gene like the genes that are involved in silencing fetal hemoglobin. But if you want to precisely fix a mutation that causes a genetic disease and convert it, for example, back into a healthy DNA sequence, that's very hard to do in a patient using DNA cutting scissors because the scissors themselves of course don't include any information that allows you to control what sequence comes out of that repair process. You can add a DNA template to this cutting process in a process called HDR or Homology Directed Repair (figure below from the Wang and Doudna 10-year Science review), and sometimes that template will end up replacing the DNA sequence around the cut site. But unfortunately, we now know that that HDR process is very inefficient in most of the types of cells that are relevant for human therapy.

(07:12):

And that explains why if you look at the 50 plus nuclease gene editing clinical trials that are underway or have taken place, all but one use nucleases for gene disruption rather than for gene correction. And so that's really what inspired us to develop base editing in 2016 and then prime editing in 2019. These are methods that allow you to change a DNA sequence of your choosing into a different sequence of your choosing, where you get to specify the sequence that comes out of the editing process. And that means you can, for the first time in a general way, programmable change a DNA sequence, a mutation that causes a genetic disease, for example, into a healthy sequence back into the normal, the so-called wild type sequence, for example. So base editors work by actually performing chemistry on an individual DNA base, rearranging the atoms of that base to become a different base.

(08:22):

So base editors can efficiently and robustly change A's into G's G's, into A's T's into C's or C's into T's. Those four changes. And those four changes for interesting biochemical reasons turn out to be four of the most common ways that our DNA mutates to cause disease. So base editors can be used and have been used in animals and now in six clinical trials to treat a wide variety of diseases, high cholesterol and sickle cell disease, and T-cell leukemia for example. And then in prime editors we developed a few years later to try to address the types of changes in our genomes that caused genetic disease that can't be fixed with a base editor, for example. You can't use a base editor to efficiently and selectively change an A into a T. You can't use a base editor to perform an insertion of missing DNA letters like the three missing letters, CTT, that's the most common cause of cystic fibrosis accounting for maybe 70% of cystic fibrosis patients.

(09:42):

You can't use a base editor to insert missing DNA letters like the missing TATC. That is the most common cause of Tay-Sachs disease. So we develop prime editors as a third gene editing technology to complement nucleases and base editors. And prime editors work by yet another mechanism. They don't, again, they don't cut the DNA double helix, at least they don't cause that as the required mechanism of editing. They don't perform chemistry on an individual base. Instead, prime editors take a target DNA sequence and then write a new DNA sequence onto the end of one of the DNA strands and then sort of help the cell navigate the DNA repair processes to have that newly written DNA sequence replace the original DNA sequence. And in the process it's sort of true search and replace gene editing. So you can basically take any DNA sequence of up to now hundreds of base pairs and replace it with any other sequence of your choosing of up to hundreds of base pairs. And if you integrate prime editing with other enzymes like recombinase, you can actually perform whole gene integration of five or 10,000 base pairs, for example, this way. So prime editing's hallmark is really its versatility. And even though it's the newest of the three ways that have been robu

This is one of the most enthralling and fun interviews I’ve ever done (in 2 decades of doing them) and I hope that you’ll find it stimulating and provocative. If you did, please share with your network.

And thanks for listening, reading, and subscribing to Ground Truths.

Recorded 4 December 2023

Transcript below with external links to relevant material along with links to the audio

ERIC TOPOL (00:00):

This is for me a real delight to have the chance to have a conversation with Geoffrey Hinton. I followed his work for years, but this is the first time we've actually had a chance to meet. And so this is for me, one of the real highlights of our Ground Truths podcast. So welcome Geoff.

GEOFFREY HINTON (00:21):

Thank you very much. It's a real opportunity for me too. You're an expert in one area. I'm an expert in another and it's great to meet up.

ERIC TOPOL (00:29):

Well, this is a real point of conversion if there ever was one. And I guess maybe I'd start off with, you've been in the news a lot lately, of course, but what piqued my interest to connect with you was your interview on 60 Minutes with Scott Pelley. You said: “An obvious area where there's huge benefits is healthcare. AI is already comparable with radiologists understanding what's going on in medical images. It's going to be very good at designing drugs. It already is designing drugs. So that's an area where it's almost entirely going to do good. I like that area.”

I love that quote Geoff, and I thought maybe we could start with that.

GEOFFREY HINTON (01:14):

Yeah. Back in 2012, one of my graduate students called George Dahl who did speech recognition in 2009, made a big difference there. Entered a competition by Merck Frost to predict how well particular chemicals would bind to something. He knew nothing about the science of it. All he had was a few thousand descriptors of each of these chemicals and 15 targets that things might bind to. And he used the same network as we used for speech recognition. So he treated the 2000 descriptors of chemicals as if they were things in a spectrogram for speech. And he won the competition. And after he'd won the competition, he wasn't allowed to collect the $20,000 prize until he told Merck how he did it. And one of their questions was, what qsar did you use? So, he said, what's qsar? Now qsar is a field, it has a journal, it's had a conference, it's been going for many years, and it's the field of quantitative structural activity relationships. And that's the field that tries to predict whether some chemical is going to bind to something. And basically he'd wiped out that field without knowing its name.

ERIC TOPOL (02:46):

Well, it's striking how healthcare, medicine, life science has had somewhat of a separate path in recent AI with transformer models and also going back of course to the phenomenal work you did with the era of bringing in deep learning and deep neural networks. But I guess what I thought I'd start with here with that healthcare may have a special edge versus its use in other areas because, of course, there's concerns which you and others have raised regarding safety, the potential, not just hallucinations and confabulation of course a better term or the negative consequences of where AI is headed. But would you say that the medical life science AlphaFold2 is another example of from your colleagues Demis Hassabis and others at Google DeepMind where this is something that has a much more optimistic look?

GEOFFREY HINTON (04:00):

Absolutely. I mean, I always pivot to medicine as an example of all the good it can do because almost everything it's going to do there is going to be good. There are some bad uses like trying to figure out who to not insure, but they're relatively limited almost certainly it's going to be extremely helpful. We're going to have a family doctor who's seen a hundred million patients and they're going to be a much better family doctor.

ERIC TOPOL (04:27):

Well, that's really an important note. And that gets us to a paper preprint that was just published yesterday, on arXiv, which interestingly isn't usually the one that publishes a lot of medical preprints, but it was done by folks at Google who later informed me was a model large language model that hadn't yet been publicized. They wouldn't disclose the name and it wasn't MedPaLM2. But nonetheless, it was a very unique study because it randomized their LLM in 20 internists with about nine years of experience in medical practice for answering over 300 clinical pathologic conferences of the New England Journal. These are the case reports where the master clinician is brought in to try to come up with a differential diagnosis. And the striking thing on that report, which is perhaps the best yet about medical diagnoses, and it gets back Geoff to your hundred million visits, is that the LLM exceeded the clinicians in this randomized study for coming up with a differential diagnosis. I wonder what your thoughts are on this.

GEOFFREY HINTON (05:59):

So in 2016, I made a daring and incorrect prediction was that within five years, the neural nets were going to be better than radiologists that interpreting medical scans, it was sometimes taken out of context. I meant it for interpreting medical scans, not for doing everything a radiologist does, and I was wrong about that. But at the present time, they're comparable. This is like seven years later. They're comparable with radiologists for many different kinds of medical scans. And I believe that in 10 years they'll be routinely used to give a second opinion and maybe in 15 years they'll be so good at giving second opinions that the doctor's opinion will be the second one. And so I think I was off by about a factor of three, but I'm still convinced I was completely right in the long term.

(06:55):

So this paper that you're referring to, there are actually two people from the Toronto Google Lab as authors of that paper. And like you say, it was based on the large language PaLM2 model that was then fine-tuned. It was fine-tuned slightly differently from MedPaLM2 I believe, but the LLM [large language model] by themselves seemed to be better than the internists. But what was more interesting was the LLMs when used by the internists made the internists much better. If I remember right, they were like 15% better when they used the LLMs and only 8% better when they used Google search and the medical literature. So certainly the case that as a second opinion, they're really already extremely useful.

ERIC TOPOL (07:48):

It gets again, to your point about that corpus of knowledge that is incorporated in the LLM is providing a differential diagnosis that might not come to the mind of the physician. And this is of course the edge of having ingested so much and being able to play back those possibilities and the differential diagnosis. If it isn't in your list, it's certainly not going to be your final diagnosis. I do want to get back to the radiologist because we're talking just after the annual massive Chicago Radiologic Society of North America RSNA meeting. And at those meetings, I wasn't there, but talking to my radiology colleagues, they say that your projection is already happening. Now that is the ability to not just read, make the report. I mean the whole works. So it may not have been five years when you said that, which is one of the most frequent quotes in all of AI and medicine of course, as you probably know, but it's approximating your prognosis. Even now

GEOFFREY HINTON (09:02):

I've learned one thing about medicine, which is just like other academics, doctors have egos and saying this stuff is going to replace them is not the right move. The right move is to say it's going to be very good at giving second opinions, but the doctor's still going to be in charge. And that's clearly the way to sell things. And that's fine, just I actually believe that after a while of that, you'll be listening to the AI system, not the doctors. And of course there's dangers in that. So we've seen the dangers in face recognition where if you train on a database that contains very few black people, you'll get something that's very good at recognizing faces. And the people who use it, the police will think this is good at recognizing faces. And when it gives you the wrong identity for a person of color, then the policemen are going to believe it. And that's a disaster. And we might get the same with medicine. If there's some small minority group that has some distinctly different probabilities of different diseases, it's quite dangerous for doctors to get to trust these things if they haven't been very carefully controlled for the training data.

ERIC TOPOL (10:17):

Right. And actually I did want to get back to you. Is it possible for the reason why in this new report that the LLMs did so well is that some of these case studies from New England Journal were part of the pre-training?

GEOFFREY HINTON (10:32):

That is always a big worry. It's worried me a lot and it's worried other people a lot because these things have pulled in so much data. There is now a way round that at least for showing that the LLMs are genuinely creative. So he's a very good computer science theorist at Princeton called Sanjeev Arora, and I'm going to attribute all this to him, but of course, all the work was done by his students and postdocs and collaborators. And the idea is you can get these language models to generate stuff, but you can then put constraints on what they generate by saying, so I tried an example recently, I took two Toronto newspapers and said, compare these two newspapers using three or four sentences, and in your answer demonstrate sarcasm, a red herring em