The Peter Attia Drive - #112 - Ned David, Ph.D.： How cellular senescence influences aging, and what we can do about it

00:00:00.000 Hey everyone, welcome to the drive podcast. I'm your host, Peter Atiyah. This podcast,

00:00:15.480 my website and my weekly newsletter all focus on the goal of translating the science of longevity

00:00:19.800 into something accessible for everyone. Our goal is to provide the best content in health

00:00:24.600 and wellness full stop. And we've assembled a great team of analysts to make this happen.

00:00:28.880 If you enjoy this podcast, we've created a membership program that brings you far more

00:00:33.280 in-depth content. If you want to take your knowledge of the space to the next level at

00:00:37.320 the end of this episode, I'll explain what those benefits are. Or if you want to learn more now,

00:00:41.720 head over to peteratiyahmd.com forward slash subscribe. Now, without further delay,

00:00:47.740 here's today's episode. My guest this week is Ned David. Ned is the president of unity biotechnology,

00:00:55.780 a company that he co-founded in 2011. In this episode, we're going to talk a little bit about

00:01:00.740 unity, but more importantly, we talk about the science upon which unity is developed. And that

00:01:06.880 is the science of senescence or cellular senescence. We begin this episode by putting that into context.

00:01:13.660 I think if you're interested in the space of longevity, you're listening to this podcast,

00:01:17.240 you've undoubtedly heard the term senescence, but it might not mean much to you. And I think what Ned

00:01:22.560 does a great job of here is explaining not just senescence, but where it fits in the overall

00:01:28.640 lattice, if you will, of all of the different cellular processes that occur during aging.

00:01:34.400 So where does it fit in with the role of mTOR inhibition, caloric restriction, mitochondrial

00:01:40.320 dysfunction, stem cell exhaustion, methylation of DNA and epigenetic change, all of these sorts of

00:01:46.940 things. We also talk quite a bit about Ned's background as a serial entrepreneur in biotechnology.

00:01:52.860 Ned is at least partly responsible for the development of a number of FDA approved compounds.

00:01:58.420 And he talks openly about what he learned, good, bad, and indifferent along the way and how that

00:02:03.440 helped him at the beginning of this process of creating this company, unity, define sort of the

00:02:09.360 jugular questions that needed to be addressed as sort of a proof of principle. In fact, I found that part

00:02:14.080 of the discussion most interesting, and I think you'll, you'll know what I'm talking about when we

00:02:17.280 get to that part of the episode, because I found that exercise that he and his co-founders went

00:02:22.260 through between about 2011, 2015, to be some of the most interesting thinking around this idea of

00:02:29.360 company creation. Ned is a prolific innovator. I think when he was about 30, he was named to the top

00:02:34.980 100 innovators in the world by MIT technology reviews. He holds a PhD from Cal Berkeley, and he got his

00:02:42.040 undergrad degree in molecular biology at Harvard. So without getting into too much more of the

00:02:45.900 details of this, please enjoy my discussion with Ned David.

00:02:55.080 Ned, awesome to see you today. Thank you for coming by.

00:02:58.240 Oh, thank you for having me.

00:02:59.540 I feel like I haven't seen you in person in way too long.

00:03:02.620 Is this true? We tried to get you to come to dinner.

00:03:04.840 Oh God, I feel horrible about that. Yes, sorry.

00:03:07.640 Yeah, you were distracted. It's all good.

00:03:09.400 No, you know what I was doing? You were doing a podcast.

00:03:12.120 I was doing a podcast. Yeah.

00:03:13.340 That I thought would take two and a half hours. It ended up taking six hours.

00:03:18.140 That must have been an intense experience.

00:03:21.020 It was a very intense experience, and I felt horrible after. I'm sorry. So I owe you an apology

00:03:25.640 for that. It's okay. We love you.

00:03:28.400 There's so much I want to talk with about you. In fact, I've been bugging you for about a year to come

00:03:32.620 on the podcast because what I was hoping we could do, and we can never do it when there's a microphone in

00:03:37.640 front of us, but was sort of reproduce a dinner conversation that we would have ordinarily,

00:03:42.240 which would be pick some topic and we go way off on a tangent about it. And it could be energy.

00:03:48.460 It could be biotech. It could be whatever. But some of the most interesting discussions we've had

00:03:53.860 over the past five years have been around longevity, something that we both share an interest in.

00:03:59.300 And God, it's been probably nine years since you talked to me about what became that which you're

00:04:09.100 doing today. Is that about right? Yeah, that's about right. It was 2011 towards the end. Yeah.

00:04:14.980 Actually, we're in December now, so it's actually been somewhat more than eight years.

00:04:19.440 So I want to talk a lot about your work. You're always a guy who's got a lot of hats on. Right now,

00:04:24.540 you're mostly wearing one hat, and that's a hat that puts you squarely at the center of a pretty

00:04:29.480 big circle with a bunch of folks working at a company called Unity Biotech, which we'll certainly

00:04:34.820 get to. But I think for listeners to really understand what you do today, what your company

00:04:41.640 does, they need to sort of understand the journey you've gone through over the past several years and

00:04:47.700 how that's sort of shaped your thinking about an interesting problem. You're a serial entrepreneur,

00:04:51.920 and we will likely discuss some of that in your background. But let's just start with

00:04:56.460 how you think about this problem that I think about, which is how do we help people live longer?

00:05:01.940 How do we help people live better? How do you define that, by the way? I don't want to impose

00:05:05.140 my definitions on you. What is longevity to you? Well, longevity for me would be being able to

00:05:11.040 live without the indignities that I've witnessed, we all witness in our lives, all of these features of

00:05:19.540 aging that seem to be inescapable. So my father, for example, he has profound degenerative disc disease,

00:05:27.080 which makes him functionally immobile. It defines his life. My stepfather died of Alzheimer's at 87.

00:05:34.420 That defined the final chapter of his life. So I've been up close to these things in other people.

00:05:41.300 In addition, we all, and you and I are now, I'm a little older than you, I'm in my 50s now,

00:05:46.940 we are witnessing these changes in ourselves. And for me, longevity would be the ability to use

00:05:53.540 what I know how to do, which is science and biology, and be able to change how we get to live our lives,

00:06:04.380 to be free of these indignities. And I think it's something that we can do, and we are doing it.

00:06:10.260 And so for me, it sort of began as a dream, but now we're generating human data that says

00:06:17.100 this, we can actually do this. So I remember when this started, Ned, circa 2011, how you really

00:06:23.580 steeped yourself in this field, which at the time was kind of new. You had already done a number of

00:06:30.220 interesting things in biology that had led to the development of drugs, but none quite in this space.

00:06:37.520 I mean, I literally remember many nights I'd be driving home and we'd be talking on the phone and

00:06:42.900 you would be telling me about things that you had just read about. And I saw this paper and it's 10

00:06:47.860 years old, but it was almost like you were going through a second postdoc or something, getting up

00:06:52.320 to speed in this. I've always liked the way you've sort of explained your evolution of thought there.

00:06:57.180 Do you want to maybe for the listeners, give them a sense of how you developed a framework around this?

00:07:02.460 About sort of aging biology.

00:07:05.060 Yeah. Why do we age? And what can you do about it ultimately?

00:07:07.460 So maybe it's helpful for me to just create some context. And so the way I think about this and the

00:07:13.940 way I explain it to people that are new to this or people that aren't scientists is I, I always talk

00:07:20.880 about there are these sort of three principles that if you remember nothing else from listening,

00:07:25.580 if you're awakened two weeks from now in the middle of the night, it's these three principles

00:07:29.960 that I think are a good way to think about this. And the first principle is that aging is not a rigid

00:07:37.500 thing. So it's this flexible, malleable thing. And nature has throughout evolutionary history sort of

00:07:46.560 bent and twisted aging for its own purposes to create creatures that have very different lifespans.

00:07:52.020 And this kind of takes us to the second principle, which is that nature has done this with these

00:07:58.980 control knobs. These are sort of biochemical systems that nature twists. And these systems

00:08:07.160 are now something we're learning. And this takes the third principle to turn as people that do drug

00:08:13.640 hunting like us. And so it's these three principles that nature is flexible or that aging is a flexible

00:08:18.620 thing. And nature has then bent and twisted control knobs to make it flexible, and that these are

00:08:25.480 turnable stuff. So if you take those three principles as the context, you may ask, well, why do we as

00:08:34.540 scientists, why do I believe these things are true? So if you just look outside at nature and creatures

00:08:42.920 across phylogenetic history, you see very similar creatures with very different lifespans. So if you

00:08:50.340 take the hard clam, so this is this thing you can eat at a clam bake. If you don't eat it, it will live

00:08:56.800 about 40 years. And it has a deep ocean dwelling relative called a quahog that lives at least 500 years.

00:09:05.780 And these are very similar creatures. They look very much the same. And no one really knows how long

00:09:11.580 this quahog lives. This 500-year thing was just because that was what the age of the wreck off of which 1.00

00:09:17.180 it was pulled was. So these are very similar creatures on the order of 12-fold different lifespan.

00:09:24.480 Similar story if you flip over to, let's say, the order of rodents. Actually, let's not even go to rodents

00:09:31.000 here. Let's just talk about mammals, because that's us. Of course, rodents are mammals too, but we are mammals

00:09:35.800 as humans. So if you look at the shortest living mammal, which is this thing called the tiny shrew,

00:09:40.580 lives less than a year. And then you look at the longest living mammal. So it's called the bowhead

00:09:45.340 whale. And again, no one knows how long a bowhead whale actually lives. So the estimate is at least

00:09:50.280 200 years. So again, you've got this 200-fold difference, at least, in creatures that kind of

00:09:57.580 have the same biochemistry. We're both mammals. Now, if you flip over to rodents, it's the same

00:10:02.180 phenomenon. Okay, you can take a field mouse, which can live somewhere between two and three years,

00:10:07.960 and then you find this naked mole rat, which is this very unusual-looking subterranean creature

00:10:14.780 that's blind and hairless. This creature lives 10 times longer, lives on the order of even more than

00:10:20.260 10 times longer, lives on the order of 30 years if it doesn't die of some trauma. And again, these are

00:10:25.980 in the same order. And so these are closely related creatures, very, very different lifespans.

00:10:32.700 So nature has clearly gone to town on this and created these marvelous examples of disparate

00:10:40.460 lifespan, but sort of the same creature, the same biochemistry. So this kind of takes us to this idea

00:10:45.880 of that nature has done this. How has it done this? And it's done it with these control knobs. And I'll

00:10:53.900 walk through why we thought they were control knobs and how we turn them. So the first control knob

00:11:01.000 that got discovered molecularly was the work of Cynthia Kenyon. And this was a paper that blew me

00:11:07.020 away. I was a first year in graduate school, maybe second year. And Cynthia published a paper that was

00:11:12.080 pretty heretical at the time. And the paper showed that you could knock down the function of a single

00:11:19.900 gene in a worm, C. elegans. So it's about a millimeter size little worm. And when you knock down the function

00:11:25.440 of this thing, it doubled the animal's lifespan. And this was a wild result. Because at the time,

00:11:31.480 aging was thought of as this kind of decay process. But the notion that you could essentially break the

00:11:36.420 function of something or at least make it function a lot less well and double the lifespan of a creature

00:11:41.500 was just a total mind warp for me and other people. Just like, oh my gosh, you can dent something

00:11:48.980 and it gets better. Now what's happened in the intervening decades is this has been repeated in

00:11:54.060 flies. It's even been repeated in mice. You can knock down the function of a single gene and calorie

00:11:58.520 restrict an animal and you can double its lifespan. That's like you or me living to 160. Okay. Because

00:12:05.900 this is a creature with really similar biochemistry to ourselves. So this just said that you could take

00:12:12.160 single genes, turn them and do radical things to lifespan. Now, in fairness, with Cynthia's work 0.63

00:12:18.500 in C. elegans, I always felt that it wasn't the best model to extrapolate from because of the

00:12:25.060 division stopping at the germ cell. And therefore, if you take DAF2 and DAF16 and project them forward

00:12:32.640 to mammals, I don't think we've seen the same magnitude of reduction with attenuation of those,

00:12:37.140 have we? No. And maybe it's worth rather you explaining it than me. Do you want to tell folks

00:12:41.060 what DAF2 and DAF16 do just to bring them along for the ride? So these were genes that were discovered

00:12:45.920 in Cynthia's screen as receptors and transcription factors associated with what we now talk about as

00:12:54.140 the IGF-1 receptor signaling pathway. And if you go in, so this is something that now has been explored

00:13:02.380 in great detail in mice, for example. And the analogous knockdowns in mice of the receptors don't

00:13:10.360 produce the same degree of longevity that you can produce in a worm. Now, interestingly, a paper

00:13:16.660 published somewhat recently, this was on Nir Barzilai and actually Pedro Beltran, who I work with now,

00:13:22.160 with an antibody that antagonizes the IGF receptor. Yeah. It extends life, but there's a sex difference.

00:13:28.820 So it only benefits the females. And it's not something that translates perfectly, but I think

00:13:37.340 that's not the interesting piece in terms of the whole kind of story of what we're learning about

00:13:43.140 aging. It was to us and to me, it was, oh my gosh, you can damage the function of a single gene.

00:13:47.980 And the fact that it still works in mammals is remarkable. The fact that it works less,

00:13:53.720 it's not as cool. It's a little bit of a bummer. But to me, it was really the, it was the DAF-2,

00:13:59.820 DAF-16 observation that opened up the field and said that there were these knobs that you could turn

00:14:06.060 and you could get these radical impacts on longevity. And the additive value of DAF-16 plus

00:14:11.000 caloric restriction was also very interesting in Cynthia's worms. And I actually thought that that was

00:14:17.120 perhaps the most extrapolable insight to mammals, but to your point, it was a proof of concept,

00:14:23.300 right? It was, this thing is malleable. Yeah. And to me at the time, it was just like,

00:14:29.460 it was as though the mask tipped a little bit and you could see the face underneath like,

00:14:34.540 oh my gosh, there's something almost programmatic going along here as opposed to just

00:14:40.140 this idea that you just decay. And that was exciting. And that paper stuck

00:14:47.020 with me for several decades. It was not something that I chose to work on. Although I became friends

00:14:52.420 with Cynthia at some point in the early 2000s, and we continue to be friends and talk all the time.

00:14:57.580 But that particular biology is not, it's more of a marker people use now, as opposed to something

00:15:02.660 that people are actively trying to perturb. So going back to this question about, of knobs,

00:15:08.500 this goes to our, this third principle, which is that we can turn some of these knobs.

00:15:13.280 The first knob actually that got turned in a super aggressive way goes way back to the 1930s.

00:15:22.840 And it was this guy at Cornell named Clive McKay. And he had a colony of rodents. And this is shortly

00:15:29.620 after the great depression and his lab didn't have a ton of money. And in an effort to keep portions of

00:15:35.920 his colony alive, he fed some of the animals all the time. And then other animals he fed infrequently.

00:15:43.600 And what happened was not what he expected. The animals that were restricted with food lived longer.

00:15:50.760 And it was the first time it's ever got experimentally demonstrated, albeit by accident.

00:15:56.060 And for me, that was the first time that scientists had sort of set out, I'll play it for a very funny

00:16:02.140 reason, and turned a knob that extended life. And it brought forward this idea that as we live,

00:16:09.900 aging or your biology is making a decision about the rate at which you are aging. Now, flash forward

00:16:15.640 several decades, a molecule is discovered on Easter Island made by soil bacteria. It's one of your

00:16:22.160 favorite molecules, rapamycin. Now, the aging effects of rapamycin were not figured out until about 2003.

00:16:28.420 And then the awesome work of David Sabatini eventually linked this calorie restriction

00:16:35.320 observation from decades and decades earlier to rapamycin. And it turns out that rapamycin and

00:16:41.800 calorie restriction are both turning the same control knob. And this is pretty wild because biology

00:16:47.800 typically only invents a few ways to do things. And what was really neat was the way in which these

00:16:53.280 two totally different approaches were both touching the same biology. So it's my belief that this

00:17:00.900 biology, so there's this complex called mTOR, which is this megadalton complex. It's this master

00:17:06.400 decision maker of whether or not we are choosing to divide and make more biomass because there's enough

00:17:12.700 resources, or are we choosing to fight another day? And this biology, that rapamycin and calorie

00:17:18.600 restriction and restriction of various other things like methionine is the single best validated biology

00:17:25.680 across labs, across species. And one of my coworkers, who's a sort of incredibly brilliant but cynical

00:17:33.860 guy, always says, if you want to know what the real biology is, it's the stuff that works in every lab,

00:17:39.940 no matter who's doing it. Okay. Rapamycin and calorie restriction definitely have that feature.

00:17:45.760 It works in all these labs. Do you know that I tried to get a personalized license plate

00:17:50.800 for rapamycin and it got rejected by the DMV in California because rapamycin is a drug and you

00:17:57.620 can't have a license plate that's a drug. So then I tried again in vain to get rapalog as a license plate

00:18:05.020 and I got thwarted again because apparently someone at the DMV was smart enough to Google rapalog and figure

00:18:11.120 out that it's an analog of rapamycin, which is also a drug. And I got the same rejection letter

00:18:16.280 from the DMV. I would just assume knowing you, you would have done rapaman, which sounds vaguely

00:18:22.340 like a supervillain. Okay. But no, we didn't go. You stopped there. I'll tell you over dinner what I

00:18:27.520 eventually went with. Okay. Gotcha. So anyway, there's this mTOR thing, which is the best validated

00:18:32.840 control knob. The next knob that got turned is this thing with young blood. So this is something that's

00:18:39.880 sort of popularized if you watch things like Silicon Valley, but this goes back a ways.

00:18:45.000 But the first time that the lifespan effects of this were demonstrated was in the 1970s. And so

00:18:50.400 what you do is you take two animals, one's young and one's old, and you literally suture them together

00:18:55.440 and their circulatory systems join. And it's not a giant mental leap to imagine that crap in old 0.98

00:19:02.940 people is bad for you. So the old blood would be bad for the young. The awesome result was that stuff 1.00

00:19:08.980 in the young was good for the old and it could extend life by as much as 15%. These are with animals

00:19:14.920 that are sutured together, which is not exactly an optimal way to live. So that tells you that's

00:19:19.660 another knob. There's stuff in young blood that is able to exert biology and do good things. And I think

00:19:26.520 that's going to be an area that will be fruitful. The next sort of control knobs we've learned about.

00:19:30.860 So in mitochondria age, so mitochondria are the descendants of ancient bacteria. So they have

00:19:37.560 their own DNA. They have their own genomes. And mitochondria live a really hard life. And so

00:19:44.000 their mutation rate of their own DNA is somewhere between a hundred and a thousand fold faster than

00:19:49.980 our nuclear mutation rate. So their DNA just gets all chewed up. And as a consequence, by the time

00:19:56.700 you are in your 70s, you can go into a human being and look in their colonic crypts, and on the order

00:20:04.200 of half of the cells in their colonic crypts, don't have functional mitochondria. They can't utilize

00:20:10.880 oxygen to make energy or do all the other things that mitochondria do that are important. And so

00:20:16.460 that's something that we've learned that is clearly a part of this aging story. Another thing that could be

00:20:22.480 a knob, and this is an area of some intense debate these days, is this thing called the methylation

00:20:27.780 clock. So this is this ever-ticking clock in which these methyl groups, this is a little carbon atom with

00:20:35.400 three hydrogens around it, are attached or detached over time to the various Cs in your GATC code. And when

00:20:45.140 you're young, there's essentially these five methyl Cs, these methyl groups on your DNA. It's like a crisp

00:20:51.140 pegboard where everything's lined up perfectly and you've got areas where they are and crisp areas

00:20:56.800 where they are not. And gene expression is not noisy. And as we get old, it's like someone drops

00:21:04.520 the pegboard and suddenly you've got methylations where you're not supposed to have them and you're

00:21:08.860 missing them in other places. And it looks as though that might, and I stress the word might, give rise

00:21:15.040 to sort of noisy gene expression. And the metaphor I often think of for this is imagine you have a

00:21:21.000 vinyl LP if you're old enough to know what that is or hip enough to know what that is. It's kind

00:21:25.800 of a hipster thing now. And you play an LP a lot. The actual act of playing that LP takes the little

00:21:32.240 knobs and grooves and wears them down so the LP doesn't sound as good. And it's possible that the act

00:21:40.640 of being alive and that aging actually is modifying these little methyl groups in such a fashion as to

00:21:47.480 cause aging. No one knows if these changes in this epigenetic clock are the cause of features of aging

00:21:53.820 or the result of features of aging or even a little bit of both. And that's going to be a super

00:21:58.940 interesting area of biology. Yeah, I've spoken with David Sinclair quite a bit about this. He uses an

00:22:03.920 analogy that's similar but has a couple differences, which is the methylation could be thought of as the

00:22:10.060 scratch as one gets in a CD surface. And the question, of course, being can you restore the CD

00:22:15.580 back to its original form if you have that master template? And that's not exactly the same as the

00:22:20.900 analogy you've used, but it certainly speaks to the music and the disc. Yeah, it's interesting. So I

00:22:25.780 always think of CDs getting scratched because you're negligent. Right, as opposed to just putting

00:22:30.220 them in the CD player and using them. Yeah. And so there's a nuance. I mean, it's funny. Those two

00:22:34.880 metaphors may even, as you know, may speak to different intuitions between me and David about

00:22:40.040 what's going on here. But I don't really have an intuition. I'm sort of reporting out what's sort of

00:22:45.140 known and what it's going to be just very exciting to see when we can actually do experiments that can

00:22:50.900 tease this apart and ask questions like, if you take these methylation groups away, do you in fact

00:22:58.340 start making cells that are younger functionally? And that's going to be an awesome experiment.

00:23:04.640 David did an experiment recently, and I think you have to be a little cautious with interpretation.

00:23:09.340 So twice in our lives, we press a factory reset your methylome. So you do it once at conception. And so a

00:23:19.020 bunch of gene products get made, which erase and reset your methylome back to day zero, because you're

00:23:26.320 now an embryo. And the program is beginning again. Now, a few weeks into your life, it happens a

00:23:34.060 second time, but only in a little bit of your cells. Yeah. So let's explain that. So when you

00:23:39.020 start at conception, you have one cell. It goes through about 50 divisions. So I don't know exactly.

00:23:46.160 So are you talking about when we do the second reset? First of all, it's different in humans and

00:23:50.140 mice. And I apologize. I don't know. But it's some relatively small number. It's a small number.

00:23:54.280 And it happens within, I believe it's weeks in the human. I think it's a little over a week in a

00:23:59.620 mouse. But anyone out there that works on this, and I have friends I could call, but not right this

00:24:04.240 second. What happens is there's a second wave of these gene products that are made. And for a second

00:24:11.540 time in your life, and the only second time in your life, you will reset. Because now you are making

00:24:19.220 germ cells for your children. And they are kept in a special spot in your sex organs until you make

00:24:27.160 babies. So those are the only two times that this factory reset gets pressed. And then going back to

00:24:32.120 where I was saying, you have to be a little cautious with interpretation. We now know how to do this

00:24:36.700 artificially. So we can express four factors called the Yamanaka factors. And these four factors by

00:24:45.260 themselves, if you express them in cells, it's super inefficient. But one in every few thousand

00:24:50.660 cells, you can cause to become a reset cell with a prime, you know, a...

00:24:57.820 When did Yamanaka do this? This is in our lifetime. I mean, this is quite recent.

00:25:01.580 Yeah, he got the Nobel Prize. At last, I don't have the date. I know a lot about actually how he isolated

00:25:06.160 them. I just don't know the precise dates. But basically, he did this. It was a beautiful

00:25:10.560 experiment with Yamanaka did. He initially identified genes, that candidate list of genes

00:25:16.620 that were turned on in this sort of embryological state. And then he started using, he was delivering

00:25:23.040 the genes in various combinations. He ultimately wound up paired down to about two dozen. And then

00:25:28.960 he then would take various pairwise combinations and figured out how few he could do and still managed

00:25:35.140 to turn on one of these embryonic genes. And he managed to weed it down to four. And he got the

00:25:41.040 Nobel Prize for that. And so what's happened in the intervening years is there's been some efforts to

00:25:47.000 see, could you actually impose features of youth? Because it's sort of a simple-minded leap to say,

00:25:54.660 oh, if you look embryonic, that's young. In fact, and I think that's the little dangerous leap.

00:26:00.500 Right. Just because you don't have the methyl groups on your DNA doesn't mean they translate

00:26:05.600 into the phenotype that is youth. That's correct. What I think you do when you do

00:26:10.420 Yamanaka four-factor reprogramming is it's a self-fate decision. You've entered a very unique

00:26:16.940 cellular state. Now, it just so happens to have things that are in common with youth. And so I think

00:26:23.540 that when you do a reprogramming event and you're able to exert something that looks like youthful

00:26:28.220 biology, you may have actually just created new young stem cells that are able to somehow pick up

00:26:35.140 the music developmentally and participate. That's sort of different than youth. It's a bit of a sort

00:26:42.100 of cell transplant-y sort of thing, although done by the delivery of genes. So anyway, this whole

00:26:48.660 question about this methyl clock is something that it's an intense area of interest in the

00:26:55.280 biological community. It's evolving a lot. There's a bunch of labs that are working on this,

00:27:00.240 and it's going to be a very interesting next five years to see where this drops out.

00:27:05.860 We sort of covered a bunch of things. We've talked about-

00:27:08.900 Yeah, you've got mTOR reduction slash caloric restriction, loss of the circulating youth

00:27:13.920 factors, mitochondrial dysfunction, methylation clock, which sort of ties into sort of stem cell

00:27:19.620 exhaustion. And then we've got another big one. What I wanted to do was I've spent the last

00:27:24.880 now just a little over eight years working on one of these last knobs that we now know how to turn,

00:27:31.320 which is cellular senescence. But I wanted to place that particular knob in the context of the others,

00:27:39.780 because I don't want to sort of overhype this idea. But I do want to explain why I've dedicated a

00:27:47.000 large chunk of my life on planet Earth to this. Because if you look at all of these aging mechanisms

00:27:52.980 and these knobs that I've described, they all play a role in this. We don't yet know what the sort of

00:28:00.140 underlying primordial clock that ticks, that makes us age. What we do know is that you have a series of

00:28:07.460 these component mechanisms that work collectively, and they create this phenomenon that we call aging.

00:28:12.760 And I want to talk about cellular senescence now, because I work on it because of all of the

00:28:19.260 mechanisms, it strikes me and struck me as the simplest mechanism to perturb and make medicines

00:28:28.140 that you can use to treat human beings. And I'll get into why that is in a sec, but I'll just tell you

00:28:34.180 about the knob. So we all began life as a single cell at conception. And then over the arc of our lives,

00:28:41.880 we, the single cell, will divide as many as 50 times. And on the road to 50 divisions,

00:28:50.100 the cells that make us up will encounter some form of stress they cannot resolve. They will pull an

00:28:55.620 emergency brake, and they will stop dividing forever. And we call this state, when you do this,

00:29:00.420 we call cells that do this, a senescent cell. And this is something that we find these cells

00:29:06.180 cells in all of the tissues we've examined. And these cells, they're a very small number most of

00:29:13.260 the time. Typically, it's on the order of hovering around 1% or less. Now, in some disease states we

00:29:20.300 have looked at, that number is dramatically higher. And I'll talk about that a little bit later.

00:29:24.820 Children don't have these cells. Now, they can make them, but they don't accumulate. And we don't know

00:29:30.020 why. But as we age, these cells accumulate. Now, before the work of my collaborators and my co-founders

00:29:38.460 of the company I work at, no one knew if these accumulating senescent cells were good for you, bad for you,

00:29:46.260 or neither. And so my two co-founders, and now almost a decade of collaborators, it's Judy Campisi

00:29:53.360 at the Buck Institute, and Jan Van Dursen at the Mayo Clinic. Both of their laboratories,

00:29:59.960 genetically engineered mice, where we could eliminate senescent cells from these animals

00:30:04.900 whenever we wanted. And for the first time, we got to ask what happens. And what happens is pretty

00:30:12.120 eye-popping. So normally, I carry this picture around on my iPhone, so I always show people this

00:30:18.360 is what I work on. But essentially, by a podcast, you have to use a little bit of imagination.

00:30:23.800 So imagine two animals, siblings, born within seconds of each other from the same mom. These

00:30:29.720 are animals that are 24 months old, so it's kind of equivalent to a 70-year-old person.

00:30:34.860 And one of these littermates is blind. It's osteoporotic. It's frail. It has kidney dysfunction.

00:30:42.080 It has cardiac dysfunction. It looks visibly old. It's littermate from whom we eliminated senescent

00:30:48.360 cells from midlife until it dies. This animal lives, on average, about 30% longer, which is cool.

00:30:57.000 But what's more awesome than that is that when this animal dies, and when it's similarly treated

00:31:03.960 animals die, they die with many of the features of aging either absent or reduced. And that is awesome.

00:31:13.760 And so we're trying to do it at Unity, the company I work at.

00:31:16.600 Now, before you say that, there's something interesting you said, which is they have this

00:31:19.540 knockout of senescence that only occurs in midlife.

00:31:22.680 No. What we do is we insert into every cell of that animal's body. We genetically engineer the 0.74

00:31:27.140 animals so that we can administer a drug weekly, which kills senescent cells within that animal.

00:31:34.340 But do you begin that administration?

00:31:36.300 At midlife.

00:31:36.920 And what happens if you begin that administration at birth?

00:31:39.840 We have not done that experiment in large part because we don't see senescent cells at birth.

00:31:44.680 I see. Is there a chance that those animals would prematurely get cancer if we eradicated their

00:31:49.660 ability to create stem cells at the beginning of life? Didn't Judith have some data that suggested 1.00

00:31:55.460 the importance of senescence early on as a sort of break against cancer as well?

00:32:01.980 So you bring up a really important point that I should clarify.

00:32:05.180 So whenever I tell people about this result, people often ask, this is an important system.

00:32:10.040 Why are you messing with it? So it turns out the ability to make senescent cells is vitally

00:32:16.400 important to our survival. So if you genetically engineer a mouse that cannot make senescent cells,

00:32:22.580 this animal is born normally. It's got 10 fingers and 10 toes, but it statistically winds up full of

00:32:29.340 tumors before it's in reproductive age. So it tells you that this emergency break, the cellular

00:32:33.640 senescent system is an anti-cancer system. And you do not want to mess with the emergency break.

00:32:41.240 You have to leave it alone. And so in the experiment I described where we eliminate these cells,

00:32:47.460 we don't touch the break. We allow the break to pull completely. And for the cell cycle to stop,

00:32:54.980 we only show up after. Now, I think it's still a reasonable question. You'd say, dude, it's like,

00:33:01.560 this is an anti-cancer system. And you're kind of screwing around with it. Are you sure this is okay?

00:33:06.240 Let's take a step way back. So in medical school, we spent all this time learning about

00:33:10.480 different cellular insults. And perhaps the most obvious one, and the one we could use as an example,

00:33:17.540 is DNA damage. So when a cell's DNA is damaged, presumably its first choice is, can I fix this?

00:33:24.640 But if a cell's fate is such that it can't repair its DNA, that's sort of when it has this bifurcated

00:33:31.760 pathway of going down the pathway of apoptosis, which means for the listener that it has a programmed

00:33:38.060 form of cell death. So it basically kills itself in a programmatic structured way. Or it goes down this

00:33:45.240 pathway that you've described, which is senescence, effectively saying, I will no longer do anything.

00:33:50.400 I'm still alive, but I don't grow anymore. I don't replicate. I don't reproduce anymore. My DNA is

00:33:55.840 dysfunctional. As we'll talk about, I suspect it still does things. It still secretes factors that

00:34:02.040 wreak havoc. But is that a fair assessment that I've just made of sort of using an example of a

00:34:07.940 cell with DNA damage has that sort of choice? It does. And we don't understand how cells make the

00:34:15.240 choice between the two. Yeah. I was going to ask you that. Do we have any sense of

00:34:19.540 what sways that decision? Well, I'll tell you some anecdotes. So what we do is we in the lab

00:34:26.140 will irradiate cells or treat them with a DNA damaging agent. These are both things that

00:34:32.640 do exactly. Yeah. Known things to destroy DNA. Yeah. So we have a setting on our irradiator,

00:34:36.760 which does both. So it both causes cells to die. And then 99% of the cells that remain

00:34:43.780 will enter senescence. So you know that all those cells in the dish got the same radiation,

00:34:48.340 they're all the same clone of the cell. Yet some of them chose death and some of them chose

00:34:54.080 senescence. I think it's a really interesting question for us to try to understand

00:34:58.700 about why did some die and why did some live? Presumably it's driven by some explicit cell

00:35:04.380 state at the time. Is there a way post hoc to examine the DNA damage of the cells that died and

00:35:09.640 compare it to the DNA damage of the cells that underwent senescence? Because presumably

00:35:13.980 the actual DNA damage is going to be stochastically distributed across those cells. And even though

00:35:20.140 they are the same cells receiving the same radiation, they probably undergo different

00:35:24.000 degrees of DNA damage. And could it be something as simple as if you damage this portion of the DNA,

00:35:30.720 you lose the ability to undergo apoptosis, but you still retain the ability to undergo senescence or

00:35:35.320 vice versa? Yeah. I mean, you have to be a little tricky because if you're... So the challenge with

00:35:40.760 asking cells that died, what happened is that it requires a bit of a cellular seance, okay?

00:35:46.420 I have a cellular Ouija board. We can... No, I'm just kidding. Yeah.

00:35:49.300 Since we've been talking about this as I've been thinking kind of creatively, there are these...

00:35:52.600 There's moments on the order of minutes to hours after the insult happens in which the cell is kind

00:35:58.840 of going through its sort of pre-death throes. And you still don't know which direction it's going

00:36:04.240 to go? This is not something I've ever seen studied. So this could be something that just

00:36:07.820 outside of my personal experience. But I just think about what we do every day in which we

00:36:13.020 insult cells in an effort to make cells become senescent, but a portion of them choose death.

00:36:18.780 There's got to be a dozen postdocs thinking about this problem, right?

00:36:21.640 Yeah. You'd have to be in a kind of hardcore apoptosis. The thing is, yeah, you want to be

00:36:25.780 in a sort of apoptosis lab. I don't know who is actually working on this problem. It's a pretty

00:36:30.580 interesting problem though. I think the way you'd want to get at it would be you might want to be

00:36:35.180 looking quantitatively. For example, if you had some sort of mark of dye that could tell you

00:36:40.460 how much DNA damage a cell got, you had some sort of high content imaging, and you could look down

00:36:46.400 at a plate and say, oh, these guys that accumulated in the minutes of radiation, three times more of

00:36:54.980 the... They incorporated three times more of the dye. These are the guys that chose death. You could do

00:36:59.120 something like that. And so then you don't need a seance. You just need somebody... It's kind of a...

00:37:03.380 I guess it's more of like a cellular death watch. There's ways that you could start to get at that.

00:37:07.780 It's kind of cool.

00:37:08.780 Sorry to take us off that track, but now let's focus on the guys that didn't undergo apoptosis.

00:37:13.180 So you've got all these cells that have undergone some insult, and they've now committed to sort of

00:37:19.760 a celibate life. They're not going to reproduce anymore. And if they just left it at that, it wouldn't 0.95

00:37:24.400 be such a problem. It's this other thing they do that's problematic, which is they sort of

00:37:29.100 poison the well for cells around them that are otherwise not damaged. Is that a fair assessment

00:37:35.140 of senescence?

00:37:35.920 That is. So in fact, early on in the history of this field, which we can get into in a little bit,

00:37:42.020 it was a real theoretical problem. How could something that's 1% of your cells damage the

00:37:49.220 function of 99% of your cells? Well, as it turns out, these cells have a very active secretome

00:37:54.720 that my co-founder, Judy Campisi, discovered in 2008. And it's, this is the means by which

00:38:02.140 these cells exert bad biology. They secrete into the microenvironment around them all this

00:38:08.700 crap that distorts tissue function. And that is how these cells contribute both to features

00:38:16.600 of natural aging, as well as very particular diseases of aging, is via this secretome called

00:38:22.100 SASP, which stands for senescence-associated secretory phenotype. And I know it's a mouthful.

00:38:28.540 Well, SASP is pretty cool.

00:38:30.120 Yeah.

00:38:30.520 And so it's really a paracrine feature. So in medicine, we learn about endocrine versus paracrine

00:38:35.980 versus exocrine features. But paracrine is when a gland can actually secrete its hormone,

00:38:42.140 and it goes directly to the tissue that it's operating on without having to even pass through

00:38:48.660 the circulation and go around. Like the way, for example, insulin, when it's secreted by the

00:38:53.160 pancreas, becomes systemic. But this paracrine effect is to secrete it and poison your neighbor

00:38:58.760 right next to you. And I guess your point now explains why, if you don't have the ability to

00:39:05.660 undergo senescence, that would be bad. Because you take all those cells that are damaged, that for

00:39:11.460 whatever reason, don't choose the apoptotic pathway, are now, they should be celibate,

00:39:17.360 but they're not. And they keep reproducing. And that's why you might be born looking normal. But

00:39:21.620 as you accumulate injury and insult, boom, you're going to develop into tumors. But that's very

00:39:25.880 different from saying, we're going to let you undergo senescence, but we might potentially block

00:39:30.320 your ability to cause toxicity after the fact, or other means of targeting those cells. I mean,

00:39:35.700 there's many ways to do this, but yeah.

00:39:36.700 Well, let me just go back to this whole thing, which is the cancer thing. So we know that if

00:39:41.880 you can't make cells become senescent, you dramatically increase tumors in mice. And that's

00:39:49.360 a powerful lesson about what not to do, which is that don't mess with the underlying biology

00:39:55.980 here. Showing up after, however, is a pretty cool idea. And the fact that this paper, and I'll

00:40:03.500 come back to how I got involved in all this, but the fact that eliminating these cells can

00:40:09.340 produce a series of youth effects while not increasing cancer risk was very awesome and

00:40:16.980 was actually kind of a theoretical validation of the picture in our minds about how this was

00:40:21.760 all working. So it might be useful to actually sort of at this moment sort of talk a little

00:40:28.160 bit about the history of where this actually all came from. And then we'll come back to what

00:40:32.440 we're doing with this now and how we're making medicines based on it. But I think it's always

00:40:36.980 helpful to place what you're doing in historical context. So this whole idea of cellular senescence

00:40:43.460 traces itself back to the early 1960s. There was a very clever guy named Leonard Hayflick,

00:40:52.000 who I've actually had the pleasure of meeting very randomly on an airplane. But at the time,

00:40:57.220 in the early 1960s, it was widely believed that cells from us, from mammals, could have infinite

00:41:04.980 capacity to divide. And this was made famous by Alexis Carell, who goes back decades before that.

00:41:12.760 The leaders in the field all believed this. And it turns out this was the result of the fact that

00:41:18.200 the food that was being fed to these cells, which was derived from chickens, was contaminated with

00:41:26.080 chicken cells. So every time cells were fed food, they were also fed other cells. And it gave this

00:41:33.460 impression that these cells were dividing forever. And this Leonard Hayflick showed up and-

00:41:38.400 And for someone listening, who's listening to this thinking, how can years, decades of dogma come from

00:41:43.760 such sloppiness? It is important to understand that they're basically relying on light microscopes.

00:41:49.580 Today, you wouldn't make that mistake because you'd be able to look at the DNA of the different

00:41:54.620 cells and realize that, hey, these are chicken cells, those are mouse cells, et cetera. But

00:41:59.280 it's a beautiful example of both how far biology has come, but also how the simplest mistakes can

00:42:08.000 lead to catastrophic misinterpretation. I write about this in my book, this story. And I remember

00:42:13.020 the first time I came across it like you, I was sort of like, my first thought was, how the hell is that

00:42:17.720 possible? But then upon further reflection, you're like, you're imposing too much of your current

00:42:22.200 worldview on the problem. Yeah. I think it's important to show a little bit of compassion

00:42:27.080 as you have by pointing out the limitations of their technology at the time and the context that

00:42:33.900 they were living in. What's funny is I will report an interesting mistake that turns out we ordered

00:42:38.760 from ATCC what we thought were mouse cells and they were rat. And as a consequence, they weren't

00:42:43.840 behaving properly. And this is, this was about a year ago. Okay. So these types of things are still

00:42:48.620 happening. So Hayflick turns over this 60 year belief basically. And so he published this paper

00:42:54.300 in 1961 and it's pretty heretical. And he coins this term. He said, he calls them senescent cells,

00:43:03.420 meaning cells that lose the ability to divide in culture. And he says something very prophetic. He

00:43:08.600 says, this may contribute to features of aging. Isn't that sort of like Watson and Crick's DNA?

00:43:14.540 Yeah. The thing about it does suggest a means for replication. Yeah, exactly. Yeah. The most

00:43:18.600 understated. Yeah, exactly. Which I love these things. Here's the great story is that when I ran

00:43:23.900 into Leonard Hayflick at the Portland airport, having gotten off the plane, I pull out my phone

00:43:30.000 and I show him, and he's very old now. All right. And I show him a picture of the mice with and without

00:43:36.320 senescent cells, the term that he functionally coined. And I said, this is what happened when we delete

00:43:43.000 these cells that you said may cause aging. And it was just a mind warp for him. He hadn't heard of

00:43:49.660 the paper. You know, at the time, the paper had been out for six years. You know, he's very old.

00:43:55.960 Then his daughter walks up who lives in Portland and- She thinks you're accosting him or something.

00:44:01.140 No, it was immediately clear that we were- You were fans.

00:44:03.840 Exactly. Yeah. We were fanboys. And I show her the picture and I said, this is where your dad's work

00:44:10.560 went. She was like, oh my God. And then she like turns to her dad and said, dad, you realize what

00:44:15.840 this is, right? And so there was this moment, it was one of these great moments where you get to see

00:44:21.140 this person who really architected this major insight in biology, confronted with the historical result,

00:44:28.600 right, to his face and watching him trying to process it all. It was pretty neat. Anyway,

00:44:33.600 so that's the kind of backstory of where this all came from. Now, flash forward decades. Now,

00:44:39.280 Judy Campisi was the first individual to find a marker that we could look at in vivo. This was

00:44:47.460 about 1995. So you could actually figure out, because senescence was something that you did in

00:44:53.820 plates, like plastic plates in an incubator. But it was Judy who figured out the first biomarker where

00:45:00.040 you could go into a living creature, into a tissue and say, aha, there are senescent cells. It was the

00:45:04.900 first time we got to know how many there are, right? It's not a lot. And her work really raised this

00:45:11.580 larger question, which I mentioned earlier, which is how can so few cells do so much bad for so many

00:45:18.540 cells? And it wasn't until 2008 when Judy described for the first time this thing that I mentioned

00:45:25.720 earlier called the SASP, the so-called senescence-associated secretory phenotype,

00:45:30.100 that are how these cells exert bad stuff. And it's over a hundred factors that have been

00:45:36.240 characterized that these cells make now that drive their negative biology. Now, it turns out that a

00:45:42.060 lot of what we know about the SASP was all figured out in cell culture and plastic. And what we've been

00:45:48.220 working on the last eight years is doing a lot of this more in vivo. And it turns out the SASP is very

00:45:54.080 different in different tissues. It turns out to be very different from different cells within those

00:45:58.800 tissues and different disease states. Give folks some of the, I mean, I know there are so many of

00:46:03.640 the SASPs, but what's a handful of things that they have in their playbook? How do they sort of wreak

00:46:09.500 this havoc on their peers? So if you look at the sort of the totality of the SASP, everything that's been

00:46:15.720 labeled a SASP factor, what you will see are kind of usual suspects of badness. Okay. So things like TNF

00:46:23.420 alpha, which is the target of the most successful drug in the world today, you know, Humira, and it's

00:46:29.080 an anti-TNF drug. You see a factor like VEGF alpha, the target of things like Lucentis and ILEA,

00:46:37.120 multi-multi-billion dollar drugs for the treatment of diseases of the eye and also used in cancer,

00:46:42.180 but that's a sort of a separate thing. So those are both SAS factors that are the targets of existing

00:46:48.540 massive drugs. There's also some SAS factors that have been clinically validated. So there's one

00:46:54.220 called CTGF, which if you make an antibody against that, you have efficacy in a rare lung disease

00:47:00.380 that's pro-fibrotic that we now know is driven by senescence. And that's not a marketed drug yet,

00:47:05.680 but it makes the same point, which is antibody. So we're taking a step back from this. I mean,

00:47:09.720 let's put some broader strokes on this. Inflammation would result from SASP, right? So there's lots of

00:47:16.040 things that go out there that basically tell immune cells, hey, let's recruit you to this area.

00:47:21.300 You mentioned fibrosis. What other types of broad destructive categories of things go on out there?

00:47:27.580 And I'm fascinated by this work because it's just a little counterintuitive. I totally get the,

00:47:35.920 and again, this is just my need to teleologicalize everything, if that's even a word, but

00:47:39.680 it totally makes sense why senescence exists. It's a little harder for me to accept the breadth of

00:47:45.740 its destructive capacity to its peers. My general thinking about teleological discussions.

00:47:51.500 I know you and I could debate this all night long, which is get over it.

00:47:54.380 Yeah. Because these things are not testable, any teleological explanation should be evaluated

00:47:59.620 solely on its entertainment value. Okay. Right. But that's not why we're here. I will say that

00:48:06.020 there's an idea that is also teleological, but which I use and is a framework that I think is helpful.

00:48:12.140 Well, so many things that we see in aging biology in particular, they're kind of head scratchers. Like,

00:48:18.780 why is the system doing this? This seems bad for the individual. Didn't evolution get a vote?

00:48:24.800 Isn't that the whole point of evolution? Well, it turns out that features of aging are

00:48:30.740 manifest post-reproductively. Oh yeah. No, no. I don't think there's going to be an evolutionary

00:48:35.160 explanation for this for exactly that reason. So is your argument that this occurred,

00:48:42.420 call it stochastically, and evolution was never there to fix it or catch it, so it just ran amok?

00:48:47.540 No, it's actually something, it's a cooler idea. So it turns out these decisions, if you want to call

00:48:53.620 them that, these seemingly irrational decisions made by evolution, of course, evolution doesn't make

00:48:58.380 decisions. It votes with the survival of a species to reproduce or an organism to reproduce. Many of

00:49:06.520 these so-called decisions are things which benefit the young. Yeah. So it's basically, let's get the

00:49:14.380 post-reproductive folks out of the way to conserve resources for the young. No, I don't think anything

00:49:18.440 so active or sort of sinister. I think mostly it's, let's take cellular senescence, for example.

00:49:24.660 So it's an awesome mechanism to block tumor formation in young, reproductively competent

00:49:31.800 organisms. So if you can make more of those organisms to make more babies. Yeah, I guess I'm

00:49:37.080 not arguing this in essence. It's the SASP that I'm struggling with. Right. Well, so it turns out,

00:49:41.700 this is interesting. So it turns out the SASP does some other things that are useful. So a paper published

00:49:47.680 by Judy Campisi again, and Marco Di Maria, who is a professor in the Netherlands now, they got interested

00:49:54.520 in wound healing. And what they demonstrated was that if you make a wound in an animal, a week or

00:50:00.500 two after you make a laceration wound, at the wound site, you will see an accumulation of senescent

00:50:06.600 cells. And they're dumping SAS factors into the wound. And the question Judy and Marco asked was,

00:50:15.440 is this important? And so in this experiment, they went in and they eliminated senescent cells

00:50:22.060 in the course of wound healing. And what they saw was that the wound closed less well and more slowly.

00:50:28.140 And it suggested that the SASP had another role. It wasn't just, well, the SASP, this is a role for

00:50:33.500 the SASP here. Its role was to facilitate wound closure, which if you think about it for most organisms

00:50:39.640 in a naturally evolving ecosystem is super freaking important. And so essentially senescence got co-opted

00:50:45.880 into suppressing tumors in one case, and in a very different case was co-opted into helping heal wound.

00:50:54.220 And so I think that what you see in these diseases of aging is often the sort of unintended consequences

00:51:02.220 of a system that was absolutely awesome for the young at the expense of the old. One metaphor for 0.99

00:51:09.080 this that I sort of like is that, let's say you innocently write a computer program to fill your

00:51:15.000 bathtub. It also unintentionally becomes a computer program designed to overflow your apartment.

00:51:20.920 It wasn't why you wrote it, but it is the net effect. Yeah, no, that's an interesting point.

00:51:25.240 Certainly things that favor wound healing could be quite beneficial. Some have even argued that LP

00:51:30.940 little a lipoprotein, apolipoprotein little a, which is attached to an LDL. So it makes this thing

00:51:35.740 called LP little a, has very potent pro-thrombotic properties. And even though it's wildly atherosclerotic,

00:51:43.060 you could argue, well, frankly, that's a very favorable phenotype to carry for most of our

00:51:49.680 evolution. It's only recently that we need to concern ourselves with atherosclerosis,

00:51:53.240 but there was certainly an era where having a pro-thrombotic factor to help you in periods of

00:51:59.900 trauma would far outweigh the tail effect of knocking off a few elders who managed to survive

00:52:05.480 and die of aortic stenosis or atherosclerosis. So that's an interesting tidbit with respect to

00:52:10.480 the wound stuff. I wasn't aware of that. By the way, the other example, this LP little a is a

00:52:14.380 perfect comparator to the, whether or not either of these teleological explanations matters not.

00:52:20.880 The point is, is that those are the same logic. So let's go back to something you said at the

00:52:25.240 outset, which you've now given us enough to put in context. You said, look, you, you sort of,

00:52:30.280 you went on this journey of thinking through final common pathways of aging. You sort of arrived at

00:52:36.220 this mTOR-CR thing probably matters a whole heck of a lot. In fact, it seems to be the only one that

00:52:41.780 works across organisms, across labs, no ifs, ands, or buts. You've got this loss of circulating youth

00:52:47.980 factors. Clearly something is happening to mitochondria as we age that is very problematic.

00:52:53.780 We clearly see this observation of methylation, whether it's causal or not is unclear. We get a sense of

00:53:00.280 happening to stem cells exhausting and senescent cells. Your view, which you very briefly touched

00:53:06.700 on, is the first four things I mentioned are probably harder to drug than the last one. Is

00:53:15.120 that a fair summation of how you kind of arrived at this? Yeah, that's how I wound up focusing the

00:53:21.140 last eight years of what myself and my colleagues have been working on was that. It was how do we make

00:53:27.340 medicines that target a fundamental mechanism of aging? I was not clever enough or knowledgeable

00:53:33.300 enough, or both, to figure out how to do mTOR and CR. I mean, there are a lot of people who are trying

00:53:41.540 that by trying to get more and more selective molecules, but it was not an area where I had

00:53:46.400 anything creative to add. But the backstory of how I kind of got involved in this was I was sitting

00:53:52.580 around actually in the Calgary airport. A lot of these important things are happening in airports,

00:53:57.960 okay, for some reason. And I was eating french fries. And in this 20-minute interval of french fry

00:54:04.100 consumption, five different people sent me the same PDF. And they were subject lines like,

00:54:10.100 holy shit, you have to read this. This has to be your next company. And I opened up the paper and 0.99

00:54:15.940 right on the front page was a figure showing these two animals, siblings, one of which had

00:54:21.840 senescent cells. The other one- These are the mice you talked about earlier.

00:54:25.300 Yes. Now this was Jan or this was Judy? Were they collaborating at the time?

00:54:29.060 They were collaborating though. Judy was not on this particular paper. And this paper was

00:54:33.740 not exactly those same mice I described. So it had a very similar or the same means of senescent

00:54:39.860 cell elimination, but there was a difference. This is 2011. These animals also contained a

00:54:46.920 mutation where they made less, far less of a protein called BubR1, which is an anaphase

00:54:54.000 checkpoint protein. So this is a protein that makes sure that your chromosomes line up on the spindle.

00:54:58.500 And it sort of is like a bit of a schoolmaster. It sort of says, you're not dividing until everything's

00:55:03.460 lined up. And so if you don't have BubR1, you just blow right through the checkpoint and you become

00:55:08.500 polyploid. And Jan van Derzen, my other co-founder and long-term collaborator, he created this animal

00:55:14.680 thinking this animal was going to get tons of cancer because he was a cancer researcher.

00:55:19.680 So let's just make sure we understand why that's the case. If BubR1 is there to make sure your

00:55:24.580 spindles line up, it means that you should not be able to divide until all your chromatin is perfectly

00:55:30.100 aligned. Everything is in perfect tip-top order and you're ready to undergo mitosis.

00:55:33.960 That's right. So if you take away your master, a lot of faulty division should follow.

00:55:40.420 That's correct. So Jan's hypothesis is we either attenuate or knock out BubR1, 0.94

00:55:46.480 we should get a lot of cancer. That's right. Okay. What happened?

00:55:49.760 That wasn't what happened. The animals were super polyploid, meaning that they wound up with

00:55:54.480 super weird numbers of chromosomes, too many, too few, all that good stuff. The animals aged quickly.

00:56:00.760 So the animals wound up full of senescent cells. Okay. So there was something about

00:56:05.960 having very wrong numbers of chromosomes and all of the effects of that that caused the animals to

00:56:15.180 be full of senescent cells and to age very rapidly, have features of aging that were onset very rapidly

00:56:21.660 and die. 0.95

00:56:23.080 So it accelerated the speed with which the record played, but it didn't change the music. In other

00:56:29.160 words, they didn't die sooner because they got cancer. They just died sooner because they sped up

00:56:33.080 their aging.

00:56:33.820 We don't know the answer to that. So these animals are extremely sick animals. So trying to make a 0.80

00:56:40.220 larger statement, I think it's too much to say with an animal like that of the whole record metaphor.

00:56:46.580 What I can say with, I think, conviction is that clearly a whole bunch of things that go along with

00:56:54.620 being old happen very quickly in these animals. And these animals die very quickly. They look visibly

00:57:01.160 old. For me, it was an experiment that suggested that it was consistent with the notion that the

00:57:07.620 senescent cells were actually driving these features of aging. And what's so cool about this

00:57:13.680 paper and why it caught so many people's attention, even though there was a somewhat earlier paper that

00:57:18.060 Jan also did, that for many allowed one to draw the same inference, but it just didn't smell like

00:57:24.340 a drug. I won't get into that paper today. But what was so compelling was that you added senescent

00:57:29.380 cells through this mutation that gave rise to the production of less bub-R1 protein. And then you

00:57:35.980 eliminated those cells or a subset of those cells with this insert into the DNA of every cell of that

00:57:43.140 animal. And you could ameliorate a subset of the effects of senescent cells. Now, you didn't get

00:57:49.460 rid of all the senescent cells. By the way, interestingly, the animals don't live longer

00:57:53.060 in that particular mutant because it's so sick when you eliminate the cells. But a whole bunch

00:57:58.360 of the features of aging, for example, their bent spines, their organ atrophy, their cataracts in

00:58:04.500 their eyes, these were all severely blunted when the senescent cells were eliminated. And when I saw

00:58:11.500 this paper, I thought, this feels like a drug. You just eliminate cells that are driving bad things.

00:58:24.320 And so I contacted Jan and Mayo Clinic. And 72 hours later, we had agreed to meet. And a little time

00:58:32.380 after that, I set up the company. But this was an observation in a mutant animal. And I remember

00:58:39.660 the sort of early blogosphere talking about how this paper didn't mean anything. And this was

00:58:47.660 exaggerating the role of senescence in humans. And it's a mouse and all this stuff. But to me,

00:58:54.300 it really felt important. Now, there was a long way between that paper and when you could ever

00:59:03.960 credibly claim there could be a drug discovery program based on it. And we spent the next four

00:59:09.820 years in stealth mode asking for biological questions. And I had some very patient investors

00:59:17.240 who gave us sort of a spoon-fed or dropper-fed money into the company. We did not have lots of

00:59:24.940 excess resources. But I will confess that some of the best times in my life have been under resource

00:59:29.820 constraint because your decision-making is very high quality because every single resource decision

00:59:35.120 matters. So we had four questions we were trying to ask and answer. And the first question, and none

00:59:40.580 of this stuff got answered in the first 2011 paper, was do senescent cells contribute to natural aging

00:59:47.640 as opposed to some genetically contrived mouse? Second question, could we find a disease, any human

00:59:55.360 disease, that we could model in an animal where we could eliminate senescent cells and take that

01:00:01.960 disease and either stop it or even better send it backwards? Third question, could we find a molecule

01:00:11.140 that could trigger the elimination of senescent cells from a living creature safely? And last question,

01:00:20.780 which is, was getting rid of these cells safe? We know that kids don't have them, but that doesn't

01:00:27.780 mean that getting rid of them from an adult, like what if adults need them for some reason, but kids

01:00:31.960 don't. So that was a formal question that really was on our minds. Now, today I look at question four

01:00:37.960 and realize that it's not an equivalence between children and adults because of the following reason.

01:00:43.780 We know that if you prevent kids from being able to make them, horrible things happen. We don't know

01:00:51.440 that if kids make them and you nuke them, horrible things happen, which is what you're trying to do

01:00:55.820 with adults, correct? We don't know that. Those could be two different things. Yeah. So kids naturally

01:01:01.040 nuke them. So we know kids can make them because you can do experiments in young animals and they make

01:01:06.580 senescent cells and then the senescent cells are gone. So in effect, that is your living, breathing

01:01:11.280 experiment, right? That's right. The superpower of kids is the ability to make them, which prevents

01:01:16.440 the replication of a cell that shouldn't replicate. And then it has the good sense to get rid of it

01:01:21.600 before it harms anyone. The adult retains the ability to make it, but keeps it around. I want to go

01:01:26.860 back to your second question. It seems obvious now. Your second question, just for the listener,

01:01:32.300 if I recall, was could we find the right disease in which to study this? Could we find any disease?

01:01:37.360 Because if you think about it, that first mouse, all the first and second mouse experiments that

01:01:42.580 Jan did, both 2008 and then 2011 publications in Nature, those weren't diseases the way-

01:01:50.060 Right. They were sort of poly-death conditions.

01:01:52.740 Well, they were sort of like, this animal has these features of being old. They weren't things that the

01:01:57.480 FDA would label a disease, like a bent spine. Like who's going to develop a drug for that?

01:02:02.400 But your intuition was, of course, you can't study lifespan here. It has to be healthspan related

01:02:08.500 because if you have an interest in taking this to humans, mortality is a very difficult endpoint.

01:02:15.940 Yeah. It's not the way we develop drugs.

01:02:17.740 I mean, unless you're tackling conditions that kill people quickly, mortality is a reasonable

01:02:21.100 endpoint for cancer and heart disease drugs.

01:02:23.140 Sure. So in fact, I would even make us even ruthlessly more practical. We wanted to find

01:02:28.380 very diseases of being old. Diseases. Things that the FDA has regulatory paths for and

01:02:34.080 endpoints that you can get drugs approved on. And so we wanted to do that. And that's what we did.

01:02:41.020 So tell me about number one. Number one, again, we take that for granted today,

01:02:45.440 but it was a very reasonable question eight years ago, right?

01:02:48.860 Yeah. So again, Jan Van Derzen to the rescue here.

01:02:52.200 So Ned, at this point, it's just you and Jan and Judy?

01:02:58.260 So it was myself, Jan Van Derzen, Judy Campisi, and a guy named Daohang Zhou who helped us discover

01:03:05.360 one of the early senolytic molecules. Also, Darren Baker, who's a professor at Mayo Clinic,

01:03:11.540 was also instrumental. He was first author on the 2011 paper. And he also was the first author on the

01:03:19.700 paper I'm about to describe as it relates to question one, which is how did we figure out

01:03:25.260 that senescence was driving features of natural aging? And so that was what Darren Baker and Jan

01:03:32.460 Van Derzen spent the next few years after 2011 doing. And that paper took five years before it

01:03:39.740 published in 2016. And what they did in this paper was they took natural mice, mostly natural.

01:03:47.760 They contained one difference was we mated into natural mice, that same construct, that piece

01:03:54.320 of DNA that was wound up in every cell of their body. So we could eliminate senescent cells whenever

01:03:58.700 we want it. And so now we got to do that in a very slow multi-year experiment, but these were

01:04:04.800 naturally aging creatures. And that paper took as a half decade to happen. And when we got that result,

01:04:13.420 which was in 2016, we managed to also ask the other three questions by then. But that was really the

01:04:18.940 trigger for us de-stealthing the effort and speaking openly about what we were doing. Because

01:04:24.940 if it didn't contribute to natural aging and it didn't cause a particular disease of aging and we

01:04:30.100 couldn't find molecules, there was no company. So you basically were processing question one and

01:04:35.900 two, three, four in parallel. That's correct. We had a small team. It was relatively virtual.

01:04:40.180 For a few years, we were just living at Arch, the venture fund. And then we moved and we got too

01:04:46.820 big. We moved up to the Buck Institute. So we were camped out right over Judy's lab or a floor up and

01:04:52.260 a few doors over. No pun intended. Yeah. And we essentially went to town with a small group. It

01:04:59.340 was just a few people supported very intimately by Jan's group with Darren's support and with Judy's

01:05:06.540 group. And so we had this sort of virtual research group where every week we'd get together and review

01:05:10.680 all the data for each of the four questions. And we had work streams for each of questions one through

01:05:15.880 four. And it was a great time. So I want to kind of pause on this journey and take a parallel or

01:05:21.420 orthogonal journey for a moment, because I suspect that at least one person listening and likely more

01:05:27.380 also have some sort of entrepreneurial bent in them. And as it pertains to biology or biotechnology,

01:05:34.100 I think I'd love to figure out a way to extract from you some of the lessons that you've learned

01:05:38.740 along the way. You've done this over and over and over again in many companies. And I'm just sort of

01:05:44.440 wondering what, well, let's start with a broad question. You've already alluded to the fact that

01:05:48.680 you enjoyed that period of time, right? It was a, it was really less of a company and more of a lab at

01:05:53.720 that point. It was sort of a virtual postdoc that you guys were doing.

01:05:57.040 Yeah. It was a coalition of the willing doing an incredibly hard project. And it was far more likely

01:06:03.440 that we were going to fail than succeed. Well, that's actually the question I wanted to pose

01:06:06.880 to you. So we don't talk enough about sort of what the graveyard looks like of these ideas gone wrong.

01:06:13.200 It's quite possible that 2011 to 2016, you spent all of that time working on something and it didn't

01:06:20.340 work. In fact, that would be the expected outcome of this type of endeavor. But that said,

01:06:26.120 no one goes into something with a belief that it's going to fail. Without getting too much into the

01:06:31.780 rose colored glasses that inevitably bias our ability to look back, what proof points did you

01:06:37.980 look to, to say, there's a biological plausibility here that if it wasn't there would have had you

01:06:43.940 less likely to consider it. And if you can, Ned, try to answer it in a broad enough way that it would

01:06:48.580 be applicable to other endeavors that maybe don't deal with senescent cells. In other words, I'm asking

01:06:53.220 this through the lens of somebody who's considering finishing their lab or finishing their time in their

01:06:58.860 lab or even graduating from undergrad and wanting to join a company, how can they think through

01:07:03.660 handicapping the odds of success, which you've done a number of times?

01:07:08.240 So the approach I take to a bold idea like this idea of making drugs that eliminates senescent cells

01:07:15.360 is I try to picture, it's a simple, beautiful idea. And I try to picture the end state. Like,

01:07:22.520 what does this look like at the end? And then I say, okay, I can see this in my mind. And then I say,

01:07:29.220 well, what are the existential risks to that beautiful, simple dream? And I try to write them

01:07:36.800 down in the most primitive way I can. And so those four questions, they were the four risks, which is

01:07:45.900 that senescent cells don't contribute to natural aging. I'll use this particular example and I'll back

01:07:49.960 out in a sec. That was a risk. They don't contribute to natural aging. They don't cause

01:07:53.720 a particular disease of aging. You can never find a molecule to eliminate or eliminating them is

01:07:58.840 unsafe. And those were the four risks we came up with. And you can apply this risk to a beautiful,

01:08:06.300 simple idea in any technological endeavor that I've been involved in.

01:08:12.240 What's nice about that is one of those is a very clear market risk. Three of those are biologic

01:08:18.420 or technology risks. Your second question is actually a market risk question. In other words,

01:08:23.380 if number one worked, so if number one is this absolutely explains natural aging. If number three

01:08:29.200 was true, you could come up with a molecule that could block it. And number four was true,

01:08:33.340 you could do so safely. But there was no way there was a regulatory pathway that was going to allow you

01:08:39.300 to go from A to B because hook spine disease was never going to show up on the FDA's list of

01:08:45.140 druggable things to think about. It's not to say that it couldn't work, but now you've taken on an

01:08:50.360 enormous regulatory risk that would be almost crushing to any one company. So what I'm hearing

01:08:56.900 you say is embedded in your questions was actually a very thoughtful risk analysis that reverse engineered

01:09:03.600 success. I think so. It turned out to be a good way for us to focus ourselves. How long did it take

01:09:11.560 you guys to come up with those four questions? Because it was just five of you if I did my math

01:09:15.440 right. There was a few orbiting. So my long-term business partner, Keith Leonard, he was at the

01:09:20.880 time, we'd founded a previous company together and he was still being the CEO of that company

01:09:26.320 where I'd previously been the chief science officer. But he was involved in this and a few

01:09:31.200 other just very smart people that I will pull into these things. I need very simple thinkers for this

01:09:38.480 sort of thing because you must, in a very disciplined way, distill out to these primitive 1.00

01:09:44.120 risks that can animate behavior over many years. They have to be durable. If you can summon the

01:09:52.520 discipline and the simplicity to distill the next, say, four to five years of your life down to call it

01:10:00.000 two to four risks, because any more than that, then your life gets too complicated. And then you can

01:10:05.400 build work plans that systematically remove each of those risks or do not remove them.

01:10:12.540 But at least identify them.

01:10:13.980 Yeah. Identify them and then build plans. Everything we did at the company for four years,

01:10:19.400 everything was those four risks. That was it. And the money we raised was pretty paltry at the time.

01:10:25.480 It was all budgeted around those four risks. And whenever I go after a new thing, I always try to just

01:10:33.120 go back to that simple, beautiful idea and then say, can I write down the risks to this beautiful

01:10:40.280 idea? Then you build plans from them and then you build budgets and then you politely ask people for

01:10:45.060 money, which by the way, almost no one would provide because this idea seems relatively preposterous

01:10:51.800 at the time. The people that stepped forward were people like Bob Nelson from Arch, who has been a

01:10:59.420 visionary across decades of biotech. And it was he who, when I explained this to him, very rapidly said,

01:11:08.420 yeah, let's do this. There's some other people at Arch like Christina Burrow, who was very supportive

01:11:12.500 early on. So that really was the prime anchor early on in the life of the company and animated all of

01:11:18.700 our behavior. And that was four years of my life. How long did it take you guys to define those

01:11:24.540 questions now? Because the way you rattled them off, they're so logical. They're so obvious. But

01:11:29.200 my guess is those were not immediately apparent to you in 2011, that those were the four things that

01:11:34.720 had to be wrestled to the ground. No, we kind of did know them. Really? Yeah. So it's funny. I

01:11:40.160 recently went through the early, I was looking through some old files and I found an old PowerPoint where

01:11:46.480 there were six risks. And it was clear that we'd done some kind of pruning down to the most primitive

01:11:51.920 because I'm kind of a tyrant when it comes down to like, we must have them simplest plan sort of

01:11:57.880 thing. So there had been some pruning. But I know that we initially generated the list of a half dozen

01:12:03.820 or so. And then over a period of weeks, I think we managed to just in a disciplined way, shear them down

01:12:11.100 to this relatively streamlined plan. Now, did you learn this the hard way? Was there a time when you

01:12:17.700 went down the path of trying to do something, lacked this discipline, and found yourself sort of

01:12:25.400 looking back sometime later thinking, we wasted time, we wasted money, we didn't pursue this as

01:12:31.260 linearly as we could have? Absolutely. No one really taught me this. This is something I sort of learned

01:12:36.260 a bit on the fly, kind of in the eras in which I was building companies like Achaogen and Kythera.

01:12:44.760 So those were all early 2000s. So Achaogen was 2003, Kythera was around 2005. My first company,

01:12:51.860 which I founded my last year of graduate school, we were strategy free. This was a high throughput

01:12:57.080 structural biology company called Cyrix, and we never had any of these intellectual tools. And so these

01:13:03.640 were tools that were really developed through feeling as though we wasted time, which is your

01:13:08.820 most valuable resource. And you wasted people's money, which was also a valuable resource. And

01:13:15.340 the net effect of this has been now this disciplined streamlining of idea and risk. Yeah, so it's

01:13:21.320 something that just picked up over time. And now it's reflexive. And everybody in my group,

01:13:26.020 everyone thinks this way. And so whenever a new thing comes up, immediately everyone goes to the board

01:13:31.100 and starts writing out risks and then starts writing out, well, this is the de-risking plan

01:13:35.720 for this, this, and this. And okay, that means it's 18 months and this is kind of the budget.

01:13:40.100 Do you think that that type of thinking is productive or counterproductive in academia?

01:13:44.760 It's clearly productive in industry. If you tomorrow decided, I want to go back and start a lab and go back

01:13:52.480 to Berkeley and apply for a grant, would you encourage your graduate students and postdocs

01:13:57.760 to think that way? Or would you modify the thinking slightly?

01:14:00.860 I would be modified, but I think there is kind of a value to this. I think in any setting where you

01:14:05.760 have a problem or a technological thing you're trying to solve, which is being explicit about

01:14:11.100 the failure modes, even if you're in an academic setting, and I'm speaking specifically of biology

01:14:17.800 or something akin to it, like I can't talk to you about math or something, because I think it drives a

01:14:23.880 certain degree of honesty when things are failing. So you know what to see. And so if you're looking

01:14:31.060 out for this isn't working and this isn't working and only this is working, and all three have to

01:14:36.360 work for this project to go, you can make a go, no-go decision in an academic setting and pivot to

01:14:43.280 another project because there's no end of creative ideas that you could be working on. They're limitless.

01:14:49.140 And so the important thing is to conserve your most valuable resource, which is your time.

01:14:53.960 And so I encourage my academic friends, some of whom listen and some of whom do not,

01:14:58.880 to think similarly and to take a sort of portfolio approach to academic projects, which is,

01:15:05.360 it's okay to prosecute this question for 18 months, but there are certain things we want to have in 18

01:15:11.140 months. And if you do not have them, we make a formal go, no-go decision. Meanwhile, you had two other

01:15:16.860 projects going. And so what you need is only one to raise its hand and say, I'm working.

01:15:21.940 Yeah. That was super helpful. Let's go back to kind of, you now have the answer to these four

01:15:26.620 questions. Is it safe to say there are still a number of ways to do this? In other words,

01:15:31.960 if you know that senescence is a natural part of aging, if you know that there are specific diseases

01:15:38.320 for which it plays a role, if you know that there are molecules that can be developed that can target

01:15:44.760 it and you know that it can be done safely, that third question really has many heads.

01:15:49.880 You can have molecules that kill senescent cells. You could have molecules that target

01:15:56.180 the factors they secrete. Presumably there are other ways you could machinate around this.

01:16:02.640 Had you fixated early on, you alluded to it that killing cells is something we really know how to do

01:16:07.600 pretty well in biology. Was that the path you guys were on from the beginning or was that a pivot?

01:16:11.340 Early on, we honored the possibility that we could either eliminate the cells because that's what

01:16:16.380 was achieved genetically and was our proof of concept, but we could also come up with ways to

01:16:23.720 reduce the pathological SASP that these cells were creating. It wasn't so much as a pivot,

01:16:30.220 but a decision to go one way rather than the other because in the beginning we thought you could do

01:16:35.180 either. And the decision was driven by a simple idea, which is that were you to make molecules

01:16:41.940 that would simply suppress the secretions of the cells, but not remove the source of all of these

01:16:49.180 factors, you'd have a drug you had to take all the time. And the cool thing about eliminating senescent cells-

01:16:54.000 And it also assumes you know all the factors. Your approach strikes me as the more logical approach,

01:16:58.840 but there are companies doing the other approach, correct?

01:17:01.340 Yeah. Well, you could take the position that any antibody therapeutic against one of these

01:17:05.200 pro-inflammatories in the SASP is just such an approach. But actually I was saying something

01:17:09.960 a little bit more primordial. I was suggesting if you understood, for example, the regulatory

01:17:16.840 mechanism that makes the cell decide to secrete all of these factors and you targeted that,

01:17:23.640 then, hey, what you've done is you've shut down the secretion, the cells sitting there not dividing,

01:17:28.520 how bad can it be? But what motivated us was this would be a drug you would have to take

01:17:34.080 all the time. And what we thought was so neat about the idea of making a molecule that could

01:17:39.720 eliminate senescent cells, which we then named, we called them senolytic molecules. If you could make

01:17:47.140 a senolytic molecule, you could dose it once. And once you eliminate senescent cells, the cells are not

01:17:54.680 there anymore. So these are not drugs you take every day, every week, or every month. These might

01:18:00.700 be dosed once a year until your body makes more senescent cells. And in fact, which maybe we can

01:18:07.160 talk about later, in our human clinical data, what we show is that when we eliminate senescent cells

01:18:12.860 from the painful knee joint of a human being with osteoarthritis, a single administration of

01:18:20.600 senolytic medicine eliminates pain dramatically in these humans for as long as we've looked.

01:18:27.120 And so the cells, we don't know if the cells have come back, but we don't think they've come back by

01:18:31.100 that time. This sort of validates the notion that you can now have a drug that's far safer, because

01:18:36.360 what kills most drugs, actually, is the fact that they're unsafe. That generally is a result of

01:18:41.940 treatment again and again and again. If you don't have to do that, if you can go in once,

01:18:45.940 surgically eliminate the cells. I mean, surgically, metaphorically here, it's not surgical.

01:18:50.980 You can make a safer drug, at least theoretically. And that's why we went that direction.

01:18:55.360 So there's another big challenge here, which is how do you identify which cells are the senescent

01:19:04.040 cells in vivo when you don't have the luxury you've had in the animal lab, which is you get to label

01:19:10.360 those cells, you get to put big targets on them. What was the proof of concept that you could go

01:19:15.360 into an organism and without the luxury of having the senescent cells raise their hand and say,

01:19:23.140 here we are, metaphorically, actually send out snipers to get them?

01:19:27.420 Deep profiling of human tissues for senescence exists, but it's few and far between. And so

01:19:34.980 we undertook a study. And this is after we demonstrated in animals that osteoarthritis,

01:19:40.600 which is the second most prevalent disease of aging and the primary reason it hurts to be old.

01:19:47.420 What's the first, by the way?

01:19:48.520 Type 2 diabetes.

01:19:49.700 Interesting.

01:19:50.380 So what we did is we knew from animals that if we could induce surgical trauma to an animal's knee,

01:19:56.420 a mouse, and we eliminated senescent cells, that we could eliminate pain and actually repair

01:20:01.400 cartilage in a mouse.

01:20:03.120 Wait a second. That's counterintuitive.

01:20:05.420 I know. That's what's so cool.

01:20:07.060 Well, it's counterintuitive because of the experiment you shared with me earlier that Judith's

01:20:11.260 group did. I think it was Judith's group that did this looking at wound healing. You would think

01:20:15.500 if senescent cells were necessary for wound healing, they would be necessary for the non-pathologic

01:20:22.520 healing of cartilage following surgical trauma.

01:20:25.380 And they may be. So our theory about what's going on in osteoarthritis, and this is just one of the

01:20:31.360 ideas we entertain here, and I can explain what data we have that supports this notion,

01:20:35.840 is that senescent cells accumulate at sites of osteoarthritis. In fact, they may be trying to

01:20:42.560 heal.

01:20:43.080 Got it.

01:20:43.680 But unlike in the skin, they literally fall off in the context of healing, the cells remain

01:20:50.380 and essentially continue to sound the alarm over and over and over again. And we think it could be

01:20:56.300 giving a sort of flawed attempt at wound healing that could be driving the pathophysiology of

01:21:01.820 disease.

01:21:02.500 And the margin for error, sorry to interrupt, is much smaller in a joint. In other words,

01:21:07.260 you don't have to heal a wound perfectly to achieve a functional outcome that is perfect.

01:21:11.800 You might not have a cosmetically perfect outcome, but functionally, you could have a

01:21:16.100 hypertrophic scar, you could have a keloid, you could have this, you could have that, but you've

01:21:20.700 closed the barrier to the outside world. But inside of a joint, it's a very delicate balance

01:21:26.940 of one cell layer too many, and all of a sudden you have a different outcome than you had prior

01:21:32.320 to the insult.

01:21:33.080 Yeah.

01:21:33.540 Coupled with everything you said. I mean, that's, all of these things factor together, I suppose.

01:21:37.380 So what we were able to show in animals was that when we induced trauma by cutting the ACL,

01:21:45.200 which is known to be risk factor for human osteoarthritis of the knee, we got something that looked

01:21:50.420 like osteoarthritis. And when we eliminated senescent cells either genetically or with our drug,

01:21:55.760 which we've now taken into human beings, we could, in this trauma-induced setting.

01:22:00.960 And sorry, that genetic mouse, just to be sure, that's a mouse that has a genetic tag for senescent

01:22:06.200 cells that's easy to turn off.

01:22:07.760 That's correct.

01:22:08.280 Okay. One aside question. Is there a mouse model of osteoarthritis that comes from overweight or

01:22:13.720 obesity?

01:22:14.640 No.

01:22:15.000 Is that the most common cause of osteoarthritis in humans?

01:22:18.200 It's a comorbidity. I'm not aware that anyone's established causality,

01:22:22.100 but it would make some sense that it could be causal because of weight-bearingness.

01:22:27.260 Interesting. So the animal model for osteoarthritis is an ACL injury, which is

01:22:31.120 still true in humans, but it's-

01:22:32.880 There are a few different models for osteoarthritis. And so really,

01:22:36.000 one of the things we've learned is that when you attempt to model disease, there are lots of

01:22:40.800 off-the-shelf models that people have built for things like osteoarthritis because they're making a

01:22:46.660 painkiller. So they do things like hurt the animal's knee and then give it a painkiller.

01:22:52.280 How relevant is that to osteoarthritis in a human? It's not. You inject iodoacetate into the joint.

01:23:00.140 It hurts like hell. And then you give them painkillers. So we had to search around for

01:23:05.120 models that happened fast, because that's the way we do experiments, in which senescence played a role.

01:23:11.240 And the ACLT model. So this was work we did collaboratively with Jennifer Elisif, who's a

01:23:16.760 professor at Johns Hopkins of bioengineering. And she was someone, an old friend, and she'd been

01:23:23.740 thinking about osteoarthritis and had the model in her lab. And I called her up one day and I said,

01:23:27.860 we have this crazy idea. We think that senescent cells could be driving this disease. Do you want

01:23:34.220 to try to figure this out with us? And she was an awesome collaborator. Some of the first senolytic 0.94

01:23:40.680 molecules we found, we shared with her. And she had her model in operation. And very quickly in her

01:23:46.960 laboratory, we were able to demonstrate the senolytic molecules that we'd identified at the

01:23:50.660 Buck Institute in Judy's group. One of them was active in Jennifer's model of osteoarthritis.

01:23:58.100 And so for the first time, we actually had a disease, a human disease, the second most prevalent

01:24:02.060 disease of aging. The primary reason it hurts to be old. I can't overstate that enough. It's like

01:24:07.080 being in pain sucks. And this is the reason old people are in pain mostly. And we could drive that 1.00

01:24:13.420 disease backwards in a mouse. So we were just overjoyed by this because we could achieve it

01:24:18.660 genetically, which says it's really senescent cells. And the second thing is we could also achieve the

01:24:24.300 same result with a drug-like molecule. It was the confluence of those two results that convinced us

01:24:29.160 this. This is really cool. It seems that we got the idea right.

01:24:32.880 How does the drug actually target the senescent cell when you don't have the luxury of a genetic tag?

01:24:38.080 We spent the first, remember that was question three of the big four questions. We spent the first

01:24:44.780 two and a half years of this whole effort searching fruitlessly and not finding a senolytic molecule.

01:24:51.920 And the first molecule we identified followed swiftly by the second, and then a series of others that

01:24:58.540 were very related to the second one. The first molecule identified was a MDM2-P53 interaction

01:25:05.380 inhibitor. And so normally MDM2 is this ubiquitin ligase. So it walks around and it's like a meter

01:25:12.280 maid who goes around marking cars with chalk for getting a parking ticket. But what MDM2 does is it

01:25:17.880 walks around and it marks proteins with a little molecule called ubiquitin, and it marks them for

01:25:22.800 destruction. And P53 is one of its client proteins. And so if you break up the interaction between MDM2

01:25:29.600 and P53, P53 doesn't get marked for destruction. So its concentration in the cell goes up. And

01:25:36.220 discovered in Judy's lab by a few people, including Remy Martin Laberge, who was a postdoc in Judy's lab

01:25:42.340 at the time, he discovered that when you did this to senescent cells, senescent cells died

01:25:48.260 preferentially.

01:25:49.140 By the way, is there anything on the outside of a senescent cell that identifies it?

01:25:52.960 There are some things on the inside, and we've been searching.

01:25:55.540 But it's not like you have an antibody or anything that would render it externally identified.

01:26:03.300 We are searching now, and we don't have any universal external marker of senescent cells. We do have some

01:26:09.640 that we just haven't talked about openly yet in various disease states where the senescent cells

01:26:14.100 in a disease state have a marker that's on the outside. We've not found, and people like Ned

01:26:20.040 Sharpless have been searching for over a decade for such a marker.

01:26:24.400 Wow. So just to add to the complexity of the biology, if you were on a little mini nanospaceship

01:26:29.880 and you were inside the joint of a person with osteoarthritis, you wouldn't be able to look out and see

01:26:36.200 which of the cells are actually senescent and which ones are normal, which ones are simply

01:26:42.420 injured. You wouldn't be able to make that distinction. So even if you had a special gun

01:26:46.900 to shoot senescent cells, you wouldn't know which cells to shoot based on just the observation of

01:26:51.920 the cell.

01:26:52.620 Optically, no. Now, if you had a little bit of a sniffer, eyes would not be useful if you were very

01:26:57.860 small. A nose, however, would be useful. So if you could just sort of swim up the gradient

01:27:03.660 of pro-inflammatory and pro-fibrotic markers, you would find yourself-

01:27:08.640 You'd find your source.

01:27:09.400 Yeah. Right. So I do think that the little mini spaceship thing is often very illuminating

01:27:13.740 to think about biology. I use that a lot.

01:27:16.040 Your drug then targets P53?

01:27:17.940 Oh, that's right. Yeah. So P53 goes up in these cells. So it doesn't target P53. It actually

01:27:23.060 binds on the MDM2 side, thereby-

01:27:26.660 Relieving. Yeah.

01:27:27.800 So P53 goes up in concentration. The senescent cells die selectively. And this turned out to work

01:27:37.600 very well in our trauma osteoarthritis models. And another very cool result, and this is something

01:27:45.340 else that Jennifer did wound up in our Nature medicine paper in 2017, was Jennifer got knees 0.99

01:27:53.340 from patients undergoing total knee arthroplasty. So these are people that don't need their knees

01:27:57.960 anymore because they're getting a metal one. And she would take the cartilage at the site of the 0.99

01:28:04.320 osteoarthritic lesion, and she would digest out the extracellular matrix, and she'd take the cells, 0.97

01:28:10.620 the cells that actually make your cartilage. And she would grow them in 3D culture. And she would

01:28:15.880 either expose those little blobs of baby cartilage to either vehicle or our drug that's now in phase

01:28:24.120 two clinical trials in humans. And she observed something awesome, which is that exposure to the

01:28:29.900 drug eliminated senescent cells that were very prevalent from the site of damage.

01:28:34.960 You talked earlier about how you might see 1% to 2% senescent cells in an organism. But when you talk

01:28:40.580 about a very local spot of damage like that, how highly concentrated were they?

01:28:44.560 So it's a little hard because I'll tell you a slightly different number because

01:28:48.120 the cartilage stuff, so basically there are a lot of them at the site where you have the

01:28:53.660 osteoarthritic injury, and you go millimeters away, and then it drops way down. So it's got more

01:29:00.380 puncta of it in that setting. If you look into the synovial membrane, which is, I'll talk to you

01:29:05.100 about another study we did in humans where we actually counted senescent cells in patients with

01:29:09.400 osteoarthritis, the number hovers around 1% to 2% even in patients with osteoarthritis, but it scales

01:29:16.780 with, you have more senescent cells when you have more disease.

01:29:20.180 And when you say, do you mean synovial fluid? Like if you did a joint aspiration on somebody

01:29:24.840 with osteoarthritis, you're saying 1% to 2% of the cells in the synovial fluid is senescent?

01:29:28.780 No, no. So we actually did, I'll tell you what, I'll come back to this a little bit later.

01:29:31.280 Okay.

01:29:31.720 Let me just finish up and then we'll immediately flip to that.

01:29:33.920 Sorry.

01:29:34.480 Okay. Is that we did this experiment where we soaked the cartilage from the patients that had

01:29:41.060 osteoarthritis in our drug and eliminated senescent cells. But another really cool thing

01:29:46.380 is they started growing cartilage in the plastic dish. So these are cells that came from the sick

01:29:53.080 person's knee. They had the capacity to make cartilage, but they just didn't. But once you

01:29:58.940 eliminated the cells, all of a sudden they were producing cartilage again. And so that was super

01:30:04.460 cool. We have yet to prove in a human being that this drug can grow cartilage when it's still in

01:30:10.240 your knee. But that experiment is kind of cool because it suggests that you have this innate capacity

01:30:17.340 if you are unburdened by these cells and their bad stuff they're making. So this next question about

01:30:24.960 how many senescent cells do you have in the disease?

01:30:26.980 I'll tell you why I'm asking the question, which is not just general curiosity. It's

01:30:31.160 really to get a question of how difficult is it to target these things in vivo? Are they ubiquitous

01:30:37.060 enough that you can whack these things with an injection? Because obviously when you're doing

01:30:41.220 this in a person, you're not going to have the luxury of taking their knee out to do it. That's

01:30:44.660 sort of the etiology of my question.

01:30:46.640 So I'll tell you a little bit of what we've done in the clinic. So in human beings, we have

01:30:52.020 eliminated senescent cells from patients' knees with a single injection of this drug. And we can

01:30:59.740 not only eliminate or dramatically reduce pain, but we can also eliminate factors that senescent

01:31:07.100 cells are making. So the quick answer to your question is that this molecule is safe enough,

01:31:12.800 selective enough, and potent enough that we can do it. So it doesn't just work in a plastic dish,

01:31:18.800 but it works actually within the knee of a patient suffering from osteoarthritis.

01:31:23.920 So this was your phase one trial with your first agent. This is the 101?

01:31:28.460 Yeah, that's right.

01:31:29.380 I've seen the phase one data. Maybe explain a little bit of what was done. Phase one for

01:31:35.540 people who aren't familiar with drug studies is mostly to ensure safety, but sometimes you'll see

01:31:42.060 efficacy. Sometimes you'll see that the drug actually does something beneficial,

01:31:44.920 and that's a bonus if you can see both efficacy and safety. But you're typically escalating the

01:31:50.300 dose. And again, you want to see if more drug leads to more toxicity. But if there's efficacy

01:31:56.260 and the efficacy improves with dose, that makes you a bit more confident that it's not the placebo effect

01:32:01.960 or not the effect of the vehicle that you use to deliver the drug.

01:32:05.700 That's correct. So the experiment we did, and this was all the result of what we saw in animals,

01:32:13.820 the result of what we saw in a phase zero study in which we went into patients with osteoarthritis,

01:32:18.600 no drugs, but we biopsied little bits of their synovial membrane and counted senescent cells.

01:32:24.940 And we saw that the more senescent cells they had, the worse osteoarthritis they had,

01:32:28.260 the more bone deformation they had in the knee, the more pain they had. So emboldened by these

01:32:35.720 results, we then took the drug to humans. And as you noted, phase one studies are typically

01:32:41.500 for safety. We realized though that we couldn't really do the right safety study in patients that

01:32:49.140 didn't have a senescent cell burden. Because if you're asking the question, is it safe to eliminate

01:32:53.760 senescent cells from an osteoarthritic knee, you need to have the cells to eliminate. And so

01:32:58.240 that was our kind of logic in the design. And so the way we did the study was 48 patients,

01:33:04.580 where we did a three-to-one randomization, meaning three people would get drug versus one person

01:33:09.840 getting placebo. And as you noted, we stepped up in dose. And it was a single dose of the drug

01:33:16.200 at day zero injected into the knees of patients with painful osteoarthritis. And we then followed

01:33:23.460 these patients for three months. And we checked in with them every week. And they checked in with

01:33:28.740 themselves every day on a little iPhone device. And we monitored their pain.

01:33:34.260 And the investigators were blinded as well?

01:33:36.200 Oh, yes. It's called a double-blind study, which means nobody knows who's got drug,

01:33:41.220 who's got placebo, or what dose of drug you were on.

01:33:44.160 So there's 12 placebos, 36 treatments, treated patients. And those 36 treated patients were

01:33:50.160 about six per dose?

01:33:51.820 They were six per dose. Yeah. So we had six dose levels. And so they were marched up from

01:33:57.060 essentially a group of a series of doses, three of them, which we, based on some modeling we did

01:34:02.880 in cell culture, we thought would be sub-pharmacologically active doses. So we thought

01:34:08.500 they'd be semi-inactive or inactive entirely. And then we moved into the dose range where we thought

01:34:13.820 we'd be doing biology in the knee. And we saw what happened. And we asked the patients. They

01:34:19.400 had once-a-week meetings with their physician where they answered questionnaires about their pain,

01:34:24.620 about their functional state. Then every day they would go and they'd enter on a little iPhone

01:34:29.260 device. They'd answer one question, which is, how much pain am I in on a scale of one to 10?

01:34:34.240 And what we saw when we unblinded the data was that there was a dose-dependent, meaning as you go

01:34:41.640 up in dose, you get more and more pain reduction and durable impact in pain. For as long as we

01:34:48.060 looked, when we got to three months, the pain is not returning. It's completely flat.

01:34:53.200 Do you remember what the placebo group experienced? How much of a reduction in pain did they achieve

01:34:57.880 relative to their baseline? It depends on which of the endpoints we're using. But I'll just summarize

01:35:03.060 by saying that injections into the knee, placebo effect is a big deal. In fact, if you did not see a

01:35:11.140 placebo effect that looked like other clinical trials, you would scratch your head and wonder

01:35:15.060 what's wrong. We saw a very similar placebo effect to what is seen with the injection of steroids

01:35:20.900 into the knee. And that was actually good. Meaning if you do a trial with steroids and

01:35:26.740 with saline, for example, you saw the same placebo effect. You didn't see the same effect as you saw

01:35:31.640 with steroids. Correct. You see the placebo groups in a steroid trial- Had comparable benefits.

01:35:36.440 Look like the placebo effect in our trial. And that is something that one must, the placebo effects

01:35:42.400 in pain studies is absolutely important. But what we saw was that we vastly exceeded anything

01:35:49.720 looking like the placebo effect. So in fact, at our highest dose of drug, again, because it's a phase

01:35:56.060 one study, the number of patients was small at the highest dose. We saw patients who were entering,

01:36:02.140 so this is this scale of one to 10 thing. Okay. It's actually zero to 10. Patients were entering

01:36:07.840 on average at about a 6.2. And at the highest dose, they were dropping to just over one.

01:36:13.560 Is that comparable to what patients experience with steroids? I know it's not apples to apples.

01:36:18.100 And by the way, there's a reason it's not apples to apples that I actually learned on this clinical

01:36:21.500 study. The reason you don't compare across studies is that different studies enroll different

01:36:26.680 patients, different inclusion and exclusion criteria. Even if you use the very same clinical

01:36:31.540 instrument, it's really not an apples to apples comparison. That said, okay, the effect looks very

01:36:38.620 large in terms of what we are doing in terms of approaching our highest dose, getting some patients

01:36:44.640 close to pain-free. Now, I know that the phase one trial was a three-month study. What's the average

01:36:51.120 duration that patients have now been since their injections? I don't have the answer to that question.

01:36:56.340 We are not monitoring the patients in that study going forward because we didn't consent them for

01:37:01.420 that. What we are doing now is we're actively dosing patients in our phase two study.

01:37:07.520 So presumably the duration will be extended in phase two.

01:37:11.760 That's right. So because we only went three months in the phase one study and we didn't

01:37:17.180 consent patients to keep watching them, we really wanted to answer this question, which is if you

01:37:22.160 eliminate senescent cells from somebody's knee, how long does the pain stay away? And so the phase two

01:37:28.480 study goes out to six months, so it's 24 weeks, and we're trying to replicate the phase one study as

01:37:34.800 much as we can. So we're using the same clinical instruments to measure pain and function. We're

01:37:41.340 similar dose levels, so we're doing placebo, half mig, two migs, four migs, okay, which was the highest dose

01:37:48.600 that we explored in the phase one study. And it's 45 patients per dose level as opposed to six. So we should

01:37:55.020 have sufficient statistical power to really see this really clearly. And in terms of the criteria of the

01:38:03.380 patients that are coming in, they have to have painful osteoarthritis of the knee. And so there is a score,

01:38:09.660 this thing called the NRS, which is this numerical rating scale. This is this zero to 10, so it's an

01:38:15.060 11 point scale. And you have to be between a four and a nine. And are these patients that typically have

01:38:20.680 already tried corticosteroids and only achieved limited or short-term response? Like I'm trying

01:38:27.140 to understand clinically, somebody listening to this, who should and shouldn't be excited about this type

01:38:32.000 of work on the horizon? So I think all of us should be excited about this work. Because not only is this

01:38:39.240 a solution, so if our phase two replicates the phase one, and which we all hope and believe it

01:38:46.720 ought to, not only is this a means to treat the primary reason it hurts to be old, but it's a

01:38:53.460 read through to this whole idea of medicine, which is could you treat diseases of aging by eliminating

01:38:59.700 these cells at sites of disease? This is just the beginning of something.

01:39:04.660 Well, let me go back and make sure I understand a few things, because I'm already doing what you're

01:39:07.960 doing, I think, which is sort of extrapolating to the what it means. Is radiographic evidence a way,

01:39:13.800 because you can certainly look at a person's knee on an x-ray and examine the loss of cartilage and

01:39:19.680 appreciate an osteoarthritic knee. You could do the same thing on an MRI. Do you know if the reverse is

01:39:25.360 true? Do you know if the level of cartilage that's making its way back into the knee as a result of the

01:39:31.380 loss of senescent cells is radiographically evident as well, or is it possible that some of the

01:39:38.180 amelioration of pain is due to the reduction of the circulating factors there, but not so much an

01:39:45.080 increase in the structural integrity of the knee? So what we saw in terms of the speed of pain

01:39:51.480 resolution in the phase one was that within two weeks of the injection, you achieved most of your

01:39:58.740 pain reduction. So it seems pretty implausible to me that that is the result of a structural change

01:40:04.700 to your joint. It seems far more likely that this is the result of getting rid of factors that are

01:40:10.420 driving pain acutely. And it's possible, like my questions are probably so ignorant, and I'm sure 0.91

01:40:16.260 there's some orthopedic surgeon listening to this cringing, we would assume that some of the pain that

01:40:21.580 people experience in osteoarthritis is due to the structural part of this, but your evidence would

01:40:26.240 suggest that at least part of the pain is not. It's consistent with that idea, but who's to say?

01:40:30.940 So we're going to know, because we're also doing MRI and x-ray in the phase two study. And so we're

01:40:36.000 going to be able to not only follow pain, but we will follow structure as well. But we have no idea

01:40:41.940 if you will see improvements in the structure of the joint, over what time scale you would see

01:40:47.780 changes and improvements to the structure of the joint. This is the cool thing about doing

01:40:52.240 cutting edge biology and clinical science. I mean, I would love to take a group of patients,

01:40:56.100 if resources were no object, and we could continue to be absolutely sure of the safety of this. You

01:41:01.160 know, you imagine take a bunch of people our age, and you do a bunch of T2-weighted images,

01:41:05.660 MRIs of their spine, and you look at these signal loss L4, L5, L5, S1 discs. Many people our age,

01:41:14.280 myself certainly, have these blown out blackened discs that just aren't taking up water anymore.

01:41:20.260 I have them. Yeah. It would be very interesting to note if you did directed injections of

01:41:24.940 senolytic agents like this, if you could restore this signal. Could you, by eradicating enough

01:41:30.960 senescent cell, create an environment where the existing cell could proliferate into a healthy

01:41:36.320 enough place where it takes up water? Something as simple as that, that again, it's a slippery slope

01:41:41.320 in the spine to go after indications, because it's not as clean as the knee, I don't think. But again,

01:41:47.040 just a great proof of concept. Yeah, I will say that I suffer from degenerative disc disease. It's,

01:41:52.760 well, everyone does at some point, but it's particularly prevalent in my family. There was a

01:41:57.540 result I did not mention earlier from mice, that when we eliminate senescent cells from mice from

01:42:04.340 midlife until death, and we do x-ray on their spines, we see a 41% improvement in the maintenance

01:42:13.540 of the intervertebral disc volume. Now, do you see a restoration or a reduction in decline?

01:42:19.620 We only measured reduction in decline. So we do not know if you could see restoration.

01:42:25.660 So that's still an unknown question in the senescent space.

01:42:28.780 Absolutely.

01:42:29.400 So it's possible you, me, and all the other folks that are in the senior category here

01:42:34.900 won't necessarily reap the benefits nearly as much as the people like your kid, my kid right now,

01:42:41.060 where you figure this out. And when you're 20, you start to prophylactically take these things to

01:42:46.720 reduce the glide rate of decline. That might be true structurally, but I would say that what we

01:42:52.700 saw in our phase one study was you can take people with, frank, painful osteoarthritis of the knee,

01:42:59.900 dose them, and two weeks later, they have profound pain reduction at the highest dose.

01:43:03.960 So that tells you that you are taking a big feature of that disease, the one that you get to feel on a

01:43:09.380 minute-to-minute basis, and you are sending it backwards. Now, whether or not that becomes

01:43:14.280 a structural change is something we hope to understand in the phase two study.

01:43:18.340 And the other thing we, I guess, don't understand, and it might take even more than a phase two study

01:43:22.760 to understand is with corticosteroids, we have problems, which is excessive use of corticosteroids

01:43:29.840 is not a viable option.

01:43:31.200 No.

01:43:31.760 They themselves become destructive and or potentially lose efficacy over time. And so they're a great

01:43:37.760 once-in-a-while tool, but not a great maintenance tool. It will be interesting to know if you become

01:43:43.740 tachyphylactic to this or if the efficacy increases with use as you start to increase the amount of

01:43:52.520 cartilage-laying tissue that exists and you tip the battle in favor of the chondrocyte over the

01:43:57.980 senescent cell.

01:43:59.140 Yeah, that would be the kind of prediction one would make from the sort of zero-order prediction

01:44:04.420 would be that, I think.

01:44:05.880 Is there an example in biopharmacology where the good guy ends up winning with progressive dosing?

01:44:11.920 Well, I would say oncology. If you think about cancer as a gain of function.

01:44:16.020 Only in the cancers that don't spread, though. But that's sort of, that almost seems binary,

01:44:20.280 doesn't it? I guess you're right. I guess you could say that.

01:44:22.300 So because if you think of cancer as a sort of pitted battle between you, the organism,

01:44:27.040 and cancer, which is sort of a gain of function, separate organism based on you living in you,

01:44:32.540 but not you, when you successfully treat cancer through the elimination of those cells,

01:44:38.840 the good guy won.

01:44:40.260 Yeah, I guess that's true some of the time. But most of the time, that's not true.

01:44:44.860 It's sort of a tautology, right? It's true when it's true, but when it's not, it's not. But most

01:44:48.600 of the time, it's not true is the problem. But it could certainly be the case here. There's

01:44:52.480 certainly a case. But even if the worst case scenario is you need a drug injected every

01:44:57.240 six months or every 12 months, it's certainly interesting. Let's go one step further, because

01:45:02.400 this is obviously a critical piece of health span. In fact, you could argue that along with cognition,

01:45:08.680 there's no more important piece of health span than the structural integrity of your body as you age.

01:45:14.320 Have you, in your leisure time, had the ability to think about how this might impact

01:45:19.580 lifespan vis-a-vis atherosclerosis, cancer, or other diseases that actually shorten life?

01:45:26.700 Well, we know that atherosclerosis is another disease in which senescence plays a role. Now,

01:45:32.460 we're not currently working on that. One of the things you may have noticed is that osteoarthritis,

01:45:37.720 the way we approach the disease, even though it affects many of the 360 joints in your body,

01:45:44.140 we treat the local version of the disease with local therapy. Atherosclerosis is about a

01:45:49.160 systemic as it gets. And Jan van Dersen at Mayo Clinic published a paper. Actually, it was Jan

01:45:55.420 van Dersen and a guy named, a very excellent young scientist named Bennett Childs. And this is a paper

01:46:00.620 in science in which he showed in rodents that senescent cells accumulate in the atherosclerotic

01:46:09.760 plaques that form in high-fat diet mouse, which is a model that can predict, say, statin efficacy to

01:46:17.560 some extent, although it exaggerates it somewhat compared to the human case. And what Bennett showed,

01:46:22.840 these plaques are full of senescent cells. And if you eliminate those cells...

01:46:26.760 What are the cells that have become senescent? What's their origin?

01:46:30.060 So they appear to be macrophage in origin.

01:46:33.680 I see. So it's not the endothelial cell.

01:46:35.200 Well, no, there's three cell types. It's basically, there's a senescent endothelium.

01:46:39.200 There is a myofibroblastic type of cell that's in there. And then there are macrophages.

01:46:45.580 So it's all three. It's basically the barrier, the immune cell that came to the rescue and the

01:46:50.660 fibroblast that attempted to repair the damage.

01:46:52.980 Yeah, all of them. And when you eliminate these senescent cells, either genetically or with a drug,

01:46:57.760 first of all, you can reduce plaque volume, which is cool. But what might even be cooler is that,

01:47:03.640 and again, it's in a mouse. And so you just got to wonder, what is that really saying about

01:47:08.240 the pathophysiology of the human disease? But there is this fibrotic cap that forms on the

01:47:16.940 surface of an atherosclerotic plaque. And one of the ideas in the literature is that the thinning

01:47:22.700 of that plaque gives rise to a unstable plaque that is clinically dangerous. And that interventions that

01:47:32.640 can thicken that plaque might be a therapy because you could take a plaque that you have and now convert

01:47:40.000 it into something clinically more innocuous. And what Jan showed was that genetically and with drug,

01:47:46.020 that he can thicken the plaques of atherosclerotic plaques in a mouse.

01:47:50.600 That would seem to counter the idea that you could lessen plaque volume. Do you have to pick between

01:47:54.920 one of those two strategies? Because typically, if you thicken the plaque, wouldn't you likely

01:47:58.320 increase the volume?

01:47:59.340 No. So if you look at the sort of relative partitioning of how much of the plaque volume

01:48:04.940 is kind of the cap, and then how much is this sort of bulk, sort of lipid deposit with

01:48:10.940 macrophages in it, eyeballing it, it's, you know, 90% is this kind of lipidy sort of

01:48:17.140 macrophage blob. And then there's this thin little kind of veneer on top that is the cap.

01:48:24.940 And so the thickening of that cap doesn't dramatically contribute to the overall volume

01:48:30.080 of the atherosclerotic plaque, at least in a mouse.

01:48:32.640 But both are happening. You're thickening the cap and reducing the subendothelial portion of the,

01:48:39.220 reducing the foam cell.

01:48:40.340 Yeah. So athero is, it's a very tough place to do drug development. I mean, we've seen

01:48:46.920 the PCSK9 story has been a multi-billion dollar battle that has given rise to relatively

01:48:54.060 moderate uptake of those drugs.

01:48:56.500 Could argue that's just due to the cost of the drug.

01:48:58.700 It's also the fact that, yeah, it's a relative cost issue going up against statins.

01:49:02.540 So switching people and all this sort of thing, but it's not been, it's caused a sort of downward

01:49:08.140 pressure on people's enthusiasm for doing cardiovascular drug discovery, just the, because

01:49:14.300 now we have to, you're not using circuit markers anymore. You're using outcome studies. You're

01:49:18.780 actually following people until they die or have a serious cardiac event.

01:49:23.500 Do you see an application here in oncology? I mean, it seems like there should be.

01:49:27.920 Well, it's an area of interest of ours. So one of the things we saw in Jan van Derzen's

01:49:32.540 2016 paper was that these mice from whom we eliminated senescent cells from midlife until

01:49:38.240 death, they had the same cancer prevalence. So 85% of them die of lymphoma with and without

01:49:45.980 senescent cells. The difference is that when you eliminate the cells, the animals from whom

01:49:51.340 these cells were eliminated get cancer 30% later in their lives. So you have to kind of

01:49:58.940 head scratch for a moment. Like, what does that mean?

01:50:01.120 And if I recall, you said they got a 30% lifespan extension.

01:50:05.440 Yeah.

01:50:05.960 So you basically created centenarians. You just phase shifted chronic disease by a third.

01:50:10.900 And some of the diseases were just dramatically reduced. So the effects on kidney function,

01:50:16.540 the age effects of kidney function were reduced dramatically. Cataracts were reduced dramatically.

01:50:21.560 So there's a whole bunch of behavior stuff that we don't know what that means, but these animals

01:50:25.880 seem to preserve features of youthful behavior. Their spinal lordosis looks youthful. So there's a

01:50:32.660 whole set of things. But one of the things that people often ask about the experiment is,

01:50:37.760 do they die of different things when you live 30% longer? And the answer is no, they die of the same

01:50:44.180 thing mostly. Now that could have to do with the fact that it's mice and these lab strains of mice.

01:50:49.040 Yeah. They just get cancer like crazy. And it's less interesting to me that it's

01:50:52.180 how much lymphoma they get. It's more interesting to me that you could phase shift it.

01:50:55.880 So we don't know why that is. But an explanation that makes some sense to me is that there is

01:51:03.960 something perverse that senescent cells do to the tissue microenvironment. So this system,

01:51:10.080 which is an anti-cancer system in the young, could become a pro-cancer system in the old

01:51:16.100 by doing something to the tumor microenvironment that makes it more amenable to tumor genesis.

01:51:24.600 And that would be consistent with what you see where the rate of tumor formation across the

01:51:30.100 animal's life is unchanged, but its ability to take root is delayed. So you could imagine

01:51:36.600 applications in oncology, but you would require something in which you had some sort of highly

01:51:41.240 sort of tumor-prone situation where you could intervene in it. There's also an idea where if

01:51:47.320 you could make cells senescent, which is what chemotherapy does in many cases, and then do a sort of

01:51:53.220 two-hit strategy where you drive the cells into senescence, and then you exploit a senescence

01:51:58.760 associated, you know, Achilles heel. So it could be a strategy in which you deliberately are trying

01:52:04.580 to drive tumor cells into senescence and then killing them. And then there are a series of cancers

01:52:10.480 that we think that are age-associated that may be essentially the product of a highly senescent

01:52:17.160 environment. So cancers of the skin, what if you could, or cancers of, say, the bladder or something

01:52:23.520 like this, these things that older people get, could you go in and eliminate senescent cells and 0.98

01:52:29.620 change trajectory of disease? That's an idea that's very cool that you could think about.

01:52:34.700 But I would say that the most powerful read-through from our results so far in humans are these other

01:52:43.200 diseases of aging for which we have no treatments. So let us take something like the dry version of

01:52:50.560 macular degeneration. This is a disease for which there is no treatment, and it's eight-fold more

01:52:56.720 prevalent than the wet disease that you can treat with anti-VEG-S. This is a disease that appears to

01:53:01.880 have senescence associated with it. What's the distribution of those two? So people that have

01:53:05.860 age-associated macular degeneration, one in eight has the wet disease. And only that one out of eight

01:53:11.720 people is treatable with the anti-VEGF antibodies. Seven out of eight of those people are untreatable.

01:53:18.840 Now that disease, the dry disease, moves more slowly, but it makes you just as blind. And so that is an

01:53:24.700 area in which senescence and senolytic medicines may play a role, but we'll have to find out in the

01:53:29.920 clinic. And next year, we are going with our first molecules into the eye. Do you have an animal model

01:53:37.120 for that? No. So you have animal models of the wet disease and our molecules work in that model. And

01:53:43.840 so you could say that that's pretty neat. This is a giant opportunity, but I am personally attracted.

01:53:50.060 First of all, I just want to say that when we talked about the four risks- Defined your work

01:53:53.480 stream. Yeah, for four years. We live in a different risk set now. So if you think about it, the chapter one

01:54:00.140 of this entire effort really was demonstrated in a human being that eliminating senescent cells

01:54:05.940 could take a feature of aging that otherwise was untreatable and send it backwards. Our phase one

01:54:11.440 study, certainly if the phase two replicates, that was the end of chapter one. And we're sort of living

01:54:16.960 in this chapter two moment where we're seeing how broad can we make this work? And if you think about

01:54:22.540 what risks live in chapter two, for me, and I think anybody, it would be that the risk is you have

01:54:29.360 your biological disease hypothesis wrong. And animal models, they only ask and answer the

01:54:35.300 questions they're built to ask and answer. And the only way you get to really ask and answer the

01:54:39.760 question you want to ask is in the clinical setting in human beings. So the way I think about diseases

01:54:47.020 of the eye is you have a series of these untreatable diseases that make you blind, where we think

01:54:51.520 senescence could play a role. And we're going to explore each of these in the clinic. So essentially

01:54:55.880 being able to ask in the relevant setting that can we intervene in these diseases of aging using the

01:55:03.560 very same approach that we took to osteoarthritis, albeit with different drugs. Now you won't be doing

01:55:11.620 a real phase one here. You're going to go directly into a phase two, right? So this is something that

01:55:16.360 we haven't really talked about yet, but notionally we're going to be going into patients because

01:55:22.820 obviously your eye is something that's super, super important and you can't mess with it.

01:55:27.700 We're going to be going into a very select population first, demonstrating safety, and

01:55:32.560 then branching outwards into multiple diseases of the eye. How do you deliver the vehicle? Is it an

01:55:38.060 injection? Yeah. So it's very similar to clinical practice for anti-VEGF delivery. So it goes right

01:55:44.480 into the vitreous fluid of the eye and our drug then takes up, we actually know exactly where the drug

01:55:50.220 goes into the various sub-compartments within the eye. And we know where senescent cells are in these

01:55:56.680 different diseases and our drug gets there. And the clinical hypothesis is that one or more of these

01:56:02.820 essentially progressive diseases of the aging eye, when we eliminate the cells, will stop.

01:56:09.460 And can you speak about the drug? What is it targeting? It's not targeting P53, is it?

01:56:13.220 No. So we have two molecules that we're advancing toward the clinic. Both of them are inhibitors of the

01:56:18.600 BCL2 protein family. So these are molecules that inhibit, inhibitors of apoptosis. So they cause

01:56:26.400 cells to enter programmed cell death. But interestingly, we've shown in animals that they

01:56:32.620 don't cause that to happen in normal healthy eyes. They only target cells that have been damaged by the

01:56:40.420 disease process and the stress associated with the disease process. And so-

01:56:44.780 Why is that? There's something about... So senescent cells, when they enter this state,

01:56:51.640 do a variety of things. They turn up proteins that are pro-apoptosis. And they also turn up proteins,

01:56:58.640 in terms of our expression, that are anti-apoptosis. So what we do is we go in with a drug and just give

01:57:03.800 a little shove in one direction. I see. But you're saying if you give that same shove to someone who has

01:57:10.600 not up-regulated or down-regulated either of those factors, it doesn't seem to matter too much.

01:57:15.740 Yeah. And so we have a beautiful experiment in mouse where we have a disease condition,

01:57:21.760 okay? And we have a normal condition. Mice are identical, but for this. And we put the drug in

01:57:28.640 and we can see the drug molecularly going in and engaging its target. So because it actually breaks up

01:57:35.860 two proteins that are stuck together. And you can see in these animals, both of them are equally

01:57:40.160 engaging target. But the apoptosis program only turns on in the disease state. And it was a pretty

01:57:46.800 awesome result because you can see the selectivity of the molecules in a living creature's eye.

01:57:53.160 Does that same molecule work in the osteoarthritic scenario?

01:57:57.300 No. So we couldn't get it to work either in the trauma models in mice. We couldn't get it to work

01:58:03.340 when we took the cartilage out of human knees. It didn't work there either. And so it never raised

01:58:08.920 its hand. What do you think that says about the biology?

01:58:12.060 There's something about cells and their fate that determines which apoptosis vulnerability

01:58:20.920 you have. They have different Achilles heels. And we don't understand mechanistically why that is.

01:58:28.200 We've sort of figured this out empirically and picked molecules based on their behavior against

01:58:35.800 the cells that we find to be senescent in these human diseases.

01:58:40.900 So 15 years from now or 20 years from now, we will likely look back and companies like Unity,

01:58:47.920 and presumably there will be many like Unity, you will have an entire suite of targets.

01:58:54.260 You will say, well, here's the playbook. In this type of cell or this scenario,

01:58:58.820 going after the anti-anti-apoptotic pathway is the way to go. In this case,

01:59:03.380 we're going to go after P53. And you might have a dozen of these different targets. 0.70

01:59:08.100 Well, we're not sure yet. I think it's too early to say how this is going to play out.

01:59:12.680 Oh, so you don't think it's a fait accompli? I mean, you think it might be that there's

01:59:15.840 a very small number of ways to go about doing this?

01:59:18.360 I doubt that. I doubt it will be dozens though. I think it's going to be a small number. I think

01:59:24.260 it'll be context specific. I think there'll be super creative ways that we and others come up with

01:59:31.400 to exploit these different vulnerabilities. I don't think biology is going to have made

01:59:37.300 24 different keys to this lock. I think it's going to make a half dozen. And we actively search

01:59:44.280 for these. It's just interesting to me that you can turn one key and get a benefit. I mean,

01:59:48.760 think about how often that doesn't work. Think of the futility of that in chemotherapy,

01:59:53.040 where you target this one piece of cell cycle replication and very quickly the cell mutates

02:00:03.000 away around that drug. So cancer is incredibly hard, okay? Because cancer exploits the single best

02:00:11.220 tool biology has, which is variation. Cancer cells can divide and mutate and become anything

02:00:19.620 to avoid death. They live under selective pressure, particularly in the context of drug.

02:00:25.960 Senescent cells can't divide by definition. So their ability to access variation is dramatically

02:00:34.900 reduced. Now that's not to say that there couldn't still be selection without division,

02:00:38.740 but boy, it's way harder when you can't make a million progeny with tiny variants and test them

02:00:45.800 all in parallel, which is what cancer does. Yeah. It's sort of funny, isn't it? Cancer is the

02:00:49.660 ultimate AB tester. Yeah. We like to think of senescent cells as cancer cells that can't divide is one.

02:00:57.420 I've heard that said semi-humorously, but that's actually really nice. Right. Which is sort of funny

02:01:01.620 because it's like, tape the top 100 most valuable traits of cancer and take away 99 of them.

02:01:07.540 Right. By the way, I was saying that and somewhat joking. No, I understand. Yeah. Yeah. In the sense

02:01:12.220 that senescent cells become senescent. Well, first of all, in vivo, we don't often know why they became

02:01:17.300 senescent, but one of the mechanisms by which cells can become senescent is activation of a cancer

02:01:23.500 causing gene, but that could be very rare event compared to mitochondrial failure or high concentration

02:01:31.040 glucose or telomere shortening. I mean, I was going to actually ask you exactly about that example,

02:01:37.200 which is, do we have any reason to believe that the increase in age associated type two diabetes could

02:01:44.840 be resulting from some sort of pancreatic senescence where beta cells become less and less robust due to

02:01:52.040 senescent beta cells in the proximity? There is some data in support of senescence in that cellular

02:01:58.240 niche, but it's very complicated biology. And there's even some talk about a compensatory mechanism where

02:02:04.320 senescent cells actually may improve the function of some of those cells within the niche. It's an area of

02:02:11.360 intense literature debate. The other challenge we have there is we don't have good models in animals of

02:02:17.820 studying this. And so it's been hard for us to get our heads around. What I will say about type two

02:02:22.720 diabetes in humans is that there was a large meta-analysis of a genome-wide association studies

02:02:31.600 looking for the genetic correlation between grievous illness of aging and loci. Where did these mutations

02:02:40.900 map? And what you saw is that one of the very big peaks for diseases of aging maps into the control

02:02:49.340 system for the establishment of cellular senescence is this P14 ARF locus. And you look at, this is where

02:02:57.400 this very important gene that people use for senescence all the time called P16, which is this part of the

02:03:03.060 emergency break. Its gene lives in this locus. And what you see is diseases like type two diabetes,

02:03:08.860 mutations that give rise to that live there. Mutations that give rise to late onset Alzheimer's

02:03:16.180 live there. Frailty lives there. Atherosclerosis lives there. So a whole bunch of these diseases where

02:03:22.280 we saw phenotypes in mice, you see a natural genetic variation of humans also shows a kind of tie-in to

02:03:31.900 senescence. And so type two diabetes caught my eye there because you see that in the GWAS studies.

02:03:36.960 And lastly, what about Alzheimer's disease? We have to believe that astrocytes are losing some

02:03:44.240 functionality to protect neurons, right? So this is what my group works on. And it's something that animates

02:03:52.400 a lot of us at the company. We know that astrocytes or glia, just call them glial cells. So if you guys

02:04:00.620 don't know what glia is, that's, it's actually derived from the word glue. So this is the, the cells that are not

02:04:05.400 your neurons in your brain that seem to hold the rest of it together, which outnumber neurons

02:04:10.340 tremendously. I think the number is, you know, on the order of 10 to one is eight to 10 to one or

02:04:14.860 something like that. You guys got to look that number up, but it's on that order of magnitude.

02:04:19.120 And what was known before we got involved was that as human beings aged, there's a dramatic increase

02:04:25.120 in senescent glial cells, specifically astrocytes and microglia, which are the sort of macrophages of the

02:04:32.460 brain and stuff we've been working on. If we've been looking at that trend in rodents and we see a

02:04:38.660 very similar trend. Now we have not talked about doing on this yet because it's still early days,

02:04:45.620 but it's hard to imagine that the pro-inflammatory environment created by this vast number of senescent

02:04:52.360 glial cells is good for you. One needs to ask the question, what would happen if you could eliminate

02:04:59.420 senescent glial cells? Would that enhance cognition? Would it harm cognition? And this is something that

02:05:05.100 we are actively looking into, but all of the work we've done over the last eight years, everything

02:05:10.200 going from the mutant animals to the wild type animals to human data in osteoarthritis and now

02:05:16.940 moving into human data soon with in the eye, it would just be very hard to imagine that senescence

02:05:23.620 is not playing a role in aspects of cognition loss as we age. And this is something when I think about

02:05:32.260 why we do what we do, that is something that there could be very few greater contributions that

02:05:39.460 myself and my colleagues could make than making a contribution there because it has been

02:05:44.180 an unsolved problem and a scourge to humanity.

02:05:49.340 I mean, do you get the sense that doing what you've done now, it would be very difficult for

02:05:52.760 you to go back in time and do something like Kythera over again. And I'm not saying that to

02:05:56.620 be critical or in any way, shape or form. I'm just saying like, I mean, maybe tell people what Kythera

02:06:00.860 did. That was a very successful company for you, but there's no comparing the nature of the two

02:06:06.300 problems you're trying to solve. Yeah. Actually, thank you for actually pulling me out of what would

02:06:11.040 otherwise be a kind of like overly sort of sappy discussion about contribution. Okay. So Kythera is

02:06:17.460 actually a total kind of opposite sort of thing. And Kythera was a company myself and two co-founders

02:06:23.000 founded now almost 15 years ago. It's about as opposite as you can get from doing Unity. And so

02:06:29.840 it was basically this idea that all this biology that's been explored for oncology and inflammatory

02:06:36.020 disease, taking molecules from that and applying it to the biology of aesthetics. And we did this

02:06:44.480 business. And I had these great business partners that continue to be partners with me today.

02:06:49.920 One of them is Keith Leonard, who is CEO at Unity and very close friend. And he's been my big brother

02:06:56.360 for 15 years and continues as of this morning, calling me and giving me feedback on things. And

02:07:02.540 I've just grown. And the reason I did Kythera was it was sort of a cute idea, but it was mostly because

02:07:08.500 of people was these guys said, Hey, let's do something together. And it was a great decision

02:07:14.480 because I learned so much in a sense, because so the drug, we ultimately got approved. We took four

02:07:21.000 things into the clinic and one out of the four things worked. And it was a molecule that causes

02:07:25.560 fat cells to explode. So we went from an in vitro observation all the way to a launched commercial

02:07:32.120 product called Kybella. You can go get it at your dermatologist. And Allergan acquired the company.

02:07:39.460 And it was a great learning experience of just how to develop pharmaceuticals in something that

02:07:44.600 didn't have historical resonance at all. And so it was a great moment where the stakes weren't as high

02:07:53.440 for whether you succeeded or failed. And as a consequence, it allowed you to hone your skills of

02:07:59.620 kind of making really high quality decisions on risk and making drug development decisions in a way

02:08:06.860 that's actually, frankly, a little more dispassionate. And it was a great growth experience for me.

02:08:11.800 I mean, is it safe to say that if you had not had the experience at Kythera, because then you would

02:08:16.980 have been coming from a Kagen, right? Basically. Yeah. Would you have had a more difficult time doing 0.98

02:08:22.280 what you and the early co-founders did at Unity? It would have been impossible. I would have none of the

02:08:27.460 intellectual or emotional toolkit to do what we did. I mean, that whole risk thing, that was stuff

02:08:34.160 we built. I mean, it was a little bit when we put together a Kagen, but it had become a kind of

02:08:39.920 ritualized practice by the time we were doing Kythera projects, this dimensionalization of risk and

02:08:47.320 creating work plans based on it. And so, yeah, I would say that there would be no unity without

02:08:54.100 there being Kythera. Last question, Ned. Actually, I have two questions for you. I'm sorry that I

02:08:58.580 think about it. First question is, what advice do you give for a person who's studying science

02:09:04.440 and trying to decide between the entrepreneurial pathway that you've been on versus a more academic

02:09:11.980 pathway? What would you offer them as an insight or a set of questions that they could pose to

02:09:16.200 themselves to further delineate that? Well, I would say two things. First, don't create a false

02:09:22.400 choice. All right. So I have friends that are academics that have founded as many companies as

02:09:28.720 I have. And I do from time to time have career jealousy over some of the freedoms and some of the

02:09:36.220 responsibilities, frankly, that they have as people who have academic appointments. There are wonderful

02:09:41.780 things that you can do from the seat of academia in company creation if you are so lucky and so

02:09:47.520 positioned to do so. So first, don't fall into this idea of a false choice. And I see people do that,

02:09:53.640 I think, out of some emotional discomfort, this desire to just move and confuse action with progress.

02:10:00.620 And I'd say just slow down. Take a deep breath. Don't rush. You have your life. That'd be the first

02:10:06.480 thing. The second thing is animate what you do with a single beautiful idea. So something that

02:10:15.040 moves you, that makes your, gives you, literally gives you goosebumps, that weird tingling sensation

02:10:21.040 when you think about it, when you think how cool it is. And oh my God, if we could actually make this

02:10:26.900 work. I mean, few things in life rival that feeling when a hard fought battle for data that was 18 to 24

02:10:35.780 months and the experiment finally works and you're looking at the data and you feel like the future

02:10:41.420 taking root in the present. The only thing I can compare it to is love from your kids. That is the

02:10:47.960 only thing that has the same emotional gravitas. Finally, third thing, learn from people that are

02:10:55.600 better than you, that are more experienced than you, that will have patience with you. I mean,

02:11:01.220 the Kythera experience was that. I mean, good heavens, I was, may still be sort of unemployable in

02:11:08.740 normal corporate settings. Yet the people there saw that I knew how to do certain things and I learned

02:11:17.440 from them and I learned how to do the things that allowed us to build unity because people were patient

02:11:25.560 and took time to teach. I learned from real experts and I don't claim to be an expert at much of

02:11:33.040 anything, but I do claim to have a deep appreciation for learning from people who are better and more

02:11:38.920 skilled than myself. So three things. Well, I'm going to leave it at that. I was going to ask you

02:11:42.660 another question, but I think this was the more interesting question. And so we'll leave it at that.

02:11:46.920 And grateful Ned for the time that you've set aside today to talk about this stuff. This is,

02:11:51.200 I think this is a super interesting topic. When I think about the pillars of longevity,

02:11:54.840 going back to your initial framework, right? Which is things that inhibit TOR, things that

02:12:00.140 target mitochondrial function, things that may set back the methylation clock, slow down time in that

02:12:06.540 regard in this problem. It's not a zero sum game and it's very likely we need to be pursuing all of

02:12:11.940 these strategies in parallel. But the data so far with respect to a very tangible problem like

02:12:18.900 osteoarthritis is pretty exciting. So I suspect there's going to be a lot of people listening to

02:12:23.080 this who are going to be very eager to follow the results of the technology that you guys are

02:12:29.380 trying to bring to market along with presumably some others down the line.

02:12:32.700 Thank you, Peter. It's been really a treat and an honor to share this with you. And I've always,

02:12:38.000 and shall always, I love your questions. They're awesome. Thank you.

02:12:43.020 Thank you for listening to this week's episode of The Drive. If you're interested in diving deeper

02:12:47.220 into any topics we discuss, we've created a membership program that allows us to bring you

02:12:51.620 more in-depth exclusive content without relying on paid ads. It's our goal to ensure members get

02:12:57.120 back much more than the price of the subscription. Now to that end, membership benefits include a

02:13:02.460 bunch of things. One, totally kick-ass comprehensive podcast show notes that detail every topic, paper,

02:13:08.460 person, thing we discuss on each episode. The word on the street is nobody's show notes rival these.

02:13:13.740 Monthly AMA episodes or ask me anything episodes, hearing these episodes completely access to our

02:13:20.400 private podcast feed that allows you to hear everything without having to listen to spiels

02:13:25.060 like this. The qualies, which are a super short podcast, typically less than five minutes that we

02:13:30.660 release every Tuesday through Friday, highlighting the best questions, topics, and tactics discussed

02:13:35.260 on previous episodes of The Drive. This is a great way to catch up on previous episodes without

02:13:40.580 having to go back and necessarily listen to everyone. Steep discounts on products that I

02:13:45.900 believe in, but for which I'm not getting paid to endorse, and a whole bunch of other benefits that

02:13:50.480 we continue to trickle in as time goes on. If you want to learn more and access these member-only

02:13:55.080 benefits, you can head over to peteratiamd.com forward slash subscribe. You can find me on Twitter,

02:14:01.660 Instagram, and Facebook, all with the ID peteratiamd. You can also leave us a review on Apple

02:14:07.880 Podcasts or whatever podcast player you listen on. This podcast is for general informational purposes

02:14:14.040 only and does not constitute the practice of medicine, nursing, or other professional healthcare

02:14:18.420 services, including the giving of medical advice. No doctor-patient relationship is formed. The use of

02:14:25.200 this information and the materials linked to this podcast is at the user's own risk. The content on this

02:14:31.360 podcast is not intended to be a substitute for professional medical advice, diagnosis, or

02:14:36.780 treatment. Users should not disregard or delay in obtaining medical advice from any medical condition

02:14:43.520 they have, and they should seek the assistance of their healthcare professionals for any such

02:14:48.560 conditions. Finally, I take conflicts of interest very seriously. For all of my disclosures and the

02:14:54.220 companies I invest in or advise, please visit peteratiamd.com forward slash about where I keep an up-to-date

02:15:02.700 and active list of such companies.

02:15:24.220 So,

02:15:31.040 you

02:15:33.280 you

02:15:33.780 you

02:15:35.860 you

The Peter Attia Drive - June 01, 2020

#112 - Ned David, Ph.D.： How cellular senescence influences aging, and what we can do about it

Episode Stats

Summary

Transcript