The Peter Attia Drive - #333 ‒ Longevity roundtable — the science of aging, geroprotective molecules, lifestyle interventions, challenges in research, and more ｜ Steven Austad, Matt Kaeberlein, Richard Miller

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

00:00:16.540 my website, and my weekly newsletter all focus on the goal of translating the science of longevity

00:00:21.520 into something accessible for everyone. Our goal is to provide the best content in health and

00:00:26.720 wellness, and we've established a great team of analysts to make this happen. It is extremely

00:00:31.660 important to me to provide all of this content without relying on paid ads. To do this, our work

00:00:36.960 is made entirely possible by our members, and in return, we offer exclusive member-only content

00:00:42.700 and benefits above and beyond what is available for free. If you want to take your knowledge of

00:00:47.940 this space to the next level, it's our goal to ensure members get back much more than the price

00:00:53.200 of the subscription. If you want to learn more about the benefits of our premium membership,

00:00:58.040 head over to peteratiyahmd.com forward slash subscribe. Welcome to a special episode of

00:01:06.480 The Drive. Today, we're introducing a new format to the podcast. It's our inaugural round table

00:01:12.380 conversation. For this one, we have gathered three brilliant minds, all former guests of the podcast,

00:01:20.000 to sit down and have a nuanced, funny, sometimes a little heated discussion about one of the most

00:01:27.900 fascinating and rapidly evolving areas of medicine today, gyroscience, also known, I guess, as

00:01:34.860 longevity science. So joining me for this episode are Drs. Steve Ostead, an expert in aging biology

00:01:41.780 and author of groundbreaking research on extending healthspan, Richard Miller, pioneer of the study

00:01:49.300 of anti-aging interventions through the interventions testing program, or ITP, which you hear me

00:01:54.480 reference a lot, and Matt Caberlin, whose expertise explores the intersection of genetics, aging,

00:01:59.980 and translational research. And Matt, of course, is famous for his work in the dog aging project.

00:02:06.060 So in today's round table, we discuss a number of things, such as the relationship between

00:02:10.360 healthspan and lifespan. And what does healthspan actually mean? Is it something we should try to

00:02:16.720 define? Can you improve one without improving the other? What has caused a surge in the public

00:02:23.360 interest in longevity science? And what major barriers are preventing longevity research from

00:02:28.400 reaching its full potential? This actually was one of my favorite parts of the discussion.

00:02:32.800 How do we evaluate the effectiveness of interventions like rapamycin, senolytics, or calorie

00:02:37.220 restriction in humans, where it's very difficult to study them for obvious reasons? Are there reliable

00:02:42.080 biomarkers or aging rate indicators that can measure biologic aging, which of course is a very hot

00:02:48.080 topic? What role do epigenetic changes play in aging? Specifically, are they causal? Are senescent

00:02:54.760 cells a valid target for longevity interventions, or has their role in aging been overstated? Are GLP-1

00:03:01.280 receptor agonists, for example, drugs like terzepatide and semaglutide, potentially

00:03:06.140 gyro-protective beyond just their weight loss effects? How do we overcome the funding and

00:03:11.700 political challenges that prioritize disease-specific research over foundational aging science? What

00:03:18.980 would it take to make longevity research more mainstream and gain broader support from the

00:03:22.800 public and policymakers? Anyway, this is a new format, this idea of doing a roundtable, so we really

00:03:28.400 want to hear from you. Is it something you like? If so, what are other topics you would like to see

00:03:33.880 for roundtables? So, without further delay, please enjoy this roundtable discussion with Steve

00:03:38.620 Osted, Rich Miller, and Matt Kaverland. Gentlemen, this is a lot of fun. I am excited to be sitting

00:03:49.980 down with you guys today. Where do we want to begin? Let me start by saying the following. The term

00:03:55.380 longevity, someone sent me something the other day that was like list of, I don't know, whether it was

00:04:00.400 how many times the word longevity was searched on Google or something like that, but it literally

00:04:04.680 looks like Bitcoin. So, we are clearly at peak longevity in terms of public interest, which for

00:04:11.420 all of you who have kind of devoted decades, plural, to this, I just want to kind of get a reaction from

00:04:17.340 each of you on what that means, why you think it's happening, and maybe even extending the metaphor a

00:04:23.140 little bit. Is there a bubble going on? We'll start with you, Steve.

00:04:25.700 It's a surprise to me that longevity has become so big, because for a long time, we tried to move

00:04:31.640 away from that in the aging field, because we were worried that people were thinking of longevity as,

00:04:36.840 well, we're going to keep frail, feeble, old people alive longer. That's what longevity meant, 1.00

00:04:42.340 when really what we were trying to do is extend health. So, I'm kind of surprised, but I think it's

00:04:47.120 because there are certain people of a certain age who've started to think about their own longevity,

00:04:51.620 and then I think there's a whole new generation of tech entrepreneurs that really feel like this

00:04:58.200 is a problem that will allow them to live healthily for several decades, at least longer than they are

00:05:05.420 now. So, I think it's a combination. It's a multi-generational thing. That kind of surprises me.

00:05:09.920 And you haven't seen this before, to be clear. So, 30 years ago, you didn't see glimmers of this?

00:05:14.400 30 years ago, I would have said, let's not even say the word longevity. Let's say healthspan. But

00:05:20.780 that's changed quite clearly, as more and more people have been from the outside. They're sort

00:05:25.800 of peeking in at the field. I don't think the people in the field itself have changed the way

00:05:29.860 they talk that much, but the people eavesdropping on the field certainly have.

00:05:33.800 Rich, is that your experience?

00:05:35.620 Well, I think there are two aspects that I would want to emphasize in response both to your question

00:05:40.100 and to what Steve said. In response to the question, I think people have always been

00:05:44.760 fascinated for millennia on things they could do to stay alive and healthy as long as possible.

00:05:51.660 But there were actually scientific discoveries in the 90s that showed that it could be done.

00:05:56.980 And then in the last 20 years, there's evidence that it can be done, at least in mice, with pills.

00:06:01.780 So, that naturally should lead to speculation that there could be pills you could give to people

00:06:08.480 that would postpone poor health for a substantial amount of time. 20% to 30% is what we're seeing

00:06:14.820 in mice, and 20% to 30% would be very important for people. So, I think that is a part of it.

00:06:21.720 The other part is that there are now people who are making a lot of money by selling stuff

00:06:26.640 that is untested to be polite about it or is useless to be less polite about it to gullible customers.

00:06:33.480 And so, people who want to make a lot of money have finally found that there's an impetus that will

00:06:39.480 allow them to sell stuff even if there's no evidence that it works, that they control an enormous amount

00:06:46.040 of advertising dollars, both formal and informal. That's a big part of the difference.

00:06:51.100 The one comment I wanted to make with regard to something Steve said has to do with the alleged

00:06:56.500 balance between health span and life span.

00:07:00.480 It's become fashionable for the last 20 or 30 years to imagine that you get one or the other,

00:07:06.580 that you have to make a choice, it's a decision, and that if you give up on life span, that allows

00:07:12.700 you to extend health span. I think that's ridiculous and controverted by all the available evidence.

00:07:20.480 That is, all of the drugs, at least, that extend life span in mice and could potentially do so in people,

00:07:27.920 do so by postponing diseases, both the diseases that will kill you, that's why they extend lifespan,

00:07:34.540 and the diseases that won't kill you, but which will annoy you and make you very unhappy to be old.

00:07:40.140 Which is true, by the way, of non-molecular tools as well. I mean, that's true of exercise.

00:07:44.440 Yeah, absolutely. That's a good point.

00:07:46.020 Not being insulin resistant.

00:07:47.220 I agree with you. So the notion, it's time to put behind us and to make fun of the notion that I'm

00:07:53.500 not interested in life span, don't put me on that boat. I am interested in health span because they

00:07:59.160 are linked together and they go up and down together. Getting people disabused of that

00:08:04.040 false metaphor, the seesaw metaphor, is probably an important goal for the public interface between

00:08:10.380 longevity scientists, aging scientists.

00:08:12.400 Now, I just want to push on one thing though. You talked about obviously the discoveries of

00:08:17.040 molecules. You've been personally central to that work, but there was still a lag, Rich. I mean,

00:08:24.900 it was 15 years ago, the first ITP was published showing the overwhelmingly surprising and positive

00:08:34.280 results of rapamycin. Those results were repeated. Why a decade, let's be generous and charitable and call

00:08:40.800 it a still decade long lag from that. And by the way, I'll throw one more thing in there. If you go

00:08:45.780 back to Cynthia Kenyon's work, which may have been the thin end of the wedge into the idea that

00:08:51.640 lifespan was malleable, albeit through a genetic manipulation in a less relevant model, there's

00:08:57.760 still a lag. Do you buy Steve's argument that it's a confluence of technology, tech entrepreneurs?

00:09:04.400 Let me answer your question first. Why the lag? I think there's a whole batch of reasons and

00:09:09.620 they're important and they're easy to spell out. One is the prevailing attitude is that aging is

00:09:15.280 there. There's nothing you can do about it. I'm gonna not be able to outwit aging, though I may be

00:09:21.300 able to be maybe healthier in my older years. The notion that aging is not malleable, though wrong

00:09:28.260 and provably wrong, is still the overwhelming opinion even of reasonably educated scientists and

00:09:34.460 certainly of the lay public. Then commercially, there are companies that make a ton of money

00:09:40.720 selling stuff that doesn't work by pretending with a wink and a nod and a lawyer that it might

00:09:46.580 slow the aging process down. And since they can make a lot of money, they don't actually have to spend

00:09:52.120 valuable marketing dollars on doing research and stuff to prove that it works. Some of the drugs that,

00:10:00.280 at least in the hands of our mouse group, the ITP interventions testing program, some of the drugs

00:10:06.340 are the patent is owned by another company or they're out of patent or it's a natural product.

00:10:11.300 None of that says, take me to whoever owns a big pharmaceutical firm. And also, even if you do it right

00:10:18.260 and you really want to do it and you've got a very large budget, it's not an overnight kind of thing.

00:10:23.440 Any one drug, a leading agent that, like rapamycin, which you mentioned, and the half a dozen others

00:10:29.500 that we've shown work, at least in mice, finding something in that same family that works really

00:10:35.480 well, that is safe for people, that's the member of the 20 congeners of that drug that's best and

00:10:43.780 most potent and safest, that's not at all trivial. That takes a long time and it takes a commitment

00:10:49.620 of money and time and effort and intellectual resources. We're at the place where we can start

00:10:54.400 to make an argument that that's a good idea, but making an argument that that's a good idea to people

00:11:00.520 who actually have the resources to carry it out as not so far been enormously successful, unfortunately.

00:11:06.240 Can I push back a little on what Rich said about healthspan versus lifespan? There's several papers

00:11:11.040 that have come out recently showing that the gap between healthspan and lifespan in people

00:11:15.500 is actually increasing, and it's increasing the fastest in the United States, and it's increasing

00:11:21.300 faster among women than men. So in humans, this is a very real gap, and it's a growing gap,

00:11:28.760 and I think one of the advantages of the kind of geroscience, the stuff that we do, is that Rich

00:11:33.680 is right. We don't see this in our experimental systems. So this, to me, emphasizes the fact that we

00:11:39.680 need to change the focus. I think one of the reasons that the gap exists is we're getting better and

00:11:45.040 better and better at treating heart disease and cancer and all these things and keeping people

00:11:49.120 alive when they wouldn't have been alive 10 years ago. But this is a really important factor,

00:11:54.640 I think, about thinking of public health globally. But I think you're both right. I think you're

00:11:59.300 looking at it from different angles. So Steve, you're pointing out that you can make people live

00:12:03.720 longer when they're sick. I think what Rich is saying, which I agree with, and hopefully I'm going

00:12:07.900 to paraphrase you correctly, which is if we target the biology of aging, I haven't seen anything

00:12:12.980 to make me believe that you can separate healthspan and lifespan, meaning that I haven't seen things

00:12:17.540 that slow aging, increase lifespan, don't increase healthspan. I don't actually think that's plausible.

00:12:23.800 And I think that's an important point, that if we target aging, we're doing something different

00:12:28.240 than the way that medicine is operating now, which is targeting individual diseases after they occur.

00:12:34.700 This is a very important point. It came up in a recent podcast that I did with Sam Sutaria

00:12:39.320 talking about healthcare costs. And in that discussion, one of the things that emerged,

00:12:43.980 which I think most people are sadly familiar with this statistic today, is that among the OECD

00:12:48.380 nations, the United States has the lowest life expectancy, which is ironic given that we are

00:12:52.980 spending on average about 80% more, and in some cases, double what most other developed nations

00:12:58.960 spend on healthcare. So how do you reconcile this? Well, Sam made a very interesting point, which is

00:13:03.480 aggregate life expectancy. But why is that the case? That's because the United States has by far

00:13:10.800 the greatest rate of death in middle age. So when you look at maternal and infant mortality,

00:13:16.880 we're horrible. When you look at gun violence and suicide and homicide, we're horrible. And most of all,

00:13:22.780 when you look at overdoses, we're horrible. When you kill a whole bunch of people in their 40s and 50s,

00:13:28.760 you cannot have a very high life expectancy. Understood. But what Sam pointed out was once 0.91

00:13:34.400 an American reaches the age of, and I forget the exact age, I think it was about 65, all of a sudden

00:13:39.860 they jumped to the top of the list. That was very interesting to me. In other words, if you look at

00:13:45.660 the blended life expectancy, we're not doing very well. But if you look at life expectancy, just in

00:13:51.480 measured as years alive, once you escape those big causes of death in middle age, we actually do quite

00:13:57.660 well. And it comes down to what you're saying, which is we get very good at delaying death in

00:14:01.920 chronic disease. That's what I call the medicine 2.0 machine at its absolute finest. We are going

00:14:07.680 to keep you along an extra six months. Once you have cancer, we are going to get you through that

00:14:12.440 third revascularization procedure. And so now the question is, because my intuition is where yours

00:14:18.460 is, Steve. I don't think we're getting any healthier, even if we're incrementally figuring out

00:14:23.860 ways to extend life in the face of chronic disease. I don't see it being a quality of life.

00:14:28.680 Now, part of this might be, how do we define health span?

00:14:32.920 Yeah, I agree with you. And I think it's even worse, though, than the way you laid it out.

00:14:36.780 So if you look at the statistics, if you accept that 60% of Americans have at least one chronic

00:14:41.580 disease, and the median age in the United States is 38 point something, and then you think about how

00:14:46.940 long are people living on average, that would suggest if you say that, and again, this is what you're

00:14:51.960 getting at with the definition of health span. I would not define health span as ending once you

00:14:55.460 have your first chronic disease, but that's the definition most people would use. If you use that

00:15:00.120 definition, most people are spending three decades or more in the absence of health span or in six

00:15:06.340 span. So the situation is even in the United States where life expectancy is relatively short

00:15:12.280 compared to other nations, a big chunk of that life expectancy is not spent in good health. And it's

00:15:17.800 exactly for this reason. But there are two different issues that are being confused here in the

00:15:23.680 discussion. One is the issue of whether you can help middle-aged people live longer. And everybody's

00:15:32.500 agreed that we're getting better at that. We're pretty good at it. And that certainly contributes to

00:15:37.220 whatever you think health span might mean. That's an issue, however, that is quite different from a

00:15:44.540 concoction that slows aging, do so by extending health span. Those both have the word health span in

00:15:52.360 them, but they are different and shouldn't ever be confused with one another. The other point in this

00:15:59.020 question you asked was, what is health span? My own personal answer to that is it's a useless term. That is

00:16:04.620 because no one can define it. It's not because no one is smart. It's because the term itself is vacuous

00:16:11.300 and nebulous. If you have somebody that gets a certain chronic disease here and then another

00:16:16.140 one, and then they fall down and bump their head. And by the way, they go to the hospital and with

00:16:20.580 COVID, et cetera, et cetera, defining when in that 20 to 30 year period, they flicked the switch. Now

00:16:26.900 they have gotten to the end of the health span is impossible and of no interest. The general notion

00:16:33.980 that people are interested in is whether you can do stuff to keep people healthy for a long time,

00:16:39.720 either without changing their life expectancy or by changing their life, with changing their life.

00:16:45.140 Those are interesting, but you don't have to assign a number, a health span digit.

00:16:50.160 I don't like the medical definition of health span, which I believe is quote,

00:16:53.920 the period of time in which an individual is free of disability and disease. I find that to be a very

00:16:58.360 unhelpful definition.

00:16:59.500 It's awful, yeah.

00:17:00.320 But part of the reason it's awful is it's binary. 1.00

00:17:02.500 Yeah, exactly. You got it.

00:17:04.140 But if we made it analog instead of digital, I'm not saying that makes it easy. It's still

00:17:09.240 very challenging, but now it allows us to start talking about things.

00:17:13.400 Except it's a concept. It's a qualitative concept. I think we should try to make it to where we can

00:17:18.980 actually come up with a way to measure whether we call it health span or not. That doesn't really

00:17:22.980 matter. I kind of agree with Rich. I agree with what you're saying, except I think it's a really

00:17:27.480 useful term as a concept. I think it's a really useful way to communicate to a broader audience

00:17:33.740 what one of the goals is, which is to increase the healthy period of life.

00:17:37.200 Yeah, I kind of like the term health for that. I have a way that helps you out with your health

00:17:43.780 and you don't have to pretend you can define it as a number.

00:17:47.280 But I think we all could agree there's a period of life where you are in relatively good health,

00:17:51.420 and then there's a period of life where you aren't. And so I think the idea that

00:17:54.340 we're trying to increase that component of life is really important. So I don't think we're

00:17:59.800 actually disagreeing on much other than whether we like the word.

00:18:02.840 Well, I also think there's an individualization of this that we're missing. To me, health is a state

00:18:10.100 of your physical being that you can do the things you like to do. Therefore, if you like to climb

00:18:14.880 mountains, your health span is going to be different than if you like to play golf, for instance.

00:18:20.880 And a lot of this is personal. Yeah. If you can't run a marathon anymore,

00:18:24.960 some people will say, oh, my health is... And we never pay attention to the mental health

00:18:28.340 piece, at least the biologists don't, right? So I have a question for you, Steve. What is my health

00:18:32.300 span? I would only be able to ask you that. So we do this exercise, guys, because I completely

00:18:37.260 agree with you, Steve. We call it the marginal decade exercise. So we say to every one of our

00:18:42.260 patients, and I write about this a lot in the book, everyone will have a marginal decade,

00:18:46.200 which I define as the last decade of your life. So obviously, by definition, everyone has a

00:18:50.760 marginal decade. Most people do not realize the day they enter it, but most people have a pretty

00:18:55.820 good sense when they're in it. Okay. So the exercise we do is we go through with the patient

00:19:01.400 and we say, what are the things that are most important to you to be able to do in your marginal

00:19:06.360 decade? And they generally fall into three buckets with a sub-bucket, physical, cognitive,

00:19:11.980 emotional, social. The physical bucket, we kind of divide into activities of daily living and

00:19:18.040 recreational activities. So that's where, again, most people obviously intuit that, boy, I would

00:19:24.760 really not be happy if I couldn't take care of myself. If I couldn't get out of bed, get dressed,

00:19:29.840 shave, cook, that would be disappointing to me. But then of course, you have different levels of

00:19:34.620 ambition within the recreational side. I've got patients who say, when the day comes that I can't

00:19:39.480 heli-ski, I'm going to be devastated. And other people are like, I just want to be able to garden.

00:19:44.720 That's going to create a very different standard. On the cognitive side, you have people who say,

00:19:49.260 I want to be able to run my hedge fund and still make money and make really important investment

00:19:53.220 decisions. And other people are like, I want to be able to do crossword puzzles and read the

00:19:56.640 newspaper. I agree with you. You can't define it, but it doesn't mean we shouldn't try to personalize

00:20:01.760 it. Okay. But I want to come back to you, Matt, with the original question. Why are we at a point

00:20:07.420 where... Why has longevity gone mainstream? Yeah. Yeah. Yeah. For lack of a better way to speak.

00:20:11.620 Yeah. So, I mean, I think both of the points that Steve and Rich raised are part of the equation. I

00:20:17.540 mean, I think it's a convergence of all of these factors and maybe a few others. I do think the

00:20:23.280 science has matured to the point where more people are believing that we can actually modulate the

00:20:30.060 biology of aging. I think the concept of biological aging has become popularized through a variety of

00:20:35.980 mechanisms, including some influencers, individuals who I personally think often err on the side of

00:20:43.020 being a little bit less scientific than they should be, but I think they've helped popularize

00:20:47.020 the concept. So I think it's been a combination of these factors. And why it has taken so long,

00:20:52.000 I mean, I just think that's the pace that science moves and the rate at which these concepts can sort

00:20:56.920 of permeate the public sphere. So it's frustrating in a sense that it's moved so slowly. I also wonder,

00:21:04.280 because you sort of said, are we at a longevity bubble? I don't know. I think maybe we're still

00:21:08.400 kind of in the early days of this hockey stick moment where you're getting this exponential

00:21:13.120 increase in attention. My hope is as we go forward, it will become more scientific and less

00:21:19.060 snake oily. And it's a spectrum. There's this huge gray area in the field right now of what's real and

00:21:24.600 what's not real. And I think none of us at this table actually can really define exactly where in

00:21:30.500 that gray area that line is, or is there a line? To that point, Matt, what is the collective

00:21:35.780 wisdom of the group on the funding appetite for that? Because I agree with you completely. Like

00:21:41.360 if we could channel this exuberance away from kind of the highly commercial speculative grifting

00:21:49.840 towards the budget increasing legitimate investigative, that would be awesome. What is the appetite right

00:21:58.280 now of NIA with respect to this? I think it's hard to say it. And I mean, NIH is a moving target. And as

00:22:05.020 we all know, there's going to be a lot of change coming in the near future. So cautiously optimistic,

00:22:10.640 I would say if you look historically, it's been really pretty terrible. The percent of NIH budget that

00:22:16.600 goes to biology of aging, I think is still probably around half of 1%. Sorry, just to put numbers in

00:22:22.500 perspective. NIA gets what percent of NIH? Not NIA. No, no, I understand. Within NIA, there's a sub

00:22:27.880 fraction that goes to biology of aging, right? Yes, yes, yes. Yeah. But I'm saying there are 17

00:22:31.720 groups of NIH. NIA being one of them gets what fraction of NIH budget roughly? I think it's a

00:22:36.820 roughly 3%. 3% of NIH budget is NIA. Within NIA, how much goes to this type of research? It was about

00:22:43.800 350 million a few years ago. It might be a little higher than that, but I don't think it's ticked up

00:22:48.680 any more proportional to the increase in NIH budget since then. So it reaches about half of 1%.

00:22:54.380 Wow. What's your level of optimism, Rich? You're obviously very close to this.

00:22:59.740 That NIH will wake up and start to pay attention to aging research the way they should? It's near zero.

00:23:05.540 It's been near zero for 30 years now. Even with this outside attention?

00:23:10.860 Well, it's gone up. I mean, they funded the ITP, the interventions testing program 20 years ago,

00:23:15.420 and they liked it, and they doubled our budget about 15 years ago. So that's something. And I'm

00:23:21.000 very, very grateful to them for that. But there's still an enormous untapped potential for making

00:23:28.260 progress in the basic biology of aging. And the reason is, again, a matter of defending turf. If you

00:23:34.780 are a cardiologist researcher or an oncologist researcher or an AIDS researcher or an Alzheimer's

00:23:40.560 researcher, anytime somebody says, the smart play is to reduce your budget by 10% or your institute's

00:23:47.040 budget by 10%, we're going to go there faster if we spend money on aging and its relationship to the

00:23:52.580 disease you care about. You get the porcupine defense. You don't take any of my money because

00:23:58.400 Alzheimer's is important. Little kids with leukemia are important. Breast cancer is important. You go away.

00:24:04.220 And that is the predominant feeling. Most of the people making those decisions were not trained

00:24:09.980 in aging research. They view it as something interesting. I read something about that in

00:24:14.280 Time magazine the other day. But they don't understand that to actually conquer or slow down

00:24:20.520 or affect or protect against the disease they care about, the smart play is to do aging research.

00:24:27.200 And so they view your suggestion, which I, of course, agree with 100%, as an imposition,

00:24:32.320 an invasion to be repelled at any cost. No one in a position of power has had whatever it takes to

00:24:40.280 reverse that. And if he or she tried to do that, Congress would, even a good Congress, would smack

00:24:46.880 them down. The Alzheimer's group has 100 lobbyists. The cancer group has 100 lobbyists. The AIDS group

00:24:52.600 has 100 lobbyists. The aging group has two lobbyists, one who's a lawyer and one who takes the calls.

00:24:58.820 And that's not enough to do it.

00:25:00.120 Can I just add something real quick? I agree completely. And I think as well,

00:25:04.460 the reputation of the field has hindered that transition as well. So historically,

00:25:09.560 the field was viewed as not very mechanistic, kind of phenomenological, became much more

00:25:15.080 mechanistic starting around the time of Cynthia Kenyon's work and since then, but has continued

00:25:21.440 to have a reputation problem as not being as rigorous as other areas of research. So I think it is

00:25:28.120 absolutely a turf war. And there's this overcoming the reputational problem, which makes it harder

00:25:33.380 for serious people in funding and policy circles to give it the attention it deserves, in my opinion.

00:25:41.060 So I've got a different take on this. I actually think that this is a very good time for aging research

00:25:46.400 funding. And that's not because of what's going on at the NIA, but it's what's going on in the private

00:25:51.660 sector. There's more and more money. There's even interest now in big pharma that was very spotty

00:25:57.680 in the past. So I think if we focused entirely on the National Institute on Aging, we would get a

00:26:04.000 false impression of what the funding climate is in the field now. And I think we need to take

00:26:09.840 advantage of that. Got to make sure that it doesn't get captured by the people who are doing

00:26:14.560 the flashy, but bad science. You're saying, look, Calico, Altos, other private companies,

00:26:22.540 especially within biotech and pharma that are looking at geoprotective molecules building on

00:26:26.920 the work of the ITP. Yeah, I think it's safe to say the amount of money that's being spent privately

00:26:32.380 probably outdoes public spending. I mean, in a given year, two to one easily.

00:26:38.560 It could, although how much of that is actually going to biology of aging? I think it's still an

00:26:42.780 open question. You mentioned Calico and Altos, right? Yeah, we don't know exactly.

00:26:45.480 I actually agree with Steve. I don't think what Rich and I were communicating is opposed to what

00:26:50.240 Steve was communicating. There are a lot of opportunities right now. And again, this is

00:26:53.820 sort of what I was alluding to is, are we at the beginning of this hockey stick moment? And I think

00:26:58.900 Steve's right. There are real opportunities for more resources to be focused on the scientific side

00:27:04.940 and hopefully less focused on the non-scientific aspects of what are going on. And you asked the

00:27:10.920 question of, can we shift resources from the more consumer facing, maybe not as rigorous stuff and

00:27:16.940 into the more rigorous stuff? I'm not a fan of that stuff at all, but maybe you need that stuff to kind

00:27:22.340 of move the needle and get people's attention. And at least people are talking about longevity now.

00:27:27.560 Naive question. I'm embarrassed. I don't know the answer because I spent more than two years

00:27:31.780 working there. What's the mission statement of the NIH?

00:27:35.280 It's to preserve and enhance human health. I mean, it's basically the same thing that

00:27:39.020 we do that we're supposed to be doing. Yeah. And I didn't actually get to give you my

00:27:43.360 spiel here, but what I started to say about the NIA budget is if you look at the major causes of

00:27:49.520 death and disability, and it's, again, we talked about how it's hard to define health spend. So

00:27:52.640 if we just look at causes of death, if you look at the top 10 causes of death in the United States,

00:27:57.160 nine of them have biological aging as their greatest risk factor. And it's not even close yet.

00:28:03.140 Half of 1% of the research budget that's supposed to be focused on improving human health

00:28:09.020 goes to study that risk factor. I mean, I think it is extremely frustrating to all of us sitting at

00:28:14.640 this table that that hasn't changed, but there's reason to be optimistic that maybe it will change

00:28:19.480 in the near future. Let's state that again because it is so profound. I want to make sure not a single

00:28:26.020 person missed that statement. The top 10 causes of death in the United States are well enumerated

00:28:32.680 and incredibly predictable and they increase by category by decade, three to 8% monotonically

00:28:40.240 with no exception. Point being 90% of, and more than 90% on an adjusted basis of what causes death

00:28:47.940 goes up with age. And yet a few basis points of federal R and D goes to addressing that.

00:28:55.980 Let me give you an example of what the sort of point that Matthew and you have been making

00:29:00.600 about once every five years, I give a talk and invite a talk at the University of Michigan

00:29:05.900 Cancer Center. And I point out that we have drugs now, anti-aging drugs in mice, and they extend

00:29:12.640 mouse lifespan and they do it mostly by postponing cancer because most of our mice die of cancer.

00:29:18.100 And if you look at age-adjusted cancer incidence rates, our drugs reduce these by a factor of 10.

00:29:24.580 Wouldn't they like to know why? As cancer scientists, we now have a batch of drugs that

00:29:30.220 postpone cancer. Wouldn't they like to study them? Invariably, I get one call back from somebody

00:29:37.320 who says, that's interesting. Maybe we should talk about that. And then it dies. And then five years

00:29:41.840 later, I'm asked to give the same talk or a related talk. So they know how to do cancer research.

00:29:48.020 They are cancer scientists. That's how they know how to do cancer research. And you certainly don't do it

00:29:52.580 by diverting your lab's attention to aging. That's insane. But that insanity is how medical

00:29:59.680 research is organized. And breaking that addiction to the kinds of models you grew up on because they're

00:30:06.440 a better idea, it's not an easy thing. It may not even be a possible thing to do. That's a major hassle.

00:30:13.200 I think this is because we think about health all wrong. We think, let's wait to get cancer and see

00:30:18.760 what we can do about it. That's what cancer biologists do. You have cancer, okay, how can we

00:30:23.300 better treat that? Or could we have diagnosed it earlier? What Rich is saying and what we can know

00:30:29.060 how to do in lots of model organs, it prevents you from getting cancer, delay it for a considerable

00:30:33.880 amount of time. That's a little bit harder to study if you're a cancer biologist because you want to see

00:30:39.440 the cancer before you can study it. I think that's why we need aging biologists rather than people

00:30:44.780 focused on certain disease to come and try to use what we do. You know, if we prevented the cancers,

00:30:50.760 they'd be out of the job. I guarantee these people or mice will get cancer. They'll just have 0.97

00:30:55.900 10 extra years of life if they're a person or 10 extra months of life. They'll get cancer. They'll

00:31:01.260 need specialists. It'll be all right. Yeah. I think that's important. I mean, I think the reactive

00:31:04.820 disease care component is still going to be there. Even if we're insanely successful at slowing aging,

00:31:10.540 people are still going to get sick. But I think Steve's point is really important. Like Peter,

00:31:14.420 you've been a leader in helping people recognize the need to shift the medical approach from reactive

00:31:21.120 to proactive. I think what a lot of people don't realize is that mentality goes all the way back

00:31:26.060 to pharmaceutical research, biomedical research, basic science that is ingrained at all the way

00:31:31.940 through. And I think one of the challenges with getting funding for aging research is that mentality

00:31:38.240 on the basic science world and how deeply ingrained it is. It's very interesting because you don't know

00:31:44.280 which is the tail and which is the dog. I've always assumed that the one leading the charge

00:31:50.700 is the clinical side of things. In other words, the engine, the machine of medicine 2.0 is built

00:31:56.940 around the delivery of care. The delivery of care, as you said, Steve, is built around,

00:32:02.820 I'm going to wait. I'm going to sit here and hang. We're going to wait. When you get the disease,

00:32:06.480 we're ready. You had the heart attack? Fantastic. You've got chest pain, ST elevations.

00:32:10.840 We got a stent for you. Now you have cancer. We're all in. And then the research flows from

00:32:16.640 that mindset. Of course, I don't know, not that it really matters, but it might be that it's flipped,

00:32:21.380 right? It might be that the clinical engine behaves in that way because that's how the base

00:32:26.400 of the pyramid has been built. Again, not that it necessarily matters, but if you could be health

00:32:31.680 czar and fix one of them, you might actually start with the research side of things.

00:32:37.220 I would. And I mean, the reality is the research flows from where the dollars are going.

00:32:41.420 This has been seen over and over and over at NIH. You shift resource allocation to a certain area

00:32:46.500 and the scientists will follow and they will submit grants to get grants in the place where

00:32:51.040 the funding line is the highest. So if somebody came along and said, we're going to go from 0.5%

00:32:55.480 to 50% of NIH budget is going to go to biology of aging, you'd have no shortage of people. I mean,

00:33:00.300 it'd be kind of messy at first, but you'd have no shortage of people applying for grants and

00:33:03.900 becoming experts in the biology of aging. And the system would work. You'd get the best

00:33:07.640 and the brightest that would go into that and do that. So this then begs another question that is

00:33:14.100 a tired question, but I can't help but ask it at this point. Is aging a disease? Is that even a

00:33:20.320 relevant question? Call me, call me, call me, call me.

00:33:22.780 Yes, Mr. Miller.

00:33:23.560 Call me. It's important to use words optimally and to distinguish causes from effects.

00:33:30.000 One of the bad things about aging is it's a risk factor for many diseases. Some things,

00:33:36.680 other risk factors for diseases. Aging is a risk factor for disease. And so saying that aging is

00:33:42.940 a disease confuses that discussion. It makes it impossible to see that relationship. So calling

00:33:49.660 aging a disease is a fundamental error. The question itself is incorrect.

00:33:54.060 I agree completely. I think it's the wrong question.

00:33:55.900 I agree. But I think we have that idea for marketing purposes, not for scientific purposes.

00:34:03.800 And the idea is, well, the money goes to diseases. Let's call aging a disease. Because I think what

00:34:09.760 we're trying to do is we're trying to treat aging as if it were a disease, even though I would agree

00:34:16.220 with both of you. I don't think it's a disease. I think that destroys the word disease if we include

00:34:20.800 aging in it. But I think there was a reason that suddenly this came because you thought,

00:34:25.140 oh, maybe this will get Congress to pay attention to it.

00:34:27.420 You're right. It's a marketing ploy. And if you think you can convince people of the importance

00:34:31.660 of aging research only by crossing your fingers and saying, oh, well, it's kind of a disease,

00:34:37.960 isn't it? You think you can fool them? Yes, that's what marketing is. And it's probably good for that.

00:34:42.600 I just don't like lying to people.

00:34:44.440 It also creates a negative feeling about the field in some people as well. So I think that should be 1.00

00:34:50.120 considered. The other point that people often raise, though, is we have to call aging a disease in

00:34:54.400 order for FDA to approve a drug for aging, which I think is a fundamental misunderstanding of how

00:34:59.640 FDA operates. But that is the other argument you will often hear among proponents of the idea that

00:35:05.140 aging is a disease.

00:35:07.480 Very interesting. Well, so now let's go one step deeper on that. How do you think about

00:35:13.380 biologic versus chronologic age in concept and in practice?

00:35:18.740 On the ride over here, Rich and I were talking about that. I don't believe there is one thing as

00:35:24.660 biological age. I think there is potentially an age of your heart, an age of your liver, an age of your

00:35:31.340 lungs, an age of your brain. But I don't see why we wouldn't simply call it health. In other words, I got

00:35:39.360 one of these epigenetic age clocks done on me a while ago, but I didn't know what to make out of it.

00:35:44.760 I thought, is this just flattery? Or did it really tell me something?

00:35:48.560 You must have got a good result.

00:35:49.500 I got a good result.

00:35:50.380 He's 13. 13 years old.

00:35:52.600 That may be the point of the whole thing, right? So I'm dubious about some number that is different

00:35:59.300 than, I know I'm in good health. For my age, I'm in very good health. So I knew that already. Now I

00:36:04.420 have a number for it. I don't put much credence in that.

00:36:06.960 Let me agree with Steve, but just put it in slightly different terminology. It's a matter of taking a very

00:36:13.140 rich, complex data set and trying to collapse it to a number. So if someone wants to know how healthy

00:36:19.540 I am, he or she would need information. How good is my eyesight? How good is my hearing? How good is

00:36:25.800 various kinds of cognitive activities? My aerobic endurance, my joints, all of that is pertinent to

00:36:32.280 how my health is and also about projected future health. Then there's no need, once you've got that

00:36:37.780 information, which is very rich to say, ah, there's a number, a single number, a real number on a point

00:36:43.400 on the number line that condenses that in any useful way. A notion 40, 50 years ago that biological

00:36:51.020 age was not the same as chronological age for a little while was useful. It emphasized that there

00:36:56.380 might well be 60-year-old people who were unusually like youthful people and 60-year-old people who were

00:37:02.900 unusually like 70-year-old people. Would my drug or my genetic mutant or whatever help to discriminate

00:37:10.140 those people or change them in some way? I can slow your biological aging process. That's a discussion

00:37:18.020 that was maybe of interest 40 years ago. It's now time to drop the notion, let alone the silly notion

00:37:25.760 that you can count that biological age, that number which some people, too many people still think is a

00:37:32.140 value. You can figure out what it is by measuring something, transcriptions or epigenetic markers or

00:37:38.160 something. I can do it and give you personally your personal biological age. That's a waste of everyone's

00:37:45.300 time, and it also distracts attention from things that actually are important and need to be thought about.

00:37:50.460 I got to talk because I think I disagree fundamentally, and I'm surprised, but this will be an interesting

00:37:55.320 conversation. I agree that the idea of a kit that you can buy to measure biological age, first of all,

00:38:02.320 the stuff that's out there doesn't work, and we can and should talk about that. But also, I sort of agree

00:38:07.180 with the idea that reducing it to one number, while conceptually I think it's possible, I think in reality

00:38:13.280 is going to be really, really difficult to do. But do I believe that there is a biological aging process

00:38:18.320 that is different from chronological aging? Absolutely. Oh, yes, absolutely. Okay, well, it

00:38:22.660 sounded like you guys were both saying no, you didn't think it was a real thing. No, no, no, no. I agree

00:38:26.180 with that completely. You can agree with that and not like the idea of a number that constitutes your

00:38:30.860 biological age. Sure, sure. Okay. There's two things that kind of make me feel pretty confident in this

00:38:35.760 idea. One is, and this is the example I use a lot among the general public, is just look at dogs

00:38:41.500 compared to people. Everybody's familiar with the idea that one human year is about seven dog years.

00:38:47.080 What does that mean? It means that dogs age about seven times faster than people do. But of course,

00:38:52.080 chronological time is the same between dogs and people. It's the biological aging process. And so

00:38:57.680 you can look across the animal kingdom and see this. And dogs get almost all of the same diseases

00:39:02.380 and functional declines that we do at the tissue and organ level, but also the whole body level.

00:39:07.180 We also know now there are single genes that significantly modulate what I would call the

00:39:12.720 rate of aging. Now, maybe we have a different meaning to what we mean that. No, I agree entirely.

00:39:16.360 So the fact that that's possible, DAF2, we talked about DAF2 a couple of times,

00:39:20.680 TOR. We can turn these things up, turn them down, and animals across the evolutionary spectrum

00:39:26.760 seem to age at different rates by modulating single genes. So I don't know of any other explanation

00:39:32.600 other than that there is this process, which we call biological aging, that can be changed. And the

00:39:39.840 rate can be sped up or slowed down. Can it be reversed? That's an interesting question. Maybe we'll get to

00:39:44.520 that. But I think the process is real. I think it's just really, really complicated. And we probably

00:39:48.740 only understand 5% of it at this point. Yeah. I think for me, the challenge is, I kind of land

00:39:54.800 where Rich was, which is if a patient says to me, hey, why aren't you doing this biologic age clock on

00:40:02.020 me? My response is, well, I know your VO2 max. I know your zone two. I know your muscle mass. I know

00:40:09.740 your visceral fat. We did a very complicated movement assessment on you. I understand your

00:40:14.760 balance. I understand your lipids, your insulin. Like I know these 57 things about you and I can

00:40:21.200 tell you individually on each of them how you're doing. That number doesn't tell me a single new

00:40:29.240 piece of information. But what if you were to come up, and you probably do this in your head,

00:40:33.720 you come up with some sort of composite. You probably don't sit down and wait each of those

00:40:37.320 things and come to one number. But you come up with some sort of composite picture of health

00:40:41.400 based on all of those things. That's a different biological aging clock. I think sometimes we

00:40:46.540 conflate, and in part this is because of the way that irresponsible people in the field and marketers

00:40:51.340 have done this. We conflate the epigenetic tests with biological aging clocks. There are all sorts

00:40:58.080 of flavors of biological aging clocks, including things like frailty indices or metrics of a whole

00:41:03.180 bunch of functional markers. I think those probably are pretty good readouts of biological age. Again,

00:41:10.540 can you combine them all to get to one number that's meaningful for every person? That's much

00:41:15.320 harder to do. Yeah. Tell us about your experience because you did what I wanted to do, but I've been

00:41:20.720 too lazy to do. Yeah. In fact, we exchanged emails at one point about doing this and each coming up with

00:41:25.440 different names. So what I did was I tested four different direct-to-consumer biological age kits.

00:41:31.740 They were all epigenetic biological age tests, but four different companies. And I did duplicates of

00:41:37.320 each kit, and it was from the same samples collected on the same day. Really tried to put my scientist hat

00:41:43.300 on. I only had two replicates. I didn't have three replicates, but it was about the best I could

00:41:46.700 afford at that point. And it was kind of expensive. So anyways, sent those in, got the results back,

00:41:52.600 and they were, to me, very informative. Fundamentally sort of changed my views on these epigenetic age

00:41:57.740 tests. So they ranged from 42 to 63. I was 53.75 years at the time I did the test. And the standard

00:42:07.980 deviation, I can't remember, was either seven or nine. So mean of my chronological age, standard

00:42:12.940 deviation of seven or nine, which I look at that data. I'm not a statistician, but I know enough

00:42:19.200 statistics to say that's completely useless. They converged on my chronological age, but with a

00:42:24.060 huge variation. Even intra... So that varied between the tests. So I think three of the four

00:42:30.280 were reasonably close to each other. Three of the four companies, the duplicates were reasonably close

00:42:35.220 to each other, but the individual tests were far apart. And one of the companies, the individual

00:42:39.720 replicates was 20 years apart. So to me, and some people will say, but maybe the true diagnostic test is

00:42:45.800 great, and the Elysium test is terrible, or the Tally Health test is terrible, and the other one

00:42:50.380 is great. Maybe, but how do we know? My take-home is that the direct-to-consumer biological age testing

00:42:57.960 industry is a complete mess. And I have no idea who to believe, or if any of them are actually giving

00:43:05.480 accurate data. I know some of the people at some of the companies, and I have my personal feelings

00:43:09.520 about who's trying to do it right, and who's sort of a charlatan. But across the industry, it's really

00:43:14.280 hard to know. The last thing I'll say on this is, where I've sort of landed is, I think these are

00:43:18.940 really good research tools. I think the direct-to-consumer component has gotten way ahead

00:43:24.180 of itself. And I think I align with what you were saying about the way you think about these tests.

00:43:29.720 I don't think there's a lot of value in clinical practice right now, because we don't know precision

00:43:34.860 or accuracy. And I don't think you can make actionable recommendations based on these tests.

00:43:40.520 Furthermore, they fail in the one thing that I think they're attempting to do.

00:43:45.820 And I usually use this illustration with patients. So if I have a 40-year-old patient who says,

00:43:51.000 I really want to do one of these tests, I say, if the answer comes back and says you're 20,

00:43:56.140 is your expectation that you will live another 70 years? Conversely, if the answer comes back and says

00:44:02.840 60, is it your expectation that you will live another 30 years? In other words, is this number

00:44:07.780 predictive of future years of life? Because right now we have this thing called chronologic age

00:44:13.780 that is the single best predictor of future years of life. So do we think biologic age as determined

00:44:20.620 by these tests is better as a predictor of future years of life? Which by the way, would be very

00:44:25.160 testable. How many people have contacted you to get ITP sample data to say, can we predict how much

00:44:31.340 longer these mice were going to live? The answer to the question is obvious and very well known.

00:44:36.980 You can tell if you have a, your 40-year-old patient and he or she is fat, doesn't exercise,

00:44:43.340 eats mostly cheeseburgers. You know that their life expectancy is probably not as good as the 40-year-old

00:44:49.280 patient in your next waiting room that has extremely healthful habits and whose parents live to be 100.

00:44:54.640 And there's tons of published data.

00:44:56.800 Right, but I don't need a biologic age to tell me that.

00:44:58.860 Right, right. That's what I'm saying. There are tons of things you can measure on individuals,

00:45:02.660 four or five of them are all you really need to ask of a 70-year-old.

00:45:05.620 Yeah, MetLife does this really, really, really well.

00:45:09.180 Because their buddies are the line there. They're writing life insurance policies.

00:45:12.880 So it's not at all hard to figure out a very small set of tests to tell you how long a 70-year-old

00:45:17.940 is likely to live. There's nothing to do with methylation clocks or things like that.

00:45:23.120 To me, that's the gold standard. When life insurance companies start using biologic clocks

00:45:28.960 as the cornerstone of their actuarial algorithms, I'll start to be-

00:45:33.300 I don't think we're that far away from that. I'm going to sound like a broken record here,

00:45:36.680 but you guys keep saying biological age when what you mean is epigenetic age or epigenetic test.

00:45:41.160 Not necessarily. And we should explain to people that there is a difference.

00:45:43.960 So some of these clocks use solely epigenetic measurements.

00:45:49.040 Not all. Most of the direct-to-consumer ones are epigenetic.

00:45:51.560 But some of these tests use a litany of biomarkers inclusive of epigenetics.

00:45:57.780 Yes.

00:45:58.160 So they'll say, we've sampled your methylation pattern, but we also looked at your vitamin D level,

00:46:02.200 your glucose level, your cholesterol level, and a whole bunch of other things.

00:46:05.860 And we compressed all of that into a number as well.

00:46:08.320 So I guess, let me frame it as a question to you. So let's take the epigenetic piece out.

00:46:12.320 Again, I do think we will get to a point where the technology is developed far enough and the quality

00:46:20.040 control is good enough on the consumer side that these tests will be better than just chronological

00:46:25.780 age. I think we can get there.

00:46:27.580 That's a big statement. I don't know that I'm disagreeing with you. I just want to make sure

00:46:30.080 we understand the statement.

00:46:30.540 I think it's clear from the research. Unless you think that all of the research that's been done

00:46:33.880 on these epigenetic aging clocks is somehow flawed, it's clear that you can create algorithms that

00:46:40.380 can predict specific methylation patterns that are more highly correlated with life expectancy

00:46:46.760 than chronological age.

00:46:48.420 But I think the big but here is that even if that's the case, they would not be as good

00:46:53.700 as what Peter would predict after all the tests.

00:46:57.120 Biological age. That's what I want to get to. Yes. And I think what you are actually doing 1.00

00:47:01.300 is looking at other biomarkers that have a long-term clinical history that you're using

00:47:07.060 to come up with a surrogate, but really is reflecting largely biological age. Maybe not

00:47:12.960 completely. And this is the other point I wanted to make is I don't think biological age and health

00:47:17.000 are equal. I think they're strongly overlapping. And certainly you can identify many ways to

00:47:22.680 reduce health without accelerating biological aging. I think that's easy. We can all think of

00:47:28.740 ways to do that. So let's take a minute and try.

00:47:30.840 Yeah. So let's think about this for a second. I have seen very impressive data where we can

00:47:38.680 look at tissue samples of organs and we can tell, okay, I'm going to show you a sample of

00:47:44.580 nephrons. And just based on nothing but the methylation pattern, we know that if I just said

00:47:50.980 to you, one of these is a 20-year-old, one of these is a 50-year-old, and one of these is a 70-year-old,

00:47:56.420 it's very easy to predict based on the methylation pattern, which nephron came from which person.

00:48:01.140 Completely agree with that.

00:48:02.300 There are a lot of things that change with age. The literature has 25,000 things that change with

00:48:06.580 age. Average amount of methylation at these 10 spots is number 11,407 of those. So great,

00:48:13.260 you've got another thing that changes with age.

00:48:15.400 So that's the question.

00:48:16.320 But that's not enough.

00:48:17.500 Right. So do you believe that all of the research we're seeing on the epigenetic clocks is going to be

00:48:24.140 the 78th variable that we would include in our gestalt?

00:48:29.660 I don't know. Yeah, it's a good question. So I am hopeful that epigenetic algorithms can get to

00:48:37.080 the point where they can replace many, certainly not all, but many of the other biomarkers that are

00:48:43.060 being measured. I think the thing that gives me hope is we know that epigenetic changes are part of

00:48:47.840 biological aging. This again is a different question, but if we look at the hallmarks of

00:48:50.940 aging, epigenetic dysregulation is one of the 12. Some people will argue it's the most important

00:48:56.340 one. That's a different conversation, but it's at least part. So that gives me some hope that we

00:49:01.460 are in fact measuring something that plays a causal role in the aging process. And I think what's missing,

00:49:07.940 I think what would give all of us a lot more confidence is if we had a mechanistic connection to the

00:49:12.040 specific methylation changes and some cause of aging or age-related disease. In other words,

00:49:18.460 this change in methylation changes this particular gene's expression level, which changes the rate of

00:49:23.320 biological aging. I think if we had that, we'd feel a lot more confident.

00:49:26.640 Yeah. You and I spoke about this very briefly at the end of our last podcast,

00:49:30.700 and I want to come back to it with all of us on this table because there's so much in what you just

00:49:36.220 said, Matt, that I'm going to lay out a broad question and then we can start attacking it in

00:49:40.220 different ways. So one of the things I want to address is, do we believe that it's possible

00:49:45.460 that of the hallmarks of aging, epigenetic change is the most important? Another topic I want to

00:49:50.980 address, do we believe that the epigenetic changes that we observe over time, which are undeniable,

00:49:55.760 are causal in the arrival of other states, everything from the arrival of senescent cells,

00:50:02.700 the increase in inflammation, the reduced function of the organs, which really is the hallmark of aging?

00:50:07.680 And if so, does that mean that reversing the epigenetic phenotype will undo the phenotype of

00:50:15.540 interest? And Rich, where I'm going that you and I left off was, what about the proteome? What about

00:50:20.780 the metabolome? So you made three statements there, broad general statements. And I think each of the

00:50:26.760 three deserves careful amendment. Let's do it. To be polite about it. The first has to do with

00:50:33.760 hallmarks of aging, which I think set the field back dramatically. I think when you are officially

00:50:39.900 branded a hallmark of aging by two people sitting alone at their computers and writing a review

00:50:45.780 article, a hallmark of aging... I thought they were walking around a pond when they came up with this.

00:50:50.320 Walking around a pond. All right. Okay. Okay.

00:50:51.240 It means that somebody once said, I'm interested in aging, that's kind of important, isn't it? Let's

00:50:56.280 put it on our list. You can't tell if something is a hallmark of aging. Does that mean it goes up with

00:51:02.680 age? It goes down with age? You can change it in a way that will extend lifespan. You can kill a mouse

00:51:08.540 or a worm by removing it. Basically, it's something that somebody once thought might be of interest to

00:51:14.780 aging. And the downside of that is once you're officially branded as a hallmark of aging,

00:51:19.980 anyone who wants to write a grant on that doesn't have to prove that their fundamental

00:51:26.260 cause and effect model has any merit because it's a hallmark of aging. I don't have to prove

00:51:31.460 it anymore. Someone, I don't know who or on what grounds, has decided it's important. My reviewers

00:51:36.400 know it's important because they've read the hallmark of aging paper, so I don't have to think

00:51:40.760 about whether it's important. The negative side of that coin is that there are lots of things that

00:51:46.200 didn't make it into the hallmark list. I really think it's premature to close thought off on some

00:51:52.520 of those. It's easy to come up with a dozen things that ought to be investigated, but if you want to

00:51:57.420 investigate it and it's not on the hallmarks list, what are you wasting? So deciding which of the

00:52:02.420 hallmarks is the big daddy hallmark or whatever strikes me as not the correct thing to talk about

00:52:09.520 in the hallmarks arena. So maybe we should talk about that before we go through all of these,

00:52:14.760 because I think there's a lot to unpack there. You'll remember the other ones to admit.

00:52:17.160 If you guys could afford to give me a little piece of paper and a pen,

00:52:20.800 then I'd be able to write down my... I think the hallmarks is a list, a kind of arbitrary list,

00:52:25.100 not completely arbitrary, because they had some reasons for being there. I don't think any of us

00:52:29.000 would say that those 12 things are not involved in aging, but that's a very little interest of itself.

00:52:35.560 Do any of us want to rattle them off, being that I'm the only one that's got the list sitting in

00:52:39.180 front of me? We could do a game where we each name one and see who can't...

00:52:42.120 See if we get to all 12. Yeah.

00:52:44.340 But certainly in that list, I would not consider epigenetics as the key hallmark,

00:52:51.160 assuming there are such things. I consider it to be an interesting list. It became biblically

00:52:57.100 sacrosanct almost immediately, and I've never understood why, but for some reason it did. So I'd agree

00:53:03.040 with Rich that... So conceptually beautiful. I mean, so I agree completely with Rich,

00:53:06.260 and he knows I do, because we've talked about this before. I think the flip side is, I think

00:53:10.120 the hallmarks have been immensely useful to the field. They are a very easy way to communicate

00:53:16.060 this idea of biological aging, and it helps convince some of the scientific community that

00:53:21.120 thought it was all just hocus pocus and snake oil that there is some mechanistic research happening.

00:53:25.620 We can point to specific things that are aging. So I think that part of the hallmarks has been

00:53:30.000 actually really valuable and has contributed to the popularization of longevity, and at least to

00:53:36.240 the extent the science of longevity has been popularized, has contributed to that. And it has

00:53:41.220 been extremely detrimental to the field. And the way I think about it is it just caused the field

00:53:45.660 to narrow prematurely. And this goes back to what I alluded to before. I don't know if we understand

00:53:51.580 80% of biological aging or 0.005% of biological aging. My guess is it's closer to 0.005%

00:53:59.700 and by and large, the funding to look outside of the hallmarks dried up once the hallmarks became the

00:54:07.000 dominant paradigm and people stopped looking. And I think we need to go back to more discovery

00:54:12.500 science and thinking outside the box. So I think it's been a double-edged sword.

00:54:16.580 Would that happen automatically if we could wave that magic wand and increase funding?

00:54:20.940 It would help. I don't know that it would help enough, but it would help. I mean, you also kind

00:54:25.520 of have to change the mindset about what people call fishing expeditions. That's like a bad word

00:54:30.940 in grant review panels, fishing expedition, meaning you don't really know what you're going to find,

00:54:34.780 but you got to go look before you can figure out what's important. So I think we have to kind of

00:54:38.800 change that mindset as well. One can usefully concretize this discussion. I imagine that one

00:54:44.220 of this, I don't read these papers because they upset me, but I imagine inflammation is on one or more

00:54:48.820 of these. Sure is. I'll bet. Chronic inflammation. Okay, good. Chronic inflammation. So what that does

00:54:53.960 is you say, I'm interested in chronic inflammation, so I'm doing good stuff, huh? But what could be

00:54:59.140 happening is this particular set of cytokines might be overexpressed by some glial cells and that leads

00:55:05.320 to loss of cognitive function. Whereas this other overlapping set of cytokines produced by the

00:55:11.860 macrophages in your fat may make you more prone to diabetes or metabolic syndrome. Whereas this particular

00:55:18.300 set of lymphocytes are necessary to repel COVID, and that's why you are more susceptible to COVID.

00:55:26.040 So learning what changes within the extremely broad generic idea of inflammation, what changes in what

00:55:34.440 cell types, in what people, under what pharmacological or genetic changes, how they are interacting with

00:55:40.980 other aspects of pathology, that's marvelous to do. But to say, oh, inflammation, that gets

00:55:48.020 bad when you're old, is a way of avoiding the labor of thinking. And that's why I'm against it.

00:55:55.600 And I think Matt brought up a really important point, and we scientists are to blame, is the way

00:56:01.300 that research gets reviewed. For lazy reviewers, having these 12 hallmarks is really helpful. Oh,

00:56:07.720 this has got one of the hallmarks in it. This must be good stuff. I do think reviewers need to be

00:56:13.440 more open to new ideas and new approaches. I mean, everybody knows that NIH grants are approved if

00:56:22.540 they're incremental. If they're really breakthrough, they don't get approved. In fact, a very famous

00:56:28.080 biologist, E.O. Wilson, told me years ago, he said, don't ever include your best ideas in a grant.

00:56:34.420 They won't get funded. Do standard stuff, save your best ideas for projects that you do on the side.

00:56:40.780 That's one of the reasons I left academia. Drove me nuts. Almost impossible to get the

00:56:45.120 important stuff funded. The second of your multi-partite question was, does epigenetic

00:56:51.680 change, what are the results of? Is it causal? Causal effect. And the third, which we may get to,

00:56:56.840 is can you reverse it and would that be a good thing? So let's talk about the second element here,

00:57:01.380 is it causal? The problem is what it means. There are some changes that occur in this particular set of

00:57:09.860 40 cells in the pineal, and there are other changes that occur in these cells in the bone marrow,

00:57:14.880 and there are other cells that change in the gut and villus lining cells and the crypt cells.

00:57:19.480 So they are all epigenetic in some. They are caused by some things, and we don't really know

00:57:27.560 which, if any of these, count for aging. If someone says, I'm going to prove that an epigenetic

00:57:35.100 change is responsible for aging, they haven't begun to come to grips with the nitty-gritty.

00:57:41.800 People always ask, just as you hinted, does your drug change epigenetic things? And unfortunately,

00:57:47.780 that's where they stop thinking. We're always willing to give people tissues from our drug-treated

00:57:52.080 mice. If they are keen on epigenetic changes that affect neuron regeneration, excellent. Their experts

00:57:58.200 will send them the brains and they can do that stuff. It's important. I'm not making fun of it.

00:58:02.260 But the general notion that that's aging vaguely thought of is due to epigenetic change more

00:58:09.420 vaguely thought of doesn't really get you anywhere. That's my skeptical view.

00:58:15.780 Is part of the issue that you're saying, well, what's causing the cause?

00:58:18.960 No, it's just that the concept of epigenetic change encompasses thousands of changes in hundreds

00:58:25.260 of cell types, under hundreds of influences. Of course, some of that causes other stuff.

00:58:32.540 Agreeing to that, assenting to that notion that epigenetic change is causal for all sorts of age-related

00:58:38.620 pathologies, everyone can agree to that. But it's meaningless because what counts is to say,

00:58:43.540 this specific change is really important in this disease. Here's an epigenetic alteration.

00:58:49.300 Or this specific broad spectrum change in multiple tissues causes something good or bad. You have

00:58:57.020 to define what it is before you can test it. So let's use a specific example. When you look at a

00:59:02.840 patient with type 1 diabetes and you look at their beta cells in their pancreas, they look different

00:59:09.820 epigenetically than the beta cells of an age-matched person without type 1 diabetes. And we also know that

00:59:16.740 their beta cells don't function. So they've lost function. So let's ask that question as a specific

00:59:21.980 example. What do you believe or what confidence would you assign to the notion that the epigenetic

00:59:28.480 change on the beta cells of the type 1 diabetic are indeed causal to the loss of function of the

00:59:33.620 beta cell? My last exposure to the causes of type 1 diabetes was in medical school, which is more than

00:59:39.720 five years ago. But if I vaguely remember, it was an autoimmune disease, right? So if your poor little

00:59:46.600 helpless beta cells are being attacked by antibodies and macrophages and things, those stress reactions

00:59:52.840 are going to cause epigenetic change. And whether those epigenetic changes contribute to some extent

00:59:58.760 to the ill fate of the beta cells, it's possible. And if I were an expert on diabetes pathogenesis,

01:00:05.460 I'd really want to know that. It doesn't have anything to do with aging, but it's an interesting

01:00:09.600 question.

01:00:09.800 But it's a way to address causality.

01:00:11.300 Yeah, but you might equally say, no, no, it's the mitochondria that have changed.

01:00:14.880 They're a hallmark of diabetes.

01:00:16.560 Yeah, or it's the glycated proteins. There's a ton of things in it. There's no reason in the

01:00:20.780 world at this stage, I think, to actually give epigenetics primacy over anything else.

01:00:25.820 Other than it's a nice hypothesis.

01:00:27.180 It's a hypothesis.

01:00:28.080 You can formulate these questions because a lot is known about type 1 diabetes. And I understand

01:00:33.480 0.05% of the biology of aging.

01:00:35.580 0.05, yeah.

01:00:37.020 I was giving...

01:00:38.060 You're off by an order of magnitude.

01:00:39.160 I'm Ted Foldoff.

01:00:39.880 He was raising you by a log.

01:00:41.500 Yeah, I thought, what log off? Formulating the questions in exactly the way Steve did

01:00:45.620 makes it clear how difficult it is to evaluate the concept that epigenetic change contributes

01:00:52.500 to pathogenesis and type 1 diabetes. And we know more or less what is going on in type... We don't

01:00:58.000 know what's going on in aging. We don't even know what part of the body is going on or parts,

01:01:02.180 more likely, of the body.

01:01:03.280 I at least internally reframe it a little bit and say, what would the experiment be? What

01:01:08.760 would you need to do to convince yourself that either, broadly speaking, epigenetic

01:01:14.320 dysregulation causes aging, whatever that means, or this specific epigenetic change that

01:01:20.020 is associated with chronological age causes aging? And so that's an easier way for me to

01:01:25.480 think about it because I feel like it's all a fascinating conversation, but we're never

01:01:29.320 going to get to the answer until somebody actually does the experiment.

01:01:31.920 Or decides that it can't be formulated because it's too complicated and gives up.

01:01:36.360 Yeah, that's right. But people are trying to do both of those things. I mean, people

01:01:39.580 are using partial or transient epigenetic reprogramming and asking, can that have effects

01:01:44.300 on biological aging? I'm actually cautiously optimistic it can. I don't think it's going

01:01:48.700 to be a game changer, but I think you can modulate aspects of biological aging. The technologies

01:01:53.560 are being developed for targeted epigenetic modifications. So if we think this particular epigenetic

01:01:59.480 mark at this particular location in the genome controls aging, and I don't think it's going

01:02:04.380 to be that simple, but let's say it is. You could go in, you could modify that, and then

01:02:08.820 see, do you reduce disease? Do you increase lifespan? Do you improve healthspan? So those

01:02:14.780 are the kinds of experiments that I think would get us to where we can have a lot of confidence.

01:02:19.340 If it's the case, if somebody, let's say at Altos, publishes a paper three years from

01:02:24.500 now, that they have made a mouse live six years by multiple rounds of transient epigenetic

01:02:28.800 reprogramming, I'll be like their biggest fan. They moved the needle. That convinces me

01:02:33.920 that that strategy modulates biological aging. Nobody's done that yet.

01:02:38.140 What about something far less impressive, but still worthwhile? So consider if we could get

01:02:44.060 to the point where we could locally deliver vectors that would epigenetically change chondrocytes

01:02:51.580 so that you could take osteoarthritis in the knee and just regenerate cartilage,

01:02:56.800 regenerate cartilage by changing the epigenome.

01:02:59.320 But is that biological aging? I wouldn't be convinced that's modulating the biological

01:03:03.860 aging process. I would be convinced that's a clinically useful strategy for people who

01:03:08.320 benefit from that therapy. I guess it kind of depends on why we think an individual would

01:03:13.360 be experiencing osteoarthritis. How much of that is senescence? How much of that is inflammation?

01:03:19.100 Trigger rich here before we go down that path. Is it the S one?

01:03:24.520 Yeah, yeah. Let's talk about senescence. If you think osteoarthritis of the knee requires

01:03:28.600 a knee joint replacement, and that's going to help your patient, you are not rejuvenating.

01:03:35.240 It's perfectly possible to do great things with technology, including chondrocyte regeneration,

01:03:40.660 without having to decide that that's related to aging. People don't age because they fail to have

01:03:47.660 titanium knee joints or something. And one way I think about this, and again,

01:03:51.800 this may be completely wrong, but it's a useful way for me to think about it is I think about

01:03:56.320 age-related disease as the downstream effect of biological aging. For most diseases, there becomes

01:04:03.780 a point where the pathology of that disease mechanistically is no longer the same as biological

01:04:10.020 aging. He's very good. You should listen to him.

01:04:12.480 And one of the implications of that is the interventions that slow biological aging may

01:04:16.060 not work once you get past that point, but things that do work for that disease may have nothing to

01:04:20.560 do with biological aging. Does that make sense?

01:04:23.120 Yeah. Go deeper on that idea though. Let's use the example.

01:04:26.180 I mean, what's your favorite disease?

01:04:28.060 My favorite disease.

01:04:29.480 Let's talk about cancer.

01:04:30.520 Cancer is an easy one. We know with cancer, in many cancers, the process is you have one or more

01:04:37.300 mutations, which then often lead to additional mutations. You get genome instability. Eventually

01:04:43.000 you get an oncogene that gets activated and that leads to uncontrolled cell division.

01:04:47.680 Or a tumor suppressor gene that gets deactivated.

01:04:50.060 Yeah. Right. And if we accept that immune surveillance is one important anti-cancer mechanism,

01:04:56.380 we know that immune surveillance declines with age. So early on, we're clearing a lot of our cancers.

01:05:01.340 As our immune system declines, these cancers are going to escape immune surveillance. They're going to

01:05:06.080 accumulate all these mutations. They're eventually going to go into uncontrolled cell division. That

01:05:10.860 uncontrolled cell division, at that point, you can treat the cancer, but uncontrolled cell division

01:05:15.660 is not biological aging. It's not a part of the normative aging process.

01:05:20.140 That's right. Yep.

01:05:20.300 So the treatment there, so the mechanism now is fundamentally different from normative aging.

01:05:24.960 And the treatment, let's just say the treatment in this case is chemotherapy, might benefit the cancer,

01:05:31.060 but it's not going to do with normal aging.

01:05:32.560 Yeah. Yeah. Yeah. Yeah.

01:05:33.180 And I think rapamycin is a good example here where I think we all believe that rapamycin and

01:05:37.840 inhibiting mTOR slows biological aging, at least in up to mice, hopefully in dogs, hopefully in people.

01:05:44.000 Yeah. So it's a fundamental node in the network. That's the way I think about the hallmarks of aging.

01:05:48.940 It's a node in the network that underlies the hallmarks of aging. So we can manipulate mTOR with

01:05:54.180 rapamycin, slow aging. Rapamycin is a pretty good anti-cancer drug until the cancers have evolved

01:06:00.500 to ignore the mTOR break. And then rapamycin doesn't work anymore. And we know rapamycin doesn't

01:06:05.200 work for most cancers. That's an example. That's been tested. We know that.

01:06:07.820 Yeah, absolutely. Yep. And it's because the cancers evolve to bypass the mTOR break or to

01:06:13.260 bypass the ability of rapamycin to inhibit mTOR. That's a really good point that we all take for

01:06:17.580 granted that I think is worth noting. Rapamycin can be unsuccessful as a chemotherapeutic agent

01:06:22.880 and can yet be very successful as a cancer preventive agent. And it's exactly for that

01:06:28.260 reason. Yeah. And I think this also illustrates why traditional disease-based medicine is not about

01:06:34.220 the biology of aging. It's about something that the biology of aging is distinct and it needs to be

01:06:40.340 investigated in a different way. And we know that in the aging field, but the people in the cancer field,

01:06:46.320 in the cardiology field, in the neurology field, I don't think they understand that.

01:06:49.980 This gets to, if I were health czar, this is what I would do. Because it comes back to what

01:06:55.140 Rich said at the outset, which is, why is this a zero-sum game? I mean, you didn't ask it that way,

01:07:00.280 but that's effectively the problem you're dealing with, which is, why can't we study cardiology,

01:07:05.260 oncology, and neurology, and aging without everybody feeling like they're taking them? So my way of saying

01:07:10.680 that in Peter terms is, we need to have medicine 2.0 and medicine 3.0 in parallel,

01:07:15.080 because the tools of the medicine 2.0 scientist and physician, which we see on display today,

01:07:21.420 are putting the stent in, giving the chemotherapy, lowering the cholesterol, all of these things.

01:07:28.020 The medicine 3.0 toolkit looks different. Different science. You're going to use rapamycin here.

01:07:34.680 You're not going to use it over here because it's too late. Instead of saying one or the other,

01:07:38.920 why isn't it both? Why wouldn't we want both of these running in parallel?

01:07:43.640 Well, we would. But of course, the zero-sum game is a pretty good analogy for what's actually going

01:07:49.420 on. The amount of research dollars, at least available to NIH, is not infinitely expansible.

01:07:55.100 It's set by a complex political process. And then there's a separate downstream process that

01:08:00.080 allocates it amongst institutions. So saying that it would be a good idea to have more funds,

01:08:05.480 I agree with you. And I'll bet these two guys do as well, but it's not easy to do.

01:08:10.560 Yeah. I think I misspoke. It will be a portfolio of reallocation.

01:08:13.960 But it will be worthwhile because the burden of this disease will be lower. So in other words,

01:08:18.120 it's sort of like saying, right now, I spend $100,000 a year on the barrier to my house to prevent

01:08:27.320 anybody from breaking in. And I spend $100 a year patrolling the neighborhood to make sure there

01:08:34.020 aren't too many bad guys in the neighborhood. There's a scenario where if your total budget

01:08:38.960 is $100,000 and $100, maybe you could spend $80,000 in total by spending more money patrolling

01:08:45.540 the neighborhood, less money.

01:08:47.080 I think we generally agree with you that having a greater proportion of available research dollars,

01:08:51.600 both private and public, going into the biology of aging and its impact on late life health would

01:08:57.420 be a good thing. I don't think you're going to get an argument here.

01:09:00.060 But I also think you're going to get a huge argument from anybody in the cardiology field,

01:09:06.400 the neurology field.

01:09:07.260 Yeah. Or Alzheimer's.

01:09:08.000 The Alzheimer's field.

01:09:08.920 It's their money.

01:09:09.940 But wouldn't some of those people, as the funding dollars move towards the aging side,

01:09:14.920 also want to move and say, look, I'm going to study this through the aging lens?

01:09:18.440 I was on the council for the National Aging Institute for three years. And if at any point I can swear to

01:09:24.440 this from personal testimony, somebody would say something like, I wonder if maybe a few percent

01:09:29.660 of the Alzheimer's budget might instead go to studying how slow aging models would have an

01:09:36.640 impact on late life neurodegenerative disease. The next day, the director of the Aging Institute

01:09:42.400 would get a call from two or three congresspeople who were on the appropriations committee stating

01:09:48.200 that this will not be happening because there was an Alzheimer's Association person who got the call

01:09:54.360 from the NIA staff member in charge of Alzheimer's saying, tell the congressman to call the director

01:09:59.940 and let's put a stop to that reckless idea. They're tied in to the political process in ways that-

01:10:07.860 Well, we just need to go maybe one step further because those congresspeople have a boss. 1.00

01:10:12.400 They report to somebody too.

01:10:13.920 Who would that be at this stage?

01:10:15.260 Yeah. No, I mean, come on. Maybe it's because the public doesn't understand this. Those people

01:10:19.580 answer to the public.

01:10:20.580 That's a good example.

01:10:21.180 These are our dollars that are going to work.

01:10:23.600 I'm with you.

01:10:23.820 But Alzheimer's Association, I mean, that's a patient advocacy group. That is the public.

01:10:28.360 Yes. Although let's ask the question, what have they done for those patients lately?

01:10:32.500 That's a different question, but I mean, I'm just reinforcing what you said. I think part of

01:10:36.100 this is educating people.

01:10:37.500 If you know somebody who's suffering from Alzheimer's disease, you know very well that the only thing we've

01:10:42.880 got going for us right now is prevention. We don't have too many silver bullets in the treatment gun.

01:10:49.260 Despite massive spending, massive spending on it. I was once in Congress trying to lobby

01:10:55.960 with about six people from the Alzheimer's Association in the same room. I was totally

01:11:02.480 ignored by staffers that were in there.

01:11:05.260 I agree with all of this. I agree with all of this. I think, again, though, we should be careful

01:11:09.320 not to demonize people for wanting to cure Alzheimer's. It's a good thing. It's a good

01:11:13.860 goal. I think the communication piece is about the fact that it's going to be much more efficient

01:11:18.820 and effective to keep people from getting it in the first place. This goes back to the idea that

01:11:23.120 once you've outpaced the biology of aging with the pathology of the disease, it gets a lot harder,

01:11:28.060 a lot harder to do anything about it. So I think that communication part, honestly,

01:11:32.680 I don't know why we've been so unsuccessful. Because I think a lot of us have been out there

01:11:36.080 trying to communicate this message for a long time, but it's starting to permeate.

01:11:40.860 We're at that moment, I think, where people are starting to get it, that biological aging is a

01:11:44.980 thing. It's malleable. We don't really know for sure what works in people and what doesn't work yet,

01:11:51.340 but we're getting there. It's going to take a little while, but there's reason to be optimistic.

01:11:55.560 And there's also the private sector is another reason, I think, to be optimistic.

01:11:58.680 So let's go on record right now. I think when we, if we defeat Alzheimer's disease,

01:12:04.960 it's going to be because of the biology of aging. It's not going to be because of the drugs that get

01:12:10.640 rid of beta. Absolutely. Yep. Probably cancer, probably heart disease. Although I think Peter's

01:12:15.400 more optimistic we can prevent heart disease. If you took the tools of medicine 2.0 and just

01:12:19.740 applied them 30 years earlier, we wouldn't have ASCVD. That's the one place where it's where,

01:12:23.640 but again, that's because the mechanism of action is so well understood with ASCVD compared to

01:12:27.880 Alzheimer's and cancer. A lot of infectious disease, a lot of liver disease, a lot of kidney

01:12:32.840 disease. All of those things can be improved dramatically by targeting the biology of aging.

01:12:39.220 You know, if I were to write my book again, I would add a fifth horseman because I talked about these

01:12:43.680 four horsemen of ASCVD, cancer, neurodegenerative and dementing diseases and metabolic disease,

01:12:48.260 but I would actually add a fifth hallmark. It's not really a hallmark of disease,

01:12:52.540 but it's the fifth thing that brings life to a bad close, which is immune dysfunction.

01:12:59.120 And I don't think I gave that enough attention in the book because of course, as you said,

01:13:02.780 it factors in very heavily to oncogenesis, but also as COVID showed us, what a risk factor it was

01:13:10.260 to be old. And, you know, I'm reminded of this when I see people my age get brutal pneumonias.

01:13:17.060 And like two months later, they're okay. And you realize two of my patients actually in the past

01:13:22.840 six months have had really bad pneumonias where you're looking at the CT of their chest and you

01:13:27.580 cannot believe they're alive. But of course they're fine. Three months later, four courses of antibiotics

01:13:32.440 later, they're fine. And you realize you do that to a 75 year old, it's over. And it simply comes down

01:13:38.700 to how their B cells and T cells work. That to me is an area where I'd love to see more attention,

01:13:44.300 which is what would it take to rejuvenate the immune system as a proactive statement?

01:13:50.080 That's part of the XPRIZE HealthSpan challenge, of course. I think that that's a perfect example.

01:13:55.400 Influenza pneumonia has never fallen out of the top 10 causes of death in the U.S. You know,

01:14:00.380 it used to be number two, but still now it's number eight or nine, but it's always there because

01:14:05.300 you can't really do anything about the late life immune dysfunction.

01:14:10.120 Just to follow this up, if magically you become in charge and you're able to

01:14:15.200 double the amount of research being done on the biology of aging fundamentally,

01:14:19.340 then we can afford to do, let's give some mice to start with, a batch of anti-aging drugs

01:14:25.300 and see if it makes them more resistant to infectious illnesses, including pneumonias,

01:14:31.540 but viral infections as well and many others. I'd love to know the answer to that. And no one has

01:14:36.720 actually really looked in a serious way because the ITP has enough money to just measure lifespan.

01:14:42.800 Right. None of the health spanishes.

01:14:43.840 And then we're hoping that everybody else is now going to look at the brain and the lungs and the

01:14:47.640 infection, the sensory systems. That really ought to be done and it's not being done because of a

01:14:53.000 lack of money.

01:14:54.200 You said something a while ago, Rich, that I think is timely now, which is with each generation of

01:15:00.500 these drugs, they get more efficacious and less toxic.

01:15:03.860 Not yet, but that's the hope.

01:15:05.700 Well, no, no, but I'm going to use another example. The GLP-1s are the best example of this,

01:15:09.340 right? So you go back to the very, very first generation of GLP-1 agonists, barely lost any

01:15:14.440 weight, horrible side effects. Generation two, about 10 years ago, a little bit better weight loss,

01:15:20.520 side effects, so-so. Fast forward to semaglutide, quite a bit better efficacy, still really bad side

01:15:26.980 effects. Next generation, trisepatide, better efficacy, side effects are almost gone.

01:15:32.160 Now, why haven't we been able to do that with these gyroprotective drugs? So we have this one

01:15:37.620 study using Everolimus that gives us a hint that says, hey, this might actually enhance immune

01:15:42.620 function in people in their mid-60s. But we need the follow-up study, the follow-up drug. Imagine

01:15:47.860 what the fourth generation of that drug can do where it's tuned to get better and better and

01:15:53.060 to have fewer side effects.

01:15:54.080 There's strong commercial motivations. You know you're going to sell a lot of the obesity drugs.

01:15:58.240 They're very strong commercial motivations to do those studies over and over and over

01:16:03.680 again until you find one that works better. And they're good preclinical models that you

01:16:07.520 can use so that you're not wasting too much of your time on clinical trials. That could

01:16:11.940 be done for anti-aging drugs as well, although testing anti-aging drugs in people is a whole

01:16:18.180 separate set of tangle of difficulties. I don't want to talk about that right now, but I'm saying

01:16:23.760 it won't be quite as easy as it was for anti-obesity medications. But no one's

01:16:28.240 doing even the first level of research to find the optimal compounds for efficacy without

01:16:34.220 side effects or even to begin to see if they have desirable effects on aging rate indicators

01:16:41.000 in people. That's kind of a cheap and easy study and no one has really tackled that yet.

01:16:45.680 Well, I just heard that there are over 80 senolytic studies in early clinical trials.

01:16:50.440 No, I meant anti-aging drugs. It's a joke. It's a joke. It's a joke. It's a joke.

01:16:54.320 We have to come back to this. Are any of them powered for anything other

01:16:58.200 than safety? This is, I think, the problem. Oh, it's all phase one.

01:17:00.520 Yeah, exactly. So they're underpowered. They're almost useless, in my opinion.

01:17:04.400 Well, until they get to phase two or phase three. If they get there.

01:17:07.260 How many years have we been having phase one senolytic trials now?

01:17:10.100 I don't know. At least a decade.

01:17:11.620 God, has it been that long? First one I remember was 2017. So yeah,

01:17:15.300 a decade easily because I probably wasn't paying attention in 2014, 2015.

01:17:19.300 There's lots of complicated issues here. I think endpoints for clinical trials are

01:17:22.480 really challenging, but solvable. So there are two places I wanted to go

01:17:26.080 next, and I'm going to let Rich decide because he's going to have the strongest point of view.

01:17:31.280 Can we talk about senescence or can we talk about what biomarkers would be necessary to help us study

01:17:39.700 aging in humans as we translate from your work and Matt's work?

01:17:44.240 I know what I want to talk about, and it's the second of those two.

01:17:47.520 Okay.

01:17:47.820 I don't want to spend the next three or four hours explaining why senescence is silly and

01:17:52.280 antisenolytics are untested at best.

01:17:54.880 There's no way we're not talking about that, but very well.

01:17:57.060 Let's go on to item number two.

01:17:58.360 And I think the most important thing is to make a clear distinction between

01:18:01.660 biomarkers and aging rate indicators.

01:18:05.560 Please explain the difference to people, please.

01:18:07.440 Okay. I'll do my best. So a biomarker allegedly, and in real life, is something that changes with age.

01:18:14.660 So if you have some drug that slows aging, the biomarkers, many of them in the different cell

01:18:20.300 types and in the blood, will change more slowly. They are a good way of looking at whether you're

01:18:24.880 slowing. It'll work in the dogs. Long-lived dogs and short-lived dogs will have differences in the

01:18:29.500 rate of change of biomarkers. A very established part of the literature and valuable. But you have to

01:18:35.500 wait till somebody's old, whether it's a dog or a mouse or a person, because only when they're old

01:18:40.180 has the biomarker of aging, the surrogate marker for biological aging, changed very much.

01:18:46.840 So in a clinical trial, certainly in a human situation, no one wants to wait 20 years to

01:18:52.940 see whether the biomarkers have changed. And a one year is such a tiny fraction of a human

01:18:58.780 lifespan that you don't really anticipate detectable change with an appropriately powered

01:19:04.300 study. It's like aging rate indicators, which are much less well-studied and much less well-established

01:19:11.180 in principle, are things you can measure that tell you whether you're in a slow aging state or a

01:19:15.440 normal state.

01:19:16.860 Can I just make a point for the listeners so they understand the challenge of what we're

01:19:19.860 talking about? When we study blood pressure drugs or cholesterol drugs, the biomarkers change

01:19:26.940 so rapidly and we know the relationship between the biomarker and the disease state. So if your

01:19:34.000 blood pressure is 145 over 90 on average, before I give you this ACE inhibitor and three months later,

01:19:41.320 six months later, nine months later, a year later, your blood pressure is averaging 119 over 74.

01:19:47.700 I know I've done something well. Now I will still probably in the phase three, in fact, I will in

01:19:53.440 the phase three have to make sure that I also reduce some event in you. But generally by the phase two,

01:19:59.080 I know that this drug is not toxic and that it's predictably lowering your blood pressure.

01:20:02.640 A biomarker generically is something that's easy to measure that is informative about something

01:20:08.500 that's hard to measure. A classical example, famous example is you want to know how many

01:20:12.440 cigarettes somebody smokes a day, they'll lie to you. But if you measure cotinine in their blood,

01:20:18.020 that's a byproduct of nicotine. You don't have to ask them. You can find out how many cigarettes

01:20:22.580 they had in the last couple of days by measuring blood. That's a biomarker of cigarette consumption.

01:20:28.340 Is it a marker of nicotine or carbon monoxide?

01:20:30.000 I don't know the answer to that.

01:20:31.180 Okay, okay. I wasn't sure.

01:20:32.420 In principle, a biomarker of aging is, there are many of them and they are measuring

01:20:35.720 biological aging processes and they're useful in that regard. But they don't tell you how

01:20:40.020 fast you're aging. The analogy I love to use is an odometer is like a biomarker of aging of your

01:20:45.820 car. It tells you how many miles your car has gone, but it doesn't tell you how fast the car is

01:20:50.960 going. The speedometer tells you how fast your car is going. And so what we need and what I think

01:20:57.600 we're just beginning now to document is things like the speedometer, aging rate indicators

01:21:02.840 that reliably discriminate slow aging mice or people from regular old mice or people.

01:21:09.320 We have now a dozen or so things that change in the fat, in the blood, in the liver, in the brain,

01:21:14.880 and in the muscle that are always changed in any slow aging mouse, whether it's drug A, drug B,

01:21:20.540 drug C, calorie restriction diet, or single gene mutations. We've looked now at five different

01:21:25.280 single gene mutations. And this whole set of 12 or roughly 12 aging rate indicators always changes

01:21:32.600 in every slow aging mouse. And it does so in youth, which is the key point. So if it does so quickly

01:21:40.220 after an anti-aging drug is administered, that's the transition, that's the bridge you need for

01:21:46.180 clinical studies in people. If you want to know whether metformin or conagliflozin or something

01:21:51.480 slows aging in people, and you don't want to wait 20 years, but you've got things that tell you

01:21:57.420 whether they are in a slow aging state, how fast they are aging versus normal, and that's a big if.

01:22:02.840 We don't yet have evidence we can do that. We just have hope we can do that. Then that allows you

01:22:08.120 quickly, quickly being within six months to a year, to know whether your alleged anti-aging manipulation

01:22:16.300 has moved them to a physiological status, which is associated with slower aging. A lot of that can

01:22:22.720 be done in mice, with drugs, with mutants. And are these all proteins, Rich?

01:22:26.860 No. No. Some of them are changes in the fat, different classes of macrophages. The pro-inflammatory

01:22:33.100 macrophages, the bad ones go away. The anti-inflammatory macrophages, the good ones go up.

01:22:39.220 UCP-1, I recall from our discussion. UCP-1 goes up in every one of our 10 different kinds, 11 now,

01:22:45.500 of slow aging mice. Does it go up in any of the mice that did not receive a successful drug?

01:22:52.120 Well, we compare them to controls, and the question you're asking is really important. That's what we're

01:22:57.160 doing in the next five years. We just got a grant to do that. We're going to take mice and give them

01:23:01.280 either a good drug or a different drug that doesn't work, and then make those comparisons, a really

01:23:07.180 important thing to prove. So far, our only control has been untreated mice.

01:23:12.860 At some point in this, I had to bring this up, but let's imagine that Rich is incredibly successful

01:23:18.320 at finding these things. That is a very, very long way from assuming that it's going to be the same

01:23:25.240 in people. Most things that clinically work in mice do not work in people. It might be,

01:23:33.220 and that would be wonderful. But I think ultimately, we're going to have to find this

01:23:37.800 for people. And my thought is the kind of evaluation that you do routinely of your patients.

01:23:44.160 If we took a group of 65-year-olds and we gave them a drug that we thought was an anti-aging drug

01:23:50.180 and followed them the next five or six years doing these evaluations, I think you could probably

01:23:56.580 safely say, this is slowing aging or not slowing aging. So I don't think that it's going to be

01:24:03.920 that easy to jump from mice to people in this.

01:24:07.040 I've always wondered if in people, the easiest way to do it would be to take the most obvious thing

01:24:13.340 that we know is going to reduce the rate of aging. So it'd be an interesting experiment,

01:24:17.880 but you find someone who is overweight, diabetic, and smokes and has hypertension. You get hundreds of

01:24:25.220 these folks. You put half of them on a, to be ethical, a plan where you try to get them to stop

01:24:30.040 and presumably many don't. In the other group, you pull out all the stops and you don't care because

01:24:35.620 you're interested not in testing the hypothesis, does this thing help you? You're interested in

01:24:40.740 getting them to lose weight, not have diabetes, stop smoking, exercise like crazy. The greatest

01:24:46.540 division between two groups of individuals where we would, I think, be able to agree that this group

01:24:52.640 is now aging slower. The group that we've reconciled their diabetes, quit the smoking,

01:24:57.160 et cetera, et cetera. And then I'd love to see Rich's 12 line up in that population.

01:25:02.480 That would be great. Let me just say that I think that people that study animals, myself included,

01:25:07.260 always underestimate how well we can evaluate health in people with a very, very thorough

01:25:14.380 evaluation because we don't do that in our experimental animals.

01:25:18.420 Why do you think that is, Steve? Why is it? Because I was going to ask about parabiosis later

01:25:22.540 on in the discussion. We might as well talk about it now, right? Parabiosis seems to actually kind

01:25:26.340 of work in certain mouse models. Do we have any reason to believe it's going to work in humans?

01:25:30.920 And if not, why not? Why are mice so different from people?

01:25:34.500 Wait a minute. I wouldn't say that just because we don't have evidence that it works in humans means

01:25:38.120 mice are different from people. First of all, when it comes to parabiosis, right? I mean,

01:25:41.520 that's a different discussion. But I agree that if you look at the attempts to cure cancer or other

01:25:46.720 diseases in mice and translation to people, most have failed. I actually think that's because

01:25:52.420 those are artificial mouse models where they tried to give young mice an age-related disease.

01:25:57.660 I'm more optimistic.

01:25:58.780 But riches are not.

01:25:59.620 I don't know this.

01:26:00.380 Rich doesn't have those mice. Yeah.

01:26:01.840 I know. I'm more optimistic that biological aging or normative aging is going to be much more likely to

01:26:08.300 translate to people, both interventions and biomarkers than the specific disease interventions. I might be

01:26:14.640 wrong. I don't know the answer. We would hope that's the case.

01:26:17.660 That's fair. I don't think we should rule out the mice as a useful model. In fact, I think there's

01:26:22.180 reason to be optimistic that it will. I actually am kind of bullish on parabiosis as I think it will

01:26:26.860 work to some extent in people. It's not a pragmatic approach for population gerotherapeutics.

01:26:32.020 But I'm just wondering why it wouldn't be as efficacious.

01:26:35.520 This is something that, I mean, aren't there six or eight clinical trials going on right now?

01:26:39.900 Different variants of that, yeah.

01:26:41.340 Yeah. I haven't seen them. I've seen the one that's looking at, it's not really a parabiosis

01:26:46.340 study, but it's looking at plasmapheresis for Alzheimer's. I consider that a little bit different,

01:26:50.460 but fair enough. Okay. Because they're just using albumin, I think, aren't they?

01:26:53.340 Right. But there's also studies going on of young blood.

01:26:56.460 Yeah. Okay. Okay.

01:26:57.280 But if you think of parabiosis as both taking away the bad stuff that accumulates with age and adding

01:27:02.020 in the good stuff that's in young, some sort of plasma exchange hits at least half that equation.

01:27:06.780 Okay. I want to come back to this, but my question was why the difference? You're saying, Matt,

01:27:13.540 the difference is probably amplified in disease-specific cases like heart disease,

01:27:17.380 cancer, and Alzheimer's disease. Probably less relevant when you're talking about aging because

01:27:21.540 even a flawed mouse model still ages. In fact, it's designed to age in a certain way.

01:27:26.000 Yeah. And I mean, I think normative aging looks very similar. Again, if we look from mice to dogs to

01:27:29.860 people, just broadly speaking, the process looks pretty similar. So I'm cautiously optimistic that

01:27:36.360 these things are going to translate.

01:27:37.620 Not to pay too much attention to Steve's pessimism on this point, although he's completely right,

01:27:42.800 of course, most things that do have an important effect in mice fail in human clinical trials.

01:27:48.440 And it's for a variety of reasons. Sometimes humans are different from mice. Sometimes the drug has

01:27:52.440 side effects that are tolerable in mice, not tolerable in people, et cetera. But I always like to look at

01:27:58.060 the other side of the coin. That is, if your goal is to develop a drug that blunts pain in people,

01:28:04.680 and you screen 40 or 50 drugs, and you find a couple that inhibit pain in mice, that's a really

01:28:11.380 good start. It doesn't guarantee they're going to work in people, but it gives you this category of

01:28:16.180 snail-based neurotoxins. Let's make 40 of those from 40 different snails. We'll find one that actually

01:28:22.380 in people works, can be made by a scalable process and doesn't produce serious side effects. So the

01:28:28.940 mice, it's not a one-to-one mapping. It works in mice. It doesn't. It works in people. But it's an

01:28:33.580 important, critical first step, which usually succeeds in finding a set of drugs of related

01:28:40.820 families or with related targets, at least, that are efficacious in people. Most drugs that are used

01:28:46.800 in people had useful rodent-based research somewhere in their pedigree.

01:28:52.080 Absolutely agree with that, Rich. And nobody's saying that 100% of things that work in mice do

01:28:56.740 not work. But I think there's a critical difference for aging research, which it takes four years

01:29:00.780 to do one of these in mice. And so if we have to do 40 to find one or two that works...

01:29:06.480 That's why I like aging rate indicators. Speed things up.

01:29:09.200 I'm stepping on your toes, Peter. But the question I always come back to, I agree,

01:29:12.540 we need these aging rate indicators. How do we get to the point where we're confident that

01:29:17.000 they actually work in people? And maybe more importantly, how do we get to the point that

01:29:20.900 FDA is confident that they work? Because that's the only way you're going to be able to use them

01:29:24.560 in a clinical trial. I don't see a path in the short term.

01:29:29.060 Well, I don't know that we need that, to tell you the truth. So I went to the FDA to try to get them to

01:29:34.040 approve a trial of metformin. And we didn't couch it in aging, because you're right,

01:29:38.940 as soon as you mentioned aging, their eyes glaze over and they're not interested anymore.

01:29:43.280 But we did it in terms of multimorbidity. And they were fine. They were fine with that.

01:29:49.080 But that's a different end point. That's not a biomarker.

01:29:50.520 My reply to your question is that you've merged two different difficult problems.

01:29:54.740 Problem A, can we find drugs that slow aging in people? Problem B, can we surmount the legal and

01:30:01.300 political barriers to getting them to work?

01:30:02.820 That's not what I was asking. I was asking, how do we get to the point?

01:30:05.280 That's what I'm saying is that you were focused on something I don't have any answers to,

01:30:11.800 basically, which is how do we get the FDA to develop and approve clinical trials?

01:30:16.500 I was more interested in a step before that. It'd be nice to have some drugs that actually

01:30:20.660 do work to slow aging in people.

01:30:23.620 But you have to trust the biomarker of aging rate before you can be confident that the drug

01:30:28.320 that moves the biomarker of aging rate works in people. That's fundamentally what I'm asking.

01:30:31.720 How do we get to the point where, let's just take FDA out of the equation. The four of us

01:30:36.360 would sit and look at the data and I'll be like, yep, that works.

01:30:38.560 Well, that's sort of my thought experiment. I would have to take an example in humans that is so

01:30:43.700 egregious that nobody with a straight face could say one group isn't now aging slower than the others.

01:30:50.220 Sure. Would that convince you though? So let's say we do that.

01:30:52.880 Well, it would make me worry. It would only show you the positive signal. It would show you the

01:30:57.360 specificity and not the sensitivity of the test. That's the problem. You might miss the signal.

01:31:01.960 If you found a proteomic genomic, like if you found a multimodal signal that detected a difference

01:31:08.260 in rate of aging between those two very extreme sets, you might miss it with a geroprotective drug,

01:31:13.660 which wouldn't be as dramatic as that chain.

01:31:16.020 So what if I told you that there are people who claim, there are epigenetic signatures that do

01:31:21.780 that, that correlate quite well. They claim with health outcomes, 10 year mortality, five year

01:31:27.580 mortality, three year mortality in people and are measuring the rate of biological aging because

01:31:34.280 it's out there. I mean, it's in the literature.

01:31:35.660 I mean, this is not perfect, but it would be one thing I would immediately think of it, which is I

01:31:40.040 would take a really good biobank that would have enough samples that I could sample a bunch of

01:31:45.380 human stuff and use an unbiased sample and a bias sample. So I would determine an algorithm based

01:31:52.440 on one and see how well it predicted on another based on enough samples. That would have to be

01:31:57.980 true at a minimum.

01:31:58.980 Yeah, I think it is. I mean, again, at least it depends on how much faith you put in these

01:32:03.060 research studies. But I mean, people have published epigenetic algorithms. Dunedin-Pace is the one that

01:32:08.760 most people are going to talk about that correlate seemingly pretty well, at least with mortality and with

01:32:14.740 metrics of health span, for lack of a better way of framing it. So that exists.

01:32:19.120 And Dunedin-Pace is using something besides epigenetic or is it only epigenetic? I think

01:32:23.260 it uses something else.

01:32:24.020 It was trained off of other biomarkers and then they found epigenetic marks that correlate with

01:32:29.360 those other biomarkers. So it's a correlation to a correlation, but there's still a correlation.

01:32:33.720 What do you think, Rich?

01:32:34.340 Well, I wanted to go back to the example you gave where you took a lot of people and

01:32:38.940 gave them intense exercises and dietary changes to improve their health, likely health outcomes.

01:32:45.740 And that's a good place to start a discussion because you said every sensible person would see

01:32:50.860 the treated group as aging more slowly. And I would want to ask before I agreed to that,

01:32:57.840 do they also have improved cognition? How are they doing in cataracts? How are they doing in hearing?

01:33:02.640 What happens when you give them a flu shot? Do they have a great flu shot?

01:33:06.820 So the things you've pointed to are really important for both overall health and for

01:33:11.900 cardiovascular risk and the things linked to that. So it's nice to know. But to convince me that you

01:33:17.540 now have a slow aging group of people, you need to go beyond the risk factors for specific common

01:33:24.540 human diseases. If you could show that, then for the first time, I would be convinced you had an

01:33:31.220 effective anti-aging manipulation in people. Currently, I don't know that there is any

01:33:37.240 effective anti-aging manipulation in people. If your approach got there, that would be a terrific

01:33:44.580 research model.

01:33:45.660 Well, but now we're getting into the definition of aging a little bit, which is, would you agree

01:33:51.300 that the approach I'm describing would produce a longer life?

01:33:55.960 It's easy to produce a longer life. If you happen to have a clinical condition where you're tied to

01:34:00.720 a railroad track and there's a train coming, you can extend that woman's life enormously by simply 1.00

01:34:06.660 giving her a knife and cutting the bonds and letting her walk away from the track.

01:34:10.840 Longevity promoting interventions are not anti-aging.

01:34:12.520 Well, if 80% of people died as a result of trains on train tracks, that might be a worthwhile

01:34:17.280 example. But given that 80% of people die from these four chronic diseases-

01:34:21.960 I'm all in favor of protecting people against chronic diseases.

01:34:25.680 Yes, yes, yes.

01:34:26.240 That's a good thing. And I'm glad that people are doing that. No question about it. Now,

01:34:30.320 talking about the biology of aging, there are all sorts of things that also happen when you get older

01:34:35.280 that are not part of those chronic diseases. And to make a case that you've got an anti-aging

01:34:40.660 manipulation, you need to show that those are changed too.

01:34:43.400 But do all of them have to change or just most of them?

01:34:45.600 Don't enough have to change that you increase the length and quality of your life?

01:34:49.540 And if you still get a cataract at the same rate, I'm not sure that should be disqualifying.

01:34:54.900 Right. But the important thing I think about what Rich said is all the stuff that he

01:34:57.960 pointed out could be easily done in humans. Wouldn't be hard to measure hearing.

01:35:02.980 The nice thing about the dog examples, we've got well-known, famous, long and slow aging dog

01:35:09.280 breeds, and it's true for horses too, it's certainly true for mice, is that more or less everything

01:35:15.760 slows down together. The tiny dogs that are very long-lived, it's not just that they have a delay

01:35:22.380 of cancer, they have a delay in neurodegenerative disease, a delay in digestive diseases, in joint

01:35:27.420 diseases. Aging has been slowed in those dogs. And if the dogs did your thing-

01:35:34.120 But we might not have an intervention that does that, to your point, Rich. I'm saying we might not

01:35:38.060 have a non-pharmacologic method that does that. It's not clear that even though exercise clearly

01:35:44.500 extends lifespan, it's not clear that it's doing so by slowing aging. You're right.

01:35:50.680 Those are two different things, to your point. I agree. It's not clear, but it's an interesting

01:35:54.360 question. Like, do you believe exercise slows aging? Exercise, healthy diet, sleep?

01:35:58.340 I have no idea. I think so. My intuition is I think so, but I can't point to the evidence that

01:36:03.940 tells me so. Well, there's evidence to support it. The question is, does it rise to the level

01:36:07.680 of evidence that would convince Rich? I believe it probably does too, but I'm not going to say

01:36:12.100 it with 100% certainty. I think here's where we get back into health

01:36:15.320 span versus lifespan. The effect of exercise on longevity is pretty small. Its effect on

01:36:22.160 quality of life is enormous. Somewhat depends on where you start.

01:36:27.240 I've always found these to be a little bit problematic because I don't think that defining

01:36:31.800 it by the input is as valuable as defining it by the output. In other words, to say you exercise

01:36:37.000 this many minutes a week versus that many minutes a week is a little dirty because intensity matters.

01:36:42.840 What you do matters. Sometimes the output is what matters more. How strong you are,

01:36:46.880 how high your VO2 max is, those tend to be more predictive because that's the integral of the work

01:36:52.100 that's been done. But your point is it's well taken. The impact on health span is what I tell my

01:36:57.320 patients. If this amount of exercise didn't make you live one day longer, the quality in which your life

01:37:02.960 would improve would justify it. Now, fortunately, we can move past the semantic discussions because

01:37:08.920 there's now molecular ways of checking this. Exercise, as I'll bet all of you know, increases

01:37:14.780 an enzyme called GPLD1 in the blood of exercise people and in mice. And Solvieda's lab has shown

01:37:22.200 that if you elevate GPLD1, it does great things to your brain, more neurogenesis and more brain-derived

01:37:28.220 protective factors, brain-derived neurotropic factors. Iresin also goes up in humans and in

01:37:33.600 mice. After exercise, it does great things for your fat.

01:37:36.140 As does cloth, though?

01:37:37.500 Let's leave that out for a moment.

01:37:38.720 Oh, boy. Oh, boy. I'm striking all the nerves here today. All right.

01:37:42.120 You may be quite right. I wanted to stick with the GPLD1 and Iresin to make the point that they also

01:37:48.800 go up in all of the slow-aging mice. That is, all the anti-aging drugs, the calorie-restricted diet,

01:37:56.000 the isoleucine-restricted diet, and five different single-gene mutants that extend

01:38:00.800 licepain in mice. They all elevate GPLD1.

01:38:04.160 17-alpha-estradiol?

01:38:05.820 Yes.

01:38:06.680 Kinagaflozin? Both sexes? 1.00

01:38:08.560 Well, this is the key question. Iresin is sex-specific. GPLD1 is in both sexes. This is

01:38:15.080 how one begins to answer that question. This is the exact kind of question one has to ask.

01:38:19.200 So, if you are interested in the idea that exercise regimes have a benefit beyond the obvious

01:38:26.860 exercise-linked physiological declines of age, do they improve cognition? And if so, how? These

01:38:34.360 molecular changes are the things you need to begin to investigate. The anti-aging studies in mice

01:38:41.640 show that the anti-aging drugs, at least the ones we've looked at so far,

01:38:45.720 increase the same things that exercise does.

01:38:49.560 Rich, have you done this experiment with an ITP cohort where you run in addition to a drug

01:38:54.460 parallel?

01:38:55.080 Nope.

01:38:55.320 You know what I'm going to ask?

01:38:56.660 Well, you're going to ask if we exercised our mice, right?

01:38:58.600 Yes.

01:38:58.940 Yeah, we've never done that.

01:38:59.760 So, you haven't done a sedentary versus exercise?

01:39:01.760 We have not done that.

01:39:02.520 You haven't done a obesogenic versus fasted?

01:39:06.380 We never use obesogenic diets. It's worth doing it. The ITP doesn't do it. We don't have the

01:39:11.640 resources. We have enough resources to test about five drugs a year, but if we wanted to test them

01:39:16.540 in exercise versus non-exercise...

01:39:18.780 We've got to get you a budget increase because that will now get to this question because now

01:39:23.540 we could look at the soluble thing.

01:39:24.920 It's a good question.

01:39:25.580 Yeah.

01:39:25.960 Maybe it would. Maybe it wouldn't.

01:39:27.640 In mice.

01:39:28.200 I'm very agnostic about what we can learn from exercising mice because mice are basically kept

01:39:33.340 in a jail cell, something the size of a jail cell their entire life. If you took a bunch of people

01:39:38.420 and put an exercise wheel in a jail cell, then we'd use it. Would that be the same? Would that

01:39:44.080 substitute for people that walk around, to go inside, to go outside, to go to the gym,

01:39:49.440 that do this?

01:39:50.360 It wouldn't substitute for all of it. No question.

01:39:52.800 So, to me, it's a very low level of exercise.

01:39:56.160 If you didn't see anything from it, then you wouldn't rule it out.

01:40:00.020 Right.

01:40:01.240 So, there are testable molecular hypotheses that link the biology of aging to anti-aging drugs

01:40:08.220 and to exercise and teasing out how those are interrelated and which of your exercise regimes

01:40:15.920 increase Iresin, increase GPLD1, and increase neurogenesis. That's a research agenda that could

01:40:23.320 be very valuable. And then if you want to screen drugs in people to see which ones deserve expensive

01:40:29.520 long-term testing, the ones that raise GPLD1, Iresin, and some aspect of neurobiological function,

01:40:37.240 in addition to the good stuff they're doing for the muscles. That's an approach.

01:40:41.580 I agree completely. And this gets back to what we were talking about before with the epigenetic

01:40:46.720 changes is if you had a mechanistic connection, which is what Rich is drawing there, not only this

01:40:52.340 is correlated with this outcome, but here's why. We all feel a lot more confident that this is real,

01:40:57.800 that it's important, and especially if that mechanistic connection is preserved in people.

01:41:01.520 Do any of you believe that GLP-1 agonists are geroprotective?

01:41:06.140 I'm super interested in that question. Yeah, I think we need to find that out. They look good.

01:41:11.140 I think there's two parts though. Are they geroprotective from a caloric restriction

01:41:15.920 effect, or are there caloric independent effects that could potentially be geroprotective?

01:41:20.940 I'm actually asking the second question. I'm taking the first as a given.

01:41:24.000 Okay. Yeah. That's a different question is, is chronic caloric restriction beneficial in normal

01:41:28.600 weight people? But most people taking GLP-1 agonists aren't normal.

01:41:31.700 Yes, yes, yes. And I think it's impossible at this point, because the studies are all done in obese

01:41:36.560 and patients with type two diabetes that we can't disentangle them. So we will just say that for

01:41:40.880 that patient population, the caloric restriction appears to be geroprotective. But yes, you're

01:41:44.980 right. I'm technically asking the second question, which is in an individual who is metabolically

01:41:52.200 healthy, but overweight, where there's actually no evidence that weight loss per se is necessary

01:41:57.020 outside of maybe some edge cases in orthopedic stuff. Is there a geroprotective nature to this?

01:42:02.780 And where it's most talked about is in dementia prevention right now. That's where it's at least

01:42:07.120 most complicated to tease that out. So what do you guys think?

01:42:09.980 And it clearly has neurological effects. There's effects on addiction.

01:42:13.600 Yes.

01:42:13.700 The dementia connection is not inconceivable.

01:42:15.680 It's crossing the blood-brain barrier.

01:42:16.980 Right.

01:42:17.280 Right.

01:42:18.060 I mean, Rich, this is one for you to test.

01:42:19.860 Yeah. Why hasn't the ITP tested this yet, Rich?

01:42:22.000 Is it because the oral ones are just not strong enough and we want to-

01:42:24.740 Can you break your protocol and do an ITP with an injection?

01:42:28.300 Why?

01:42:28.540 Because it's enormously laborious to do weekly injections.

01:42:33.020 That sounds like an I need more money problem.

01:42:34.900 And also you need a separate control group.

01:42:37.320 That sounds like an I need more money problem.

01:42:38.920 Because the separate control group needs to get sham injections. And yes, if you increase

01:42:42.020 our budget dramatically, I think it's a worthwhile experiment. But what we're waiting for is oral 0.98

01:42:47.460 drugs that work, that you don't have to do injections of drugs.

01:42:51.120 I mean, there is an oral semaglutide formulation that's taken daily.

01:42:53.740 It was submitted to us this year. The detailed protocol, however, is, again, technically very

01:42:59.220 laborious. Each mouse has to be food-deprived for six hours. Then the material is administered

01:43:04.740 and then they have to have a change in their water balance for the next two hours. It is

01:43:10.400 technically not an injection, but it is not any less laborious. And in addition, you have

01:43:17.420 to have your own separate control group that gets all of those different manipulations with

01:43:22.380 a sham injection. Could you do three instead of five next year and make that one of them?

01:43:27.660 Reallocate some funding? Well, I'm not in charge.

01:43:29.760 It's a heavy lift. I'd vote against it. I would vote for waiting about a year until somebody comes

01:43:35.580 up with a pill that you can just mix into mouse food or water and give it to the mice and it'll

01:43:40.480 work. And these are going to be mice that are an incredible amount of stress from all the handling

01:43:45.000 the injections. Oh, yeah. Yeah. That's why the control group is necessary. But the companies are

01:43:49.060 putting so much money into this, they understand why people don't like to inject themselves.

01:43:53.360 I'm reasonably sure. I mean, I know nothing about it, but I'm reasonably sure that in a year or two,

01:43:59.400 there'll be some agent that works when you put it in the food of a mouse or pop it as a pill as a

01:44:05.580 person. Those would be enormously important to test. Do we know if terzepatide, for instance,

01:44:12.120 if we're given to people of normal body weight, do they also lose 15% of their body weight?

01:44:19.900 I have not seen the data on that. I can tell you anecdotally, having seen patients,

01:44:26.100 it's going to be dose dependent. So as you know, that drug is dosed from as low as two and a half

01:44:31.220 milligrams weekly to as much as 15 milligrams weekly. Usually people who don't need to lose

01:44:37.160 much weight, someone who says, look, I just want to lose this last 10 pounds and I've done all the

01:44:41.820 exercising and dieting I can do. They typically just lose that 10 pounds and they take a very low

01:44:46.820 dose. Now, to your point, if they took the 15 milligrams, would they become sarcopenic? I don't

01:44:52.800 know. I think this conversation points out again, how constraining lack of resources are. I mean,

01:45:00.140 there are probably like 15 or 20, 50 amazing questions that can't answer. I mean, every time I

01:45:04.900 hear Rich talk about this stuff, it just pisses me off because there's a bunch of stuff that should be

01:45:08.820 tested, should have been tested by now, that hasn't been tested, not because it's not a good idea,

01:45:13.960 but because there just isn't any resources to do it. Well, I think what's really frustrating as well

01:45:19.880 is that these are the types of experiments that would allow us to actually start to economically

01:45:25.720 model the impact of these drugs outside of just kind of a disease state. For example, if drugs like

01:45:33.520 these are indeed gyroprotective and people can work three years longer or five years longer because 0.98

01:45:38.380 they're healthier, think of the impact on that over at OMB. What does that mean to tax take? What

01:45:45.920 does that mean to delaying Medicare? What does that mean to reduced healthcare spending at the time when

01:45:51.920 it is most expensive? So- Last estimate I saw was 38 trillion a year for every year of healthspan.

01:45:57.780 Wow. That was a McKinsey report. That's 38? I'll send you the link. Not 3.8? Nope.

01:46:03.380 No, 38. That's analysis by Andrew Scott, his British economist. That's bigger than I would

01:46:09.940 have guessed. Wow. Can we just, because I'm in the mood to see you get spicy, can we just talk

01:46:15.760 about senescence for a minute? Senescent cells, he means, Rich. You know, the things that drive aging.

01:46:21.180 What do you mean? Do you want me to talk about senescent cells? Okay, yes, I'll be glad to do that.

01:46:25.920 It's a terrible historical accident. Leonard Hayflick way back found that human cells would

01:46:30.380 only divide 50 times and stop. One of his colleagues, a guy named Vittorio Defendi,

01:46:34.900 made a joke at lunch and said to him, hey, Len, maybe they're getting old. Ha, ha, ha, ha, ha.

01:46:40.220 And Len did not understand it was a joke. He thought it was a serious scientific hypothesis.

01:46:45.260 It's clearly nuts because we don't get old in a way that is modeled by having embryonic lung

01:46:51.800 fibroblasts. Stopped growing. But at the time, the hottest technique in modern medicine was you

01:46:57.780 could grow cells in culture. That was really so cool. You could do stuff with them. So all the

01:47:02.240 cell biologists who really wanted to use the coolest new toys, wanted to have a way of studying

01:47:07.300 aging without all these messy mice and rats and having to wait and stuff, they could do it in vitro

01:47:12.960 because this was in vitro aging. This is in vitro senescence. And the field, to skip 30 or 40 years,

01:47:20.640 the field went ahead with this metaphor without ever questioning it. It's now such an industry

01:47:27.580 that the people who review these grants and papers and advise billionaires and advise startup companies,

01:47:34.820 they all were trained in labs that just do senescence for a living. So they never stopped to

01:47:38.780 question. One of the most famous and best scientists in this area is a woman named Judy Campisi, 0.99

01:47:44.420 who just died last year. She and I were assistant professors together at Boston University. She and

01:47:49.600 I were going to send in a program project with a third person, Barbara Gilchrist. I was going to

01:47:53.800 study immunity and aging. Barbara was going to study skin cells. We talked, Judy, you want to study 0.59

01:47:58.480 cells in essence? So she read the literature. She came back to us and she said, it has nothing to do

01:48:03.920 with aging. I mean, it's good cell biology. It's good about cancer biology, but of course,

01:48:08.180 it has nothing to do with aging. And we told Judy, of course, it has nothing to do with aging. We

01:48:13.040 understand that. But the reviewers think it is aging. So if you can just keep a straight face for

01:48:18.240 the three hours of the site visit, pretend you think it has to do with aging, you'll get a great

01:48:24.300 score. And that's what happened. She got a great score. We got the program project when she moved

01:48:29.080 to Berkeley. She took her grant with her. And after a year or two, she had apparently convinced

01:48:34.360 herself that it was aging. It was close enough to aging. So the notion that aging is due to senescent

01:48:41.760 cell accumulation is bad for two reasons. It's a grotesque oversimplification. The evidence for this

01:48:48.920 is awful. But even worse, it again cuts off productive thinking. There almost certainly are

01:48:56.080 changes that occur in some glial cells in the brain so that as you get older, they start making bad

01:49:01.280 cytokines is bad for your brain. There probably are changes in some bone marrow cells or some cells

01:49:06.220 in the lineage that leads to the beta cells in the pancreas that lose the ability to divide.

01:49:12.340 And that's bad for you. And finding out how it happens is really important. But once you've

01:49:18.600 convinced yourself that's all the same thing, this cytokine, this sort of proliferation, this change

01:49:25.100 in ability to make specific fibrous connective tissue, let's call that senescence. It's the same

01:49:31.200 thing. You've lost what you need to think of good, careful, well-defined experiments with well-defined

01:49:37.620 endpoints. If you say that senescence, there is a thing called a senescent cell, the thing that's

01:49:43.760 happening in this glia and in this marrow cell and this pancreas, it's due to the senescent cell

01:49:49.200 accumulating. You've blocked off productive generation of research hypotheses. The last

01:49:57.000 point I'll mention in this rant has to do with senolytic drugs. So the ITP was asked to test an

01:50:04.500 allegedly senolytic drug called phycetin. It was given to us by someone who is using this now for

01:50:11.600 clinical trials and who has a company that's interested in senolytic drugs. So we gave it to

01:50:17.180 mice, it had no beneficial effect whatsoever. What's the mechanism of this drug's action?

01:50:22.040 Oh, it has no action. Has no action or it had no effect?

01:50:25.600 It had no effect. What is it supposed to do? It's supposed to kill senescent cells or something.

01:50:30.320 So we told this guy, sorry, it had no effect. He said, well, let's prove it, whether it had any

01:50:36.540 change in senescent cells. So we gave him blind tissues from each of the treated and untreated mice

01:50:43.260 and he tried a test and there were no changes in senescent cells by his marker. He tried six

01:50:48.340 different markers. There were no changes in senescent cells. So then he said, well, send the

01:50:52.900 brain and the liver and the muscle. Maybe the senescent cells have been changed in the brain.

01:50:57.300 So we sent blind samples to a colleague of his. There were no changes in senescent cells by any of

01:51:02.780 the markers that these folks looked at. So this drug, which is now being marketed in clinical trials,

01:51:09.780 and you can buy it. I'm sure it's a natural product. Yeah. It's a supplement. There's no

01:51:14.060 evidence as far as I know that it either has an anti-aging effect or remove senescent cells.

01:51:19.880 But once you've got a commercial company pushing this stuff and your whole brand, your whole lab,

01:51:25.380 your whole program project, and all the people who are reviewing you are convinced senescent cells

01:51:30.320 exist, they're bad and drugs can kill them. It's a snowball rolling downhill and a rant of the sort

01:51:37.020 I've just delivered has no impact on the field. So can I give a counter example? Yeah, I want to

01:51:41.440 hear your take. Because there's good experimental data that these things can be at least partially

01:51:48.000 eliminated. And when you do that, there's an improvement in health. And this has been done

01:51:53.360 both in a genetic treatment, which they prime these cells to be genetically killed. And it's also been

01:52:00.220 done with drugs, not with phycetin, I hasten to say. So I think there's strong evidence that getting

01:52:07.840 rid of these P16 positive cells, which is really what it's all based on, can have an improvement in

01:52:15.160 health and in longevity.

01:52:16.800 Is the Van Dersen paper you're talking about in which they were allegedly depleted?

01:52:20.880 Yeah, yeah, yeah, yeah.

01:52:21.740 Let me tell you about that, because I was on the program project.

01:52:24.780 Two papers. Okay, one was with the short live mice.

01:52:28.060 Yes. I'm going to talk about the one that is not the short live mice. It is a paper,

01:52:33.040 a famous paper by Van Dersen, Kirkland, and several other colleagues, Darren Baker.

01:52:36.920 Darren Baker.

01:52:37.520 Are these the guys at Mayo?

01:52:38.720 Yes.

01:52:39.160 Yeah, I remember this.

01:52:40.620 They've laughed, two of them have laughed. But yes, they allege that they could remove

01:52:44.320 senescent cells by taking genetically modified mice, giving them a drug, all the senescent cells

01:52:48.600 would go away, and the mice lived longer, according to the paper.

01:52:51.560 Because it was on the cover of Nature.

01:52:52.520 It was on the cover of Nature.

01:52:53.820 I remember this one.

01:52:54.420 I was a part of the program project. So was Judy Campisi. And my job was to do the lifespan

01:53:00.660 experiment. We got the mice from Kirkland and Van Dersen. We got Campisi's mice. We got the drugs

01:53:08.200 from them. And we gave the drugs to the mice at 18 months. And you know, they had no effect on

01:53:14.400 senescent cells. Not one. We tried seven times to show depletion of senescent cells in their mice

01:53:22.740 using their drug. And went zero for seven. We then took the tissues, blinded, and sent them to

01:53:30.600 Judy's lab, Judy Campisi's lab, so she could measure P16 cells. But she didn't know which ones

01:53:36.520 were from treated and which ones were untreated. When we undid the code, there was no effect on

01:53:41.520 senescent cells whatsoever. So I remained somewhat skeptical. I asked Van Dersen, had he measured

01:53:49.140 the number of senescent cells in his treated mice? No, we're planning to do that.

01:53:54.880 But what was the phenotypic change in the mice when you did this experiment?

01:53:58.020 Oh, when I, I didn't want to do an expensive lifespan experiment

01:54:01.080 with an alleged anti-senolytic drug until I knew that it was depleting senescent cells.

01:54:06.720 So how long did you treat for? I used their protocol and I asked them, I asked Darren Baker,

01:54:12.720 what is the dose? How long do you treat the mice? And how long after you add the drug should you wait

01:54:18.620 before you detect the removal of senescent cells? And his answer, astonishingly, was, we don't know.

01:54:25.720 We've never looked at that. But the nature mice were treated for how long?

01:54:29.720 They were a long time. Their whole life. A long time.

01:54:31.920 I think they started treatment in middle age, right? And I mean, in the published papers,

01:54:36.120 they do show a reduction in P16 positive cells. And you're saying you couldn't replicate that

01:54:41.460 in your lab. But we're conflating a bunch of different issues here. We're conflating the

01:54:45.080 genetic model with the drugs and do senescent cells even exist. And I feel like, I mean, I think

01:54:50.020 Rich's skepticism is valid in many ways. And there's actually a large body of evidence that

01:54:57.460 whether we agree on the definition of senescence, what people are calling senescent cells do

01:55:02.780 accumulate in multiple tissues with age in mice and people. And if you get rid of them,

01:55:09.320 you can see some health benefits. Am I convinced they have big effects on lifespan? No, I'm not.

01:55:13.660 Because the data is mixed. And even that genetic model, other people haven't been able to reproduce.

01:55:17.580 So it's messy. But I think partly maybe start with what is the definition of a senescent cell?

01:55:22.800 Because that's where a lot of this confusion comes from.

01:55:24.540 That's what I was saying, that there is no satisfactory definition.

01:55:28.300 Satisfactory to you? I mean, is your issue, Rich, that we talk about it like it's one cell,

01:55:34.460 but in reality...

01:55:35.000 Yeah, that's a big part of it. You can't think about it clearly if you imagine that these many,

01:55:39.180 many different kinds of cell intrinsic changes with potential pathological impacts

01:55:44.180 are all aspects of the same phenomenon.

01:55:46.860 But we do that with other things. We have mitochondrial dysfunction. There's lots of

01:55:50.000 different ways to get to mitochondrial dysfunction.

01:55:51.620 So the NIH has just put about $600 million into a network of researchers to study cell senescence.

01:55:59.080 And I'm on the advisory group for that. And to the extent that Rich is saying these are many,

01:56:05.180 many different things all pretending to be the same thing, that's clearly true. But they're coming up

01:56:10.360 with bigger and bigger and broader definitions of what a senescent cell is. But on the other hand,

01:56:16.140 they're also coming up with more and more interesting things that those senescent cells do,

01:56:21.220 either in tissue culture, which I don't put much... Or in mice. I don't think the NIH would put that

01:56:28.460 kind of money into something if they didn't feel there was a valid basis. I think part of this is

01:56:33.340 we're calling it senescence. And I think none of us... To me, that's stolen a really good word out of the

01:56:40.500 vocabulary. Because senescence just means aging. And it used to be, you could talk about calendar

01:56:45.340 aging, you could talk about senescence, which is what we now think of as aging. And now you can't

01:56:49.660 use this anymore. Because anytime you do, they think you're talking about these cells.

01:56:54.900 Is this what they call the zombie cell? I keep forgetting.

01:56:56.960 I keep trying to purge this from my memory. I mean, the most common definition, I think,

01:57:02.620 is just an irreversibly arrested cell that doesn't die and typically gives off a pattern

01:57:09.640 of inflammatory cytokines and other factors, which is a catch-all for a lot of different

01:57:14.880 ways to get there and a lot of different states that these irreversibly arrested cells can exist in.

01:57:21.480 Yeah. But even neurons, they're not considering senescent neurons and neurons are post-mitotic.

01:57:26.160 Right. But they don't always give off this pattern of signals, right?

01:57:28.820 No, no. No, that's right.

01:57:29.080 I mean, again, this is part of the problem. As you mentioned P16, I think even at the

01:57:33.180 molecular level, the catalog of markers that people are using to define a senescent cell is

01:57:39.100 changing and it seems to change. Broadening, yeah.

01:57:41.380 Yeah. I agree with much of what you're saying. I just don't think we should throw the baby out with

01:57:45.460 the bathwater here and say there's nothing to this. I think there is something to it. And I think

01:57:49.700 there's lots of evidence that are there enough similarities between all the different classes of

01:57:55.280 senescent cells that people are studying now that they should be categorized as one thing. I think

01:57:59.800 that's a valid conversation to have. It's a good discussion point. I don't think we know the answer.

01:58:03.840 And they discuss this a lot in the SENDECT because even the SASP, even these things that are oozing out

01:58:09.860 of the cells, varies quite a bit depending on the nature of the cell.

01:58:14.480 That's the problem, of course. You referred to it as almost anyone would as the SASP, the set of

01:58:21.240 senescence-associated proteins, secretory proteins. And once you think of it as the SASP, you've lost

01:58:28.700 because the key point is not to do that. The key point is here's a set of cytokines that this

01:58:33.820 cell has begun to make. That's really interesting. Here's another set overlapping probably. They make

01:58:39.420 it when you've made them stop dividing for a separate reason. That's interesting. We should

01:58:43.920 study that. But to think you've proven something about this cell type when you've actually been

01:58:49.680 looking at this cell type because the SASP has been changed.

01:58:53.920 But do you think it's possible that a drug such as rapamycin has part of its effect on aging through

01:59:00.660 a broad inhibition of a subset of the SASPs?

01:59:03.760 I think it's very likely that rapamycin changes cytokine production by many different cell types

01:59:11.360 and that some of those changes probably have health benefits. I would like to know what it does to the

01:59:16.340 cytokine production from the macrophages in the fat and the glial cells in the brain and cells that

01:59:22.360 are in charge of protecting you from viral infections. But the mistake is to say, yes, it's affecting

01:59:27.940 the SASP. It's easy to see an analogy. If I said, here's a drug and it helps you because it affects

01:59:34.180 neurons, you'd laugh at me because what you really want to know is, is it motor neurons, sympathetic

01:59:41.520 neurons, parasympathetic neurons, neurons in your hypothalamus, what part of the hypothalamus,

01:59:46.820 the ones that control appetite? And I said, no, no, no, it affects neurons. I've got a drug that

01:59:52.140 affects neurons.

01:59:52.940 But I mean, people are aware of these complications and are studying these complications now. It seems

01:59:57.620 to me that it's the terminology that you object to and I can appreciate that.

02:00:02.120 It's thinking that I object to. The terminology is problematic because it makes people stop thinking

02:00:07.560 about the important details and start imagining that they've had a thought when they say, I have

02:00:13.240 a drug that removes senescent cells. The problem is that the words trap you into patterns of thought

02:00:18.380 that are in this case, nonproductive and misleading.

02:00:20.600 It may be inefficient, but the field is making problems.

02:00:22.940 I would say quite a bit of progress. And I think the way you learn about the complexity

02:00:26.500 is you start with a simple model, you study it, and then your model gets more complicated.

02:00:30.500 So I totally get the frustration, Rich, because I get as frustrated as you are about senescent

02:00:34.420 cells about other things. But I think this is also part of the natural process here.

02:00:39.460 And I think what Steve said is really important. The fraction of the NIH budget that goes to study

02:00:44.480 the biology of aging through NIA has remained tiny. But senescent cells are actually a really

02:00:49.000 good example of how a bunch of people in other institutes are studying aging,

02:00:52.240 and they don't even know it. They're studying senescence in cancer, or senescence in Alzheimer's,

02:00:56.480 or senescence in kidney disease. So it actually has had an impact in broadening the appeal and

02:01:03.000 scope of the field outside of NIA in ways that I certainly didn't anticipate.

02:01:07.800 Do you think that going back to the meta problem at the beginning of our discussion,

02:01:11.760 do you think that's maybe a better way to think about allocating funds? So for example,

02:01:16.780 the NCI obviously receives the most funding within NIH. Maybe some of the NCI funding goes

02:01:22.820 to the NCI to study cancer prevention through Gero protection, right? If the turf war is what 0.99

02:01:30.140 matters-

02:01:30.480 Sure, but that's a good idea.

02:01:31.200 No, no, no, no. We've actually saw a group of us who are lobbying Congress have actually asked

02:01:37.060 the NIH to tell us exactly this. How much work in Gero science is going on in all these other

02:01:45.560 institutes? Of course, they're going to have some motivation to minimize that or maximize it or

02:01:52.200 something, but at least it will give us an idea. Right now, we have no idea how much of the NCI budget

02:01:57.240 is going to this or NIDDK or anything else. They already have produced a report that told us how

02:02:03.260 much they were spending in the NIA, but we already knew that. We wanted to know how much they're

02:02:08.240 spending in the other institutes.

02:02:10.180 I mean, I think that could alleviate some of the turf war issues, but I think what you really need

02:02:14.360 is the change in leadership and leaders who actually recognize why this is important. And that's

02:02:20.580 where it starts. We can have a conversation about how much power does the NIH director have? How much

02:02:24.500 power does the director of HHS have? But that's a place to start. If you can get people in those

02:02:28.460 positions who get it, it's going to have an impact. Let's talk a little bit about metformin.

02:02:33.020 Rich, do you think metformin is gyroprotective in humans? I know it doesn't appear to be in your mice.

02:02:38.620 I think the evidence is uncertain. There's a famous paper from Bannister that alleged that

02:02:45.520 diabetics on metformin had lower mortality risks than age-match.

02:02:48.580 You don't listen to my podcast, do you?

02:02:50.660 I do occasionally when-

02:02:51.700 Actually, no, no. You know what? It was a different podcast. I did a very lengthy treatise

02:02:57.020 in a journal club comparing the Bannister paper to the Keyes paper and came to the conclusion that

02:03:02.600 the Bannister paper had too many methodologic flaws to be valid.

02:03:06.140 That's exactly what I was going to say. As a matter of fact, Keyes Christensen, who's the senior

02:03:11.180 member of the group, and I have just written a review article which says exactly that. That's the

02:03:16.080 title of my paper.

02:03:16.440 Is it out yet?

02:03:16.920 That's under review.

02:03:18.360 Okay.

02:03:18.680 Yes. So, you know exactly what I was going to say, and I agree completely. The question

02:03:22.840 as to whether metformin would be gyroprotective that is-

02:03:26.560 In a non-diabetic.

02:03:27.380 Slow, in a non-diabetic, in humans, I think is interesting, unanswered. It's not the drug

02:03:32.900 I would have looked at myself. If I had a big set of dogs, for instance, and I wanted to

02:03:38.900 give them a drug that modified their glucose homeostasis, I would probably start with something

02:03:44.940 like canagliflozin that actually does work in mice and which is known to be safe over

02:03:50.420 the long term in people. Metformin is safe over the long term in people, but I don't think

02:03:55.280 there's much evidence that it's anti-aging, leaving aside how great it is for diabetics 0.68

02:04:00.460 and pre-diabetic.

02:04:01.020 What do you think, Steve?

02:04:01.680 I think it's very promising. I'm skeptical because I'm always skeptical in the absence

02:04:07.140 of evidence, but the observational evidence, ignoring the Bannister paper, just the consistency

02:04:12.640 of the observational data that it reduces dementia, cancer, cardiovascular disease, suggests to

02:04:19.500 me there's enough smoke there to look to see if there's fire or not.

02:04:22.520 I'll send you the Keys review article and then you can rethink that.

02:04:25.600 Okay.

02:04:26.200 But sorry, Steve, you're saying it does all of those things in diabetics.

02:04:29.920 Well, most of the studies have been done in diabetics. Absolutely. And how much of that

02:04:33.780 is just because you're curing the diabetes is an open question.

02:04:38.740 And how much of that is a selection for people in diabetes that are progressing much less slowly

02:04:43.680 because they're the ones that stay on a single agent as opposed to the ones that progress

02:04:47.400 into...

02:04:47.940 Right. Which is why you have to do the study.

02:04:49.900 Yeah.

02:04:50.440 Where is TAME in the world of...

02:04:52.380 TAME is in a very preliminary state. There's now enough money to get it started.

02:04:59.920 It has not enrolled anything yet?

02:05:01.660 It's enrolling right now.

02:05:02.940 Okay.

02:05:03.620 Previously, they didn't want to start it until they had enough money to do the whole thing.

02:05:07.260 It's been impossible to get that. There's now a small amount of money, enough to get

02:05:11.820 it started at a small scale with the hope that that will start the pot rolling. But yeah,

02:05:18.500 it's been around for eight years now. And I was in on the original discussion about,

02:05:24.080 do we do rapamycin? Do we do metformin? And it was all about cost and safety. That was the whole

02:05:30.720 thing. I went in strongly advocating for rapamycin. I came out saying, okay, there are these cost

02:05:36.500 issues. And I think it was important because when we went to the FDA, we didn't want them to think

02:05:40.600 that we were trying to make a bunch of money with this trial. And nobody's going to get rich from

02:05:46.460 metformin.

02:05:46.920 Why is generic sirolimus so expensive still?

02:05:50.260 I think it's supply and demand, honestly.

02:05:51.980 There's no need for them to ramp up production.

02:05:53.480 Yeah, I think so. But coming back to the metformin question, I mean, I think,

02:05:57.080 first of all, we don't know the answer. I mean, Rich is right. We don't know. So what are our

02:06:00.960 opinions? My opinion is diabetes probably accelerates biological aging and metformin is

02:06:06.260 effective at reducing diabetic symptoms and probably reduces biological aging in that context. Probably

02:06:12.200 doesn't in people who are not diabetic. That's my intuition.

02:06:15.300 Let me push back on that for a second, which is diabetes is an artificial diagnosis in that we just

02:06:20.700 make a cutoff. We say your hemoglobin A1c is 6.5. You have type 2 diabetes. If your hemoglobin A1c is

02:06:25.480 5.9, you don't. But there are data that we've looked at that suggest a monotonic improvement in

02:06:31.900 all-cause mortality as average blood glucose goes down measured by hemoglobin A1c in the non-diabetic

02:06:37.500 range. Meaning people with an A1c of 5 live longer than people with an A1c of 5.5 live longer than

02:06:44.440 people with an A1c of 6, all of whom are non-diabetic. Point being, if metformin's

02:06:50.580 gyroprotection comes through the regulation of glucose in the patient with diabetes, does it stand

02:06:57.340 to reason that even in patients without diabetes, further attenuation of hepatic glucose output

02:07:03.400 is going to improve all-cause mortality? Maybe. I don't know the answer, obviously. I think the

02:07:09.200 question is, is the biomarker in this case A1c, what is that actually reflecting, right? Is that

02:07:15.200 presumably reflecting some aspect of metabolic homeostasis? And so, first of all, does metformin

02:07:20.660 in non-diabetics have the desired effect or the effect we would associate with reduced mortality in

02:07:28.060 non-diabetics consistently? It'd be question number one. I don't know the answer to that. You probably

02:07:31.800 do. And I don't want to speak for Nir because it's been a while since we've spoken. But the last time I

02:07:36.460 head near on the podcast, his rationale for why metformin was gyroprotective had nothing to do

02:07:41.600 with glucose homeostasis in a non-diabetic. It was, and I know you're going to love this. I mean,

02:07:47.240 Rich, you're really going to love this. There was a figure of the hallmarks of aging and how metformin

02:07:52.560 acted on each of them. But my point being, not to say that that's incorrect, correct, or anything,

02:07:56.620 it's that there was something much more primal about metformin's actions. Now, here's my pushback

02:08:04.580 on that. Metformin requires an organic cation transporter to get into cells, as I've learned

02:08:09.560 somewhat recently, that muscles don't have. So, if you look at the tracer studies, metformin does

02:08:15.220 not get into muscles. It gets into enterocytes and the liver. It's very concentrated in the liver,

02:08:21.480 gets in the gut, unclear from these tracer studies if it's getting into immune cells.

02:08:27.000 So, Nav Chandell tells me that he believes they are getting into immune cells as well.

02:08:31.680 So, the question is, at least I think we need to ask ourselves the question, if it's working,

02:08:37.060 which cells is it working on and how? And so, the liver part's easy. Everybody gets

02:08:41.920 big concentration of metformin shows up here. We sort of understand that that reduces hepatic

02:08:47.560 glucose output. After that, I'm sort of scratching my head going, I don't know how it works.

02:08:51.100 Well, we know it has a target in the mitochondria, complex one. It inhibits, we know-

02:08:55.480 But in which cells?

02:08:56.400 Well-

02:08:56.960 That's my point. Like, it's not in the muscle.

02:08:58.220 That's a question. And we also know that it activates AMPK.

02:09:01.980 But those mechanisms are probably related.

02:09:03.860 This is why Nir points at two of the hallmarks. I just have to tell you this. But here's an

02:09:08.860 interesting thing. A good friend of ours, George Martin, who died a couple of years ago, once

02:09:12.980 went through and cataloged all the human diseases he could and tried to look at the similarities

02:09:19.740 of their phenotypic changes relative to what happens with normal aging. He came up with diabetes

02:09:25.440 as having the most similarities to accelerated aging of any of the groups that he looked

02:09:30.980 at, which in this context-

02:09:33.080 And it makes sense. The glycosylation, the hypergrowth factors like insulin, IGF-1, all

02:09:37.800 these things. I mean, there's logic to that.

02:09:39.720 Let me agree with the emphasis you were just putting on organ-specific and tissue-specific

02:09:45.520 changes. And I think it's about time to get away from what does metformin do to the body,

02:09:52.960 or any of these drugs for that matter, and start to think what does it do to each of the

02:09:56.820 interesting players and how they talk to one another. Someone in my lab has been looking

02:10:01.740 at the enzymes related to de novo lipogenesis. And she's been looking at a couple of different

02:10:07.100 kinds of slow-aging mice. And it has major effects in the liver, and it has major effects

02:10:12.600 on white and brown adipose tissue, and they go in different directions. And which is primary,

02:10:18.400 which is reactive, whether any of these are related to the effects of the mutations on the

02:10:22.880 muscle or the brain is now an open question. So having a diagram of hallmarks which are changed

02:10:29.920 by a drug is much less useful than asking what specific changes in what cell types of which

02:10:36.580 organs that talk to each other are being changed by this drug as a primary or as a secondary or as

02:10:42.300 a compensatory effect. That's how you'll start to get into first mechanisms, but also start to be

02:10:47.640 able to think clearly about ways of targeting therapy so that it has a benefit with fewer and fewer side

02:10:53.240 effects.

02:10:54.320 Let's use konegaflozin as an example. We've demonstrated, and I use we very liberally here, 0.99

02:10:59.660 you've demonstrated that it reduces all-cause mortality in your mice, in males. And we know

02:11:05.560 exactly what konegaflozin does in the kidney, and we know that those mice lived longer. Do

02:11:11.360 you believe that the longevity benefit came through glycemic control? Because there was

02:11:16.540 no difference in weight, if I recall. They actually lost weight. Males and females lost

02:11:20.400 weight on konegaflozin. Was the difference in weight statistically significant between the

02:11:24.500 long-lived males and the normal males? The mice treated with the drug were lighter in weight

02:11:30.060 than controls, and that's true of both sexes. 0.83

02:11:32.060 Both sexes. So the weight loss wasn't necessarily what explained.

02:11:34.860 They actually lost more weight in females than in males. So the question is very valid, and we do 1.00

02:11:40.220 not know the answer. SGLT2 is on many other cell types, and it's quite possible, very plausible,

02:11:47.580 that konegaflozin had an effect principally through controlling peak daily blood glucose,

02:11:52.860 not average, but peak. And it's also possible that it had effects on cells of unknown origin in the

02:11:59.000 brain, and all of these are very valid, and I don't think anyone knows the answer. It's well worth

02:12:03.900 evaluating. There are other inhibitors of SGLT2 and SGLT1 that have differential cell specificities

02:12:11.220 and differential effects on different cell types, and looking at those would help give you glimpses

02:12:17.800 into this question. We guessed it had to do with glucose, but we might be wrong.

02:12:21.640 So what is your intuition, Steve? Going back to metformin.

02:12:24.580 My intuition is that it might work. I don't have a strong opinion. There's enough suggestive evidence

02:12:31.240 that I think it's worth a trial. I think that if we wait until we figure out exactly what each drug

02:12:36.880 does in each cell type, it will take us forever to get any therapies. And in medicine, there have been

02:12:43.200 many, many advances that came about before we understood the mechanistic underpinning. And if there's

02:12:49.660 enough suggestive evidence, and there's not a lot of side effects, suggest me that it's worth digging

02:12:55.360 into now, because the benefits are so enormous. Like we said, one year, healthy aging, $38 trillion.

02:13:02.140 That should talk to Congress if nothing else does.

02:13:04.280 Yeah. Well, and I would also say TAME could be successful independent of whether metformin is 0.65

02:13:09.220 effective at slowing biological aging necessarily.

02:13:11.700 By getting others into the field, you mean?

02:13:13.580 Well, also even just hitting the end points, right? So the end point is multimorbidity or

02:13:17.580 comorbidity. So it's quite possible that the trial will be successful, even if metformin is not

02:13:23.220 effective gerotherapeutic, which is also true. It may not succeed for a variety of reasons. The

02:13:28.840 clinical trials don't succeed. I sort of agree with Steve. Like I'm supportive of doing the trial.

02:13:33.720 I also agree, I think with probably both Steve and Rich, that it's not what I would pick if I was

02:13:38.700 going to do one trial, if we could only do one trial, but we have to start somewhere.

02:13:42.120 Matt, why do you think that the ITP studies for rapamycin always worked, regardless of start

02:13:48.820 young, start old, give it with metformin, do it by itself, always worked, and the mice are taking

02:13:53.200 rapamycin every day?

02:13:54.860 Because inhibiting mTOR increases lifespan and slows aging. I know what you're asking.

02:14:00.420 Yeah, you know where I'm going.

02:14:01.200 Because most people who are using rapamycin off-label have moved to once weekly or some

02:14:07.600 sort of cycling like that. So I think one question is, would that increase lifespan in mice as much

02:14:14.240 or more than daily? We don't effectively know the answer to that question, I don't think.

02:14:19.340 You guys do some intermittent...

02:14:21.000 It can increase lifespan, but it's never been dose-optimized, right? I think this is the question.

02:14:25.440 Is the metabolic rate of the mouse so fast that giving the mouse daily rapa is not the same as

02:14:31.940 giving the human daily rapa?

02:14:34.360 Yes, and the rapa in the ITP study is in the food. So it's not a single dose, or it's not a single...

02:14:40.460 They're just chowing on it all day.

02:14:41.840 Well, at least during the period of day that they're eating and have access. I'll let Rich

02:14:45.740 talk about what they know about the blood levels, but it is a fundamentally way of delivering the

02:14:50.040 Why did you guys decide... I mean, I guess in 2008 or 7 when you did the first study, maybe it wasn't

02:14:56.180 clear this idea of mTOR1 versus mTOR2 and the constitutive dosing.

02:15:00.720 Maybe we should ask how many people at this table actually believe that model.

02:15:03.980 Yeah, that's kind of where I want to go. I want to understand what we think is true and

02:15:08.520 not true about rapamycin based on this experience.

02:15:11.500 I guess I don't understand the question.

02:15:12.980 That the bad side effects come from mTOR2, off-target effects of rapamycin, and all the

02:15:17.960 good stuff comes from inhibiting mTOR1.

02:15:19.980 I don't know enough to say. Many of our slow-aging mice, actually, mTOR complex

02:15:26.000 1 function is down in all of them, but mTOR complex 2 is often up, and it's up in an

02:15:31.740 interesting way. Mice eat mostly at night, and they more or less fast during the day.

02:15:37.720 In our slow-aging mice, mTOR complex 2 is elevated, but it no longer responds in the

02:15:44.720 fasting period, but it doesn't respond to food in the same way. So they're complex changes

02:15:48.960 in both its baseline state and its response to food, whether these would happen in people,

02:15:54.500 what happened in people taking it every other day or every fifth day, whether they are beneficial

02:15:58.440 or harmful or a mixture. I really don't know. The mTOR complex 2 story is trickier.

02:16:03.720 The other thing that is, I think, important but not really appreciated is that not all mTOR

02:16:08.720 complex 1 drugs, like rapamycin, not only lower the overall effect, but it also changes the

02:16:14.960 substrate specificity so that the kinase that is susceptible to TORC inhibition that looks at a

02:16:23.020 ribosomal protein, S6, that goes down. It doesn't work nearly as well. It's inhibited.

02:16:28.400 For how long?

02:16:29.180 I don't know. But the other aspect of TOR downstream is on a protein called 4-EBP1 that's involved in

02:16:35.420 translation. It does not change that kinase. What it does is it changes the total amount of the protein.

02:16:41.640 So the proportion of the protein that's phosphorylated drops down, but the actual

02:16:46.880 kinase that adds the phosphate to that substrate is unchanged. So whether that's important, that it's

02:16:55.120 having at least two different pathways that are being influenced in one case by changing the

02:16:59.900 substrate and in the other case by changing the kinase, no one's really looked at that. They say it's

02:17:04.760 a drug that blocks mTOR kinase 1 function. And downstream is where a lot of the action is. I know

02:17:10.920 your lab at one point was interested in cell type specific inhibitors of the TOR complex one. I

02:17:17.280 don't know whether...

02:17:18.360 Everything you just said, rich occurs in what cell?

02:17:20.840 Mouse liver.

02:17:21.980 What about muscle?

02:17:23.200 Well, the overall decline in the ratio of phosphorylated versus substrate, we also published

02:17:28.720 that I think in muscle and kidney. I would have to go back to the papers and see whether we also

02:17:34.440 found the elevation of the substrate, the 4-EBP1 in both of those tissues. I vaguely recall that it

02:17:41.600 was the substrate that changed, not the kinase in those tissues as well, but I'd rather look it up

02:17:47.160 before I sign my name to it.

02:17:49.500 But even what Rich is saying is, wow, I mean, really important and informative. Also,

02:17:53.920 only a tiny piece of all the downstream things that mTOR affects. And I think the point is we just

02:17:58.900 really don't have a good understanding of how rapamycin or fasting or other drugs that hit

02:18:05.240 mTOR are affecting all of the things that are downstream of mTOR.

02:18:08.620 I agree completely. Let me give you an example. So Linda Partridge just published in bioarchive,

02:18:13.120 at least a nice paper, rapamycin increased lifespan of her mice. If she added an inhibitor of a 0.98

02:18:17.640 different kinase called ERK, it did better. The inhibition by ERK worked by itself, but it actually

02:18:23.400 improved on rapamycin. So two people in my lab were looking at that. And it turns out that the

02:18:29.960 ERK kinase inhibitor is working in an entirely different pathway. It's affecting the proteome by

02:18:35.920 increasing the degradation through a chavron-mediate autophagy mechanism, which is not affected by

02:18:43.320 rapamycin.

02:18:44.120 At least at the dose they used, right?

02:18:45.820 At least at the dose they used. Right. That's right.

02:18:48.380 Sorry, what model was this? This was mice.

02:18:50.240 Mice.

02:18:50.900 Probably Black 6. 0.99

02:18:52.060 No, we never used Black 6.

02:18:53.540 No, no, not you. I'm talking about the Partridge paper.

02:18:55.700 No, it was an F1 hybrid.

02:18:56.760 Oh, okay.

02:18:57.200 Actually. So it's agreeing with and amplifying the question. There may well be multiple

02:19:02.060 cell intrinsic pathways, some of which are TOR dependent, some of which are MAP kinase,

02:19:07.900 ERK dependent, which can synergize as in the Partridge case for lifespan, but also potentially

02:19:13.640 synergize for health impact.

02:19:15.180 Yeah. And here's, I mean, I think an important, again, limitation to what's been done. There are drugs

02:19:20.500 out there that hit both types of kinases. There are drugs out there that are ATP competitive

02:19:25.700 inhibitors that have different affinities for different types of kinases. Haven't been tested

02:19:29.800 for longevity. These dual kinase inhibitors. In fact, in the Restore Bio trial, the last one,

02:19:36.340 the phase three, which did not get to completion, they substituted. They took the RAPA log out and

02:19:41.920 used an ATP competitive drug.

02:19:44.800 Didn't know that.

02:19:46.140 So what is your belief, Matt, around dosing RAPA in humans then, or even in your dogs? 0.86

02:19:52.340 You're doing it.

02:19:53.120 We're doing it once a week now. We've moved to once a week. So, I mean, maybe it's worth at least

02:19:57.340 talking about how that evolved. And this is my understanding of how we got to where we are

02:20:02.140 today, which is that most people using rapamycin off-label for potential health span effects,

02:20:07.520 most doctors prescribing it are recommending once weekly dosing in the three to six, sometimes eight,

02:20:14.060 10 milligram range. So the first place I'm aware of in the literature where this was shown to have

02:20:21.780 a potential benefit for anything related to aging was Joan Mannix's work when she was first at Novartis

02:20:28.260 and then at Restore Bio looking at flu vaccine response in elderly people. And they were using

02:20:35.420 Everolimus, so a derivative of rapamycin. And they found that for vaccine response, it was most

02:20:43.100 effective and had the least side effects at once weekly dosing at five milligrams. And they tested

02:20:48.620 daily one or two migs, five migs once a week, 20 migs once a week. 0.96

02:20:52.900 Yeah. It was once, it was a milligram a day, five once a week, 20 once a week. Now I've had both

02:20:57.100 Lloyd Clickstein and Joan Mannix on the podcast. It's been so long that I don't recall if I asked

02:21:02.820 them why they designed the trial with those four arms.

02:21:05.640 So my understanding is that Novartis had internal data at that point on side effects and had an internal

02:21:12.620 hypothesis that if you let the trough levels bottom out, that reduced side effects. The side effects

02:21:19.320 in organ transplant patients were largely driven high troughs.

02:21:23.640 Yep. And then after that, they developed, based off of David Sabatini's work and then Dudley

02:21:28.340 Lamming after he left David's lab, a hypothesis that chronic treatment with rapamycin, which

02:21:34.840 maybe would be equivalent to daily dosing in people, this was all done in cells, led to off-target

02:21:39.820 effects on mTOR complex 2. And it was mTOR complex 2 effects that were driving the side effects.

02:21:45.020 So that got sort of dogmatized as the truth. Actually don't think there's a ton of evidence

02:21:51.740 beyond those initial papers to support the idea that the side effects are all through mTOR

02:21:56.120 complex 2. The idea is if you dose once a week, you let the trough levels bottom out. You don't

02:22:02.560 get the off-target effects on mTOR complex 2. You avoid the side effects. Again, we don't

02:22:07.000 have definitive data. The data I've seen seem consistent with that idea. People dosing daily

02:22:12.480 seem to be more likely to have side effects, mostly things like bacterial infections or

02:22:17.420 the really severe mouth sores, but sort of anecdotal. And I don't know for sure how strong

02:22:22.140 that data is in people. It did hold up in all of the rest or bioclinical trials that I'm

02:22:26.940 aware of. That once-weekly dosing really didn't show any side effects different from placebo.

02:22:32.140 In the dog study, you're using a slow-release formulation.

02:22:36.000 It's an enteric coded. It's a different formulation than what the ITP uses, but all of the human

02:22:42.340 sirolimus formulations have some way to get to the small intestine. So it's not substantially

02:22:49.080 different, I don't think, than rapamune or the generic sirolimus you would get.

02:22:53.760 Let's do the closest thing that a group like this could do in terms of a speed round. I'm going to go

02:22:58.400 through a couple of other ideas. I just want to get the, what are you thinking about this?

02:23:03.460 Can we say anything positive about resveratrol?

02:23:07.880 No.

02:23:08.880 Rich?

02:23:10.580 No.

02:23:11.740 Why does this thing not die? Why is there still a hundred different resveratrols being sold on

02:23:19.540 Amazon? Why do I still get people asking me, do you take resveratrol? Should I be taking resveratrol?

02:23:26.680 It has a good PR team.

02:23:28.460 I think it's really hard to prove something doesn't work. So once it gets in the consciousness

02:23:33.020 as improving health, I mean, even in the longevity field, Jesus Christ, I was saying

02:23:37.980 the resveratrol stuff was garbage for 10 years before people believed it. Now everybody believes

02:23:43.780 it, but it takes a really long time.

02:23:46.280 Well, at least in the aging field, like you never see people studying resveratrol in the aging

02:23:51.220 field anymore. I think if you went to a conference and asked scientists, what do you think about

02:23:55.320 resveratrol? You'd get the same answer here with maybe one exception. But I think, I think

02:24:00.280 it takes a really hard, just one exception.

02:24:03.160 It takes a long time to change.

02:24:06.140 Bad ideas don't die hard.

02:24:07.480 That's right. And that's true in the scientific literature. And it's especially true when there's

02:24:10.780 a profit motive to continue selling this stuff. And I'm not a hundred percent convinced that there

02:24:16.040 are no health benefits from resveratrol. Pretty convinced there's no reason to believe it affects

02:24:21.480 the biology of aging or as a longevity drug, but I can't say for sure that nobody would ever benefit

02:24:26.780 from any dose of resveratrol.

02:24:28.240 Yeah, but we couldn't say that about anything.

02:24:29.740 I agree.

02:24:30.200 Yeah, yeah, yeah. Now we could say that if you were force fed the highest fat diet in the world,

02:24:35.180 such that your liver encroached on your lungs through your diaphragm, isn't there a chance,

02:24:40.880 Rich, that under that situation, resveratrol might help?

02:24:43.900 I have no idea.

02:24:44.820 Yeah. Wasn't that the one and only one experiment that worked?

02:24:48.460 Yeah, the famous experiment, which was published as resveratrol, the first drug ever found to

02:24:53.960 extend mouse lifespan. It turns out that the mice die because they were on a 60% coconut oil diet.

02:25:00.980 It's poisonous to the extent that it causes the liver to fill with fat and compresses the thorax so

02:25:08.380 that they cannot inhale. Three or four papers later, they published as an obscure paragraph

02:25:14.360 and a discussion section on a paper. Pearson was the first author of the second paper that,

02:25:20.520 oh, by the way, all these mice on the coconut oil diet, finally, we've looked at them,

02:25:25.400 they're all dying because of lung compaction due to expansion of the liver. So the notion that

02:25:32.540 their drug had slowed aging because on the 60% coconut oil diet, it temporarily extended lifespan

02:25:40.140 was due to the prevention of this extremely bizarre phenomenon.

02:25:45.220 I just cannot get enough of that story.

02:25:47.420 Well, it's all documented in the literature.

02:25:48.820 No, no, no. I believe I know it well.

02:25:50.540 I mean, two separate papers.

02:25:52.040 All right. Let's have a word on NAD, NR, NMN. Steve, what is your point of view on this?

02:25:58.720 Well, the current state of evidence, I'm skeptical. It's one of those

02:26:01.620 things that makes a great deal of conceptual sense, but the evidence at this point is not

02:26:08.160 very compelling. And we have the ITP evidence that is, I think, the strongest. And there was

02:26:14.560 strongest negative evidence.

02:26:16.580 Yes.

02:26:17.020 Yeah.

02:26:17.220 Okay. Just to make it clear.

02:26:18.520 I assumed that people knew that. I guess I should.

02:26:21.100 And is it your view, Steve, that this stuff probably does not extend lifespan,

02:26:25.960 but maybe there is some other healthspan benefit out there that has just not been studied. The

02:26:34.140 right experiment hasn't been done. It hasn't been powered. Pick your favorite excuse.

02:26:37.760 Oh, yeah. I think NAD is a very, very interesting molecule. And I don't think we could

02:26:41.260 throw out manipulating NAD as something that could be important for aging. I just don't think the

02:26:46.100 evidence is there at this point.

02:26:47.500 Do you think if you're going to manipulate it, you would have to do it with

02:26:50.640 really, really high intravenous doses? Or do you think you could achieve those levels using

02:26:55.040 oral precursors?

02:26:56.800 That, I don't know. I will express complete ignorance on that.

02:27:00.480 Matt, what is your point of view on all of this?

02:27:02.780 Yeah. Well, I think the way you framed that question to Steve is indicative of why it's so

02:27:07.120 hard to disprove something, especially when there are people out there who have money to make,

02:27:11.400 who really want to make the case that you should buy this stuff. Because it's always possible that

02:27:15.300 there's some way that this could be beneficial. Having said that, NAD, like Steve said,

02:27:19.300 central molecule in thousands of chemical reactions, really important. Good reason. I

02:27:24.600 don't know about good reason. Some reason to believe that NAD homeostasis declines with age,

02:27:29.040 like lots of many other things. So it's plausible that if you fix that, you can get benefits from

02:27:33.460 it. The data is decidedly mixed, both in the literature, preclinical literature, and in people

02:27:39.920 as to whether or not boosting NAD increases lifespan, improves healthspan. So I think there's lots of

02:27:45.840 issues.

02:27:46.500 What's the most positive data you would point to?

02:27:48.540 Well, for lifespan, the original study by Johan Auerx's lab, where they started treating,

02:27:54.120 I think, at 20 months of age, was published in Science, I believe, showed an effect that was

02:27:59.280 reasonably good-sized, except the controls were short-lived, which is a different issue.

02:28:03.760 Yeah, it's a different issue, right? There's a number of cases where something was reported

02:28:08.360 to increase lifespan when the controls were short-lived, and then when the study was repeated

02:28:12.240 and longer-lived controls, you didn't see an effect. I don't know why there was a difference

02:28:16.260 between that study and the ITP, but that's probably the best case you can point to. There's studies in

02:28:22.020 C. elegans as well, where NAD precursors increase lifespan. So there's evidence out there, and again,

02:28:27.640 it's plausible. The biology is plausible. But then I think when you talk about the precursors,

02:28:32.060 it's even more complicated than maybe boosting NAD could slow aging, because can you get the right

02:28:38.660 doses in people? You talked about bioavailability. Is there any difference between NMN, NR, niacin,

02:28:44.620 nicotinamide? When you take it orally, the data suggests that it all gets broken down to niacin

02:28:50.040 in the gut. So why are people taking $70 NMN or NR? Yeah, why are people selling it? The people who

02:28:57.380 are selling it, some of them are scientists, dodge that question. It's complicated. I don't personally

02:29:03.300 believe there is enough evidence to think that NAD precursors, as they are being marketed today,

02:29:09.820 are likely to benefit most people. Some people, probably people who have conditions of dysregulated

02:29:16.140 NAD could get a benefit. I don't think there's any difference between the various molecules that

02:29:21.260 are being marketed right now. And there's at least one study in mice that giving NMN to aged mice

02:29:28.000 causes kidney inflammation and potentially kidney pathology. I'm not saying NMN's dangerous,

02:29:32.600 but when you try to weigh the risk-reward, if it causes kidney pathology in aged mice, at least at

02:29:38.320 high doses, could it do the same thing in dogs or people? Yeah, it could. And it bothers me,

02:29:43.760 particularly in the companion animal space, that people are marketing NMN for people's pets when

02:29:50.220 they know that it might cause kidney disease in people's dogs and cats. That's problematic to me.

02:29:55.580 We talked briefly about parabiosis and plasmapheresis. Let's come back to it a little bit.

02:30:00.400 Steve, is there going to be a day when the substance found in the blood of someone much

02:30:07.580 younger than you, when infused into you, whilst some of your old blood is removed,

02:30:14.840 is going to, assuming we figure out at what frequency that has to be done, impact your life?

02:30:20.600 Yeah. I think this is an incredibly interesting question and it really deserves to be investigated

02:30:27.780 in detail. Because if it's true, it's a real game changer because we do transfusions. I mean,

02:30:34.160 this is not exotic medicine. I think we very much need to know whether this works the same way in

02:30:40.820 people. And also it would be nice to know how much of it is due to the taking out versus how much of

02:30:47.380 it is getting rid of the old blood. But the evidence from mice is very, very compelling.

02:30:52.580 It is. Steve, if we could design the perfect experiments that would try to ask these questions,

02:30:58.120 let's just say we started by doing just the one experiment, which was the full parabiosis.

02:31:02.360 So the putting in, the taking out, we didn't try to disentangle the effect and there was no benefit

02:31:07.860 in humans. What would be your best hypothesis as to why it would have failed? Assuming it was

02:31:13.720 statistically powered correctly and there was no methodologic error. If this was a biologic result,

02:31:19.340 why would you think, given how favorable this has been in mice, it would not occur in humans?

02:31:24.920 That the products that ended up in the circulation of humans was a very different

02:31:31.760 nature than in mice. The number of things that differ between humans and mice and blood would be

02:31:37.840 enormous. So pinning it down would be. But I think there probably is some reason to suspect that it

02:31:44.680 may work. I'm very impressed. I mean, if it does work, this is an opportunity that we had the technology

02:31:50.040 to do this 50 years ago, right? Right. And it may not work in young people, but it may work in older

02:31:56.340 people. I think there's a lot of drugs that could affect aging that because young people haven't

02:32:02.420 aged as much might not have minimal effect, but you give it to somebody that, you know, 50 years

02:32:07.220 later might have a big effect. I find myself frustrated by the question rather than by the

02:32:12.580 answer because... You've got a horrible question asker here, Rich, is the problem.

02:32:16.300 I think you are well above average, but this particular one I think is illustrative because

02:32:22.020 the reason people like parabiosis is that they've seen it in a sci-fi movie. It sounds exactly like

02:32:27.660 what you do in sci-fi and they're flashing lights and it's so sexy and it's just so great. And you

02:32:33.360 can take the blood of young virgins and give it to old people and they stand up and they can get on 0.99

02:32:38.520 the road. I didn't realize they had to be virgins.

02:32:39.760 They have to be, yeah. Okay.

02:32:41.080 But none of that is pertinent. Pertinent is, is there something that is in the blood of old people 1.00

02:32:46.940 that it would be good to remove? And if so, what is it? And is there something, a cell,

02:32:53.220 a molecule, a set of three molecules that's in the blood of young people or mice that would be good

02:32:59.980 for you? The only virtue of this parabiosis circus is to suggest that, you know, the answer might be 1.00

02:33:08.000 yes. There might be something you could remove from old blood, a cell or some plasma molecule,

02:33:14.540 and there might be something good in the blood of young individuals. So the challenge now is to

02:33:19.100 find out what those things are and then you can do real life science. Real life science is not done

02:33:25.300 by taking blood from young people and putting it into old people. That's medieval science where 1.00

02:33:33.080 it's a complex mixture of dozens to hundreds of potential-

02:33:35.820 Right, but that could be the proof of principle. In other words, you might start with that and no one

02:33:40.740 thinks that if you do that experiment where you literally take blood out of an old person

02:33:45.980 and discard it and take blood out of a young person and put it in and you get a favorable result,

02:33:52.000 nobody thinks that that's what's going to the FDA. That is the proof of concept- The issue is what

02:33:56.880 experiments would be worth, you have a limited amount of volunteers, doctors, and money. What experiments

02:34:03.120 are most informative? And in my view, by far the most informative experiments are,

02:34:07.740 what is in the blood of young mice that is so good? And what is in the blood of young mice?

02:34:13.100 But I don't know, would you want to go on that fishing expedition until you at least saw a signal?

02:34:17.440 Yes. People are doing it. I mean, there are companies-

02:34:19.800 Of course they are.

02:34:20.120 Companies doing it and on the basic research side.

02:34:22.140 Of course they are. I'm asking a different question though, which is-

02:34:24.900 Yes, that's the only way you can turn your idea into science.

02:34:27.920 Well, on the other hand, if it has a positive effect, I don't think it really matters. That's

02:34:33.640 something to be investigated later. My thought is it's not simple. It's not one thing. It's not

02:34:39.320 GDF-11 for sure. If it were simple, there's enough people looking at it, they would have figured it

02:34:44.120 out. My guess is it's some combination. If there's something there, there's some combination.

02:34:48.120 I mean, why can't you do both? I think Peter and I are saying the same thing.

02:34:51.020 Would we love to understand the mechanism? Yeah, absolutely. Do we have to understand the mechanism

02:34:55.400 to figure out whether it works? And people, no. And if it works, great. That's a win too. I think

02:34:59.920 Rich's point is, we only have so much money, let's spend it on figuring out the mechanism. But

02:35:04.660 again, that's a fundraising issue. It's a scientific question. If you have a choice,

02:35:10.020 the ITP loves to test individual chemical compounds, even sometimes ones where the mechanism

02:35:15.280 of action is not known. And that's very sensible. We are very dubious about, let's take a little of

02:35:20.900 this and a little of that, a little of that. And we're really dubious about taking,

02:35:24.160 let's grind up the asparagus. Who knows what's in it? Let's see if it works.

02:35:27.480 I agree. But yet you guys have tested natural products where we have no clue what the mechanism

02:35:31.320 is. Or even metformin, you pointed to complex one inhibition. Yeah, that's one thing metformin

02:35:34.760 does and it might activate AMP kinase to that mechanism.

02:35:36.780 I'm not saying we have to know the mechanism exactly of each drug. What I'm saying is that if

02:35:41.380 you have a very complex mixture of hundreds of molecules and something happens, you don't know

02:35:47.840 what to do next because it could be any one or two or eight or 10 of those and you haven't really

02:35:53.360 decided, you have trouble then with standardization, with mechanistic tests and with transferring to

02:35:59.680 a key species like some humans.

02:36:01.500 My thought is we still wouldn't be using anesthesia if we had to wait until we figured out how it worked.

02:36:08.900 Yeah. And it doesn't have to be parabiosis. It doesn't have to be taking blood from young

02:36:13.500 people and putting it into old people, right? There are other variants of this that can be done

02:36:17.760 clinically and there's some evidence to support things like therapeutic plasma exchange or things

02:36:22.300 like that. So should we test it? I think so. And my gut feeling is, yeah, it probably will have

02:36:28.340 some benefits in people. So if you could only do one experiment, would you do a plasmapheresis

02:36:34.220 experiment? And if so, would you test the simplest one is you literally just exchange old plasma for

02:36:41.480 albumin. That's what they're typically doing in these studies.

02:36:43.920 Yeah. First of all, I don't know enough about this area to be confident in my answer, but yeah,

02:36:49.880 that's probably where I would look to start simply because it's going to be logistically easier to do

02:36:54.900 from a clinical trial perspective.

02:36:57.440 So scientifically then the hypothesis is it's the presence of something bad.

02:37:02.160 Well, it's both.

02:37:03.180 That is worse than the absence of something good because the albumin is not going to give you

02:37:08.000 the young person.

02:37:09.000 That's the problem with that experiment to me. We don't know now if it's young blood is good,

02:37:15.400 old blood is bad, or some combination. We would automatically, if we only did the plasmapheresis,

02:37:21.580 we would only be testing part of that.

02:37:23.260 I'd push back on that. I think we do have reason to believe it's a combination of both. There's data

02:37:27.580 in both directions.

02:37:28.900 That's why I proposed starting with that experiment.

02:37:31.320 I think that's, again, as much as anything's sure in this field, that's not as sure as rapamycin

02:37:35.700 increases lifespan in mice, but there's at least evidence to support that idea.

02:37:39.700 Last thing I'll say is you asked why might it fail in humans. I think Steve's answer is

02:37:43.200 valid. It's also worth mentioning, at least with the parabiosis experiments, the parabiosis

02:37:48.120 experiment itself shortens lifespan in rodents. And so just the fact that you're surgically connecting

02:37:55.600 these animals together. So it may be that the benefit from parabiosis, true parabiosis in that

02:38:01.320 context, is somehow related to the shortening of lifespan due to the procedure. I don't think

02:38:06.180 that's the case because there's other lines of evidence that argue against that, but there may

02:38:10.420 be something about the procedure itself that is impacting the outcome.

02:38:13.660 That increases muscle repair and improves cardiac function. It just seems to me that...

02:38:17.920 I agree. I'm just saying that may be an alternative explanation for something that's limiting in

02:38:23.220 those mouse experiments.

02:38:24.720 Just seems like there's not enough time and not enough money to do the work. Hopefully some of

02:38:29.400 that's changing. If we were to do another longevity round table next year, which is problematic

02:38:35.500 because this table, you guys are going to have to get awfully cozy. Any nominations for

02:38:40.060 folks you'd want to invite to a longevity round table next time? There's so many people we could

02:38:44.120 do this with, right? And I'm guessing nobody wants to give their seat up next year to make

02:38:48.640 this table bigger.

02:38:49.300 I think it would be good to invite Vadim Gladyshev because I think even though I disagree with some

02:38:58.060 of what he says, I think he always has something interesting to say.

02:39:01.960 Who's your nominee? I'd need some more time to think about it. All right. Matt, anybody jump in?

02:39:06.740 I mean, I think we would all agree there are tons of great people in the field. I mean,

02:39:10.560 I think Brian Kennedy, and I think Brian's going to be on your podcast in an upcoming date,

02:39:14.580 is somebody who also thinks broadly and deeply about the science and is fantastic. So he would

02:39:20.340 be great to have. It'd be great to have some differing, I mean, we differ sometimes on opinions,

02:39:25.740 but I think more or less are aligned. Be interesting to have some differing voices as well.

02:39:31.440 All right. So we think we'll do another longevity round table around the oval table?

02:39:36.320 Sure.

02:39:36.820 Let's do it. Let's see where we are a year from now.

02:39:38.740 In a year from now, I think there's going to be a lot of new stuff. That's what's new in aging

02:39:42.500 research. Rate of progress.

02:39:44.640 The derivative is very much positive.

02:39:47.120 You know who else I want to throw out there is Morgan Levine. I think she'd be really interesting 1.00

02:39:50.700 to have because while she is an expert in epigenetics and biomarkers, I think takes a

02:39:56.720 pretty clear-eyed view of that space. Now, is Morgan at Yale still?

02:40:01.760 She's at Altos.

02:40:02.440 She's at Altos. Yeah. Okay. I wasn't sure if she was there full-time. Got it.

02:40:04.900 Yeah. I would second that. That's an excellent idea.

02:40:07.620 All right. Well, Rich, you can get back to me on your nominees as well.

02:40:11.460 I will definitely do that.

02:40:12.900 All right. Gentlemen, thank you for making the trip.

02:40:15.380 I'll set my nominee committee onto this and I'll get back to you.

02:40:17.040 Thank you. It was fun. A lot of fun.

02:40:18.960 All right, guys. Thank you.

02:40:20.840 Thank you for listening to this week's episode of The Drive. Head over to

02:40:24.940 peteratiamd.com forward slash show notes if you want to dig deeper into this episode.

02:40:31.920 You can also find me on YouTube, Instagram, and Twitter, all with the handle peteratiamd.

02:40:37.020 You can also leave us review on Apple Podcasts or whatever podcast player you use.

02:40:43.220 This podcast is for general informational purposes only and does not constitute the

02:40:47.380 practice of medicine, nursing, or other professional healthcare services, including

02:40:51.140 the giving of medical advice. No doctor-patient relationship is formed.

02:40:56.180 The use of this information and the materials linked to this podcast is at the user's own risk.

02:41:01.560 The content on this podcast is not intended to be a substitute for professional medical advice,

02:41:06.580 diagnosis, or treatment. Users should not disregard or delay in obtaining medical advice

02:41:11.480 from any medical condition they have, and they should seek the assistance of their healthcare

02:41:15.980 professionals for any such conditions. Finally, I take all conflicts of interest very seriously.

02:41:21.460 For all of my disclosures and the companies I invest in or advise, please visit peteratiamd.com

02:41:28.460 forward slash about where I keep an up-to-date and active list of all disclosures.

The Peter Attia Drive - January 27, 2025

#333 ‒ Longevity roundtable — the science of aging, geroprotective molecules, lifestyle interventions, challenges in research, and more ｜ Steven Austad, Matt Kaeberlein, Richard Miller

Episode Stats

Summary

Transcript