The Peter Attia Drive - #357 ‒ A new era of longevity science： models of aging, human trials of rapamycin, biological clocks, promising compounds, and lifestyle interventions

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,

00:00:58.020 head over to peteratiyahmd.com forward slash subscribe. My guest this week is Brian Kennedy.

00:01:07.240 Brian is a renowned biologist and leader in the field of aging research. He's the former CEO of

00:01:12.980 the Buck Institute for Research on Aging, and he is now the director of the Center for Healthy Longevity

00:01:19.480 at the National University of Singapore. In this episode, we discuss why Brian moved his research

00:01:24.460 from the U.S. to Singapore and how that shift opened the door to running larger-scale clinical

00:01:29.380 aging studies, how the field of longevity research changed around 2017 when serious funding started

00:01:35.240 pouring in and reshaping priorities and the pace of discovery. We explore two different concepts of

00:01:41.800 aging, one being the linear accumulation of wear and tear with age, but the other being the

00:01:48.460 exponential or non-linear increase in all-cause mortality with age. And again, I think Brian's

00:01:53.880 explanation here is one of the more interesting ones I've heard. Talk about how rapamycin is being

00:01:57.960 tested in humans today, what we know so far, and why dose timing, especially around exercise,

00:02:03.460 could be critical. Why current aging biomarkers often miss the mark, and what Brian's team is doing

00:02:09.280 to build a clock that clinicians might actually find useful. Compounds that show early promise,

00:02:14.840 such as alpha-ketoglutarate, urolethan A, and sublingual NAD boosters molecule we've long

00:02:22.560 discussed and questioned and that Brian himself has been skeptical of, but nevertheless, we've found an

00:02:27.180 interesting place to discuss it here. How to combine lifestyle factors and pharmacology with

00:02:31.580 a focus on VO2 max strength training and the use of GLP-1 agonists and SGLT2 inhibitors. Lots more as

00:02:38.400 well. So without further delay, please enjoy my conversation with Brian Kennedy.

00:02:47.980 Brian, thank you so much for being here. You might actually hold the record for longest

00:02:53.560 journey taken to come to this podcast. In fact, I don't know if anybody could travel a greater

00:02:59.360 distance than from Singapore to come out here. So thank you very much.

00:03:03.520 Oh, it's my pleasure to be here. And I don't think you can get further from here to Singapore.

00:03:07.420 Let's tell folks a little bit about how you wound up in Singapore. I'll speed things through a bit

00:03:12.240 by way of background. Obviously, you used to run an institute called the Buck Institute. Tell folks

00:03:16.980 a little bit about what the Buck is and what you did there.

00:03:19.740 It was really the first institute solely devoted to understanding aging and longevity. And it started

00:03:26.020 around 2000 with some money that was donated by a woman who died in Marin County, north of San

00:03:32.660 Francisco. And I was the second CEO there in 2010. There were about 20 faculty at the time,

00:03:39.120 all devoted to either aging or aspects of aging, very basic science. And as you can imagine,

00:03:44.820 in the 2000s and around 2010, that was a significant component of the aging research field. It was still a

00:03:51.380 very small field. And so the goal was really to help that institute grow. And it was tough times in the

00:03:57.740 2010s because the funding levels were low. And it was right before the real interest in aging and

00:04:03.260 longevity happened around 2017, 2018. So we were really struggling to keep the doors open. And I

00:04:09.520 think the Buck's doing a lot better now, as well as the rest of the aging field.

00:04:13.200 At the time that you were there, how much of the funding came from NIH and how much came from either

00:04:19.940 donations or industry?

00:04:22.060 Yeah, it was very heavily oriented to NIH. And our goal was to get more industry funding. We started

00:04:27.120 seven companies when I was there, some of which are still hanging around and also really tried to

00:04:33.100 ramp up the philanthropy. But philanthropy for aging wasn't really happening until around 2017,

00:04:38.760 2018, when people started really getting the idea that you could slow aging and prevent all these

00:04:45.360 diseases and stay healthy and functional. And so that revolution happened around that time.

00:04:50.880 The faculty there are quite the star-studded cast. Eric Burden's there. Judith Campisi was there

00:04:56.280 until she passed away.

00:04:57.160 Yeah, that's unfortunate.

00:04:58.120 You really had a collection of people doing great work.

00:05:01.080 Yeah, well, Eric came after I left. He took over as the next CEO. But yeah, Judy was there,

00:05:05.960 Henry Jasper. He's gone on to Genentech. But Gordon Lithgow and a bunch of other people

00:05:11.320 working on aging. So it was a good group of people for sure.

00:05:14.120 What do you think led to this interest that you've alluded to in 2017, 2018 in this idea of,

00:05:23.560 you describe it as you see fit, but something happened in 2017, 2018 that's brought a lot more

00:05:28.360 interest into the field, however one describes it?

00:05:30.760 I think you just reach an inflection point. It's really hard to know what triggers it.

00:05:35.320 Calico started a few years before that. Google's Alphabet companies that was focused on longevity,

00:05:41.720 they've not been very open about what they're doing, but it triggered a lot of publicity for

00:05:46.760 the longevity field. It got Silicon Valley interested. And I think that Matt and I used this slide that

00:05:53.640 you have aging pointing to all these different diseases. We started using that slide around 2005,

00:06:00.040 and we were making the point around health span shortly afterwards. And really this idea of

00:06:05.320 preventing disease and keeping people healthy, interacting earlier. I don't know. I'm so sick

00:06:09.880 of that slide. I can't look at it anymore. And we weren't the only ones doing it, but I think there were

00:06:14.200 a few of us doing it. And it finally, I think helped trigger a movement, hopefully.

00:06:19.320 Do you think the field, these are sort of dumb questions as I realize, as I'm asking it, but

00:06:25.080 do you think the field would have accelerated sooner had it not been for some notable setbacks? For

00:06:29.960 example, I don't remember exactly when GSK bought resveratrol, but I believe it was like around 2006,

00:06:36.440 2007. It was clear to me, I think by about 2010, that that was not going to work. And I think it was

00:06:43.400 probably clear to GSK around that time as well. Yeah. It may be sooner. So do you think that that

00:06:48.840 type of hype with nothing to show for it was kind of a negative force in that equation? And maybe this

00:06:55.240 inflection point could have happened sooner, i.e. during your earlier tenure there, could it have

00:06:59.880 been easier to have raised funds if there had been less of those examples? I think it's hard to know.

00:07:04.680 I mean, it's unfortunate what happened. I mean, in one way, the investors made money off of that deal

00:07:09.880 since it was a success, but what was developed was not going to go anywhere. And that's unfortunate.

00:07:14.680 I think that it probably slowed things down a little bit, because there's always this doubt

00:07:19.880 about whether you can slow the aging process. And so when you have a major effort that's triggered

00:07:26.520 around trying to do that, even though they ended up focusing on disease, and we can talk about the

00:07:31.480 struggles of longevity biotech companies in that way. But when something like that fails, it probably does

00:07:37.960 slow down other investor interests. So today you're in Singapore. Tell me

00:07:44.760 what you're doing there. I kind of have one foot in academics and one foot in the private sector

00:07:49.000 these days. On the academic side, we're really focused on targeting aging. And that comes back to

00:07:55.880 what I alluded to with the biotech companies a minute ago. A lot of them are targeting aging pathways,

00:08:01.240 but to raise money and get their drugs tested, they have to turn to some disease indication,

00:08:05.640 which is understandable. And companies I'm involved with do that too. But that's not what we really

00:08:11.720 want to do. What we really want to do is slow the aging process and keep you from getting sick.

00:08:16.520 And so in an academic setting, we can test that clinically. So we basically have a whole range of

00:08:22.440 animal models, a pipeline from yeast, worms, flies, killifish, mice, and humans.

00:08:26.440 Billion years of evolution there.

00:08:28.360 Yeah. Yeah. We bring interventions in at the right place, validate them, really believing the idea that

00:08:35.720 if it works across different model organisms, it's more likely to work in humans. And then we design

00:08:41.480 human clinical intervention studies to validate that they're targeting the aging process. I don't

00:08:47.080 think anybody knows exactly how to do that yet, including us, but we're doing our best and learning as

00:08:51.640 we go. How is it funded, the institute you're at?

00:08:55.320 Mostly through the university and the government in Singapore, but we also have some philanthropy and

00:09:00.200 we do contract sponsored research to test interventions from companies as well sometimes.

00:09:04.680 How many PIs are there?

00:09:05.800 So I run a program that has about 35 PIs in it, but a lot of them are doing other things. They're not

00:09:13.000 all focused on that one concept I just told you. They have their projects around Alzheimer's disease,

00:09:18.360 or we just have this guy, Michael Chee, who's working on sleep and aging, which is so understudied.

00:09:24.520 It's kind of like an academic department. People have their own projects they're focusing on.

00:09:29.880 Is there any department at a U.S. university that brings together as many people that are focused on

00:09:37.160 specifics of aging this way?

00:09:38.920 Yeah, I don't think so. Although I would say it depends on how you define aging.

00:09:43.000 Yeah. If you branch it out to cancer.

00:09:44.840 Yeah, then definitely. But I'm not sure if it's this focused on actual aging process.

00:09:50.040 So let's kind of start with a question that I think we'll end up coming back to because it's

00:09:53.400 so fundamental. I enjoy going down the rabbit hole of fundamental questions in physics.

00:09:58.520 Sounds good.

00:09:59.320 And we're not going to do that, but I think the fundamental questions in biology,

00:10:03.160 I think some of them center around aging. What do we think is actually causing aging?

00:10:08.680 Okay. So I'm going to force you into physics since you asked that question.

00:10:11.080 All right, very well.

00:10:11.720 Because we just had the first ever international conference on gerophysics in Singapore.

00:10:17.400 I was one of the organizers and the reason I got behind it is the very question you just asked.

00:10:22.280 We've been debating what aging is for the longest time. And I think we would argue for

00:10:27.800 two hours in some conference room somewhere in the world. And at the end of it, we would come

00:10:32.360 up with the definition, shit happens and then you die. It's really just frustrating. I don't

00:10:37.640 even want to talk about it now. And Vadim Gladyshev has been on like an evangelical rant about how

00:10:44.360 do we define aging? We don't know how to define it. He asks that question at every conference.

00:10:48.520 I think it's a fair question. And we all throw our hands up in the air. And so the idea was we

00:10:53.160 have a lot of data now, a lot of human data. And aging researchers are beginning to try to model

00:10:58.840 that data. But they're not modelers. You know, I think most aging biologists, or at least myself,

00:11:04.440 if I have a skill, it's intuition. It's not writing equations and code. But the physics people,

00:11:09.480 the theoretical physicists especially, they know how to model things. And they model things based

00:11:14.360 on physical principles that are proven. And so we've been trying to bring these groups together.

00:11:20.360 Because I believe maybe the only answer to your question is that we have to write it in equations.

00:11:24.920 Early days on that. But I'm excited about where that's going.

00:11:28.440 And do you think that these will be explainable through equations? Or do you think that this

00:11:34.840 exceeds our level of intelligence to understand, and it's really going to be up to a black box that

00:11:41.240 contains a neural network to understand this? And maybe we take a step back, actually. So for the

00:11:46.440 listeners, they've heard us on this podcast talk about, quote unquote, hallmarks of aging. Maybe

00:11:52.600 explain to people what the hallmarks of aging are, which I don't mean like, list them all, but the

00:11:57.560 concept of them. Yeah, don't worry. I wouldn't put you on the spot for that. But the idea that has

00:12:03.560 been proposed is that there are hallmarks of aging. And why is stating those not the same as

00:12:09.240 answering the question that you, me, and everybody else is struggling with?

00:12:13.080 Well, this is kind of an existential crisis with me, because the hallmarks of aging came out in 2014,

00:12:19.320 or 2013. And then I wrote another paper right after that called the pillars of aging,

00:12:23.560 which is kind of the poor stepchild of the hallmarks of aging. And that was because there

00:12:28.120 was an NIH conference, and there were seven topics discussed, and they asked me to write a review,

00:12:33.080 calling them the pillars of aging. So I did. But even in that review, I had the seven pillars of aging,

00:12:39.240 but I connected them all with lines, because I don't really think that these hallmarks and pillars,

00:12:44.760 which are the pathways in the cell that are thought to be driving the aging process,

00:12:49.560 inflammation, epigenetic changes, these kinds of things. They're all interesting to aging,

00:12:54.280 and you can modify them, or they get modified if you slow the aging process. But what strikes me is

00:13:00.440 how entrained everything is. So if you take an intervention like rapamycin that slows aging,

00:13:05.720 it can impact all of the hallmarks. I think those are like outputs or ways you can look at aging,

00:13:11.160 but nobody is really just targeted. The idea that you can target each hallmark and then you'll live

00:13:15.960 forever is not going to work, because it's really the network that connects the hallmarks together.

00:13:20.840 And to me, healthy aging is about maintaining homeostasis. It's about maintaining a responsive

00:13:27.160 network in your body that sort of keeps you in equilibrium, responds to the events that are

00:13:31.720 happening during aging, the stochastic events, the damage that's happening, and it keeps you functional.

00:13:37.560 And that network is highly malleable. You can influence that network by drugs or behavior.

00:13:42.680 And if you do, you can drive benefit from it and you can read it out as an improvement of all the

00:13:47.000 hallmarks. It's not like one thing of exercise affects only this hallmark. So I think the hallmarks

00:13:52.680 was good because it drove interest in the field. It's part of the reason a lot of investment came

00:13:58.120 in the biotech sector, but it also is misleading because I think the idea that aging is 12 different

00:14:04.600 things and you just need to fix all 12 of them is completely wrong. It's really about your body

00:14:09.960 knows how to function in a healthy way. It's about trying to maintain that and maybe improve upon it.

00:14:14.440 So do you look at the hallmarks, which I believe have been modified since the original paper and a

00:14:19.880 few others have been added, do you see a rank order or a seniority of them in terms of causality?

00:14:27.320 For example, one of the hallmarks is mitochondrial dysfunction. Now, one could say that mitochondrial

00:14:33.480 dysfunction occurs independent of another hallmark of aging, which you've listed epigenetic change.

00:14:40.840 Alternatively, you could say, actually, it's the epigenetic change that occurs

00:14:45.560 stochastically and that that is driving mitochondrial dysfunction. And if you reverse

00:14:50.280 the epigenetic change to the previous epigenetic layout, you will correct the mitochondrial dysfunction.

00:14:56.200 How do you think about the interconnectedness through the lens of causality?

00:15:00.040 Dr. Yeah, I think the primacy issue is a major one. I like the idea that mitochondrial might

00:15:04.920 have been one of the primary drivers. Also, every time we do an experiment, we keep coming back to

00:15:10.360 inflammation. All of the interventions that extend lifespan reduce chronic inflammation,

00:15:15.480 almost all of them. And then every time we create a new biologic aging clock, which we're doing a lot

00:15:20.360 of now in my lab, and we do principal components and figure out what the main driver is, it's always

00:15:25.400 related to inflammation. I think there's something... It may not be... Inflammation may be a response,

00:15:31.800 but it's so central that a lot of the interventions, I think, are working by dampening inflammation.

00:15:37.720 Dr. But to your point, inflammation could easily be a readout state.

00:15:41.640 Dr. Yeah. I think when you modify it, you get an outcome. So it's not just an endpoint that you look at.

00:15:47.960 Dr. How are you measuring inflammation? And maybe walk me through how you're doing it in

00:15:52.440 different model systems. So are you studying inflammation in yeast?

00:15:55.960 Dr. Not so much in yeast, but you can study innate inflammation in worms and flies because

00:16:00.360 those pathways, the rudimentary elements of those pathways are there. And then in mice,

00:16:04.840 you have both innate and adaptive immunity that you can study. And so we look at inflammatory cytokine

00:16:11.000 panels and a range of other things in various tissues to see how that's changing over time and

00:16:16.440 how interventions impact that. And we do that in humans too, in our clinical study.

00:16:20.520 Dr. And so what do you believe is the hallmark of maladaptive inflammation? Do you think that the

00:16:26.760 hallmark of that is based on immune function, i.e. deteriorated immune function and or over

00:16:33.800 aggressive immune function? Or do you think, no, the hallmark of that could simply be found in

00:16:38.920 a cytokine profile that is not typical? I mean, this is maybe more of a technical question,

00:16:44.200 but it's going to become interesting as we move our way into humans.

00:16:46.680 Dr. I think it's central to mTOR. One of the things we were,

00:16:49.800 one of the earliest people to publish was that what's happening during aging is that

00:16:56.200 baseline levels of mTOR are creeping up. You can't turn the pathway off.

00:17:00.600 I think most of the interventions, sirtuins, mTOR, inflammation, it's not about doing anything

00:17:07.240 super physiologic with the interventions. It's about restoring the dynamic range

00:17:12.360 that you had when you're youthful. And mTOR is a great example of that.

00:17:15.640 And I'll bring it back to inflammation. You need mTOR on when you wound your skin or you get an

00:17:21.000 infection or you have a big meal in your liver, but you need it off the rest of the time. And

00:17:26.360 when you're young, you're very good at maintaining that dynamic range. But what's happening with aging,

00:17:31.560 at least in stem cells, is that the baseline levels of mTOR are creeping up.

00:17:35.880 People on this podcast are pretty familiar with mTOR. We've had David Sabatini on many times,

00:17:41.400 Matt Caberlin as well. But I just want to make sure that for anybody who's here, who's either new

00:17:46.440 or forgot, let's take a step back. mTOR is so important to this discussion. Let's go back as far

00:17:52.600 as we need to and explain mTOR. Talk about complex one, talk about complex two, talk about how one impacts

00:18:02.040 the other. We're obviously going to talk about rapamycin in that context, but take as much time

00:18:05.800 as you need to make sure listeners really understand why mTOR is so central across all of life that

00:18:12.680 we've ever known.

00:18:14.200 Our entry into the mTOR pathway was in yeast. This is Matt Caberlin and I. And so we were screening the

00:18:20.600 yeast deletion set. It's a set of strains, which each gene is deleted and looking for one.

00:18:26.200 How many genes, by the way, in yeast?

00:18:27.400 They're about 5,000. And then some of those are essential. So you can't screen those if you

00:18:31.800 knock them out, the yeast are dead. About 70% of them, the yeast are still viable. And so we were

00:18:36.840 trying to find ones where when you knock out a gene, the yeast live longer. Surprisingly, there were

00:18:42.120 like 300 genes that met that description. And that one thing I learned from that is that extending

00:18:47.720 lifespan, at least in a simple organism like yeast, and there's also data in worms, is much easier

00:18:52.920 than anybody would have ever expected.

00:18:55.480 That's still a pretty cool tour de force in the 90s. You guys were doing this in the early 90s.

00:19:00.520 We did the sirtuin stuff in the 90s. We did the full genome screen in the 2000s when I was in

00:19:05.400 Seattle with Matt. It was a lot of work and we did it brute force. We had people sitting at microscopes

00:19:11.080 dissecting yeast all the time. And I think there were 80 some authors on that paper. You know,

00:19:15.800 for a yeast paper, that might've been a record. But anyway, one of the main things we hit was the

00:19:20.440 mTOR pathway. And downstream of it is protein translation. We hit a lot of things in protein

00:19:25.880 translation. And mTOR is a nutrient responsive kinase. So it responds to the levels of carbohydrates

00:19:32.440 and also amino acids that the cell encounters. And so that fit into the calorie restriction data,

00:19:37.960 which I'm sure you've talked about that reducing calories can extend lifespan. So it seemed like it

00:19:43.320 was going to be very central from the beginning in modifying aging. And I think that's proven true.

00:19:48.920 Turning down mTOR signaling across a wide range of species extends lifespan. And I think the data

00:19:54.600 in humans, it's not fully validated, but I think that if you alter mTOR signaling in the right way,

00:20:00.040 you can probably slow aging in humans too. Now, why is it then that the first human brush

00:20:06.760 with mTOR modulation shows up in the form of immune suppression? That's a bit unfortunate in a way.

00:20:13.960 Rapamycin is discovered on Easter Island and you've probably gone through the whole story of this.

00:20:19.320 It almost died.

00:20:20.040 I feel like I'm overdue for a retell of it, but yes, probably top five stories in science, right?

00:20:24.920 It's pretty amazing, actually. I think I read you went to Easter Island too. I haven't even been there.

00:20:31.160 I'm impressed.

00:20:31.640 We're going to go back in 2016. But this drug had the ability to kill bacterial cells that was

00:20:39.960 bactericidal, you know, they'd start with, and then they discovered it had impacts on human cells,

00:20:45.480 but almost got thrown away and then gets restored and it makes a new life as an immune suppressant.

00:20:50.840 And certainly if you use it at high enough doses and you really dampen the ability to activate TOR,

00:20:56.200 you can impair the immune system. And especially if you combine it with cyclosporine or some other

00:21:01.480 anti-inflammatory or immune suppressant, then it's used for after organ transplant in those contexts.

00:21:07.480 And do you think that it's that necessary combination with one to two other immune suppressants

00:21:15.000 that allows it to shine? And I'm actually not aware of literature that looked at

00:21:19.400 rapimmune or rapamycin in isolation as a potential treatment for organ transplant patients.

00:21:27.160 I'm not either. At high enough doses, it may have that impact, but I think the side effect

00:21:31.000 profile would be too extreme by combining it with other drugs at lower doses. I think you get a bigger

00:21:36.200 effect. But I think the main thing for aging is that it's not immune suppressant, I think,

00:21:40.440 at the levels that people are taking it for longevity, which is once a week, let the trough levels come

00:21:46.600 down. I don't think we're seeing immune suppression in that context, at least not above background.

00:21:51.880 What do you think of the window we got into this idea of immune modulation and maybe immune enhancement

00:21:58.280 with that type of a dosing regimen vis-a-vis the paper that Joan Manik and Lloyd Clickstein published

00:22:03.960 about 12 years ago?

00:22:05.000 Yeah, I like those papers. I think there's definitely a nugget of truth in there, and I think it can protect

00:22:10.040 from respiratory infections if used correctly. So I still think rapamycin is the gold standard for

00:22:16.120 a small molecule impacting aging. At the end of the day, it may not be the best,

00:22:19.800 but right now I think the evidence is still the best. And I think coming back to the earlier

00:22:24.120 thought, when you have mTOR creeping up when it shouldn't be, that's driving chronic inflammation,

00:22:29.400 and then chronic inflammation is continuing to drive mTOR. So it's this feed-forward circle of

00:22:34.600 disruption that connects this nutrient pathway to inflammatory signaling. I think that's one of the

00:22:40.360 earliest events that's happening.

00:22:41.800 So let's go back to something you said a second ago, which is absent the equations,

00:22:46.760 biologists have to rely on their intuition. And so if we believe in the primacy of that

00:22:52.680 deterioration, that homeostatic deterioration that you just described, what is driving it? Is this

00:22:59.320 entropy? What is it that is changing?

00:23:02.680 Entropy has been used off and on ever since I started in the aging field.

00:23:07.320 You know, it's just a lazy term we use because we don't have something better to say.

00:23:10.440 No, but I think there's a nugget of truth in that. And I think ultimately,

00:23:15.320 I work with Peter Fedichev at Gero, and I think he's probably the deepest thinker

00:23:21.080 in the aging field in terms of understanding this process of aging at a mathematical level.

00:23:26.600 My only role in working with him is, I was a math major. Fortunately, I didn't try to pursue that.

00:23:31.800 But I survived enough to get a math degree in college. So I'm kind of his not whisperer or

00:23:38.680 translator. I take what he says and try to help him reframe it in a way that the rest of the world

00:23:43.960 can understand. He really thinks about things deeply. I think his view on it is that really

00:23:50.280 what we're talking about is resilience. From an image, you could imagine that when you're young,

00:23:55.160 you're living in this deep valley, and you do all kinds of crazy things. You get too much fast food,

00:24:00.200 you get sick, you've diverged off the bottom of the valley, the lowest activation energy,

00:24:05.560 and biologically, you look older, but you're in a deep valley. You keep getting pulled back

00:24:10.520 to health. And so almost no matter what you do, at least in the short term, you're coming back into

00:24:15.880 health. There are hills you can go over into what you would call failure states, which could be

00:24:21.320 chronic diseases, or could be some major functional decline. And that doesn't happen when you're young,

00:24:26.760 because it's a very steep hill. But these hills are coming down as you get older. And so when you

00:24:32.360 diverge off the healthy state, it's harder to get you back. And occasionally, you go over the side,

00:24:37.320 and then you're in a frailty mode once you have a later stage disease. So the question is how to

00:24:43.320 mathematically model that. And it's more of a dynamic systems type of modeling. And when you look at the

00:24:49.160 large data, and I'm getting to the point where you're going to understand why I need the physicists,

00:24:53.880 but when you look at the large data, it almost looks like there's a linear accumulation of damage.

00:24:59.960 So what we're doing now is we're trying to measure biologic age from large data sets like UK Biobank.

00:25:05.320 And then we're breaking it down into principal components. And when you do that, a lot of the

00:25:10.200 principal components don't track with aging, they track with sex or smoking or something else.

00:25:15.480 The ones that track with aging, usually there are a couple of them. One of them is the primary

00:25:20.280 driver, and it's kind of going up linearly. And that looks like damage. And that's probably the

00:25:25.960 main driver of aging. It doesn't have to be damage. It could be stochastic events. It could be subtle

00:25:30.920 changes.

00:25:31.400 When you say damage, do you mean, what damage are you tracking? DNA damage?

00:25:35.160 That's why I was trying to back off a bit. It may be a cluster of different types of damage. It may

00:25:40.280 be stochasticity not really defined as damage. Subtle changes here and there that each on their

00:25:46.280 own have tiny little impacts. But when they start to add up, they put stress on this network and

00:25:52.040 eventually it starts to break down. So that looks linear. And the problem is that mortality looks

00:25:57.720 exponential. And so if you model it into a mathematical equation that talks about these

00:26:03.800 valleys going down, that's a linear change. But the chance the ball's going to go over the hill

00:26:09.640 is an exponential change. And so what I like about it is you can fit human data into an equation that

00:26:16.920 is compatible with Gompert's equations and exponential increase in mortality. I think it's on the right

00:26:23.000 track. I mean, there's probably a lot of changes. There's also another component, usually in these

00:26:27.800 biologic aging clocks that's age-related. And it's oscillating. It sort of oscillates around that

00:26:34.120 first component going up. And I think that's why you see these methylation clocks are going up and

00:26:39.400 down and changing and everything. And I think of that component is how well you're functioning at the

00:26:45.640 damage state you're at. So you have all these events that are going bad and that's defining your age.

00:26:53.480 Maybe it says you're 50. But then you can be somewhere between 40 and 60, depending on what

00:26:59.160 your behavioral patterns, what supplements or drugs you might be taking. And those things are going up

00:27:05.160 and down. So that's why when you get sick, you look older and then you get better. You come back down.

00:27:10.200 That first driver is not changing. It's the second one that's oscillating. And most of the interventions

00:27:15.080 seem to affect the second one, which suggests that what we're doing right now is, all right,

00:27:20.600 I'm going to say this and then I'm going to qualify it. It suggests that what we're doing right now is

00:27:25.000 sort of working around the edges. We're doing things that may have five or 10 years impact on

00:27:29.720 healthspan, which by the way is a revolution if that's successful. I think that's a major breakthrough

00:27:35.560 in medicine if we can give everybody five or 10 years of extra healthspan. But these things may not

00:27:40.520 impact maximum lifespan in humans and they may not get us to 150 or 200. And the kinds of ways to get

00:27:46.680 there may be totally different kinds of interventions. So we're thinking about that

00:27:50.760 a lot with Peter right now. And I translated my research and translated as a better term,

00:27:54.760 I switched my research into translation about 10 years ago because I was like, I don't want to be

00:27:59.720 retired and 80 on a porch somewhere and not have any impact on humans. But now I'm starting to think

00:28:05.560 back a little bit to basic science because I'm starting to think that the interventions that we need to

00:28:10.920 develop if we really want to have the big changes are not being done yet. And we have to go back to

00:28:15.880 some discovery science to do that. So Brian, there's a lot you said there that's really interesting.

00:28:20.760 And I'd like to unpack it both for myself and for the listener. The first thing you said,

00:28:25.640 actually, I'm not sure I understand is we have a linear process of, we're going to use the word

00:28:32.520 damage loosely. And over time that is increasing monotonically and linearly and not alterably.

00:28:41.320 No, exactly. Superimposed on that. So we have damage occurring this way. Superimposed on that,

00:28:49.400 we have cyclic, episodic, volatile change that probably explains a lot of the difference between

00:28:58.680 two 50 year olds that you might see. You might see a 50 year old, they look great. You see a 50

00:29:03.000 year old, they look like they're 75. Why? Or you might see it even within an individual.

00:29:08.600 Boy, at 50, I looked horrible, but I got my act together. And by 55, I actually looked like I did

00:29:13.800 10 years sooner. So that's the superimposed curve. And you're saying, look, everything we're doing

00:29:21.480 from a translational perspective, all the stuff Peter talks about, by the way, wrote a whole book on this

00:29:27.000 topic is how do you impact the oscillation? Well, if you sleep this way, if you exercise this way,

00:29:33.000 if you eat this way, if you take this supplement, this drug, manage all of these factors, you are

00:29:40.040 absolutely going to put yourself on the better wave here, but you are not impacting this guy.

00:29:45.640 That's what I'm afraid is happening. Yeah. But then you said something a moment ago that I just want

00:29:50.520 to make sure I understand. Is your explanation for why aging follows a Gompertz curve as opposed to a

00:29:57.960 linear curve, is that all due to the superimposed wave that goes on top of the linear curve? Or was

00:30:04.440 there another reason that aging follows exponential Gompertz law?

00:30:08.520 No, I think I had conflated two ideas at once there. So let me associate that. You've described the

00:30:14.200 wave better than I do. So we'll leave there that. I think that you're right. It's almost like we're trying

00:30:20.040 to get to the best state we can be at for the damaged state we're in.

00:30:23.240 By the way, I have never thought of it that way, Brian. I love that description. And it's actually

00:30:28.920 what I say to my patients, because I get people that come to me and they say, Peter, I really want

00:30:33.880 to live to 150. I'm told you're the guy. And I say, actually, I'm not the guy. I don't believe it's

00:30:39.960 possible. What I believe is possible is seven to 10 more years of infinitely higher quality life.

00:30:47.320 And if that's not what you want, if you want something that is far in excess of that, you're

00:30:53.800 going to have to go to somebody who's got proof that they can do different.

00:30:57.800 Yeah. And I don't think there is any proof right now.

00:30:59.640 I obviously don't think there is either.

00:31:01.400 I do leave the idea open that it could be possible. I think it may be feasible to do that,

00:31:05.480 but I don't think anything-

00:31:06.520 It's just, I don't see any evidence that there's anything that's doing it now.

00:31:08.920 Yeah. Translationally, I agree with you. I think the linear to exponential is the idea of the hills.

00:31:14.600 You've got a ball. If you do it at two dimension, you've got a curve that looks like this and you've

00:31:18.520 got a ball here and the damage is causing the thing to come down. And so the chances the ball

00:31:24.280 going over is actually exponential.

00:31:26.280 I see. Now I get it. Yep. Yep. Yep. Yep. So even if you have a linear reduction in the height of the

00:31:31.240 hills, the activation energy will increase probability exponentially over the curve.

00:31:38.920 And by the way, I'm just trying to think through why that's the case. Is that the case

00:31:43.800 because of, is it a V squared problem and getting over the hill if we were to model it out as actual

00:31:49.880 balls? Yeah. I have a graph. I only have a model, a movie that really helps people that are not

00:31:56.040 mathematicians. I think both of us have some understanding of the math, but a lot of people

00:32:00.280 don't. And I think that, yeah, it's something like that. I think also it explains very well why

00:32:06.840 treating disease doesn't work because you have 50 failure states you can go into.

00:32:12.040 Each person based on their genetics and their lifestyle, the hill to go over that failure

00:32:16.040 state may be a little bit different. Sometimes person's not going to go over this one, but

00:32:20.200 there's a chance you're going to go over a lot of different failures. And if you block one of them,

00:32:24.520 say you treat diabetes or something, you're still going to go over the other ones. The only way to

00:32:29.480 really slow aging is to keep the hill higher. Yeah. Again, that's both a beautiful model,

00:32:37.400 a mental model for how to think about it. And yet still equally infuriating because I don't know why

00:32:42.280 the walls are coming down. Yeah. Why are the hills coming down? What is the fundamental reason? What's

00:32:49.960 the particle reason for the wall height coming down? I think the hills are resilience. Even still,

00:32:55.800 resilience is the most important term in aging and nobody understands it. Yeah, exactly.

00:33:00.280 All right. I'll give you a half-baked answer because it's all I can give you. I think that

00:33:04.360 what's happening is this damage is impacting this network that's keeping you healthy,

00:33:08.360 this homeostatic network. And it's in little ways here and there and here and there,

00:33:13.240 and the network compensates for that and does okay. But when enough damage happens,

00:33:18.280 you just can't compensate anymore for events that are happening. So when you get sick,

00:33:23.240 you get some viral infection or you fall down when you're 80 and break your hip, you just don't have

00:33:29.240 that homeostasis pathways in place to allow you to recover and compensate for that. I mean,

00:33:35.080 that's sort of the best answer I can give you right now. So if you had to guess,

00:33:39.800 and I'm hopeful that the listeners are with us because this idea of the linear and monotonic

00:33:45.400 increasing in damage is the thing that has to be addressed if we're going to make a step function

00:33:51.320 change in human longevity. I like how you described it. We're really tinkering around the edges.

00:33:57.080 Everything we do is tinkering around the edges. But if we fundamentally want to get to a point where

00:34:01.960 maximal human lifespan is changed and health span is fundamentally altered, we have to bend the slope

00:34:08.600 of that line. So my first question for you is, what is the probability in your mind that rapamycin is

00:34:15.480 doing that based on what you've seen in animal models?

00:34:18.920 First of all, if you look at something like a worm, I think the modeling is very different there.

00:34:24.200 Worms are just already in a failure state. They're like designed to last for two weeks. They don't

00:34:29.320 have that homeostasis that humans have. And so you can get huge impacts-

00:34:33.240 You can get fooled by interventions there.

00:34:35.240 Yeah. In worms or other organisms that the pathway may translate to humans, but the effect size in

00:34:41.480 humans is going to be much smaller. I suspect that's where rapamycin is, that it's going to give

00:34:45.800 you a healthier period if you take it the right way. I don't think any intervention is going to

00:34:50.360 affect everybody. Okay. But a majority of people may benefit from that. But I think it's in the

00:34:55.720 modest effect in size, not in the change the healing.

00:34:58.760 Not in change slope. Okay. I want to come back and talk more about RAPA and mTOR because you're,

00:35:03.480 again, one of the few people along with Matt, David, people who can really talk in depth about it. But

00:35:08.360 let's now stay in the world of speculation. If you had to even imagine something that can change

00:35:14.680 the slope of the line, i.e., I guess we define maximal lifespan as the 90th percentile of lifespan.

00:35:20.680 So let's just make a number up and say maximal human lifespan today is, I don't know.

00:35:25.240 105, maybe 90.

00:35:27.160 Something like that. Okay.

00:35:27.880 120 is the absolute max.

00:35:29.560 Is the 99.9999. Yeah, exactly. So we're going to take 90th percentile human lifespan up by 25 years.

00:35:36.520 If I told you, Brian, you might not be alive to see it, nor will I, but I have a crystal ball.

00:35:41.720 And in the year 2100, 90% of humans will live to be 130. And now I say, give me your best guess

00:35:50.040 as to what did this. Is your guess going to be small molecules, genetic engineering,

00:35:57.240 epigenetic engineering, like just go down the pathway or multimodal. It's going to have to

00:36:01.960 be 10 different things. Like this is just kind of like the fun sci-fi game.

00:36:04.920 Well, I think the scary thing is that that linear accumulation, it does look like entropy, which

00:36:10.520 reversing the second law of thermodynamics is...

00:36:12.920 We don't have to reverse it. We just have to slow it.

00:36:15.000 Slow it. Even slowing it is a challenge, right? And if you think about it from a physics standpoint,

00:36:19.960 which I'm getting further and further and deeper in a pool, I shouldn't be in. I'm just

00:36:23.880 telling you that right now. But if you think...

00:36:25.160 I'm giving you a life jacket. Just keep treading water.

00:36:27.880 My motto, by the way, with the consulting I do is that I know what I don't know.

00:36:32.440 It's not a good model for consulting and I've learned. But anyway, as you get into the

00:36:37.160 physics of it, it's really about temperature. And I don't mean temperature in terms of the

00:36:40.920 temperature in the room. I mean the energy in the system that's driving the changes or damage.

00:36:46.520 And the question is, how do you lower that? And so maybe what you need to do when you're looking

00:36:51.560 for longevity interventions is not looking for how long a worm lives, because the worm is dying

00:36:57.400 for a different reason. It's already in the failure mode. It's about how to lower the noise

00:37:04.520 in the system. And lowering the noise in the system might be a way of changing that slope.

00:37:10.440 So that could be transcriptional noise. It could be anything you can measure as noise that happens

00:37:15.400 over time. You might want to try to lower that noise. That's an interesting concept. Now that may

00:37:20.360 also come with secondary effects that people don't want.

00:37:23.240 Right. There might be a retardation of growth and development early in life. And it might be one

00:37:28.200 of those things where you don't want to touch this slope for the first 30 years of life.

00:37:32.120 Yeah. Yeah.

00:37:32.840 Where's the point at which you want to intervene?

00:37:34.840 Yeah. And I think that temporal component is really important. It's taking us into a different

00:37:38.440 concept, antagonistic pleiotropy. And it is true that if you look at all the yeast mutants that extend

00:37:43.560 lifespan, most of them would not make happy yeast in the wild. They slow growth or they do something else,

00:37:48.440 affect some property like mating that is not going to make for a yeast that survives through natural

00:37:53.560 selection. But if you put them in a lab, they can divide more times. So a lot of long-lived mutants

00:37:58.440 have fitness costs. And so the question would be that if you target this noise in the system,

00:38:03.480 which is a completely different way of thinking about interventions, what will the fitness cost be

00:38:08.120 with that? And you're right, maybe you can get around it by temporal things. Like the mTOR pathway,

00:38:12.520 you probably don't want to impact as a child. But as an adult, it's more important early in life than

00:38:17.640 it is later in life. It's only important at certain times. So if you impact it the right way,

00:38:22.360 you can get the benefits without the cost. And maybe that's possible at these interventions,

00:38:27.080 but we're so early, I can't even with any confidence tell you what kinds of interventions

00:38:31.960 would have that impact. I will say one other thing is that I think reprogramming is potentially a way

00:38:39.000 to mitigate some of this inter-entropic change. Because if you can replace the cells with new

00:38:46.120 cells, those new cells may have some of the damage because they come from the old cells,

00:38:50.280 but they would probably get rid of a lot of the damage too. And so that may be a way of changing

00:38:54.680 the slope. So like reprogramming, which I think is still very early stage, that may be a feasible

00:39:00.920 strategy. So as a thought experiment, if I could clone you right now, are you at a twin? Let's just

00:39:05.720 say you've got a twin. One of me is enough in the world. Yeah, but you're in Singapore. So we're

00:39:09.720 going to have a North American version. We're going to have the Singaporean version. So we have two of

00:39:13.720 you and in one of you, we're just going to act as the control. We're going to give you some vehicle.

00:39:18.760 In the second one, let's just assume I can use the fidelity of CRISPR to revert your entire

00:39:28.840 epigenome in every cell of your body to what it looked like when you were 20. And anytime it gets

00:39:36.200 out of whack, I smack it right back to 20 year old Brian, epigenome only, not genome, not proteome.

00:39:43.320 I don't change anything, but epigenome. Well, those things are good. Exactly. They're going to affect

00:39:48.120 everything else. What is your guess as to the difference in lifespan and health span of those two

00:39:53.800 versions of you? That's an interesting question. I think there would be a difference for sure.

00:39:58.840 I'm not sure it'd be a huge difference. You don't think it'd be huge.

00:40:01.400 That's interesting. Well, I'm not as sold on the primacy of the epigenetics.

00:40:05.400 So what would go wrong? I don't know the answer. Of course, I'm just kind of thinking

00:40:09.080 through data that I've seen. So if you look at epigenetic code for two different

00:40:17.160 hepatocytes, liver cells, one from a 20 year old, one from a 50 year old. And I tell you which one's 20,

00:40:23.640 which one's 50. And then I show you a bunch of others. You can always tell which one's the older one,

00:40:27.480 which one's the younger one. Probably tell that from mitochondria or a lot of other things too.

00:40:31.160 Yes. So the question is, is your belief system that just because you revert the epigenome back

00:40:37.160 to what it looked like when it's 20, it's not going to change gene expression enough to move

00:40:41.960 the needle? I think it will definitely influence gene expression, but there's also DNA damage that's

00:40:46.440 happened. There are mitochondrial changes. The question we're asking is, if you revert that,

00:40:51.560 how many of these other things can we fix and restore? And that's the unknown answer. I suspect

00:40:56.840 you would have a significant impact on those things, but not fully restore them. I think

00:41:02.280 the question of the primacy of the epigenome is an open question. Nobody knows the answer to this.

00:41:07.720 How testable is the hypothesis? How would you design the experiment to test that?

00:41:11.800 I think that's difficult because if you want to do it in a very direct way, you really need to

00:41:17.320 modify the factors that are controlling the epigenetic regulation. But there are a lot of

00:41:21.960 factors doing that. It's not just DNA methylation. It's histone modification. There's nuclear packaging

00:41:27.160 and nuclear lamins, which are linked to aging as well. And there's not just one pathway to change.

00:41:31.800 So I think from a real life standpoint, it's hard to think about how you would do that effectively.

00:41:36.680 I think people have really jumped on it. And maybe when talking about aging clocks at some point,

00:41:40.520 people have really jumped on this idea of epigenome being a D driver of aging,

00:41:46.040 because you can get a biologic age by measuring the DNA methylation changes across the genome.

00:41:52.840 Well, I mean, they're asserting that that's the case. I don't see any evidence that that's

00:41:57.320 the case. Yeah, we can come to that. We definitely want to dig into this.

00:41:59.880 You can get to that same point by measuring the proteomic changes, by measuring the microbiome changes,

00:42:05.560 by looking at facial structural changes, and Jackie Han's got great data on that in China.

00:42:10.760 So anything in a human where you have a deep enough data set that's enriched enough,

00:42:16.120 and you have samples across a wide enough age range, you can make a clock that predicts their age.

00:42:21.560 And the facial clock is about as accurate as the methylation clock. So I think that a lot of

00:42:26.840 people have jumped on this methylation or epigenetic bandwagon, but they're taking association

00:42:32.680 and causality, and they're making a big leap there. Now, we know that you can modify epigenetic factors

00:42:37.720 and extend the lifespan of yeast and worms and flies, maybe even mice. So it does have a role,

00:42:44.200 but you can do that with senolytic factors or nutritional regulators or calorie restriction

00:42:49.400 or a lot of other things. It's not clear to me that that's a bigger effect than you're going to

00:42:53.480 get from targeting these other interventions. To be fair, none of them are completely reversing

00:42:58.120 things either. So they're not addressing your question.

00:43:00.680 Yeah. Do you believe, again, we're in the philosophical, I'll bring it back to reality

00:43:05.720 at some point today. Do you think that immortality is impossible unless we define it through AI copying

00:43:13.720 your brain? I mean, physical immortality. Do you believe that that is impossible?

00:43:18.680 Well, I like to tell people that I'm immortal because I think the mindset that it gives me

00:43:23.160 is a very healthy mindset for me. You talk a lot about the emotional aspects of aging in your book.

00:43:28.840 I love that chapter. We can come to that later. But I don't really believe it's true. I think the

00:43:33.880 odds that you could achieve that level of change in aging is non-zero, but close. So I'm skeptical

00:43:42.040 that that can be done. I think it's fair to say, but I wouldn't rule out the possibility. Of course,

00:43:46.200 nobody's ultimately going to be immortal because you're going to get hit by a bus sooner or later.

00:43:49.960 But what you're really talking about is being immortal in terms of dying from aging.

00:43:54.360 I guess what I'm really saying is, can one ever get to the point where resilience is high enough

00:44:01.160 that you cannot die from disease? I have seen nothing so far

00:44:06.680 that suggests that's possible, but that doesn't mean it isn't possible.

00:44:10.200 Yeah. And then that gets even to physical frailty and sarcopenia and things like that,

00:44:15.720 where even when we see centenarians and super centenarians, their frailty is still pretty

00:44:22.200 remarkable. Meaning they still look pretty feeble and frail. Age adjusted, they're great. But at the

00:44:28.440 end of the day, when they're 110, they still look like someone who's in the final years of their life,

00:44:35.240 just as someone would if they were 84, 85.

00:44:37.640 I mean, I had two grandmothers that lived to almost 100. One died at 99 and the other 101.

00:44:43.560 The one at 101, I would say that she was driving at 95. I think she quit driving. She bowled a 238

00:44:49.880 game at 93. She looked like a 70 something year old.

00:44:52.760 Yeah, exactly. My point is she just had a phase shift of 20 years, but it didn't undo

00:44:57.880 the inevitability of that decline.

00:44:59.560 No, I agree with that. Getting us to 100 is a good goal, I think.

00:45:03.080 I agree completely. Do you think that we're spending too much time worrying about finding

00:45:10.680 immortality, escape velocity, understanding the core of aging when maybe we should be spending

00:45:15.320 more time on how do we preserve health span in the last decade of life? Why is it that most people

00:45:23.160 in the final decade of their life are physically too frail to enjoy life, are cognitively just even

00:45:29.080 absent Alzheimer's disease. They're just not cognitively sharp enough. They're in pain. They're

00:45:33.720 fracturing their hips. They're not doing what gave them joy through most of their life.

00:45:38.680 Yeah, I agree with that. We should fund aging somewhere closer to the level we're funding

00:45:43.960 cancer and answer both of those questions at the same time. I think one of them is a translational

00:45:48.680 question about how do we slow aging as much as we can right now and improve the health of the

00:45:53.480 population as much as possible. And the other one is a basic science question. Can we stop aging?

00:45:58.760 Can we reverse aging? If anybody tells you they have the answers to that, they're lying to you

00:46:03.320 or they're lying to themselves. We don't know. It's maybe the most important question in biology,

00:46:08.920 and we should be throwing money at it. So we've seen all this money go into the private sector side,

00:46:15.400 biotech companies, supplement companies, longevity clinics, and on and on. And I think that's great,

00:46:21.160 by the way. And I've spent a lot of my time working with those groups because I think it's important.

00:46:26.200 But we're not seeing the academic funding that's going into the basic science of aging and longevity.

00:46:32.840 And the big questions that you just raised are still not answered. I'm changing your question

00:46:37.960 to a plea for more funding. And unfortunately, the kind of funding that supports that is usually

00:46:43.080 government funding, foundation funding, and that's under major threat right now. I'm really worried

00:46:48.200 that we're not going to answer those questions.

00:46:49.960 Yeah. This was a discussion that came up on a longevity round table. I think most people,

00:46:55.560 myself included, were really surprised to hear how disparate the funding differences are and how,

00:47:01.960 if you could put, I don't know, if you could reallocate 10% of funding from the disease specific

00:47:09.160 pools to the age pools, it could have an enormous difference.

00:47:13.160 That enormous is an understatement.

00:47:14.840 Yeah, yeah. So when you think about the big chronic diseases, cardiovascular disease, cancer,

00:47:20.760 dementing diseases, and metabolic diseases, those would be the big four. I've often maintained

00:47:26.360 that the least inevitable of them is ironically the one that is the most deadly today, which is

00:47:30.920 cardiovascular disease. Atherosclerotic diseases, so cerebral and cardiovascular, ironically, the most

00:47:36.680 preventable, both because we have the best understanding of what causes them and we couple

00:47:41.560 that with the most tools to prevent them, whether it be tools to combat hypertension, dyslipidemia,

00:47:46.040 et cetera. And they're responsive to lifestyle modification.

00:47:50.040 Which of those major diseases of the other three, dementing, cancer, metabolic, do you believe

00:47:57.160 is the most inevitable to our species? I wouldn't put metabolic in that category for sure,

00:48:02.920 because I see that more like cardiovascular. I agree. So of the other two, cancer and dementing

00:48:08.440 or neurodegenerative diseases, which one is just seemingly inevitable? We don't know enough about

00:48:13.880 dementia to answer, but I will say that cancer is a little bit different than these other diseases,

00:48:18.120 I think. And it may be less modifiable by longevity interventions. Dementia, we just don't know. My guess

00:48:24.840 is it's highly modifiable too, but there's not enough data to be sure of that like there is for

00:48:29.320 metabolic and cardiovascular disease. But cancer is an accumulation of mutations. So it's a more defined

00:48:35.560 event that's happening. It's also a impact on the immune system that's different a little bit than

00:48:40.600 normal aging. So it may be less approachable from a longevity viewpoint. It's funny. That's exactly my

00:48:47.720 view that cancer is the most inevitable of these diseases. Do you think that the inevitability

00:48:55.080 or the age-related component stems more from the accumulation of mutations or the weakening of the

00:49:02.680 immune system? It's probably both. You don't get to cancer without the right mutations happening.

00:49:08.280 But I think we're learning more and more that the immune system is playing a major role in it.

00:49:13.000 We can see that very clearly from the interventions that improve immune function, and they're having a big

00:49:18.320 role in certain types of tumors. But I think that's going to be true for Alzheimer's and dementia as well.

00:49:23.080 We've completely underestimated the role of inflammation in the immune system and those

00:49:27.480 diseases as well. And they may be the primary drivers. I'm very frustrated by some of these

00:49:33.320 fields. One of them is Alzheimer's. I kind of feel like one of these Alzheimer's researchers,

00:49:38.120 they're going to die at some point of 90. And on their tombstone, it's going to be like

00:49:42.440 major accomplishment was to completely remove plaques from the brain, died of Alzheimer's at 94.

00:49:49.160 There's been so much focus on one or two mechanisms of disease that we spent 30 years

00:49:53.880 not studying the others, which may be more important.

00:49:56.120 Why do you think that? I mean, I write about it in the book. I'm really curious as to why you think

00:50:01.080 that's happened. Unfortunately, that's not an isolated incident in science. So why do you think

00:50:05.480 it's happening in a field where the results are otherwise so dismal?

00:50:09.640 What's the saying that scientific progress happens one funeral at a time?

00:50:13.560 Yeah. I think that's part of it. You get people that have successful research programs and

00:50:18.920 their postdocs get hired in all the jobs. And so when you take a field and it grows from a small

00:50:23.880 field to a bigger field, everybody can draw their lineage back to four or five different PIs. And so

00:50:29.720 whatever models, and those PIs get really focused on those models and they see that as their ticket to

00:50:36.360 prizes and things like that. And so then you focus on a subset of the disease mechanisms at the

00:50:42.200 exclusion of all others. And I don't want to single Alzheimer's out. I think a lot of

00:50:46.680 diseases meet that category, but it's unfortunate because what we're realizing is that there's a

00:50:52.680 lot of factors that contribute to any disease. And I think longevity may be an interesting way of

00:50:58.120 looking at it. Like, I think it's better, take a mouse. We've tried to make Alzheimer's models in

00:51:03.320 mice and they don't prove that informative. Why? You're creating a disease a mouse doesn't get

00:51:08.520 genetically in a young mouse and comparing that to a natural disease in an old human.

00:51:14.600 I think you learn more about Alzheimer's. If you look at the brain neurodegenerative changes

00:51:19.000 that happen in the mouse normally with aging, the downstream things are different, but the drivers

00:51:25.320 may be the very similar to the ones that are driving Alzheimer's. And that may be a better model

00:51:30.120 of Alzheimer's than trying to artificially create something that a mouse doesn't get. So

00:51:34.440 I think aging is helping change that perspective. The drivers of aging I think are very similar

00:51:40.440 between a mouse and a human. The downstream events can be different, but the drivers are what we care

00:51:45.960 about. Let's go back to rapamycin for a moment. Do you believe that the primary effect of rapamycin

00:51:54.200 is tamping down maladaptive inflammation through obviously the intermittent blunting of mTOR?

00:52:00.920 I think that's one of the major things, certainly. There's good evidence for enhancement of autophagy,

00:52:06.200 there's good evidence for changes in protein translation, and those things are not mutually

00:52:10.120 exclusive to inflammatory changes anyway. But I think those are the three things that we have

00:52:15.240 pretty good evidence for. I really think that all of these interventions that we're looking at

00:52:19.640 are restoring dynamic range. It may take super physiologic changes to change that linear line,

00:52:25.560 but I don't think that's what we're looking at right now. We're restoring things that happened when

00:52:29.320 you were young. So given that we're not likely to have human clinical trials of rapamycin that study

00:52:36.120 aging for the simple fact that we don't even know what an aging biomarker is, we're going to largely be

00:52:44.360 extrapolating from animal data if we have to make decisions about humans using rapamycin for gyro

00:52:50.760 protection. I'll push back a little bit. I know where we're going with this, but we're doing a study like

00:52:55.800 this in Singapore on humans. Six-month intervention with rapamycin. And we're looking at as many

00:53:01.560 different parameters of age. We're not doing disease. In fact, we're not taking people with

00:53:05.160 disease. We're taking people that are 40 to 60. They may have a precondition for a disease,

00:53:11.320 but they don't have anything that would be defined as a disease. So it could be high glucose. And then

00:53:16.760 we're looking at changes in a wide range of different biomarkers, clocks.

00:53:20.840 So tell me a little bit more about the study. So how many subjects?

00:53:24.520 We're doing these, I don't remember the exact numbers. It's somewhere around 150 to 200. So it's

00:53:29.720 not huge. And how you're dosing it?

00:53:31.480 Intermittently.

00:53:32.520 Once a week?

00:53:33.320 Yeah.

00:53:33.720 How much?

00:53:34.360 I think it's five milligrams is what's in the protocol. It's being run by Andrea Meyer,

00:53:38.600 which is one of my collaborators in Singapore.

00:53:40.200 Got it. So five milligrams of rapamune once a week for six months. And then let's go through all

00:53:47.480 the different measurements. So a range of clocks, inflammatory cytokine panels,

00:53:52.920 functional measures, pulse wave velocity, DEXA, strength measurements, cognitive measurements.

00:53:59.800 I'm still missing a couple of them. Do you expect to see changes in strength or

00:54:06.520 cognition or things like that? I mean, do you worry that those are kind of the wrong outcomes to look

00:54:11.480 for in a six month study of people that are young?

00:54:13.800 Yeah. This comes back to what do you measure? And this is where I knew we were going with this.

00:54:18.120 I don't think we know. I mean, certainly you can change those parameters. I think if you exercise,

00:54:22.680 you're going to change your strength.

00:54:24.040 Absolutely.

00:54:24.440 For sure.

00:54:24.920 If this was a six month exercise trial, fill your boots.

00:54:27.960 I think six month exercise trial might also change the cognitive parameters as possible.

00:54:32.440 Yep. A six month sleep correction trial would undoubtedly change cognitive parameters.

00:54:37.800 Yeah.

00:54:38.200 So I don't think it's unreasonable that a drug could do these things as well. I think

00:54:41.720 rapamycin is complicated when it comes to muscle. And I know that partly because

00:54:45.720 I'm going to be non-scientific for a minute. I've become my own best model organism.

00:54:50.440 So I try all kinds of different things on myself now. I know it's N equals one. I'm not sure rapamycin

00:54:56.840 and skeletal muscle, you know, without exercise, I'm not sure what it's going to do.

00:55:02.040 One of the things I notice is when I take rapamycin, if I do like a hard run,

00:55:06.920 I'm a runner. I've gone to more lifting the last three or four years. I've always been a runner.

00:55:12.600 I don't have good runs within 24 hours of taking rapamycin. And it may be because

00:55:17.480 you have to activate mTOR in a context or something like that.

00:55:20.840 Sorry, if you take rapa 24 hours prior to a run.

00:55:24.360 Within 24 hours of running.

00:55:26.040 Got it. Now, what about if you take rapa after a run? Is your recovery better?

00:55:31.880 I don't have a sense of that. What I do know is three or four days after I take rap,

00:55:35.880 I have really good training. I think what's happening is that maybe in that short window

00:55:40.360 after you take it, you can't activate the pathway enough. But in the long term,

00:55:44.520 what you're doing is dampening the basal signaling and you're getting the better dynamic range.

00:55:48.920 So if the trough levels are low, I think that's me.

00:55:51.480 I'm going to experiment with that, Brian, because I always take rapa the same day of the week.

00:55:57.000 I do the same workouts on the same days of the week. I'm going to do an adjustment on that and see.

00:56:02.520 I didn't try it with resistance either. So I don't know what it's going to do there.

00:56:06.280 What tools do we have to measure autophagy in humans?

00:56:09.720 Well, you can pull out blood cells. We're talking about the limitations of what you can get from a

00:56:15.720 human, right? Blood, saliva, that's where we're going, I think. You can pull out blood cells and you can

00:56:20.200 look at white cells and see whether autophagy pathways are induced or not. You can take muscle

00:56:25.160 biopsies. We don't really like doing that in our clinical studies because it makes it harder

00:56:30.840 to get volunteers. I think if you do muscle biopsies the right way, they're probably not

00:56:35.480 that painful to people, but people have that perception and we need healthy volunteers for

00:56:40.600 our studies. I would love to look at muscle. I think that autophagy is another one of these dual

00:56:45.880 edged swords though, right? You don't want autophagy on all the time. You want it on at

00:56:50.040 the appropriate levels at the right time. If it's on all the time, you're going to get muscle atrophy

00:56:54.600 probably. So it's about dynamic range. What would you need to see in this study

00:56:59.640 to feel that rapamycin is gyroprotective? Because my concern, I suppose, would be you're not going to

00:57:06.600 see a difference in DEXA. You're not going to see a difference in physical performance. You're not

00:57:09.800 going to see a difference in cognitive function. You might see a reduction in certain cytokines,

00:57:15.000 but not all cytokines. I forget what the other markers that you said were, but I guess my concern

00:57:20.440 is since we can't measure aging, we're not going to see enough of a signal. The other,

00:57:26.120 oh, you mentioned epigenetic clocks. In an ideal world, that would be the perfect

00:57:30.120 tool to measure them except for the Kaberlin experiment. I bring Matt to all my conferences

00:57:35.400 now because I used that slide. Where he did like eight of them at once. But there are multiple issues

00:57:40.600 here. One issue is, are the consumer testing companies, do they have a standardized enough

00:57:46.520 protocol that it's reliable? And I think I'm skeptical that that's the case.

00:57:50.120 You're doing in-house measurement.

00:57:51.640 We're doing everything in-house. We can control everything. So we get around a lot of those

00:57:55.880 problems.

00:57:56.360 Just for the listeners so they understand what we're talking about, Matt Kaberlin bought four of

00:58:01.400 the top commercial tests, did them in duplicate simultaneously. So it took eight tests at the same

00:58:07.880 moment in time and he has a funny graph that shows how pathetic they are. Not only do none of the

00:58:14.280 tests agree with each other, the identical tests rarely agree with each other. So just if you're

00:58:19.960 listening to this and you want to go out and get a commercial test that tells you how old you are

00:58:24.760 biologically, reconsider it. I want to tell you we have the potential for a better clock.

00:58:30.280 Want to talk about that? Yeah. Yeah. So consumer-wise, I think there's concerns.

00:58:34.440 What are you using as a control when you talk about, given how inflammation and epigenetic

00:58:41.720 change might be the only two signals that you find here? And again, this is just me being

00:58:47.560 pessimistic nanny. This is just my prediction. I do not think there will be a finding in any of

00:58:52.760 those other measurements, but you might have a chance with inflammation and epigenome if you're

00:58:57.560 measuring it. Maybe pulse wave too.

00:58:58.600 Here's my problem with pulse wave, Brian. It is so user dependent in terms of the technician who is

00:59:05.240 doing, like we don't use it clinically at all because we think it's a useless test.

00:59:09.800 I think the carotid intimal thickening is a useless test and that's an easier test to do

00:59:15.400 because unless you have a tech who basically has a PhD in how to do vascular imaging and they're the

00:59:23.640 only one that does it every minute of every day. If my patients come in with a CIMT, I open a bird

00:59:29.560 cage, I take out the poopy bottom paper, I put their CIMT in there, I close the bird cage. That's

00:59:35.320 how useless it is. So I just worry that all of those tests, they're just going to be noise,

00:59:39.880 no signal, but these other two might have signal. What's your control for accelerated aging or

00:59:45.080 something else? In other words, it'd be really interesting if you did a six month parallel fasting

00:59:48.840 trial where if you took people and you rendered them hypocaloric, you put them on some draconian

00:59:55.960 60% calorie diet for six months where you really think you would tamp down inflammation and autophagy.

01:00:03.640 If anything's going to reprogram the epigenome in six months, you think that would be it. That would

01:00:07.240 be a very interesting control, even if you'd had a fraction of the number of subjects.

01:00:11.080 Yeah. I mean, we're doing multiple. The first study was with a time-release version of AKG

01:00:16.200 with PDL Health. Alpha ketoglutarate? Yeah.

01:00:18.520 Okay. I'd like to hear more about that as well.

01:00:19.880 That's finished and we're just analyzing data. So I can't tell you much data, but we finished the

01:00:24.360 trial. We did six month intervention and three month follow-up. So we want to see if their changes,

01:00:28.840 do they maintain if you stop taking the compound? Okay.

01:00:31.880 Because in mice, oddly they do for aging, but I'm skeptical that's going to happen in humans.

01:00:37.720 We're doing multiple studies. They're not all at the same time, but we will be doing the same

01:00:43.400 study over and over and over again with different interventions.

01:00:45.640 Yeah. Yeah. Yeah. It's like, I was just saying your ITP equivalent in US.

01:00:48.600 Yeah. Yeah. By the way, I don't know if six months is long enough. I don't know if we're

01:00:52.360 doing the right tests. I don't think anybody knows these answers.

01:00:54.520 What is the cost of doing that experiment you just described?

01:00:56.840 It's about a million and a half sinks. So that's a million to a million and a half US dollars.

01:01:01.640 Yeah. This is really interesting. And not that I'm here to do this, but it's just,

01:01:05.240 I get asked so much by people, where can I put money? Where can I put money? So I'd like to make

01:01:10.520 sure that people who are listening to this, who are thinking about, hey, how do I fund

01:01:14.760 insanely high levered research? This to me strikes a great way to be funded.

01:01:19.880 Thank you. We would love to have that because it's really hard to raise money for this still.

01:01:23.480 Yeah. But if you think about it, if you're sitting there and you're listening to this and you're

01:01:26.120 saying, look, I'd like to put a million dollars to work on something that would dramatically change

01:01:31.080 adding a decade to life. In my opinion, you really are better off putting money into this type of

01:01:35.960 translational research than you are into basic science or into pharma research that tends to

01:01:42.520 be a little bit more, you're not going to move the needle.

01:01:44.520 Well, pharma is still not doing aging. They're thinking about it. They're starting to,

01:01:47.640 but they've been saying that for a long time.

01:01:49.400 No, this is great because you can tell I have strong opinions on ideas for how to do some of

01:01:53.800 these experiments. So let's talk a little bit about alpha-ketoglutarate. Walk us through the

01:01:59.320 rationale for how that came to be something that you would put through this type of rigorous program.

01:02:03.480 Yeah. So a company, full disclosure that I'm involved with, PDL Health, they have a product,

01:02:08.520 Rejuven now. They came to us many years ago at the Buck and working with Gordon Lithgow and I,

01:02:13.240 and they were like, let's green natural products that would have an impact on aging. Because the

01:02:18.360 mindset of what became the CEO of this company was, I can't get drugs approved for aging at any

01:02:23.960 time in the near future. Let's work with natural products.

01:02:27.080 Things that the FDA just calls grass out of the gate.

01:02:30.440 Right. And let's look for combinations of things that work together. Because we can get IP around

01:02:35.960 combinations. We can't get IP around single natural products. So we screened a lot of things in worms,

01:02:41.240 and AKG came out as one of the best things in worms. And then we started testing interventions in

01:02:47.000 mice. And that's led now to, we're testing, we do four or five different intervention studies in mice

01:02:52.120 every six months now. Not just with natural products and not just for this company, but we built

01:02:57.160 that into our own kind of ITP in mice too. Although we do it differently. But anyway,

01:03:02.280 AKG came out as one of the biggest effects. And then in the mouse studies, we found that for male

01:03:07.400 mice, there was a combined effect with vitamin A. Vitamins are an interesting discussion too.

01:03:14.280 There was also a combined effect with both sexes with vitamin D. And so that led to this product

01:03:19.000 Rejuven, which is AKG, time release AKG. Very important because otherwise it goes away in five

01:03:25.160 minutes if you don't have a time release version. Tell folks what alpha-ketoglutarate is.

01:03:29.480 Is it part of the Krebs cycle? Yeah, it is. It's a TCA cycle or

01:03:32.680 Krebs cycle, central metabolite involved in hundreds of reactions. It's a lot like NAD.

01:03:37.480 They're doing different things, but they're both central metabolites. They're both going down with

01:03:41.400 aging in organisms. And the idea is supplementing them back up would be beneficial. And so when we did

01:03:48.360 that with AKG in mice, we see about a five to 10% increase in lifespan, but a dramatic increase in

01:03:54.680 frailty or decrease in frailty. Yeah. Yeah. I know what you mean.

01:03:57.000 Yeah. So that mice, I would argue that you're squaring the longevity curve as we talk about it.

01:04:01.400 Yep. Yeah. So if that translated to humans, it would be a big impact. And so that's what led to the human

01:04:06.680 product. By the way, I think that if you did not extend lifespan by a day, but you just improved

01:04:13.640 health span, that's a home run. Yeah. I agree. For most people, that is all they actually want.

01:04:19.240 Yeah. I agree. I think if they get that, they'll want the other two.

01:04:22.520 Never the reverse. No.

01:04:23.960 Nobody wants more lifespan. The reverse is what we're doing now.

01:04:26.440 I know. I know. It's called medicine 2.0.

01:04:29.720 But yeah, I completely agree with that. So we're still excited about this product,

01:04:34.440 about AKG and especially the time release version. And there's been some studies that have been

01:04:39.880 published, one of which we helped analyze the data for using rudimentary methylation clocks,

01:04:45.240 showing that it reverses aging by a few years. Again, it's that oscillation thing we were talking

01:04:50.600 about earlier, I think. And that's why we wanted to do a controlled placebo double-blinded clinical

01:04:55.720 trial at the university. In that case, we're just testing the time release AKG. We didn't include the

01:05:01.160 vitamins because we're trying to get some mechanistic information. We don't want confounders in there.

01:05:06.760 But I think the data is still pretty good on AKG that it's going to have an impact in people.

01:05:11.640 That human trial has been completed. You're evaluating the data. You're going to have to see a

01:05:15.800 signal. But in each of these trials, do you do the same measurements, the same outcomes?

01:05:20.680 Generally, yes. But we sometimes modify the primary endpoint because we want to choose a

01:05:26.280 primary endpoint that's most likely impacted by the intervention. In a way, it doesn't matter

01:05:30.920 that much because we're measuring as many things as possible anyway. Of course, we learn over

01:05:35.880 time as we do things. So we add things or take things out that are not working well,

01:05:39.720 that sort of thing.

01:05:40.360 Did you guys do a study on urolithin A as well?

01:05:43.480 We haven't published it yet, but I'm happy to talk a little bit about it. We don't have human

01:05:47.400 data, but the mouse data is really good on urolithin A. This is why we haven't published

01:05:51.240 yet because when we did it, it dramatically reduced frailty in male mice, but not females. And so we're

01:05:57.720 repeating that. There may be specific differences, but we're the first ones to see that. So we want to go

01:06:02.440 back and see what's going on. I feel like I've written a newsletter on this where I

01:06:06.920 came down on the side of this isn't doing anything. Am I mistaking this for a different

01:06:12.360 molecule? Is this the one that in theory enhances mitochondrial function?

01:06:16.120 Yeah. The idea is that it enhances mitophagy, so turnover of damaged mitochondria. I'm not sure

01:06:22.200 I completely agree with that. We do see that in cell culture, but we also see mitochondrial

01:06:27.080 biogenesis. And so these two things are connected. If you induce mitochondrial biogenesis,

01:06:32.120 you'll also induce mitophagy. And I'm not sure where's the chicken and where's the egg in this,

01:06:36.200 but it does seem to induce mitochondrial turnover. But we also find other pathways.

01:06:41.560 When we look at a molecule, if we don't know enough about it, like rapamycin, we know it binds to

01:06:46.200 mTOR. Urolithin, we don't know what it does. And AKG also. We know a lot of things it can do. We don't

01:06:51.800 know which ones are relevant for aging. For urolithin, we went back and did this screening

01:06:56.360 assay to proteomics thermal shift assay to look for binding partners for the compound. And we have

01:07:02.120 some... Tell me about the history that led to using it.

01:07:05.080 Well, for us, Johan Ower published this data that it was slowing aging in mice, and that's

01:07:12.200 led to a lot of research. So we weren't the first ones to study urolithin. And a lot of what we do is

01:07:17.160 testing interventions that come from other labs. Because if I'm going to do a human study, I want

01:07:22.360 to at least see it repeat in my hands in an animal first. Absolutely, yeah.

01:07:25.800 If it doesn't repeat in my hands, that doesn't mean it doesn't work. It may mean the conditions

01:07:29.560 are different. But I think if it does repeat, it makes an argument for robustness. And so that's why

01:07:35.560 we started urolithin. We also see positive effects of spermidine and glycine and other things.

01:07:41.240 But we didn't make the initial discovery on urolithin.

01:07:44.440 So we don't know the mechanism of action.

01:07:46.440 Well, they would argue it's increased mitochondrial turnover, but we have...

01:07:49.080 But we don't have a target.

01:07:50.120 We have targets we haven't published yet.

01:07:51.960 Oh, you do? Okay. Did you guys discover the target?

01:07:54.520 Yeah, we've got a couple of new targets that we haven't published yet. So that could explain

01:07:58.680 some of the effects.

01:07:59.480 And have you studied this in humans yet?

01:08:02.360 No, not yet. That's planned, but hasn't been done yet.

01:08:05.480 Okay. This might be actually one of the ones where I would say, unlike in the RAPA trial,

01:08:11.480 you actually want to come up with primary outcomes that are quite different.

01:08:14.920 This is where, for example, I'd want to see fat oxidation.

01:08:17.880 Yeah.

01:08:18.600 I'd want to see what we talk about is zone two efficiency. If this is indeed improving

01:08:23.400 mitochondrial function, I'm not aware of a better test of mitochondrial function.

01:08:27.720 Yeah, no, I agree with you. And that raises the point of whether we should couple these

01:08:31.080 interventions to exercise.

01:08:32.440 That one for sure.

01:08:33.960 That adds complication, but it may be worth doing in this context.

01:08:37.000 And not to get in your business, you must muscle biopsy these people.

01:08:41.080 Yeah, I know.

01:08:41.640 You're going to have to get athlete. You're going to have to get people that are willing

01:08:43.880 to-

01:08:44.120 For that study, we're going to have to do it.

01:08:45.240 To get a little punch biopsy of the quads.

01:08:47.480 Yeah.

01:08:47.800 Because it's just too important to understand what's happening.

01:08:51.080 I think a lot of these supplements, I'm going to go back to an N equals one story here.

01:08:54.680 Yeah.

01:08:55.000 I think a lot of these supplements are impacting exercise. And so, it's kind of a win-win. You

01:08:59.960 take something, you exercise better, you drive benefits from that exercise. So, it's kind of,

01:09:04.680 you win twice with some of these. I'll tell you what happened recently. I've been very skeptical of

01:09:09.480 NAD. Not that going down and restoring it won't be good, but I'm skeptical it's going through sirtuins.

01:09:15.960 I think it could be doing a lot of other things. But I'm also skeptical that NRNMN are really

01:09:21.160 changing NAD levels that much. And our mice don't really respond in our studies with NRNMN.

01:09:27.000 I've never noticed anything taking NR. Again, that's just one person, but that's me.

01:09:32.200 When I last looked at this, and I know I'm going to get a lot of hate mail because I always get

01:09:36.680 hate mail when I talk about this. The only study I've seen in humans that shows a real benefit of

01:09:44.840 NR was in patients with ALS that had the patients in the NR group had a longer time before requiring

01:09:53.080 ventilation than the patients on the placebo. Is there another study I'm missing?

01:09:58.040 Well, there have been healthspan studies arguing improved healthspan.

01:10:01.240 But by what metric?

01:10:02.680 I can tell you we've done them as well. And we do see

01:10:06.680 very subtle changes that are aspects of the frailty measurements we do.

01:10:10.760 In mice or in humans?

01:10:11.640 In mice. But it's not enough to really convince us it's statistically significant.

01:10:15.720 I wouldn't be shocked if there's a tiny effect is what I'm saying.

01:10:18.840 And why do you think that is? Do you think it's that NR and NMN are not efficient vehicles

01:10:25.320 to generate enough NAD?

01:10:26.360 Yeah, I think that's one thing for sure. Well, let me tell you the rest of the story because I've

01:10:30.680 pretty much given up on the pathway and a company that came to me and wanted to test one of their

01:10:35.720 products. This company is called IX Biopharma anyway. They have a new product that's

01:10:40.200 sublingual NAD. They specialize in technology for sublingual delivery and they make other things

01:10:46.440 too. Just explain to folks why that matters. You can't take NAD orally because it just gets

01:10:51.720 destroyed. Typically, people take NAD intravenously. But if you take something under the tongue,

01:10:58.360 you get this magical property where it dissolves and it enters the circulation without passing through

01:11:04.920 the digestion and obviously the liver where these things get chewed up.

01:11:08.360 And you can do this every day.

01:11:10.120 Lots of drugs are given in this man.

01:11:12.040 Yeah. IV is just not practical on a repetitive basis.

01:11:16.280 How many milligrams of NAD?

01:11:18.920 100 milligrams. The one I'm taking also has apigenin in it, which is a CD38 inhibitor.

01:11:25.080 So CD38 is a consumer of NAD. So if you block that enzyme, this is another natural product. If you

01:11:31.880 block that enzyme, then you also effectively increase NAD.

01:11:35.480 Okay.

01:11:36.120 So I was taking that along with the rejuvant, which is the AKG plus vitamin A and also the B complex,

01:11:41.800 but mainly AKG effect. And when I take them together, I notice this acute effect on my exercise

01:11:48.040 performance. So I'm running, my heart rate goes up. My respiratory rate doesn't go up as much. It goes

01:11:53.720 up a little bit, but normally if my heart rate's at 150 or 155, I'm breathing hard. I'm breathing

01:12:00.600 closer to normally when I take these two things together. I've gone off. I've gone back on. I've

01:12:05.080 taken one off.

01:12:05.800 Is your rate of perceived exertion tracking more with your respiratory rate or your heart rate?

01:12:10.680 It's tracking more with my respiratory rate. I don't perceive that I'm exerting. I run faster

01:12:15.480 when I do it.

01:12:16.360 You know what would be so cool to see, Brian, is measuring your lactate levels

01:12:20.360 at a fixed heart rate under those two different respiratory rates, but under the same load. Do it on a

01:12:25.720 treadmill just to make it unambiguous. I'd be super curious to see a lactate performance curve.

01:12:30.680 I should do that. I don't think you can imagine this. I've got things telling me these two

01:12:35.480 parameters.

01:12:36.280 I'd like to just quit my job and do nothing other than these experiments.

01:12:42.680 Yeah. And this may happen for no one else. I don't know. But for me, I can go off. It sort of

01:12:47.480 goes away. It's starting to not go away now because it's improving my exercise performance and now I'm

01:12:53.080 getting more fit. So it's getting harder to see the effect, but it still comes and goes when I

01:12:57.880 go on and off either the AKG or the... The other thing I'd be interested to see is,

01:13:02.520 let's pretend off drug, heart rate 150, RPE 7, velocity X. On drug, I want you to go back to the same RPE,

01:13:14.600 let heart rate go higher and let velocity go higher. And I'm curious as to how long you can

01:13:20.840 sustain that relative to what you were doing before. In other words, does some other system

01:13:27.560 get in the way that ultimately reduces your capacity? I can't answer. But the reason I noticed

01:13:33.640 is that I'd taken it three days. I didn't even think about it. It's like every morning I do it.

01:13:39.080 I was on the treadmill and I'm running. I keep pushing the speed up and I'm not getting out of

01:13:45.640 breath. And I was going to run 5K. I ran 12K. And I still didn't feel that tired at the end of it.

01:13:52.200 And then I started thinking, how did this happen? Because I know how my body performs normally.

01:13:57.560 Is this molecule in trials yet with humans?

01:14:00.280 It's not even have to be in trials. It's the natural products. It's on the market.

01:14:04.040 Yes, yes, yes. But for those of us who want to actually know if it works.

01:14:07.080 Oh, come on. Yeah, yeah, yeah. Those of us who believe in this pesky thing called evidence.

01:14:11.080 Yeah, yeah, yeah, yeah, yeah. I'm getting past that now. I'm joking. There have been a lot of

01:14:16.040 animal studies and they can show that when they, I'm not even sure this is published. I think I can

01:14:20.600 say this. When they add sublingually in a rat model, you have to anesthetize the rats to get the

01:14:26.120 sublingual delivery. You see it incorporated very highly in red blood cells, the NAD. This was just with

01:14:32.040 the NAD, not epigen. There is literature to support this and it's not my field, but I'll just

01:14:36.840 say what it is, is that they're channels that can take up NAD directly in certain cell types.

01:14:43.160 There is literature out there for that. Yeah, I had Josh Rabinowitz on the podcast

01:14:46.840 and he talked at length about this. What I took away from that discussion was intravenous NAD will

01:14:52.200 work. The question is, what's it working for? And so would your belief be that the effect you

01:14:57.480 experience here should be mirrored by what somebody experiences with intravenous NAD,

01:15:02.600 notwithstanding the limitations of the frequency that they could do it?

01:15:05.240 Yes, I think so. But the dose is so much higher on NAD and it could be too high. I don't know. But

01:15:11.400 in theory, yes. But this is so simple. You just, every morning it's gone in 30 seconds. For me,

01:15:17.240 again, it's enhanced by the AKG too. So you're adding two metabolites that are both going down with aging,

01:15:23.320 that are both involved in hundreds of different, they're giving you cellular metabolic flexibility.

01:15:27.880 I think that's what they're doing. I don't think it's one pathway they're activating.

01:15:31.400 And is the alpha-ketoglutarate, it's vitamin E or A that it's combined?

01:15:35.320 For a male, it's vitamin A and the new product has some B-complex in it as well.

01:15:39.320 And these are commercially available products?

01:15:41.080 Yeah.

01:15:41.640 And you're about to publish on them in humans?

01:15:43.880 The rejuvenate has already been published on the AKG.

01:15:46.680 By itself?

01:15:47.640 Yeah, an uncontrolled, not placebo controlled, methylation study of users. Okay,

01:15:53.160 so take what you want out of that.

01:15:54.680 Not much. I'm just going to say. You can talk me off my-

01:15:57.560 No, no, no. I don't want to try to, but I will say that I think community data is valuable.

01:16:01.320 And I think if you go back and look at the paper, we didn't do any of the analysis. We just analyzed

01:16:06.520 the data. But if you go back and look at the paper, I think we were very clear in that paper,

01:16:10.680 even in the abstract of the limitations of the finding. So I think that for community-based data,

01:16:16.120 it's better to have it out there and published, but you want authors that are willing to be

01:16:21.320 honest about what the data says and what the data doesn't say.

01:16:24.680 But yeah, the study that's been completed is just AKG time release, so none of the vitamins,

01:16:30.040 and it's placebo controlled in as many parameters as we can measure. So hopefully it'll show something.

01:16:35.560 Okay. And then you're doing the one with vitamin A and E?

01:16:39.000 We're not doing that right now. We'd love to do that, but we financed this study ourselves.

01:16:44.440 The only thing the company did was supply the time release AKG. It would be good to go back and

01:16:49.240 look at the actual product in combined with this NAD now. But I just sort of figured that out myself

01:16:54.600 five months ago. I don't even have any animal data that that's true. It would be interesting to go

01:16:58.040 back and add these things in animals, but you can anesthetize a rat and do sublingual delivery

01:17:02.840 one time and measure the PKPD.

01:17:05.480 And doing all the time is going to be a nightmare. Yeah.

01:17:08.120 You mentioned spermidine a moment ago. Let's talk about that. It's getting quite a bit of buzz.

01:17:12.200 Tell us everything about spermidine.

01:17:13.640 Well, so we studied that a while ago when I was at the Buck and we wanted to look at,

01:17:19.240 we were confused because at that time, the data was spermidine could extend lifespan,

01:17:23.240 but it didn't impact metabolism. And I'm like, almost everything in a mouse that extends lifespan

01:17:28.280 has some impact on metabolism. So we did a high fat study with spermidine,

01:17:32.840 and we showed that in old animals, spermidine could suppress the metabolic dysfunction that came from

01:17:39.560 a high fat diet in mice. But we had some control mice too, and actually not many, but the study

01:17:45.400 wasn't designed to do lifespan, but we left them alive and looked at survival and the spermidine

01:17:49.320 extended the lifespan of the mice too. Has spermidine been studied in the ITP?

01:17:53.160 I don't think it has, does it? I don't think it has. Don't quote me,

01:17:57.960 but I don't think it has. We were able to repeat the lifespan effect in mice,

01:18:01.880 even though that wasn't the goal of our study, and show that it restores metabolism in a

01:18:07.160 calorie-challenged context. And so that's what made me believe that it's a robust molecule,

01:18:12.520 that we can see that effect as well. So I would say I'm optimistic about spermidine as well,

01:18:16.920 that we haven't done a lot with it. Where does spermidine occur in nature?

01:18:19.720 I actually don't remember the answer to that. But it's naturally occurring?

01:18:22.920 It does naturally occur, yes. I think it's in certain foods or habit, but I'm worried I'm

01:18:27.960 conflating that with the urolithin. Okay. Have you guys looked at alpha-estradiol-17?

01:18:33.240 No, we haven't looked at that. That would be an interesting one.

01:18:36.680 Yeah, I agree. And I'm also director of an Asian center for reproductive longevity and equality in

01:18:43.080 US and Singapore, where we're looking at ovarian aging. It's a PhD biologist. All of a sudden,

01:18:49.320 I'm getting asked a million questions about HRT, so I decided I better learn something.

01:18:53.480 But there's two things that I've learned from this. One is that geroprotectors

01:18:57.080 tend to extend fertility in mice. And it's not just one thing. It's spermidine, it's AKG,

01:19:02.600 it's been shown for metformin and rapamycin.

01:19:04.760 There's a RAPA trial going on, I think, in Columbia or Cornell.

01:19:08.600 So that's one really potentially useful avenue of translation for these geroprotectors. I'm

01:19:15.000 really excited about that. The other thing I think, you have more expertise than I do,

01:19:19.880 please disagree, is that the question should not be, should a woman do HRT? The question should be,

01:19:27.000 is there any reason a woman should not do HRT? Because I just see the value of hormone

01:19:32.520 replacement to far outweigh the risk in the majority of women. I don't know if you feel that

01:19:36.600 way. But 17-alpha is interesting, right? Because it worked in males.

01:19:40.280 And not females.

01:19:41.240 Yeah. It's also non-feminizing. I don't know what that really means. Not sure how much I believe that.

01:19:47.320 And so it could be triggering some of these same pathways. And I know testosterone is another one

01:19:52.360 that the NIE did the study on it and showed it's a little bit of an increased risk in prostate

01:19:56.600 cancer. And then everybody said, don't do testosterone replacement.

01:19:59.640 That's been fully, fully debunked now.

01:20:01.880 I agree. And then most men are not doing that. Or if they're doing it, they're doing it in

01:20:07.720 uncontrolled ways that taking levels maybe to too high that could be dangerous.

01:20:11.880 I feel like men are in a better position today because one, more trials have been done to undo

01:20:19.080 the bad ones. So the Traverse trial, which was published probably been a year now,

01:20:22.600 even though it was the Traverse had many problems with it, but it undid a lot of the damage of some

01:20:26.280 of the fear mongering from really bad studies that suggested testosterone causing prostate cancer.

01:20:30.680 Turns out it's the exact opposite. Women, unfortunately, are still

01:20:34.840 struggling under the dark cloud of the Women's Health Initiative, which was apocryphal. No one's

01:20:40.760 undone it.

01:20:41.640 I travel to 25 countries a year and a lot of countries in Southeast Asia. And so now everywhere

01:20:45.640 I go, if I see a doctor, I'm like, how many women are doing HRT?

01:20:48.520 And what are you finding?

01:20:49.240 It's extremely low.

01:20:50.200 And because they're deferential to US recommendations or?

01:20:54.040 I think the changes and what doctors tell people is slower in these countries, but maybe there's

01:21:01.480 less risk. It's also cost money. If you have somebody that doesn't have a lot of access to

01:21:05.960 finances, they may be less able to do it in these countries. I don't know all the reasons,

01:21:10.440 but I'm just looking into it now, but it's stunning how low the HRT is around the world still.

01:21:16.920 And we're just missing an easy opportunity, I think, to help people.

01:21:21.320 I agree. I mean, I think there's a part of me that is just so interested in the frontier.

01:21:26.840 How do you push the boundaries of this stuff? I truly could be infinitely happy working

01:21:32.360 nowhere but at the frontiers of thinking about the molecules. But at the same time,

01:21:37.480 I feel just as excited about trying to figure out how to make sure people aren't scoring own goals

01:21:43.320 all day, right? Like there's so many ways to just help a person without anything magical,

01:21:52.440 find another five years of life and dramatically improve health span through judicious use of

01:21:58.200 everything from HRT to correct exercise, reasonable nutrition. There's just so many great ways to do

01:22:05.400 it. But obviously the idea of doing both of these is so appealing.

01:22:08.600 I come back every three months to the US. You know, I lived here 50 years and then I left. And

01:22:12.520 I come back every three months. You leave someplace, you come back, you notice things you didn't notice.

01:22:17.320 And the same two things every time I come back, I get it in the airport. I don't even have to like

01:22:22.040 leave the airport. One is so many people are not in shape. There's so much obesity. It's striking

01:22:29.400 compared to almost anywhere else in the world. The second one is people just seem so stressed here.

01:22:34.440 It's Singapore is a pretty stressed country. And I still feel that people are more stressed here

01:22:39.400 when I come back.

01:22:40.280 What are the obesity rates in Singapore?

01:22:42.200 Relatively low, but in Asia, you've got this challenge with skinny diabetes. So you have a lot

01:22:47.160 of people that can't build the adiposity that they need.

01:22:50.760 Right. But they're storing all the visceral fat.

01:22:52.520 So they're storing the fat in the wrong places. And this may be even worse. So the diet is moving

01:22:57.240 more Western in Asia and it's creating problems. It's just not as obesity associated.

01:23:01.960 Yeah. And now that you're coming back, do you notice a difference at all based on GLP-1 agonists?

01:23:08.680 I think they're making a difference. I guess-

01:23:11.000 It's not passing the airport test yet.

01:23:12.280 It's not passing the airport test yet. Yeah. I don't know how widespread they're being used

01:23:16.200 either. You probably know that.

01:23:18.120 I don't think I'm familiar with the latest stats on how widely they're being used. Let's go back to

01:23:22.040 the epigenetic clock. Do you think that there could be value in this as a tool? And let me hold the

01:23:27.240 bar as high as I think it would need to be to justify their use. Right now we have this thing

01:23:32.280 called chronologic age. I can look at your birth certificate. I can know how old you are. And based

01:23:37.400 on that, I can make an estimate of how much longer you will live. So if I look at a person who's 40 years

01:23:43.800 old and I know nothing else about them, and then I see another person who is 65 years old and I know

01:23:50.440 nothing else about them, I can say with a high degree of confidence that the 65 year old person

01:23:57.640 will live, I'm making this up because I'm not an actuary, but somewhere between 20 and 30 more years.

01:24:05.320 And the 40 year old person, I can say with a pretty high degree of confidence, will live

01:24:09.560 somewhere between, call it 30 to 50 more years or something like that. Now, to me, that's a pretty

01:24:15.080 good test. I know a knowable measurable thing about them. It's measurable by knowing their birth date

01:24:20.840 and it predicts future life. Do you think biologic clocks will ever serve a purpose like that where I

01:24:28.040 could take two 50 year olds and one of them has a biologic age of 40 and one of them has a biologic

01:24:34.840 age of 60 according to the clock and that those numbers will actually be better at predicting future

01:24:42.360 life than their chronologic age of 50. Or do you put yourself in the camp that says, no, Peter,

01:24:47.800 that's a ridiculous standard that no biologic clock could ever come to, but it might tell me

01:24:53.800 about their health. It might be yet another biomarker that says, hey, the guy at 40 is just healthier

01:25:00.440 than the guy at 60. And somehow, by the way, it's picking that signal out of a data field that I can't

01:25:07.240 pick out anywhere else because they otherwise look identical. So we measured this recently

01:25:13.080 because collaborators of mine at NUS, Jan Gruber and Feng Shung, and I had this small role in this

01:25:18.520 project. They decided, Feng Shung's a geriatrician. He sees people all the time frustrated. The

01:25:24.840 geriatricians have limited things they can do. They're seeing people that already have multimorbidity

01:25:29.640 and the clocks are not that useful. And so he's wanted to like, how do we create a reliable clock

01:25:35.000 that a doctor can understand? For what purpose?

01:25:37.720 For biologic age. I'll tell you what I think the purpose is for it in a minute.

01:25:42.360 So the first generation and second generation clocks, the first generation clocks try to predict

01:25:46.520 your chronologic age and the second ones predict some outcome. So the question is, we want to predict

01:25:51.560 mortality. We don't want to predict your chronologic age. So intrinsically, if it works, it's going to

01:25:56.600 do better than the chronologic age for the second generation clocks. So he took NHANES,

01:26:01.800 data collected around 1999, 2000 mortality data for 200 months. And these parameters are nice because

01:26:09.400 you can actually do a consumer test of HbA1c. There are many labs that do that. It's reproducible

01:26:15.160 to a large extent, much better than DNA methylation. And doctors use all these parameters. So the things

01:26:21.080 that are in NHANES, LDL, all the things in your book, inflammatory markers, medical tests, some

01:26:27.080 cognitive self-reported stuff. So we just took everything as a feature and used AI, a linear

01:26:33.160 model, to try to predict mortality. And we're on the second generation of this clock now. And it

01:26:39.800 predicts mortality better than any other parameter in NHANES. It's way better than ASCVD. There's

01:26:46.440 cardiovascular disease measurement. And so recently, the methylation data came out on NHANES. So we could go

01:26:53.080 back and compare the mortality prediction for methylation clocks. Some of the first generation

01:26:59.000 clocks are worse than chronologic age. Your passport is better than they are at predicting mortality,

01:27:04.760 which to me means that they're not useful because even if they're not designed to predict mortality,

01:27:09.960 if they don't capture some element of that, what are they measuring? The second generation clocks,

01:27:15.000 like grim age and pheno age, they do better job than chronologic age for sure.

01:27:20.120 At predicting mortality?

01:27:21.560 Predicting mortality, yeah.

01:27:22.440 Just to be clear, we've captured that out of the NHANES database?

01:27:26.440 That's from NHANES. That's the only thing we've looked at right now.

01:27:28.920 Okay. Let me make sure I understand that. You had 200 months of forward-looking data.

01:27:34.440 Yeah.

01:27:35.000 You've got 18 years of data. And you're saying, if we know the methylation of somebody at that time

01:27:42.120 in the cohort, 1999 to 2000, we could predict their date of death better than the actuarial

01:27:49.800 data of their age.

01:27:50.680 Yeah.

01:27:50.920 Okay. I wasn't aware of that.

01:27:51.880 It's been accepted for publication. I'm not even sure it's online yet.

01:27:55.000 Okay.

01:27:55.320 So it's not your fault.

01:27:55.960 That's very interesting. Okay. Yeah. So that basically answers a question that I've

01:27:59.640 never seen answered.

01:28:00.520 Yeah. But our clinical chemistry clock does better than those.

01:28:03.320 Okay. What is included in that clinical clock?

01:28:05.960 Well, there are about 50 parameters that we measure now, but complete blood count gives you about 30

01:28:10.520 of those parameters. So it's not as elaborate as you think it would be.

01:28:13.240 Yeah. Yeah. Yeah. No, I believe that.

01:28:14.120 And then it's a lot of standard markers that you already measure. You probably measure all of them.

01:28:18.040 Yeah. So that, I guess that was going to be my question is-

01:28:21.400 It's about $300 in Singapore if you did it all de novo, but anybody going to a doctor's office

01:28:26.600 has most of those parameters measured anyway. And if they're going to a wellness longevity center,

01:28:31.800 all of them are probably-

01:28:32.600 Of course. Yeah. The question I suppose is this, if you're MetLife, you are better at predicting mortality

01:28:39.240 than anybody on the planet. And I don't know if it's MetLife by the way, but pick the best life

01:28:43.320 insurance company. This is their business. They're so good at predicting mortality. It's frightening.

01:28:49.000 I don't know how they're doing that. So I can't comment directly.

01:28:51.400 Well, and nobody does, right? Yeah.

01:28:52.520 But my point is they're looking at age. They're looking at a whole bunch of things in your medical

01:28:57.080 history. They're looking at a whole bunch of blood tests, your blood pressure, your weight,

01:29:01.800 your waist, your circumference, all those things. And they're coming up with an exceptional

01:29:06.120 prediction of remaining years in life. The real question is, do you believe that they will

01:29:11.880 incorporate a second generation epigenetic clock? Or do you believe that they've already got that

01:29:16.920 captured in their dataset? They may, I don't know. It's an unanswerable question. We're focusing on

01:29:21.880 the clinical chemistry anyway. We're not doing any methylation. And what we're finding is hospitals

01:29:26.280 want to use this now. Clinical chemistry or methylation?

01:29:28.840 The clinical chemistry. I think it's because it resonates. When you show them the list of parameters,

01:29:33.640 doctors, a doctor doesn't have to be an expert in epigenetics to figure out what's going on.

01:29:38.600 And the other thing is they're all actionable. So we have principal components that we can break

01:29:43.000 it down in, and we can see smoking in one component, and we can see metabolic disease

01:29:46.840 in another one and obesity in another one. And we find cases, a few conclusions from this are really

01:29:52.600 interesting. One is we find cases where nothing's out of the reference range, okay? So a doctor that's

01:29:58.680 looking at things, especially if they have a few minutes to look at, they're not going to prescribe

01:30:02.840 anything for this person. But these four parameters in this principal component are increasing their

01:30:09.160 biologic age by four years, which means it's 50% increase in mortality risk. These are actionable

01:30:14.440 things. You can treat LDL, you can treat high blood pressure, you can treat these markers, right? And so

01:30:20.440 clinicians are actually willing to then be a little bit more aggressive and try to prescribe

01:30:25.560 something or lifestyle modification or something to treat these markers. It's actionable.

01:30:30.360 Yeah, exactly. That makes sense. The reason that you prefer this is it doesn't just give you an

01:30:35.480 answer, it gives you a solution. Yeah, that's what we're working toward. And

01:30:39.000 the other thing we did is we broke in Haines down, really interesting. And Haines also has all the

01:30:43.640 medications people are taking. But the weakness of it is a snapshot, it's a cross-sectional measurement

01:30:49.160 of all these things, but they measured a lot of stuff. And so we could look at people's clinical

01:30:54.360 parameters out of the reference range, should they be being prescribed some drug and they're not

01:30:59.400 being given it. And that goes up to about 20% when you're 65 years old in the year 2000 in the US,

01:31:05.720 meaning 25% of people should be treated for something, but they're not being treated.

01:31:10.200 Those people have a higher biologic age and faster mortality, not surprising.

01:31:14.040 There's also the group at 65, every clinical parameter looks good. They have a lower biologic

01:31:19.240 age, they live longer. But you can break that group down and you can say one group is not taking

01:31:24.760 any medication and the other group is taking medication. It's just their clinical parameters

01:31:29.560 are managed well. The people taking the medication have a lower biologic age and live longer.

01:31:34.680 And it doesn't really matter what the medication is. It's true for the major medications you would

01:31:39.640 give for metabolic and cardiovascular disease in the year 2000. So I think that suggests that being

01:31:45.400 more aggressive and getting people optimized earlier is better than just being, oh, I'm pretty healthy

01:31:51.960 and my blood pressure is a little bit high, but I don't need to take, you know, and so it's suggesting

01:31:56.920 we need to be more aggressive. I think the other thing is that suggesting that these drugs that were

01:32:01.160 around in 2000 for these treating preconditions are actually aging drugs. They're actually extending lifespan.

01:32:08.280 And what are the classes that we see the most common lipid and hypertension?

01:32:12.840 Hypertension, Metformin. It's the standard things.

01:32:15.640 So do you think Metformin will have

01:32:19.400 gyroprotective properties in people who are metabolically healthy?

01:32:23.320 Skeptical. I think what it is saying is that if you catch preconditions early enough,

01:32:28.040 you protect against the other failure states a little bit too.

01:32:30.760 So you're more optimistic that rapamycin would be gyroprotective in humans than Metformin?

01:32:35.720 Yeah, that would be my prediction. Yeah.

01:32:37.560 Do you think there is a drug out there that you think is more

01:32:40.920 likely to be gyroprotective in humans than rapamycin at this time?

01:32:44.600 The GLP drugs and SGLT2, I think are interesting. I don't think we have the data right now.

01:32:51.880 Again, I don't know in people that are, you know, if you're obese, I need...

01:32:55.960 Let me restate the question. If we just took a population of middle-aged,

01:33:00.520 healthy individuals, we could design an experiment. You had the placebo arm, the Metformin arm,

01:33:06.760 the rapamycin arm, the SGLT2 inhibitor arm, and the GLP-1 agonist arm. What is your prediction

01:33:14.360 in length of life or additional years of life given in that six-arm study or whatever it is?

01:33:20.360 I think the last three would be comparable Metformin I'm skeptical of.

01:33:24.200 So RAPA, SGLT2, GLP-1.

01:33:27.080 We don't have data and healthy people that much with SGLT2 and GLP-1. I mean...

01:33:32.040 So mechanism of action is what? If the GLP-1 group and the SGLT2 group are metabolically healthy,

01:33:39.720 they don't have glucose excursions that are high, they're insulin. What do you believe is the

01:33:43.800 mechanism of action?

01:33:44.680 I guess I'm going by a different statement, which is most people we're calling healthy are

01:33:48.440 not totally metabolically healthy.

01:33:50.040 Fair.

01:33:50.680 In those cases, I think there would be a benefit. I don't know in the perfectly optimized person

01:33:55.480 whether there'd be a benefit. Every time I talk to a doctor, I ask them,

01:33:59.320 are you losing more lean muscle mass with these drugs than you are just by fasting or

01:34:03.640 lifestyle? And I get... If I ask 10 doctors, I get 10 answers. So I don't know what the answer

01:34:08.120 to that question is.

01:34:09.000 I think the reason, at least as a thought experiment, it's an interesting question is

01:34:15.080 because if the GLP-1 agonists and SGLT2 inhibitors only work, if you have some degree of glucose

01:34:24.520 irregularity, then it begs the question, well, it says, look, glucose homeostasis is one of the

01:34:31.240 most important features of living. Great. But if you could correct that with diet, sleep,

01:34:36.840 and exercise, which you can.

01:34:38.840 Yeah, I agree.

01:34:39.560 It's hard, could be done.

01:34:40.920 Yeah.

01:34:41.480 Then those things aren't going after fundamental pillars of aging because people who eat well,

01:34:47.960 who exercise well and sleep well, still age.

01:34:50.360 Matt and I debate this all the time about... We published a study in mice recently where we

01:34:55.240 analyzed all the data that's out there in mice. And we tried to determine where's reality on

01:35:01.720 interventions that extend lifespan in mice. Because if the control mice are really short-lived,

01:35:06.680 which happens a lot.

01:35:07.560 A lot. Why is that, you think?

01:35:08.920 It's just really hard to control. I mean, if you look at the ITP data, the control mice are all over

01:35:13.320 the board and they're very well-controlled, best scientists doing the experiment. We see a lot of

01:35:18.760 variation too. There are some cases they're bad vivariums and that causes a problem. But even

01:35:24.760 in good vivariums, I don't know. It's true in every organism in yeast and worms. One cohort of

01:35:31.320 worms will all live a little bit shorter. One cohort of worms will live a little bit longer.

01:35:35.480 It's cohort dependent, but I don't know why. But anyway, if the mice are short-lived, if your extension

01:35:41.560 is there, all you can say is it's longevity normalizing. You don't know that it's slowing

01:35:47.800 aging. It's only when the controls are really long-lived and you're getting extension that

01:35:51.400 you can really make the argument it's longevity extending. And so that gets to your question.

01:35:56.520 The real answer though is dependent on how many people you believe that are optimized right now.

01:36:01.880 No. Because it's impossible to do the study. Yeah, no, no, of course.

01:36:05.720 It is few. Longevity normalizing works in this population,

01:36:08.440 trust me, at least for keeping people healthy. But whether it's really slowing aging is an

01:36:12.680 open question, I think. I think the best case would be rapamycin there.

01:36:15.800 Yeah. And then of course, it begs the question, which is, could these effects be additive?

01:36:20.760 So would there be a benefit to a person who is on balance quite healthy, but let's say their

01:36:27.400 hemoglobin A1c, if it is indeed an accurate representation of their average glucose,

01:36:31.480 is 5.4%. So I don't know exactly what that translates to. It probably translates to an

01:36:36.840 average blood glucose of 110 or so milligrams per deciliter. But the data, there are data that

01:36:44.360 show based on hemoglobin A1c that lower is always better. So 5.0 is better than 5.4,

01:36:50.360 even though 5.4 is deemed completely healthy. That's all cause mortality data. So we're saying,

01:36:55.720 we take a person who's at 5.4, they're not even pre-diabetic. They can barely see where pre-diabetic

01:37:01.160 starts, let alone diabetic, but we give them an SGLT2 inhibitor.

01:37:05.560 I'm at 5.4. I can volunteer for this.

01:37:07.800 There you go. So you go from 5.4 down to 5.1, just on the basis of that drug. We throw a GLP-1

01:37:14.520 agonist on top of that. Now you're at 4.9. Then we give you rapamycin, which really doesn't impact

01:37:19.960 your glucose, but we think it's going to do something a little bit different. You would say

01:37:24.120 in that situation, you might believe that there's some actual

01:37:27.160 geroprotection, but not adding metformin. I don't want to lose muscle mass though.

01:37:31.160 And which of those drugs would you be most afraid of? The GLP-1 agonist or RAPA?

01:37:34.360 GLP-1 agonist. I think lean muscle mass is super important. It's probably better to have

01:37:40.520 high lean muscle mass and be a little bit more fat than it is to be low on both is my best guess.

01:37:45.640 Now, what do you make of the data that I talk about all the time, which look at the hazard

01:37:54.120 ratios for mortality based on high VO2 max and high muscle mass and high strength and how those

01:38:01.000 three things stand out so far above anything else? Meaning, when you look at hazard ratios associated

01:38:08.600 with smoking, type 2 diabetes, even cancer, they are not as lethal as being incredibly weak,

01:38:17.720 incredibly low in muscle mass, and incredibly low in fitness. How much causality do you think is there

01:38:24.120 versus how much of that is just, those are just such good markers of health?

01:38:28.040 GLP-1 I think there's causality there. I think it's super important, but it may only be important

01:38:31.640 for squaring the curve. I don't think there's much evidence that maximum lifespan is extended

01:38:36.040 by these things. We don't have the human data, of course. The animal data-

01:38:39.400 GLP-1 We can't do the experiment.

01:38:40.440 GLP-1 The animal data is pretty much with exercise says you square the curve. Again, I agree with you,

01:38:45.400 it's a revolution if we can do that. I'm not being negative about it, but I don't know about

01:38:50.280 maximum lifespan, whether there be an effect or not. I believe in so much that I put a lot of

01:38:54.840 effort in increasing my lean mass. That's why I started resistance training, because I wasn't

01:39:00.040 getting as much from running. I get more mindfulness from running, but I did a lot more resistance

01:39:04.760 training too, so.

01:39:05.640 I typically will tell patients that you should really think of exercise. I do actually think

01:39:11.960 it's reducing your risk of chronic disease, but then you still get into the whack-a-mole game.

01:39:15.880 If it lowers your risk of Alzheimer's disease, it might not have much of an impact on cancer risk,

01:39:20.120 it's unclear. But if it lowered your life expectancy by six months, it would still be worth

01:39:25.800 it based on the health span benefits that you get and the quality of life that you would enjoy,

01:39:30.200 especially in that final decade of life.

01:39:31.800 Dr. Probably. Let me come back to one point though,

01:39:33.960 because I'm a bit of a rant about this. Combining interventions. First of all, I will say two things

01:39:40.120 before I say what I'm going to say. One is that I believe we need to empower people to make decisions

01:39:46.600 on their own health. And so I support hackers. If they want to educate themselves and try different

01:39:51.400 things and they know what the benefits and risks might be and what we know and we don't know,

01:39:55.800 more power to them. I feel like part of the reasons we get such low compliance

01:40:00.200 in medications is that we don't empower people. We don't give them choices. They don't know why they're

01:40:06.280 doing things. We just tell them what to do and people don't respond well to that. Having said

01:40:10.600 that, I can't pick three interventions that work well together and a mouse. And we do these studies

01:40:17.080 all the time. They're more likely to cancel each other out than to have additive effects.

01:40:21.400 If you're taking 20 pills, it's like mixing 20 colors of paint together. You're going to get some

01:40:25.720 ugly gray outcome. Or at best, you're going to get an unknown outcome that we can't predict. So I'm

01:40:31.000 really cautious and want to tell people that there are a lot of people out there promoting,

01:40:35.480 doing a million different things at the same time. I try one or two things at the same time.

01:40:41.000 I try to see how my body responds. I measure things. Even simple measures are useful. I think that

01:40:47.400 if you're doing 10 things, you don't have any idea what's working and what's not working and whether

01:40:52.040 things might be impairing each other. And I really think that's a scary path to go down.

01:40:56.680 Now, what about, for example, how are you deciding to use the alpha-ketoglutarate and the NAD?

01:41:03.960 You've seen at least some evidence that each of those individually works.

01:41:08.120 Yeah. I've been using the rejuvant with the AKG for a long time. I was involved in the research.

01:41:12.920 I'm actually on the board of the company. It's something I've done and I've just taken for years.

01:41:17.400 And I add one thing to it and take it away. So I've tried astaxanthin. I've tried

01:41:22.600 fucoidin. I've tried rapamycin. I've tried a bunch of other things too. So I try to measure before and

01:41:29.080 after. I'm getting better at that. When I first started doing it, I wasn't measuring that much.

01:41:32.600 But that's the approach I take. I never take six things at one time. It's kind of intuitive. If

01:41:38.440 something's interesting to me, you see an effect on the mice, I want to try it. I want to do urolithin

01:41:43.960 and astaxanthin next. So if you had significantly more resources, so let's say you had a budget.

01:41:52.120 What is your current budget this year for both animal and human research?

01:41:55.240 It's complicated because we have multiple streams. I would say we probably spend about 4 million a

01:42:00.200 year. Okay. So if that number were multiplied by 10 or 20, 25, you had a hundred million dollar

01:42:08.120 annual budget to do world-changing translational gyroscience. What would you be doing different?

01:42:15.400 Scale would be one. When you do a combination in mice, you've got four groups. If we had more money,

01:42:21.240 we could design multifactorial clinical studies and preclinical studies where we're testing many

01:42:26.040 compounds at the same time. We're sort of doing nested groups and we could get an indication for

01:42:33.000 things that actually could be additive together, not just things that are working on their own.

01:42:37.560 Right now, we struggle to get to that next step for finances. And we could apply those to human

01:42:42.200 studies and combine it with lifestyle interventions too. The other thing we really believe is that when you

01:42:48.120 have a compound like urolithin, you're never going to really know what to combine it with unless you

01:42:54.040 know what the compound's doing. So we do a lot of discovery stuff now trying to figure out not what

01:42:59.800 pathway. If you take a drug like rapamycin, it affects every hallmark. That doesn't tell you

01:43:05.000 primary thing the drug is doing. In this case, we know it binds tor. Urolithin, we don't know. So we need

01:43:10.360 to know what that molecule is binding to in the cell. And if we understand the mechanism at that level,

01:43:15.960 we can combine it better with other interventions and start to understand how to put the puzzle

01:43:20.920 together of what we need to combine to get the biggest effect. Do you have enough human resources

01:43:28.120 to deploy that kind of capital if it were available? It would take a center to do it,

01:43:32.280 but yeah, we could build it. And Singapore is very motivated by the way. I have to give the

01:43:36.280 government credit. They've understood the aging problem before almost anybody. It's taken them a long

01:43:41.960 time to really figure out what to do about it. And they went through kind of early stages of putting

01:43:47.560 roofs on sidewalks. So people walk an extra hundred steps in the hot sun, you know, that sort of helps.

01:43:52.760 But now I think they're really motivated to commit to targeting healthspan. If they do,

01:43:58.760 it's the right place to be because it's a small island. It's a compliant population. They believe in

01:44:03.800 their government. What's the population?

01:44:04.920 5 million and then 1 million workers and 5 million people that live there permanently.

01:44:09.720 I think that it's a good place to really take studies, not just in the clinic, but move them

01:44:14.920 into the community and actually get validation in large populations. And that's why I like being in

01:44:20.520 Singapore. I think the opportunity is there to make it an example for how to do healthspan.

01:44:25.720 It's already very long lived, by the way. It's among the top three in the world,

01:44:29.640 depending on what statistics you want. But the healthspan, people still have morbidity there. They still

01:44:34.600 have 10, 12 years of sickness and decline. There's a lot of frailty there. You can see

01:44:39.640 that just walking around. So there's room for improvement for sure.

01:44:42.440 How will AI help in this field? Do you think it will allow for more intelligent experiments? Will

01:44:49.400 it allow for better signal detection in messy data? How do we unleash AI on this problem?

01:44:55.320 We're using it now already for the clocks and signal detection. We're using it to pick drugs now.

01:45:00.040 We just published a paper with Gehrig Fullen in Germany, where we're trying to

01:45:04.760 improve how to ask large language models, medical questions related to longevity. I talked to a lot

01:45:10.520 of doctors and they said that you shouldn't be asking perplexity, these questions. And I'm like,

01:45:14.040 I'm a realist. People are asking. So this figure out how to get the questions asked in the right way

01:45:19.080 to get the right answer. But I'm thinking more in terms of the research.

01:45:21.640 Now, yeah, I'm just giving you examples there. I think that what's going to happen next is AI is going to

01:45:26.520 start telling us what questions to ask. Right now, it's telling us how to analyze our data.

01:45:31.240 It's still not very good at telling us what the next question is. And I think that's the

01:45:35.080 threshold. How do we get it there?

01:45:36.280 It's not one that I'm qualified to answer. But I think that the question is,

01:45:40.520 the trajectory that it's on now is amazing. Is there a barrier to go to that next step? Or is it

01:45:46.600 just a matter of getting computational power and slightly modifying neural network algorithms or

01:45:52.520 something? And I don't know the answer to that. But I do think there's a reasonable chance that

01:45:57.000 I'm not going to be needed in 10 years. I think about this question a lot when it comes to

01:46:02.600 experimental topics, because I still don't have a good enough sense of how many training cases and

01:46:09.560 AI needs to learn this. We know for language what it took. We understand how many tokens were needed

01:46:17.240 to allow the neural networks to do what they do today. And it was enormous. So there are some

01:46:23.800 problems that might require far less input. You might be able to do it with 10,000 hours of data,

01:46:30.120 as opposed to billions of hours of data. To me, I think that's the question more than anything else.

01:46:36.200 I was wondering if you had a point of view on the answer.

01:46:38.440 Yeah, I don't really know. All I know is half my lab is doing it now. So I never would have guessed

01:46:43.640 that five years ago. Meaning they're trying to use AI to help them ask experimental questions.

01:46:49.000 Or interpret data. Okay. The latter I can understand.

01:46:51.400 Yeah, yeah. We've become half dry lab, which I never would have guessed in my lab.

01:46:55.480 Meaning that half the people just sit at computers and aren't doing experiments in animals.

01:46:59.720 Yeah. So what are you most excited about trying to uncover truth or high probability of truth

01:47:09.160 in the next five years? I think experimentally, at least pre-clinically,

01:47:12.680 we want to find interventions that really combine together to have synergistic impacts. I don't

01:47:17.800 think there's much out there. There's rapamycin and metformin and a couple other things from the

01:47:21.720 ITP, but they're small effects. Can we break through a barrier and get 50, 60% effects in mice by

01:47:27.960 combining things together? I think that's what pre-clinically we're excited about. And the other

01:47:32.600 thing I'm excited about is going back to this entropy question. Are there new classes of interventions

01:47:37.480 that change that primary linear accumulation of quote-unquote damage?

01:47:42.600 I'm also really excited now that I'm working with longevity clinics in various countries.

01:47:48.280 So we're applying the clock we built. We're also helping them try to decide which interventions to

01:47:54.360 do and hopefully collect data so it can be analyzed. Because I think there's so much going on. I tell

01:47:59.640 this joke all the time about getting scared to tell it more, but it used to be that yeast and worms and

01:48:04.280 flies were the model organisms for aging research, and now billionaires are the model organisms.

01:48:08.600 Because they're doing all kinds of stuff. I can't even test. I'm really curious to see what's

01:48:13.720 happening. Some of it, I guess, might work. Some of it, I guess, might not work, but we can't find

01:48:18.600 out any other way. And I don't want to see these clinics working in isolation and nobody's ever at

01:48:23.400 least learning what's coming out of the data, even if it's not perfect. I'm excited to work with these

01:48:28.040 clinics and go see what they're doing right now. That's maybe not the answer to your question, but

01:48:33.000 there's a lot of cutting edge stuff going on. I will pretty much work with people if they're doing

01:48:38.120 something I consider safe. And if they're honest about the data on efficacy, those are the two

01:48:43.640 things I ask, transparency and safety. And then I'm happy to try to interact.

01:48:48.440 What are you seeing that you're worried about? What trends do you see that people are doing from

01:48:53.080 a biohacking longevity standpoint that have you concerned? Either, I'll put this in two buckets,

01:48:58.680 the higher bucket would be safety. Second bucket would be predatory behavior around basically people

01:49:04.440 having their money wasted, even if the agents that are being sold are not necessarily harmful.

01:49:09.000 I think the first one, I'm excited about gene therapy. Don't get me wrong. I think it's

01:49:14.040 interesting and it may really change the field going forward. I even kind of like folistatin.

01:49:18.040 Folistatin as a protein.

01:49:19.480 Yeah. Yeah. But I think that those treatments are not very well proven yet and I would not do it.

01:49:25.240 You wouldn't spend $100,000 for folistatin gene therapy?

01:49:28.840 I probably don't have to spend the money and I still haven't done it. I've done MSCs though,

01:49:35.240 IV to try that. And I think stem cells, it's a different question there. I think if you're

01:49:40.840 repairing soft tissue damage or something like that and injecting them directly, it probably works.

01:49:45.800 For aging, I have no idea, but I think it's probably safe if you have somebody that's a

01:49:50.360 good practitioner that knows what they're doing. But I think there's a lot of, this is a problem with

01:49:55.800 stem cells. You go places and you really don't know who you're working with. And if they're

01:49:59.960 really treating the cells correctly, if you're putting the right things in your system. And so

01:50:04.200 there's a safety concern there based on the practitioner, I think. So those are things that

01:50:08.920 concern me. I haven't gotten totally on board with growth hormone yet. I think probably used correctly

01:50:14.040 might be okay. The data is interesting.

01:50:16.920 Are you aware of human data that, and I'm in your camp by the way, which is, I actually had a

01:50:22.200 bunch of friends over for dinner last night and this came up and I said, look, I can't point to

01:50:26.520 a study that tells you this is a bad idea. And I've never spoken to a person who takes

01:50:32.680 a modest judicious dose of growth hormone who doesn't tell me they feel better.

01:50:36.600 Yeah, I agree.

01:50:37.240 So it's hard to believe it's not making people feel better.

01:50:40.680 I also have never seen data to suggest it initiates cancer, but it seems very biologically

01:50:48.680 plausible that if you have cancer, small amounts of cancer, your probability that this becomes

01:50:53.960 clinically significant is higher in that camp.

01:50:56.280 Seems like it. I agree.

01:50:57.400 Again, I say that with no data.

01:50:58.600 Yeah, me too.

01:50:59.240 And so my view has just been, despite my own interest in trying things, I've left the

01:51:05.080 growth hormone one off the list. Is that an answerable question, do you think?

01:51:09.160 I think we could do clinical studies. Some are being done.

01:51:12.120 How would we address the safety concern? You really need to be able to try to track people

01:51:16.360 for quite a long period of time who are cancer susceptible.

01:51:19.320 In all of these things, we don't know what the long-term is. And I guess it's a gray area.

01:51:24.600 This is a weird thing with these clinics, because I think from what I read in your book,

01:51:29.880 you're doing sort of validated stuff. You're not really out there in the stratosphere doing

01:51:34.840 crazy stuff that we don't know about.

01:51:36.920 I would like to think I'm not, but I'll tell you, there are people who are very critical of my use

01:51:41.000 of rapamycin in patients for geroprotective reasons. There are people who might think I'm crazy

01:51:45.960 for giving people SGLT2 inhibitors who don't have diabetes. So I think there's things that we do

01:51:51.240 not for all of our patients. Fewer than 10% of our patients take rapamycin because my view is,

01:51:57.560 unless you're willing to have a very lengthy discussion about the pros, the cons, the risks,

01:52:03.240 the uncertainties, and I don't give people an answer that says, oh, this stuff's amazing. My answer is,

01:52:08.600 I don't know. Here's how I think about it probabilistically. Here are the trade-offs.

01:52:12.920 You can tell I'm not a good salesman if only 10% of the patients are taking it.

01:52:16.680 I think that's perfectly reasonable. They're cutting stewing some really out there stuff.

01:52:21.800 How do we know the long-term safety on it? I think if you're going to go do that stuff,

01:52:25.640 you need to go in with your eyes open. You're taking a risk.

01:52:27.960 I've just seen so many horror stories of people that have come back from,

01:52:31.160 because you can't do this, a lot of the stuff you can't even do in the United States. So they're

01:52:33.960 coming back from South America or Mexico, places in Asia, having done folistatin therapy or other

01:52:40.360 very questionable stem cell therapies. And I mean, people that have had horrible

01:52:44.280 infections, literally just artifacts of the treatment.

01:52:47.080 And that's a practitioner problem with this.

01:52:49.160 Yeah, exactly.

01:52:49.800 There's some great stuff happening too. I work with Bumrun Grad Hospital in Thailand,

01:52:53.880 and they've got a longevity clinic now, and they're very grounded in good science. And so

01:52:58.120 the problem is if you're a consumer for these products, it's really hard to know.

01:53:02.840 Someone listening to us who's saying, guys, can you give me some rules of thumb,

01:53:10.040 some heuristics for navigating the never-ending landscape of longevity hacks that keep showing

01:53:18.520 up on my Instagram feed, my TikTok feed, and at cocktail parties?

01:53:22.360 That could be diet books and not just going to clinics around the world. I think it's really

01:53:27.160 difficult to sort that out. What I was going to say is it's clinical practice and research at the

01:53:31.560 same time. It's a very unique situation, right? And there aren't many examples of that that I know

01:53:37.400 of that are really maybe some functional medicine is a little bit like that too, but it's interesting.

01:53:42.120 And I feel like it's better for scientists to engage with where it's possible to engage with

01:53:48.280 these clinics and try to help them than it is to just let people do things. If you can provide

01:53:53.960 oversight that's helpful, you should be doing it. That's kind of how I feel about it. A lot of

01:53:58.280 academics don't even want to work with these clinics at all. I get criticized for working

01:54:02.520 with them sometimes. So it's an interesting world right now.

01:54:05.000 Well, there are a lot of things we talked about today, Brian, that I can't wait to

01:54:09.720 follow up on myself. So I'm really looking forward to the NHANES second-gen clock paper. That'll be

01:54:14.840 interesting. Again, my personal curiosity there will be, is that clock providing value over all the

01:54:20.440 other data we have? My intuition is it won't, but the fact that we now at least have a clock that can

01:54:25.720 outperform chronologic age is a step in the right direction. It'd be very interesting to see some of

01:54:31.000 the data that you've talked about as far as alpha-ketoglutarate. You've also piqued my curiosity

01:54:35.720 with the sublingual NAD. And so that's really interesting. Have you been able to measure NAD

01:54:41.320 levels in your RBCs? I haven't done it myself. There have been some studies done.

01:54:45.480 On that exact supplement? Yes. I don't think anything is published. It's done by the company, but yeah.

01:54:51.160 And I do like this idea of taking ITP winners and combining them and seeing if we can get,

01:54:59.240 I mean, if you could get accretive value, I mean, that would be remarkable. But even if you could

01:55:04.440 just get additive benefit, it would be pretty amazing. Yeah. I think so too right now. I think

01:55:08.440 that's still a hangup for the field. Yeah. Well, this was super interesting. Brian,

01:55:13.240 I really appreciate you making the time to come out here. I know that being on a plane for that long

01:55:16.920 is not always fun, but. That's where I live anyway, so it's fine. Thank you. Thanks a lot.

01:55:22.200 Thank you for listening to this week's episode of The Drive. Head over to peteratiamd.com

01:55:28.440 forward slash show notes if you want to dig deeper into this episode. You can also find me on YouTube,

01:55:34.840 Instagram, and Twitter, all with the handle peteratiamd. You can also leave us review on Apple

01:55:40.920 podcasts or whatever podcast player you use. This podcast is for general informational purposes only

01:55:47.400 and does not constitute the practice of medicine, nursing, or other professional healthcare services,

01:55:51.880 including the giving of medical advice. No doctor patient relationship is formed. The use of this

01:55:57.960 information and the materials linked to this podcast is at the user's own risk. The content on this

01:56:04.040 podcast is not intended to be a substitute for professional medical advice, diagnosis, or treatment.

01:56:09.240 Users should not disregard or delay in obtaining medical advice from any medical condition they have,

01:56:14.760 and they should seek the assistance of their healthcare professionals for any such conditions.

01:56:19.640 Finally, I take all conflicts of interest very seriously. For all of my disclosures and the companies I

01:56:25.400 invest in or advise, please visit peteratiamd.com forward slash about where I keep an up-to-date

01:56:32.680 and active list of all disclosures.

01:56:39.480 Thank you.

The Peter Attia Drive - July 21, 2025

#357 ‒ A new era of longevity science： models of aging, human trials of rapamycin, biological clocks, promising compounds, and lifestyle interventions ｜ Brian Kennedy, Ph.D.

Episode Stats

Length

Words per Minute

Word Count

Sentence Count

Misogynist Sentences

Hate Speech Sentences

Summary

Transcript