The Peter Attia Drive - #204 - Centenarians, metformin, and longevity ｜ Nir Barzilai, M.D.

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

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

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

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

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

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

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

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

00:00:47.740 here's today's episode. My guest this week is Nir Barzlai. Nir is making his third appearance on

00:00:54.840 the podcast, the previous one being in August, 2020 with Joan Manik, and then originally back

00:01:00.420 in January of 2019. In this episode, Nir and I speak mainly about two topics, just in much,

00:01:07.320 much more detail than we've ever spoken about them before, at least publicly, centenarians and

00:01:11.800 metformin. We start the conversation speaking about centenarians with a focus on what can the majority

00:01:17.340 of us who are not centenarians learn from them. We talk about longevity genes such as GH, IGF-1,

00:01:24.040 CTEP, FOXO, TSHR, and ApoE. We talk about whether or not environment matters at all in these

00:01:30.840 individuals or whether it's all genetic. We talk about what we can learn about them from the

00:01:34.620 importance of preventing diseases. And we talk about what we can learn from centenarians around

00:01:39.140 extending lifespan while also trying to improve healthspan. From there, we get into a deep dive on

00:01:44.160 metformin. We talk about the TAME trial. Now, this is something that we did speak about briefly in our

00:01:48.460 first podcast, but we get into much more detail. And I actually found myself learning some details

00:01:52.480 of the study design that I didn't understand previously. Talk about Nir's thoughts on why the

00:01:57.560 Rich Miller ITP program found metformin to be unsuccessful in that model and why he thinks

00:02:03.040 that may or may not apply to humans. We talk about the impact metformin can have on exercise,

00:02:08.080 both strength training and cardiovascular training. Lastly, we speak a little bit about

00:02:13.420 epigenetic clocks and end with a conversation around NAD precursors. As a reminder, Nir is a

00:02:19.880 director of the Institute for Aging Research at Albert Einstein College of Medicine, spearheading

00:02:24.520 the Longevity Genes Project, conducting genetic research on more than 500 healthy elderly people

00:02:30.920 between the age of 95 and 112 and on their offspring. He is also the director of the Paul F. Glenn Center

00:02:37.520 for the Biology of Human Aging Research and of the National Institutes of Health, Nathan Schock

00:02:43.340 Centers for Excellence in the Basic Biology of Aging. So without further delay, please enjoy my

00:02:48.680 conversation with Dr. Nir Barzalai.

00:02:56.240 Hey Nir, it's great to have you back. I was thinking about this when I was preparing for the

00:03:00.760 podcast today. There's so much I want to cover that I don't think it's actually going to be possible.

00:03:05.940 I'm pretty sure that we're going to talk as much as we talk today. And I'm going to be saying to the

00:03:10.740 team, all right, let's talk about when we're going to have Nir back, because there's really just too

00:03:14.960 many things I want to go through. So anyway, thank you for making time. And let's just get right into

00:03:20.260 it. Thanks, Peter. I'm happy to come back. But it's you who's coming to me every week. And I'm

00:03:25.880 so grateful for what you're doing for this field and for helping all of us catching this field of

00:03:32.460 longevity that's going to come true very rapidly, I hope.

00:03:36.480 Well, just thinking about the first place I wanted to start, and there's really no good

00:03:40.240 one place to start because there's just so much I want to talk about. But let's start with

00:03:44.100 centenarians. You, along with Dr. Pearls, are probably two of the people who have spent the most

00:03:50.340 time studying this very, very unique subset of the population. So I think everybody knows what a

00:03:58.420 centenarian is, someone who lives to be a hundred or more. But there's so much nuance about what it is

00:04:05.560 about these special people. And then there's sort of the pop culture view of this, which is people

00:04:10.380 love to talk about all of the bad behaviors that centenarians engage in, how much more they smoke,

00:04:16.020 how much less they exercise, how much whiskey they drink, and all of those things, which are really

00:04:20.820 cute. But when you study them scientifically, and when you study their offspring scientifically,

00:04:25.940 as you've both done, we learn a lot of things. And if my interpretation of the literature

00:04:32.140 is at least partially correct, it appears that genes play a significant role. So genes don't seem

00:04:38.680 to play a big role in people living to 70 versus 80. But boy, when you start to talk about living to

00:04:44.080 90 versus 100 relative to 70 or 80, genes play a pretty big role. So tell me a little bit about what

00:04:51.280 we understand about the role that the parents play in determining the lifespan of offsprings.

00:04:59.140 How fortunate were these people to pick their parents?

00:05:02.000 So let me just tackle one of the things you said, that there's no much genetic impact in people

00:05:07.940 between 70 and 80. And it's true if you compare the lifespan of fathers and sons, okay, or mothers 0.52

00:05:17.980 and daughters or sons. And let me tell you why it's problematic. My grandfather got a heart attack when he

00:05:24.980 was 68. And he died. That's my grandfather. My father got a heart attack at 68. And he had triple bypass

00:05:32.920 and he died at 84. So the correlation between age of death in different cohorts is not much revealing.

00:05:41.820 But let me say it now differently. Let's say it's 20%. If we understand this 20%, understand it really,

00:05:51.160 we can use that in order to prevent the 80% of the environment.

00:05:57.080 Well, maybe another way to think about it, Nir, because that example is a great example,

00:06:00.260 which makes it very difficult over discrete generations to make a comparison. Do we have

00:06:05.960 twin data? Because it seems to me that if you had monozygotic twin data, that would be the gold

00:06:14.020 standard for looking at the discordance and concordance between the role of genes in separate

00:06:18.700 environments, right? You would think so. So let me tell you the problems with twins. Twins are usually

00:06:25.800 born small for their gestational age. In fact, it's more true that one of the twins is small for their

00:06:35.420 gestational age. Now, I've been doing studies with rats from before. When you ligate the uterine

00:06:43.180 artery and make them small, they get diabetes, which they never get at three months. We know that twins

00:06:51.580 or that babies that are born small for age develop age-related disease very rapidly. It's called the

00:06:58.060 Barger hypothesis. It's observation from Holland in World War II. And we actually determined

00:07:05.100 some of the epigenetic manifestation of what happens epigenetically when you do that. So I don't think

00:07:12.660 twins are the right model unless you understand that and account for that. Nir, this is super

00:07:19.840 interesting. Can you tell me a little bit more about that? I actually was not aware of the relationship

00:07:25.580 between low birth weight and the epigenetic imprint of that on reduced lifespan. And I assume health

00:07:32.960 span or is it just lifespan? Well, it's mainly health span in humans that we know. First thing that's

00:07:38.860 obvious, when you have a small for gestational age twin, there's the catch-up growth of the small baby.

00:07:47.900 And those twins born in the same day, in few years, one of them is an obese child and one is a normal

00:07:55.920 child. And as you know, obesity drives aging very rapidly. So that's one mechanism. The changes in

00:08:03.920 imprinting of epigenetic, I would say for now that it's more of a description than a mechanism.

00:08:11.900 There are many genes that are involved and I'm not aware of a recent paper that says this is what

00:08:18.420 happens. Okay. So let's get back to the broader question, which is when an individual or when a

00:08:26.480 cohort of individuals lives to 100 and we compare them with a cohort of individuals that lives to 80,

00:08:34.700 what are the types of genes that seem to be offering protection to that group that lives to 100? What is it

00:08:44.540 that the centenarians have in a polygenic sense that the rest of us schmucks don't have?

00:08:51.480 When we went to the centenarians, we had three hypotheses that we had to take care of. One is 0.57

00:08:57.580 that it's all the environment. Okay. It happens that they did exactly the right thing, what the

00:09:03.540 doctors tell us to do now. The second, and it's not true, as you mentioned, it's not true. 60% of the

00:09:09.600 men are smoking and 30% of the women, 50% of them are overweight and obese and older and not exercising 0.99

00:09:16.600 and not vegetarians. The second hypothesis is that they have perfect genome. We know that we have a

00:09:24.300 lot of genotypes that are putting us at risk for variety of age-related disease. So maybe one out of

00:09:31.180 10,000 doesn't have that. And that's why they're flying in so gracefully.

00:09:35.360 So to be clear, Nir, part of that hypothesis is the absence of bad genes, not necessarily the

00:09:42.560 presence of good genes. Exactly. That only the absence of bad genes will allow them just to get

00:09:48.320 without diseases. And sorry, I took you off your track, but what was the third hypothesis?

00:09:53.400 The third hypothesis is that there are genes that slows their aging. Longevity genes, we call them.

00:09:58.980 Okay. Fair enough. As for the second hypothesis, that they have what we call the perfect genome, 0.86

00:10:04.520 we took our first 44 centenarians and did whole genome sequencing at the time. Huge expense. But we

00:10:13.680 only had those centenarians. We don't have had the control. But we had a great instrument, we thought.

00:10:22.220 It's called CleanVar. It's an accumulation of all the genes that have shown to be causing diseases.

00:10:31.120 If you had a clean variant, you're very likely to have a disease. So we simply asked,

00:10:38.660 do our centenarians have any of those variants?

00:10:42.660 And how many of those variants are there?

00:10:44.940 At the time, there were 15,000. Now there are many more.

00:10:49.060 So just to put that in perspective for the listener, we have between 20,000 and 30,000 coding

00:10:54.180 genes, correct? So these are variants of how many genes?

00:10:58.040 Well, I don't remember how many genes are in the variants, but those are variants that were found

00:11:05.440 to be compelling. And by the way, a lot of them are not. That's another story. So let's keep it

00:11:11.680 simple. We had 15,000 variants. And we asked, do our 44 centenarians have variants? And the answer was,

00:11:19.540 each centenarian had between five and six bad variants.

00:11:24.580 Five and six out of 15,000 possible?

00:11:27.080 Right.

00:11:28.140 And we didn't have a control, so we don't know how many the average person had.

00:11:32.360 Right. But think of it. Those centenarians, each one of had five variants that will probably

00:11:38.380 cause a disease. And none of them had it. And if you're asking, are those variants important?

00:11:45.960 Well, we have two centenarians who have the ApoE4 homozygosity that puts them at major risk,

00:11:53.780 one of the best genetic risks for Alzheimer's. That the textbook says they would be demented at 70

00:12:01.700 and dead at 80. And they're at 100 and not demented. Genes for Parkinson's, for cancers,

00:12:07.520 for other diseases. So basically, the centenarians don't have the perfect genome, which left us

00:12:15.660 with finding genes... That are protective.

00:12:19.740 Right. That slows their aging, are protectives even against genes that are thought, at least,

00:12:27.220 as I'm saying, it's probably not totally true, thought to most probably cause a disease.

00:12:33.540 And tell me, Nir, approximately what year did you arrive at that conclusion?

00:12:37.580 This paper is more than 10 years old, I think. We had several papers since then confirming this.

00:12:45.840 And look, like always, there is a decrease also in bad genotypes in centenarians, some decrease.

00:12:53.940 But really, the majority of the study shows that it's not that. It's not the perfect genome.

00:13:00.760 It's something else.

00:13:02.100 So we've very easily eliminated hypothesis one, which is centenarians live to 100 because of what

00:13:09.560 they do, their behaviors, their environment. You now make a very compelling case that it's not number

00:13:15.560 two either. It's not that they lack any disease-driving genes. So it is, in fact, this third hypothesis.

00:13:25.300 When did you actually demonstrate that? That's obviously a much harder one to demonstrate because

00:13:30.640 you probably have a far smaller library of disease-sparing variants as opposed to disease-causing

00:13:38.000 variants. So when did you start to arrive on what some of those variants were?

00:13:42.140 The story was interesting because, remember, we started the study in 1998. And there were other

00:13:48.880 parts of the studies that I didn't talk with you about. But one of the things we had is we start to

00:13:54.900 establish the phenotype. And one of the phenotype that came up is high level of HDL cholesterol.

00:14:03.080 Actually, very high level of HDL cholesterol that was more obvious in the offspring of our

00:14:09.200 centenarians even than our centenarians. So we have offsprings that have HDL cholesterol 130, 140,

00:14:16.720 folds higher than it should be.

00:14:19.400 And what we could do when the methods were poor and we were poor, we were going about genes that are

00:14:25.740 involved in those phenotypes. And we actually got a very compelling data on two genotypes that seems

00:14:33.700 to be functional, important, that are controlling lipid metabolism. One of them was a CTP genotype and

00:14:42.620 one of them was an APOS C3 genotypes. And those genotypes increased from about 8 to 9 percent of the

00:14:51.380 homozygosity in control to almost 20 percent in centenarians. And it's not only that. Look, when you

00:15:00.580 have people of all ages, unrelated, you can look at the trend of the genotypes. If the genotype is

00:15:09.320 killing you, then as you go closer to 80, 90 and above, those genotypes will decrease. And if they

00:15:17.540 are going up, and by the way, the slope that's going up is a very important statistical tool, then

00:15:24.220 they are very likely to be longevity genotypes. And this is what we found. And it's so interesting.

00:15:31.700 It was our initial discovery. And drug companies were at our doorsteps immediately, not because

00:15:39.140 they're interested in aging, but they said, just a minute, if we make a good drug, okay, if the drug

00:15:44.320 is really good, it targets exactly that without side effects, then we have safety because those guys

00:15:52.660 for 100 years had, in both cases, by the way, suppression of the expression of those genes. It must

00:15:59.900 be safe. So let's develop the drug. And isn't it amazing how big the graveyard is of CTEP inhibitors?

00:16:07.480 Yes. And I don't totally understand it. I listened to one of your podcasts. I forgot the name.

00:16:15.780 Yeah. Tom Dayspring and I have discussed this in great detail. I think a lot of it comes down to

00:16:20.260 not understanding the biology of HDL. I mean, the biology of LDL is relatively straightforward. The

00:16:25.260 biology of HDL, I think it's safe to say we don't understand at all. I mean, that would be putting it

00:16:30.600 mildly. And I think the challenge with the CTEP inhibitors is they raised HDL cholesterol. In one

00:16:36.920 sense, they reproduced the phenotype in its most crude sense. But you can think of it very

00:16:43.560 simplistically, I think, which is how do you raise HDL cholesterol? Do you raise it by putting more

00:16:48.640 cholesterol into HDL? Do you raise it by impairing HDL from conducting reverse cholesterol transport and

00:16:56.420 getting rid of cholesterol? Those are two very different approaches to raising HDL. And when you

00:17:02.980 look at the centenarians and examine their phenotype, you don't really know what it is. You would speculate

00:17:09.700 that they have better HDL function. But the reality of it is we have other phenotypes that exist in nature.

00:17:16.380 I'm not sure if you're aware of this, but I've seen a number of papers that examine people with

00:17:21.540 HDL cholesterol that's very elevated who have very advanced atherosclerosis. And in fact, the elevated

00:17:27.760 level of HDL cholesterol they have suggests impaired HDL function and impaired reverse cholesterol

00:17:34.980 transport. And so I think that's probably the issue around why a lot of these drugs have failed. And it

00:17:42.060 obviously speaks to the humility which all of us need to be able to examine these phenotypes. And

00:17:48.880 it's a clever way to go about doing it. A very clever thing is look for phenotypes that are different

00:17:53.440 and work backwards to find genotypes. Right. So let me add two points to the trade-offs of this pathway

00:17:59.820 are amazing. On one hand, you don't clear cholesterol. On the other hand, you have high HDL. But for me,

00:18:05.680 it's not about the HDL because all the particle size are significantly bigger. So you could say it's all about

00:18:13.860 high dense LDL, right? That's the phenotype that we're depicting. So I think this part is important. The other

00:18:22.000 thing that was really striking, I told you about those two centenarians with APOE4 genotypes. They both had

00:18:30.580 very high HDL cholesterol and they were homozygous for CETP. So it is possible. In fact, the major

00:18:38.540 phenotype for us of the HDL wasn't cardiovascular. It was cognitive function. So when I'm saying

00:18:46.740 cognitive function, maybe we're talking about physiology. We are thinking of this physiology

00:18:51.380 from a heart perspective. And maybe there's a physiology of that from a brain perspective that

00:18:56.660 we don't totally understand. And by the way, I did ask Merck. I said to Merck, do cognitive function.

00:19:03.900 Okay. And they did. But the people they got to the study were between 50 and 70 years old.

00:19:10.620 So there's no results for that.

00:19:13.280 I know this off the top of your head, but even offline, I'd love to hear what you learned about

00:19:17.160 the variants of clotho that those people had, especially the ones who were homozygous for APOE4.

00:19:25.100 You're probably aware, but there are variants of clotho that seem to completely abrogate the

00:19:32.100 effect of E4. Meaning you take people with APOE4, either hetero or homozygous. And if they have this

00:19:40.400 particular variant of clotho, they behave as though they are E3. This is so interesting. So we published

00:19:46.980 on clotho and clotho was the example of what we call a V-shape or a U-shape genotype.

00:19:53.760 Clotho basically seemed to kill 50% of our subjects by age 85. The genotype has disappeared

00:20:04.820 by age 85. And all of a sudden, after that, at age 100, it was the same. And our interpretation

00:20:14.220 was that those centenarians were born with clotho, but they were also born with longevity

00:20:19.760 genes, which made the clotho not significant. But another hypothesis is that the actual clotho

00:20:26.880 that they had is a clotho that was protective in another mechanism.

00:20:31.700 Yeah. So what about some of these other genes that have now come to light? So FOXO, for example,

00:20:39.360 tell us a little bit about FOXO. How did you arrive at that? What was the phenotype that

00:20:44.220 tipped you in that direction? Or did you arrive at FOXO through a pure genetic analysis?

00:20:50.800 Well, I didn't come up with FOXO. FOXO came out from Japan and Okinawa.

00:20:55.340 I'm using you in the very liberal sense, Nir. I'm giving you and the field credit. Yes.

00:21:01.740 And me is not only me, right? It's a big team. And in fact, the research now has changed very much.

00:21:07.300 The teams are really teams because you need the computational and the AI people and the doctors

00:21:12.520 and the physiology. So it's we. But let me start it differently. Because why did I start centenarians?

00:21:20.140 I started centenarians because all of a sudden in the mid-90s, it became apparent that you change

00:21:27.140 just one gene in a nematode and they can live 10 times longer. By the way, the gene was bugging me

00:21:34.440 because it's the insulin receptor or the IGF insulin receptor gene. And the nematodes were insulin

00:21:41.000 resistant. And they also had abdominal obesity. They accumulated fat in their intestinal cell. It

00:21:47.600 wasn't the right example, but the concept was right. And so I started the centenarians. I told

00:21:52.840 you about the first genes that I saw. And then I wrote a grant. Most of the grants that I wrote,

00:21:58.280 the hypothesis ended up being wrong, though we found the right explanations mostly.

00:22:05.200 There was one grant that I wrote and I said, I'm writing this hypothesis knowing it's wrong. And that

00:22:11.740 is that growth hormone IGF signaling pathway have anything to do with human longevity. I said,

00:22:17.860 it's great. We have the nematodes. We have this and that. But this is wrong. And I was totally wrong

00:22:23.540 about it. And I'm telling you the bottom line. 60% of our centenarians have genes that impairs

00:22:32.420 growth hormone IGF signaling pathway, including the FOXO3A. 60%. It's the most common genotype. It's

00:22:42.720 not only genotypes, by the way. It's microRNA. It's genotypes on the IGF receptor. It's deletions

00:22:49.100 of the growth hormone receptor. There's lots of ways to get to this. But this is very common.

00:22:55.340 And sorry, just to double click on that a little bit, Nir, I think it's worth maybe giving people

00:22:59.320 a little bit of a primer on the relationship between growth hormone and IGF-1. And if you want,

00:23:06.820 we can talk about IGF-BP3. Okay. So I said growth hormone IGF, and I'll stand by picking those two.

00:23:14.280 But for growth, there are actually hundreds of genes. Okay. There are hundreds of genes. If you only look

00:23:20.140 at what determines height, there's hundreds of genes that are determining heights. But the growth hormone

00:23:27.620 is the growth hormone pathway or the growth hormone signaling pathway that really comes from the

00:23:33.860 pituitary, controlled by the hypothalamus, but comes from the pituitary and makes you grow when you need

00:23:41.420 to grow. And it has its own actions, but also it has a specific action of binding to its receptor in

00:23:50.880 liver and releasing the second important growth hormone, which is called IGF-1. It stands for

00:23:57.040 insulin growth factor 1. And so those are the hormones that we're talking about. The advantage

00:24:03.460 of IGF-1 is that it's measurable. Growth hormone is measurable in provocation and in young people

00:24:13.060 and not much in old people. So IGF is kind of the biomarker for this action. IGF itself is really

00:24:21.280 complicated because there are five. Actually, I hear that there are seven binding proteins and there's

00:24:27.240 a lot of regulation in between. But it's true to say that levels of IGF are really very good biomarkers,

00:24:35.840 even if you don't agree that they're causative for variety of health outcomes. And I want to say

00:24:41.960 another thing that in nature, the dwarfs are doing better as far as longevity. The little dogs are doing 1.00

00:24:48.960 better. The ponies are living longer. And in animals, everyone, no matter how you interfere,

00:24:55.640 you're getting better longevity. Even the Laron dwarfs that have deletion of the growth hormone

00:25:02.500 receptor, we don't know if they live longer, but at least they have less age-related diseases like

00:25:08.880 cancer or diabetes. But they have more alcoholism and suicide. And as you said, it's not clear that they

00:25:15.980 live any longer. They just seem to be spared of some of the chronic diseases in exchange for others,

00:25:21.820 correct?

00:25:22.660 Right. I'm always describing it that they're still unhappy to be short. They drink, they cross the road,

00:25:29.020 and nobody sees them when they're drunk, and they're being run over, right? It's not natural.

00:25:35.200 But let me tell you a much more optimistic story about that. We discovered, and when I say we,

00:25:42.260 it's Gil Etzmon, who was a fellow and then a faculty with us, he's still involved, he's now

00:25:47.620 in Israel. But he, we looked at all the pathway, and we discovered that our centenarians have deletion

00:25:55.960 of exome 3 in the growth hormone receptor. And Gil came to me and showed also the trend, okay? And as I

00:26:05.120 said, the trend is important. So the homozygosity was 3% in our population and 12% in centenarians.

00:26:12.260 And I said, Gil, what's their IGF-1 level? And he showed me it's significantly lower. I said,

00:26:20.280 terrific. What was their maximal height? And he said, that's the problem. And he shows me

00:26:25.700 they're significantly taller than the rest of the people, two, three inches taller.

00:26:31.560 Wait, the people that were homozygos were taller than the controls? 1.00

00:26:36.260 Yes. Significantly taller.

00:26:39.180 And two to three inches is not subtle. What was the relative Z-score difference between them in IGF?

00:26:45.380 I don't remember. I'll get you the reference. It was like a 20% decrease.

00:26:50.680 Okay. And what about their IGF-binding proteins?

00:26:53.880 There is nothing special there.

00:26:55.540 No major difference. Okay.

00:26:57.160 No. But you're on the right track. So I said to Gil, what do you do with genetics? First of all,

00:27:03.120 you need to do validation. The paper is not accepted without validation. I said, do validation.

00:27:09.080 So Gil got three other studies from all over the world, actually. And all of them were showing the

00:27:16.060 same thing. The oldest old have false difference in their growth hormone in this deletion, which was

00:27:24.440 very compelling on its own. But still, why do they have lower IGF-1?

00:27:29.780 Sorry, just to be clear, they had lower growth hormone. They should have lower IGF-1. You're saying,

00:27:35.100 but they were taller.

00:27:36.320 Why they were taller? Sorry. I meant why they were taller.

00:27:39.140 Yes. Yeah.

00:27:39.840 So Hasey Cohen, who's the dean of the gerontology school at USC and one of my collaborator,

00:27:46.380 is a growth hormone expert. And he took the cells. And by the way, that's another thing that we have.

00:27:52.680 We have the lymphoblasts of our centenarians. So if you have a genotype, you can actually look at

00:27:58.740 its action with lymphoblasts. And he took this lymphoblasts and he incubated them with and without

00:28:05.180 growth hormone. And something really interesting happened. When they were not incubated with growth

00:28:10.540 hormone, for relation of this receptor, the activity of this receptor was lower, as we thought it would

00:28:17.500 be because they have deletion of an exome. But when he incubated with growth hormone, it was almost

00:28:24.120 like an amplifying switch. They were phosphorylated three times as much. The same was with

00:28:32.580 proliferation. Proliferation was lower and with growth hormone was increased. So what we understood that

00:28:40.500 although we don't understand this switch mechanism, but when they go through puberty, they are activating

00:28:49.480 growth hormone. They're very sensitive. Yeah. They're sensitive. They grow taller. And once their

00:28:56.080 growth hormone decrease after puberty, they are tall, but their IGF stays low for the rest of their lives.

00:29:03.080 Very interesting. What fraction of centenarians share this genotype?

00:29:11.380 12%.

00:29:12.060 Okay. I want to talk more about the genotypes and I'm going to come back to a question around this.

00:29:17.580 So hopefully I'll remember. So this is now another big major axis is the genotype cluster around GH.

00:29:25.220 We've already addressed the genotype cluster around CTEP. Let's talk about FOXO. Let's talk about

00:29:32.060 TSHR. Let's talk about some of the others. We recently published a paper because we didn't

00:29:38.740 understand so well why the literature is so confusing us with the growth hormone and IGF.

00:29:45.440 We went to the UK Biobank, which has really changed our ability to validate and to learn and to get

00:29:53.940 hypotheses. They have 440,000 people who have actually IGF-1 measurements. We looked

00:30:01.740 at the young people that had high IGF-1 level and we saw that for young people, high IGF-1 was

00:30:11.400 protective from variety of age-related diseases and from mortality, although not from cancer.

00:30:19.920 Yeah. I was just about to say it's a very complex relationship. IGF seems to protect from everything

00:30:25.660 but cancer, right? Right. Except that, by the way, I'll get to cancer in a second. On the other hand,

00:30:32.880 people over the age of 60, it's exactly the opposite. They had more of every age-related

00:30:39.680 disease except cancer, and they also had increase in mortality. Sorry, was that a linear relationship?

00:30:47.480 Totally linear. It was. Okay.

00:30:49.420 Totally linear relationship. What I'm describing to you is what we call the antagonistic pleiotropy

00:30:56.420 hypothesis of aging. The things that are good for you when you're young can turn against you when

00:31:02.680 you're old. Or in this case, yeah, in order to do reproduction and to do evolution, you need a lot of

00:31:11.080 growth hormone to get there. But after that, you have to switch the energy because now you're going to

00:31:17.620 have breakdown. It doesn't make sense for you to expand growth in any way. And I think it was

00:31:24.400 beautifully demonstrated, and it also explained the confusion in the literature. It depends if you

00:31:30.720 looked at young people or old people or took care of that at all. One thing that's confusing about

00:31:36.400 that pleiotrophic relationship is that the cutoff is quite old. So if you really think about this just

00:31:43.080 through kind of a Dawkins lens, you would think that that cutoff would be done by 25 or 30. At that

00:31:51.460 point, evolution is done with you. You've served your purpose. Now, anything that you get after that

00:31:57.940 is gravy and really not under the purview of evolution. Evolution sort of stopped caring about

00:32:02.800 you. Unless you buy the argument, and I think this is an argument, that there's an evolutionary benefit

00:32:07.760 to you being a caregiver beyond your reproductive capacity. So maybe there's something to be said

00:32:13.000 for that. But are you surprised by how late in life we see that switch flip? Look, you made this

00:32:19.540 argument. By the way, I would say I believe in the grandparent theory. But still, if you had your kid

00:32:25.740 and died the next year, the next day, or survived for 100 more years, I think it's too late for

00:32:33.380 evolution. Number one. Number two, one of the most interesting thing that we show, and it's true

00:32:38.820 around the world, people with longevity, with exceptional longevity, have less offspring. From

00:32:44.800 an evolutionary point of view, you should be losing longevity genes. By the way, anybody with kids will

00:32:52.140 immediately find that completely intuitive. They just suck all the life out of you. Especially in

00:32:59.760 COVID times. They're incredible. I love them more than anything. But I believe that they are indeed

00:33:05.940 shortening my life on some level. I think there's hormesis. They'll come back and you'll be more

00:33:11.860 resilient. It's okay. But I actually wrote a paper saying people in the Old Testament are quoting

00:33:20.080 Methuselah to be 969 and Moses to be 120. Maybe they were right. Maybe we had this capacity because we're 0.91

00:33:27.620 losing longevity genes because of reproduction. That's really interesting. Has there ever been

00:33:33.440 a serious study about the biblical stories and if there's any way to assess any validity to some of

00:33:41.760 the age-related claims that were made merely 2,000 years ago? You know why it's a problem? Because

00:33:48.000 the Orthodox people believe that every word in the Bible is true. You ask them, and I ask many,

00:33:55.300 believe me, believe me, I ask many, you think really that's happening? No, they didn't know

00:34:01.280 how to count. They're very skeptical. That's why I wrote the paper that I'm not skeptical. I think

00:34:07.640 that's true. I want to ask you another question about growth hormone. It's a hormone that I've

00:34:13.700 prescribed to patients when they're healing from injuries. So I've seen pretty good literature that says

00:34:20.680 you tear a bicep, you have surgery to repair it, growth hormone for eight weeks, fosters

00:34:27.280 rehabilitation better than if you did nothing. So that's the very narrow window in which I've

00:34:34.180 prescribed growth hormone is typically around the healing from orthopedic injuries. A couple of things

00:34:40.940 I'll say, every patient I've prescribed it to, and there hasn't been many, maybe half a dozen over the

00:34:45.820 last 10 years. They all say, I've never felt better, which then helps me understand why this

00:34:53.020 cottage industry of doctors out there exists who run longevity clinics prescribing growth hormone.

00:35:00.740 I've drawn a hard line in the sand with my patients that I don't believe in the literature that would

00:35:06.420 suggest that prescribing growth hormone is a pro longevity tool. But if I'm being brutally honest,

00:35:11.940 and I tell them this as well, I can't tell you that it's killing you either. I can come up with

00:35:17.440 theoretical arguments why prescribing growth hormone is going to make you feel better,

00:35:23.460 but is going to shorten your life. But I don't really see any data one way or the other.

00:35:28.480 Even when you look at extreme cases, which are basically athletes who use growth hormone is the

00:35:36.100 most abused drug in all of sports because we don't have a test for it. You'd think that the

00:35:41.360 morgue of athletes would be much bigger. What is your view on exogenous growth hormone

00:35:48.260 as a, not necessarily a pro longevity tool, but as an agent that clearly helps health span,

00:35:55.820 but might be not as destructive to lifespan as I believe it could be?

00:36:00.040 I think you're absolutely right. Look, we're talking about chronic environment,

00:36:05.960 and that has nothing to do with the fact that there could be indication for growth hormone. You

00:36:12.140 mentioned one. There is a paper about growth hormone after strokes. We were actually interested

00:36:18.960 in growth hormone in the brain. There are more examples like that. So I don't think those are

00:36:24.340 mutually exclusive. Right. And I've often wondered, by the way, I haven't seen the literature. If you

00:36:29.160 know of it, I'd love to see it. Would growth hormone be protective in stages of early cognitive decline?

00:36:35.120 But anyway, putting that aside for a moment, what about this idea of the 50-year-old who goes to the

00:36:42.440 longevity clinic and they're being given low doses of growth hormone every single day? Typically,

00:36:49.180 it's somewhere between 0.4 and one milligram daily. I'm not going to answer you about the dose,

00:36:55.820 but I'm going to make the thing that I think is very important, and it's getting us back to this

00:37:00.940 antagonistic pleiotropy, and it's relevant to tame and metformin. It is possible that things that

00:37:08.140 you're doing are good for you when you're young and against you when you're old. So when people are

00:37:14.360 asking me on any of those geroprotectors, gerotherapeutics, vitamins, when do we start them?

00:37:22.680 The answer is, I really don't know. I think you shouldn't get senolytics before you're 70 or 80

00:37:28.040 years old. I think probably metformin, although we start the study at 65, most of the studies so far

00:37:35.520 that showed really large effect of metformin are people who were recruited above the age of 50. So

00:37:42.140 I think it's 50. So I don't know to tell you for a singular patient who chronological age is 50,

00:37:51.160 and biological age, I don't know what it is. You can determine it better. What do I tell them?

00:37:57.240 I don't really know based on literature, based on clinical trials.

00:38:01.260 So last question on this, Nir, based on the number of people that are taking growth hormone out there,

00:38:06.100 and I don't know how you would quantify this, but presumably it's not rocket science to figure it

00:38:10.880 out. But let's just say that there are hundreds of thousands of people in the United States alone

00:38:16.320 who are taking growth hormone daily as part of a geroprotective regimen. Why aren't they all

00:38:23.660 dying prematurely? Assuming that they're older. We don't know that.

00:38:27.920 Well, that's true. We don't know it, but wouldn't we see a signal of it?

00:38:32.060 I think not, because the people who are taking growth hormone are probably taking also metformin

00:38:36.940 and exercising and doing other things. Yeah. So fair point. There's too many confounders.

00:38:41.800 Yeah. I mean, again, my view is still, and probably will remain for the foreseeable future,

00:38:47.700 that it is not a great geroprotective agent because I do have these concerns. But this is an

00:38:54.700 example of something where I really wish we had data. So I want to tell you another thing that also

00:38:59.760 is a gray zone for me. Most of the negative effect of growth hormone IGF in humans, we see in females,

00:39:08.500 not in males, actually in animals too. So I'll describe two things. When we look at our

00:39:14.740 centenarians, that's done by Sophia Millman, who's running the longevity studies now. 0.92

00:39:19.880 And she measured IGF-1 in all our patients. And she looked at our centenarian. So they're already

00:39:28.100 a hundred years old. Is IGF-1 level predicts their longevity? And remember, centenarians are likely to

00:39:35.100 die. 30% of them are going to die each year. So those with the lowest half of IGF-1 lived twice as long

00:39:44.480 as those with the highest level of IGF-1. And that's females only or both sexes?

00:39:50.660 That's females only. Okay. 0.99

00:39:53.160 Males, the ratio of female-male centenarians around the world is there are 85 females for every 15 men.

00:40:01.040 We have better results because a lot of the female centenarians have never got married. 1.00

00:40:07.140 Nuns and other things, going back to your problem with your kids. But we need people with offspring. 1.00

00:40:13.420 Our study is based on offspring because the phenotype you capture in offspring and not in

00:40:18.640 centenarians. In centenarians, the phenotype is going down already. Those women also have better 0.95

00:40:24.660 cognitive function. And as far as muscle function, it's not different. In other words, I think a lot

00:40:32.800 of our problem, when we come from sports, we're trying to preserve muscle. I'm not sure that low IGF is

00:40:40.220 the best way to preserve muscle. I think maybe it's making the muscle biology better, but it's not

00:40:45.500 making the muscle any better. Sorry, just to be clear, you're saying that the women in the bottom 1.00

00:40:51.320 quartile, say, of IGF are no more likely to be sarcopenic than the women with higher IGF,

00:40:58.840 but they do tend to live longer. And smarter.

00:41:01.880 They have better cognitive function.

00:41:03.220 Right. With male, there is a trend, but it's not significant. So I think that if you have male,

00:41:13.180 and that's my way out of that, a lot of it is males that are taking it. I haven't convinced myself in

00:41:19.280 my study that it's not a major sex differences, the sensitivity to growth hormone.

00:41:25.020 Yeah. I'm just reflecting on the few patients in my practice who do take growth hormone. It is

00:41:29.840 prescribed by other doctors. I've made it clear that I'm not thrilled about it, but they feel

00:41:35.660 strongly about it, and it's their choice. I think it's an equal, it's a very small number,

00:41:41.000 and it's an equal mix of male and female. And which again, gets to a question that we are going

00:41:44.980 to talk about today, which is healthspan, lifespan trade-offs. So let's continue down this path of

00:41:50.100 double-clicking on the centenarians and their bucket three genes. That is to say, their good genes,

00:41:56.680 rather than their absence of bad genes. Are centenarians more likely to have APOE2

00:42:02.020 than non-centenarians, which is the protective variant of that gene?

00:42:07.120 That's the most common general longevity genotypes that we have. And APOE2, either an allele frequency

00:42:17.160 or the homozygosity of APOE2 is the most validated longevity genotype that we have. First of all,

00:42:26.620 it's the truth if you measure genotype. It was hard for me to accept it, and it's only recently that I

00:42:34.340 convinced myself. And the reason is that APOE2 genotype is also associated with diseases too.

00:42:42.120 It's not so simple. I thought for a long time that what we have actually, we have a problem,

00:42:51.300 it's ascertainment bias. That to our study, we don't get people with APOE4 because they are demented

00:42:59.540 and they're not getting to our study or to any study. So that we increase the proportion of other

00:43:07.720 genotypes. But what convinced me is that it should have been equally distributed between APOE2 and

00:43:16.760 APOE3. And it's not. It's really an APOE2 phenomenon. And so although I don't understand

00:43:25.700 totally the mechanism, I'm convinced myself that there is something true in this and that APOE2 by

00:43:31.940 mechanism that I don't totally get is the longevity genotype. Presumably, this is backtracking a

00:43:38.620 little bit on something you said, I'm guessing centenarians have less LPA genotype? So LPA genotype

00:43:46.240 is really interesting. Remember, we talked about how we are losing cloto with age. Until we regain it.

00:43:53.660 Yeah. Right. We are losing a little LPA and regaining that. Now, what we've done with computational

00:44:01.120 biologists, the system biologists, we interacted every one of those bad genotypes with a longevity

00:44:08.020 genotype. And actually, the centenarians with high LP little a, they're homozygous for the CTP

00:44:15.180 genotype. So they have some protection to counter it. And by the way, they might be getting some benefit

00:44:21.680 from LP little a. There may be something that LPA, because you have to think, LPA is common. It's about

00:44:28.560 10% of the population. So 10% of the population overexpresses this thing. 10% of the population

00:44:34.280 has an elevated phenotype for LP little a. And it's true that evolution wasn't selecting against

00:44:39.720 atherosclerosis. But there are arguments to be made that LP little a could have played a role in

00:44:47.720 managing infections, for example. That this could have been a manner in which we fought oxidative stress.

00:44:54.800 So it begs to at least question the idea, do the centenarians who have it get some benefit from it

00:45:02.640 while having genes that offset some of its negatives?

00:45:06.580 Yeah. But you have to explain on a population base how all of a sudden at age 85, you switch and

00:45:14.220 this genotype becomes protective.

00:45:16.560 Yeah. Is that really necessary? Or is it just a denominator problem where at 85, you've really

00:45:23.020 eliminated so much of the population that you're now concentrating the people who were never harmed 0.97

00:45:28.900 from it? In those people, they've never been harmed by the gene. And now they are disproportionately

00:45:35.720 rising to the surface because the population around them has withered away so quickly.

00:45:40.220 So those are the two explanations, either what you just said, or that they're protected by other

00:45:48.020 longevity gene and it makes it irrelevant. What is LP little a doing?

00:45:52.300 Although I don't think those are mutually exclusive. I think the latter is an explanation for the

00:45:55.780 former. It's the amplifier of it. It's what lets them get there in the first place. Using CTEP as an

00:46:01.340 example, they happen to have a CTEP mutation or a CTEP variant that offers remarkable protection

00:46:10.320 against atherosclerosis, of which an interesting but kind of irrelevant phenotype is high HDL cholesterol.

00:46:17.900 And that's offsetting the damage of their LP little a. And then eventually at some point when

00:46:22.500 everybody else has died because of their LP little a, they're still standing. And they might even be

00:46:27.320 getting some benefit from LP little a that everybody gets, but it's in other people's cases, it's so

00:46:33.240 dwarfed by the damage of LP little a. Again, total hypothesis or speculation, but it's plausible.

00:46:39.480 It is.

00:46:40.140 Okay. Tell me about TSHR. I've never understood that one fully.

00:46:44.660 So the thyroid story is interesting because we found a correlation between high TSH and longevity.

00:46:54.380 As an endocrinologist, Nir, maybe give people the two-minute story on what TSH is and how it

00:47:00.700 functions.

00:47:01.840 Sure. My sisters are listening to you and I promised I'm going to be so simple. You won't need to call

00:47:08.640 me again and ask me what did you mean? And now I'm falling into...

00:47:12.160 I know. I'm doing a bad job of this. I'm sorry.

00:47:14.720 So TSH is really your control of thyroid function in the sense that if you become hypothyroid,

00:47:25.080 then this TSH, this hormone from the pituitary will increase in order to get those thyroid hormones 0.57

00:47:33.080 to be normal again. And they might fail and then you'll be hypothyroid, but there's an effort to get

00:47:39.620 those thyroid out of your glands. Okay. So that's TSH.

00:47:42.900 So when we see a high TSH in a normal person, we ask the first question, is their thyroid gland not

00:47:52.040 making enough T4 and or converting enough of that T4 to the active hormone T3, which is the feedback

00:47:58.560 loop that tells the pituitary how much TSH to make. So what you're saying is we see a higher amount of

00:48:06.260 TSH in long-lived people.

00:48:09.180 And their children.

00:48:10.020 Yeah. So maybe suggesting a subclinical or potentially a clinical degree of low thyroid

00:48:16.760 function or hypothyroidism. Is that a safe summary?

00:48:20.040 Correct. And I'll tell you, our discovery has led to several papers and to change in the thyroid

00:48:28.820 association recommendation of what to do with old people. You know, the endocrinologists or the

00:48:35.200 thyroidologists were looking for business. So once TSH was above 10 and then they said it might be

00:48:41.940 subclinical at seven and we went to five. And now if you're three and symptomatic, which all we are

00:48:48.880 always because we're tired. So that's where the science came. And we pointed out to the fact we did

00:48:56.380 it in our study, then we did it in a national study. And we said, you have to leave those older people 1.00

00:49:02.620 because maybe this is a physiological way for them to be well.

00:49:08.120 And tell me how high you're seeing the TSH. What is the difference when you age match them between

00:49:14.560 centenarians?

00:49:15.900 It's like five to eight.

00:49:18.180 Wow.

00:49:18.880 So normal is like until five.

00:49:21.280 Yeah. Or even 4.2 on our lab. Yeah.

00:49:24.800 Yeah. And changing baseline. And look, the thyroid hormones themselves are normal. It's really only the

00:49:32.620 TSH. So normal free T4, normal free T3, but they walk around with a TSH of five to eight.

00:49:39.880 Now here's the kicker, and this is where I'm hoping the offspring can help us. Do you know what

00:49:44.860 their TSH would have been in their 30s and 40s? Would they have also had a TSH of five, six, or seven?

00:49:51.360 Well, I think that's why we have the offspring. And I don't have the answer for this yet, but that's why

00:49:57.180 we have the offspring, because we want to see the effects of those longevity gene as they get all.

00:50:04.060 And to see really what phenotype and what measurements are changed throughout. I don't have the answer,

00:50:12.240 but a lot of them do have high TSH when they're 60.

00:50:17.920 So they're euthyroid, high TSH patients.

00:50:21.060 Right. Right.

00:50:21.960 What's the hypothesis for that? Is this simply a biomarker of something else that is unrelated to

00:50:29.540 thyroid function? Or does it suggest that these people live right on the edge of hypothyroidism

00:50:37.320 without actually becoming clinically symptomatic? And that there's something protective within running

00:50:43.080 a lower RPM, running the engine a little bit lower, a little bit slower.

00:50:47.980 So, by the way, the growth hormone deficiency models are also hypothyroid. It might be part

00:50:55.820 of the physiology of what we're seeing anyhow.

00:51:00.040 Have you looked at prolactin in these people and other pituitary hormones that tend to move with

00:51:05.080 TSH? 0.68

00:51:06.440 We have prolactin. I'm not sure that we looked at TSH. This TSH paper came years ago. I don't remember

00:51:13.540 that we're doing anything. It's a good question. I have two endocrinologists that are looking at

00:51:17.960 our data and analyzing it. I'll be happy to come again or have them talk more about the

00:51:23.680 endocrinology of those people. But we're just looking and I don't have association with others,

00:51:30.300 but the hypothesis was that their metabolism is maybe slow. And although they're compensating by

00:51:37.860 higher TSH, still their metabolism, you know, it's like insulin resistant. You don't totally

00:51:43.640 normalize the glucose, although you have enough insulin for that, that there's a metabolic

00:51:49.420 advantages. I'm saying it like that because I don't know that it's true. I don't really think

00:51:54.680 that our data supports low metabolism necessarily. But I do think as an endocrinologist that if an

00:52:02.280 elderly person comes by incidental finding, has TSH between five and eight, you don't have to go and

00:52:10.240 treat it straight away. Has anyone done Mendelian randomization on any of these clusters of genes?

00:52:18.180 Because we haven't really established a causal relationship here, have we, in these genes?

00:52:23.440 No. By the way, Peter, you just said another word that nobody knows.

00:52:27.780 Okay. Sorry. Do you want to explain a Mendelian randomization?

00:52:31.580 Well, I don't want to go long into that, but because we have so much genetic data,

00:52:36.360 and by the way, genetic data, this is disappointing because they're not so predictive. Most of them,

00:52:42.680 you can find lots of genes for obesity and you'll be lean for lipids and you'll have normal lipids,

00:52:47.760 but we have lots of genotype and we try to integrate those genotypes in order to assess how much they are

00:52:55.620 increasing the risk of us of getting a disease. And usually it's not by much. What we're trying to do,

00:53:02.800 we're trying to have an instrument that will do Mendelian randomization to longevity. Okay. So we can

00:53:10.980 see what's your genotype that fits longevity more than anything else that's in evolution now. But

00:53:19.520 otherwise, look, I have to tell you, the reason this is not good is very simple. We're doing something

00:53:26.160 so stupid in genetics. We do lots of genotypes and we take each one of these genotypes and we ask, 0.99

00:53:33.860 is this associated with obesity or diabetes or not? And we found many things. And in order to make

00:53:42.660 them statistic significance without increasing their power much, we just need to get hundred thousand

00:53:48.540 more people in the study. Okay. But we're not built with one genotype at a time. We're built with

00:53:55.320 numerous genotypes. What we've started doing lately, and that's our latest nature paper,

00:54:02.560 was to say, just a minute, we're looking at the difference between centenarians and people without

00:54:08.740 longevity. And we're going to take all those rare genotypes, less than 1%, because to be centenarians,

00:54:16.880 one out of 10,000. So we need to find rare genotypes. And by the way, there are 80,000 rare genotypes

00:54:23.600 just in our population. Then we take those genotypes and put them in pathways and look at the enrichment

00:54:32.280 of the pathways rather with a specific genotype. And then we get really important information. And by the

00:54:40.900 way, the important information that we got in the study, in the summary statement, is that the genetics

00:54:48.000 of longevity in humans is exactly what we learned from animals. It's the insulin signaling pathway.

00:54:55.020 It's the mTOR signaling pathway. It's the MAP kinase pathway. It's exactly the same genotype,

00:55:01.820 the distinct centenarians and other people. And I think this is really very important for us because

00:55:08.640 people have blamed us that our animal models, animals' models are not good. Yeah, they're not good

00:55:14.320 for diabetes. They're not good for Alzheimer's. They're not good for other things. But for aging,

00:55:18.700 actually, our models are really good because it's so conserving evolution. All our animals, their skin,

00:55:25.040 their hair, their skeletal, they get cancers, they get diseases. So this is the same, but it's really the

00:55:32.160 same pathways. And the first that comes out is the IGF insulin signaling pathway.

00:55:37.640 All of this really points towards the importance of polygenic risk scores. And when I talk to someone

00:55:47.800 like Richard Isaacson, who is so focused on understanding Alzheimer's disease, and Richard

00:55:54.680 and I work so closely on this because it's obviously one of the most important things we think about in

00:55:58.900 our practice, I think of the evolution that we have had in our thinking over the past five years. In fact,

00:56:05.980 we're working on a paper now that's looking at this polygenic assessment of risk in dementia.

00:56:12.900 You brought up earlier APOE4. Obviously, any listener to this podcast is no stranger to it.

00:56:18.840 We've spent so much time talking about it. One of the things that I would say is 10 years ago,

00:56:24.420 we thought that being homozygous for E4, so being in that roughly 1% of the population that has two

00:56:31.780 copies of the E4 gene, that was a death sentence. As you noted earlier, that's a person who's going

00:56:36.780 to have Alzheimer's disease by 60, and they're going to be dead by 70. There's no way out of that.

00:56:41.660 Their risk is deemed at about 20-fold that of the general population. And then five years ago,

00:56:49.500 we looked at the data again and said, it's still awful, but it's not 20 times the risk of the E3,

00:56:56.460 it's 12 times the risk. Then we look at the data again two years ago, and you know, it's still bad,

00:57:04.340 but maybe it's five times the risk. And now we are looking at patients who have E4, E4. Not only do

00:57:13.420 they not have Alzheimer's disease, more importantly, they don't have any of the early signs of it based

00:57:19.060 on really, really advanced cognitive testing that shows very subtle signs decades before.

00:57:24.560 And we're starting to look at other genes that are abrogating some of the effects of this.

00:57:30.540 And so now the focus has been less at looking at APOE4 and making a determination. And it's looking at

00:57:37.220 E4 plus TOM40 plus mitochondrial haplotype plus cloth though, plus a whole bunch of genes

00:57:43.800 and taking a polygenic approach to risk. It seems to me that this would be the most logical way

00:57:50.740 to do the same thing with respect to longevity. I think this is true with cardiovascular disease.

00:57:57.360 LPA is part of the story, but it's not the whole story, right?

00:58:00.880 Absolutely right. You described the problem and the solution absolutely right. I would just tell you

00:58:06.320 that looking for those longevity genes that we just published, 12 of them are associated,

00:58:13.500 we're funded for that, but they're associated with resiliency to Alzheimer's. And I think that

00:58:20.880 not enough of the genetics is explaining not only the genetics, but the resiliency or looking at the

00:58:28.600 genetics of resiliency, which is what we're calling longevity in this case, but the resiliency

00:58:34.180 to diseases, which happens to diseases that you get in young age too. Not every genotype causes

00:58:39.880 this disease. So I think you're right. Would you put FOXO in that category as kind of a general

00:58:45.280 resilience gene that is less disease specific and more broadly protective? Yeah. I actually count

00:58:53.200 FOXO3A as in the insulin IGF singling pathway. We can argue about it, but that's my bucket.

00:58:59.940 What are the genes that you would put in the general resilience pathway, not necessarily a disease

00:59:06.140 specific or system specific as in the endocrine system? Well, I think we mentioned a lot of them

00:59:13.360 now, but I think more important to realize that we have 750 centenarians in our study. Now they're all

00:59:22.520 Ashkenazi Jews. Why is that? Not because religious is important, but because Ashkenazi Jews are a genetically 1.00

00:59:31.520 homogeneous. They went through a bottleneck, an expansion, and then a bottleneck and very few

00:59:38.360 survived. And they lived in isolation and intermarriage and their genetic pool is much

00:59:44.960 more homogeneous. In fact, we can measure it. It's genotype specific, but we need between 20 and 50 times

00:59:52.680 less people in order to get the same data. So think about it that not 750, but few thousands,

01:00:01.820 but it's still not enough to find all longevity genes. And one of the things that we realize

01:00:07.720 is that there are 50 ways to leave your lover. There are 50 ways probably to get to longevity.

01:00:14.840 It's possible there's only one person who's 150 years old somewhere. And if we knew his genotype,

01:00:20.680 he will solve the problem for anyone. And the genotypes we are discovering are things like that.

01:00:28.220 We have 20 genotypes in centenarians and one in control. But those for geneticists who are dealing

01:00:36.180 with 400,000 people in study, they say, well, in 750 centenarians, this can change any minute.

01:00:43.320 So what can we do? Well, what we can do is look at the function of those genotypes. And you know,

01:00:50.160 our friend Yuxin Su, that's what she's doing with our studies. She takes those discoveries and either

01:00:57.620 takes the lymphoblast of our patient or she does construct and express those genes in cells and see

01:01:04.980 how resilient they are to injury. So there are a lot of genes. And I think the biggest news for me now

01:01:13.360 is that American Federation of Aging Research has gotten contribution from a single guy at $2.8 million

01:01:21.540 to recruit 10,000 centenarians across the United States and their offspring and control. And I think

01:01:30.780 that this will give us such an acceleration of understanding longevity and such an acceleration

01:01:36.640 of getting drugs that probably are likely to work. And also things that we don't know yet. Like,

01:01:44.440 who knows? Some of the things that we discovered are in pathways that were not in our longevity lexicons.

01:01:54.480 Nir, that's interesting. Only $2.8 million is required to recruit 10,000 centenarians plus their

01:02:01.680 offsprings plus appropriate controls. Well, let's see. Let's see where we get.

01:02:07.320 That seems like a very high ROI for that donor.

01:02:11.340 Well, we hired two companies who have approaches. And actually, there's a preliminary study to see

01:02:20.720 what's the best way to approach these people, because it's all going to be basically web-based.

01:02:27.320 So we need their children and grandchildren. And we're going to send them a swab so that they can 1.00

01:02:34.400 do it with help. And we'll get the swab, at least on the first pass, and be able to rapidly do the

01:02:41.860 genetics and post it, by the way. Just immediately post it so people can start looking at it and making

01:02:48.800 sense of the data.

01:02:49.980 And you're going to do whole genome sequence or exome?

01:02:53.780 Both. We're going to do mainly exome sequencing and maybe switch as the whole genome sequencing

01:03:01.580 price comes down, which it does rapidly.

01:03:05.180 What's the current price of whole exome?

01:03:07.740 I don't totally remember that. The reason I don't remember is that our genotypes was done by

01:03:13.680 Regeneron, and they paid for that. So I don't remember how much it is now.

01:03:19.040 So let's talk about a thought experiment. You mentioned something earlier, which is what

01:03:22.420 made me think of this. So you said there's somebody out there maybe who's going to be

01:03:25.460 150 years old, one person that will concentrate all of these genes in perfection, and they'll

01:03:30.840 get to be 150. And if we could look at that person's genes, we might have an answer. I would

01:03:36.280 push back and say, or not. It might be that we're discounting the stochastic nature of this.

01:03:43.480 And even if you found that 150 year old person, and even if you identified which genes played a

01:03:50.680 role, the likelihood that you'll identify which environmental factors turned on those genes or

01:03:57.300 amplified some and attenuated others seems very low. It begs this thought experiment, right? If you

01:04:03.760 took 10,000 identical people with whatever program you have of the perfect genotype. So this is 10,000

01:04:15.800 people that have as many of the good genes as possible and is none of the bad genes. We put them

01:04:22.980 in a time capsule and we let them live their lives. But now we randomize them to, I'm making this up,

01:04:29.220 but three groups. One group is the base case. Go and live a normal life. One group is the do

01:04:36.700 everything bad that you possibly can. So I want you to start smoking when you're 15. I want you

01:04:43.000 drinking three drinks a day. Never exercise. You're only allowed to eat at McDonald's. And then the third

01:04:49.740 group becomes the do everything right. We do the opposite. Give me your prediction of how long each of

01:04:56.640 those three groups lives. So Peter, I was trying to do things simple and you complicates me again.

01:05:04.620 And I'm taking back. First of all, when I said 150, I think the maximal of human lifespan as a species

01:05:11.600 for us is about 115 years, even if we argue, even if there's 122 somewhere. And we die before the age

01:05:19.220 of 80. So we're talking about 35 years that we can realize. It's a lot of years. We should realize

01:05:25.240 that. But I think aging will improve by other methods and mechanisms that can break eventually

01:05:32.960 this 115 years and maybe get us to 150. When people are asking me, when will we leave 150?

01:05:41.220 I said, oh, in 150 years, because even if we start the experiment now, it'll take 150 years

01:05:48.420 to get there, right? We wouldn't know. I'll tell you what I'm getting at in the experiment. I want to

01:05:53.820 understand how in the perfect genetic makeup, how much can environment hurt or better what is already

01:06:05.080 a genetic lottery? I totally understand. And I would take you back to, and I meant to ask you that when

01:06:11.640 you describe that ApoE4 was a 20-fold risk and became smaller and smaller, well, became smaller

01:06:20.500 at whom? And how their lives were different than those that we knew 20 and 40 years ago that have

01:06:28.420 the ApoE4 genotype? Yes, it gets back to the point you raised about your grandfather versus your father

01:06:34.400 versus you. Right. The environment, both in what we do, both in medical treatment, in surgery,

01:06:40.400 and all those things. Absolutely. And so you're right. We think that they're a master switch to

01:06:46.660 longevity, and some of them we're doing with exercise and food. So you don't really have a

01:06:52.600 sense of how much longer group three could live than the group one when you basically put amazing

01:06:59.620 genes in everybody? Yeah, you know, it's not the things that I'm doing predictions on. No, it's interesting

01:07:05.920 to me because where I'm really going at a macro level is most people don't have these genes. So

01:07:12.280 the only interesting question is the contrapositive of that, which is once you have a sense of what

01:07:18.720 that could be, now for the rest of us, the 9,999 of us who don't have centenarian genes,

01:07:26.420 how much does environment make a difference? I know the answer is significant, but I'm curious

01:07:31.240 as to how much you think it is. For me, some centenarians are coming to me and said,

01:07:36.720 okay, what can you do for me? And I'm like, you're it. I don't do anything for you. So for me,

01:07:47.340 the question, if those centenarians are willing and I'm starting to exercise them and to change their 0.99

01:07:52.960 diet, am I going to kill them or help them? That's how I'm thinking about it. And I don't know the answer. 0.99

01:07:59.320 Fair enough. So Thomas Pearls has written quite a bit about this idea of dividing centenarians

01:08:05.760 into buckets. I think we have the escapers, the delayers, and the survivors. Are you familiar

01:08:12.420 with that terminology? Yeah, yeah, of course. I'll tell you two things about it. First of all,

01:08:17.320 Tom Pearls and Paolo Sebastiani did a terrific job and we're collaborating a lot and they've been

01:08:24.160 great. And I don't hear Tom talking about it because this is the point. Average life expectancy

01:08:30.880 is 80. Some people are going to be 81, 82, 83. It's not that 100 years old is any special. You're

01:08:38.640 picking at one time people who actually aged before or will age later. And what we've done together

01:08:46.600 to kind of overcome that. And that's a very important thing. We looked at our data in both

01:08:52.440 studies and harmonized them and asked what is the health span of centenarians versus, by the way,

01:09:01.340 cohort that lives now. It's not their cohort. Their cohort died before. And the answer is that it's not

01:09:08.420 only that they live long. They get variety of age-related disease 20, 30 years later. There are control

01:09:15.460 groups between 60 and 80, accumulate lots of diseases. At 80, only 10% of them don't have a

01:09:21.560 disease. In our centenarians, after the age of 100, 30% don't have a disease and are not treated with

01:09:28.620 anything. But there's a 20, 30 years of health span. This is not really the important part. The

01:09:36.020 important part that at the end of their life, they get sick and die. Some of them don't wake up in the

01:09:42.400 morning. But they have such a compression of morbidity. They are sick for months at the end

01:09:48.480 of their life, unlike us that are sick for years at the end of our lives. And I think this is really

01:09:56.160 the boundaries. And so first of all, there's example of humans that can live healthy and long and have

01:10:02.360 contraction of morbidity. And for me, it really says it's not that they didn't get older. Of course,

01:10:07.840 they got older, but they had great life. They have great health. And at some point,

01:10:12.580 they're checking out much quicker without diseases at the end of their lives. And this is what we're

01:10:19.060 trying to imitate. I completely agree, although I'll throw a wrinkle into it, Nir, which is,

01:10:25.320 could it be that the average person has their first heart attack at 70, languishes with congestive

01:10:31.760 heart failure for five years and dies at 75? The centenarian has their first heart attack at 101,

01:10:38.780 and they die six months later. Could it be that the reason that the first guy languished for five

01:10:44.540 years is he really technically had more resilience given that he was 25 years younger? And if so,

01:10:52.120 you can now do a very grim thought experiment, which is we could easily replicate the contraction of

01:10:59.220 morbidity and non-centenarians with a rule that says the moment you get a disease, you get a pillow

01:11:04.960 over the head. The moment you get your first heart attack, we're not going to cath you. We're going to

01:11:09.440 kill you, right? Thought experiment, please. Let's be clear. But you would immediately square everybody's 1.00

01:11:15.500 longevity curve for the most part. You would contract that period of morbidity to replicate that

01:11:21.920 of the centenarians, but you would dramatically truncate lifespan as well now. And now you would even

01:11:27.520 widen the gap between the non-centenarian and the centenarian. So in my mind, it's very difficult

01:11:33.320 to disentangle the objective for someone like me, which is I do want to lengthen lifespan. I do want

01:11:41.620 to delay the onset of chronic disease. And I want to compress the period of morbidity, not by having you

01:11:49.880 die quicker, but by having you live better longer, if that makes sense.

01:11:54.740 It absolutely makes sense, Peter. And of course, your point is very important, but it's like the

01:12:01.080 podcast last week on the COVID, you had a scientist and the advocates.

01:12:08.940 Which by the way, our mutual friend, David Allison, he was really the one who I think

01:12:13.220 made me first realize we need to talk about these two things totally separately.

01:12:17.760 Absolutely true. And what you did, you really stopped me from being advocate because you're

01:12:23.980 absolutely right. I think that the reason that centenarians are dying so fast is because they

01:12:30.160 are already frail. They are aging and so they are less resilient. But on the other hand, without

01:12:37.260 diseases, I mean, some of them are working as hedge fund managers until 107. Some of them are

01:12:45.100 painting. As long as they have no pain and mobility, their life is good. Not like when they were 20,

01:12:51.540 but their life is good. So I totally agree. But as an advocate, I wanted to hide that, Peter. I didn't

01:12:58.500 want to. But you're right. Why they are dying. But you know what? I think it's still important for

01:13:05.520 people. And I think this concept is probably true. If you prevent aging and age-related disease,

01:13:11.180 you're going to compress morbidity too. One of the biggest challenges I have with health

01:13:15.600 span is I don't think that we have great ways of describing this in medicine. I think there's

01:13:20.200 several issues with it. The first is our definitions, I think, are not wonderful. We talk about freedom

01:13:25.880 from disability and disease, but that doesn't really capture it. I can tell you a lot of people

01:13:30.500 who don't qualify for having disability or disease, but their health span is still poor. And by the way,

01:13:36.820 this completely excludes a very important element of health span, which I think is emotional health.

01:13:41.660 So let's put that aside because it's not particularly age dependent and it's outside of the purview of

01:13:46.860 what we're talking about today. But if you just limit it to physical and cognitive, in fact, if you

01:13:50.620 just limit it to physical, you can have people who can still carry on activities of daily living,

01:13:57.080 but one of them has a VO2 max of 50 mils per minute per kg. And the other is 30 mils per minute per

01:14:04.260 kg. By the way, neither of those people would ever qualify as disabled because whether you're

01:14:08.960 at 30 or 50 in VO2 max, you can still do any activity of daily living. But one of those people

01:14:15.480 can clearly get more out of life. You can take someone who has the grip strength to hang on to

01:14:21.480 a bar for 30 seconds versus hang on to a bar for two minutes. Both of those people will see no

01:14:27.980 immediate difference in their day to day activities. But one person can do far more should they choose

01:14:35.000 to. One person can sit on the floor for half an hour and play with kids and feel nothing. Another

01:14:42.520 person, their back will ache for the rest of the day. Neither of those people are debilitated or

01:14:49.160 disabled. And therefore, by traditional health span metrics, they're both equal, but they're nowhere near 0.95

01:14:55.340 equal. So that's problem one that I have with the way we as a community talk about health span.

01:15:00.980 The second thing is nothing that you or I learned in our medical training even remotely prepares us

01:15:07.640 for how to help people be truly stronger late in life. It just wasn't part of our training. There's

01:15:14.700 nothing about that. So I guess my point here is it's hard for me to really interpret the data

01:15:21.680 and get at something I'm very interested in, which is do centenarians truly have better health span

01:15:29.460 or are they just dying later? And for the most part, they have this period of compressed morbidity.

01:15:38.400 So it looks like they have better health span. But do we really know that? I mean, I think they do

01:15:44.880 based on the literature, like I think a 90 year old who will become a centenarian is functioning

01:15:51.180 more like a 70 year old who will not, but it's still very difficult to quantify. I don't think

01:15:57.000 we have great metrics here. Do you disagree with me? No, I don't think that health span is well defined,

01:16:04.100 definitely. So there are two comments that I will make. For the NIH, back to 10, for them,

01:16:10.980 what is aging is if we can prevent diseases. That's their measurements, which is not satisfying,

01:16:18.600 not to geriatrician and not to physician, but that's how you get drug approved. For the economist,

01:16:25.260 there are two issues. One is what we're counting as the medical cost of the last two years of life.

01:16:31.900 By the way, in centenarians, the CDC has data. In centenarians, the last two years life,

01:16:37.820 are third the cost in a hundred years old than when you die in 70.

01:16:43.220 That is fascinating. I mean, to me, that is something that speaks to what we're really

01:16:48.980 aspiring to. And actually, that statistic captures much more about the quality of a person's life.

01:16:54.780 Right. But Andrew Scott had a paper in Nature. I think I sent you guys, right?

01:17:01.220 I read it. Yes, you sent it to me. Yep.

01:17:03.140 He describes an economical term that's called the value. And by the way, it's very hard. If you want

01:17:09.640 to know more about this paper, we should talk. I had to sit with him in order to understand

01:17:14.340 what he was saying, but basically saying, hey, if you increase the health spend of someone,

01:17:20.300 it's not only medical costs, because this guy is going to travel and spend money traveling and buy

01:17:26.320 gadgets and buy houses for his kids, their value of the person life is going to be increased.

01:17:34.120 And some of those people, by the way, are going to still work. They're not just playing golf.

01:17:38.360 Right. So the economical value is huge. So you can do it economically. And the third thing I wanted

01:17:44.080 to tell you is this story, the real story about this 102-year-old guy that I met. And I'm sitting

01:17:51.700 and talking with him and he's the nicest guy I've ever met. He's so considerate. He's thinking about

01:17:57.540 life in such a nice way. Nothing bad about his daughter-in-law. Right. And so I'm spending time

01:18:05.460 with him. I'm going out of the room and I'm bumping into his son, who's 80 years old. Yeah.

01:18:11.580 And I'm telling this 80 years old what I'm telling you, you know, your father is just the nicest guy I've

01:18:17.680 ever met. So he looks in my eyes and says, you should have seen the son of a bitch when he was 0.99

01:18:22.540 my age. He was a terrible, terrible person. Okay. So then I realized, and you'll see why it's 1.00

01:18:30.640 connected to what you just told me. Then I realized that because we wrote paper about the personality

01:18:37.160 of centenarians and you have to be positive and stuff like that. But apparently he became positive

01:18:41.400 only when he was 100 years old, not when he was 80 years old. And then you see papers like

01:18:46.920 in University of Pennsylvania, they took two groups of people, young and old people and show

01:18:52.160 them bad slides, cockroaches in pizza and good slides like islands in the Caribbean. And they

01:18:59.360 asked them to repeat what they've seen. And the young people knew, remembered a lot from both bad

01:19:06.120 and good. The old people remember less, but mainly the good things. So I'm waiting for this part 1.00

01:19:13.400 in my physiology that I remember. Only the good thing. But what I'm telling you is that think of

01:19:21.200 the complexity now, because your brain age, you also retired, you lost your spouse, you moved from

01:19:29.520 independence or you moved to somewhere else. You moved several times, you get to 100 years old.

01:19:35.120 So apparently there's changes, not only in your environment, but in your physiology,

01:19:40.320 where life can still look good to you. Is this wrong?

01:19:44.820 Actually, that's such an interesting point, Nir. I know you mentioned already that a number of the

01:19:48.620 centenarians in your cohort were not married. What fraction were married?

01:19:54.340 No, my centenarians had to have offspring.

01:19:57.660 Oh, okay. So all of them at least had a partner.

01:20:00.220 At least married for a while.

01:20:02.320 Okay. And then what fraction of them have lost their spouse?

01:20:06.380 Most of them.

01:20:07.080 And sorry, one other question, Nir. What fraction of them have also lost children?

01:20:11.240 Also a lot of them.

01:20:12.720 This gets to something very interesting. Someone asked me once at a dinner party,

01:20:17.400 if you could wave a magic wand, how long would you want to live? And I said, not that long,

01:20:22.500 truthfully, because unless I could wave that magic wand for the people around me,

01:20:26.980 I think it would be an awful life. Could you imagine if I waved a wand and said, Nir,

01:20:31.300 you have immortality? And not just immortality, I'm going to let you preserve your quality of life

01:20:36.880 today. So as smart and healthy and able as you are now, I'm going to let you do this for the next

01:20:42.460 500 years. I would view that as an awful curse. You'd have to watch your wife die. You'd have to

01:20:47.940 watch your children die, your grandchildren die. Sure, you could remarry and do it all over again, 0.97

01:20:52.440 but then they're going to die. That's a silly thought experiment, but I also think it gets to 0.59

01:20:56.920 something I've never really considered with respect to actual centenarians, which is the

01:21:03.020 price of being a centenarian is giving up and losing most of your friends. And how many funerals do you

01:21:11.000 go to? The most common thing that centenarian tells me, the thing that underlined their age is when

01:21:18.400 the children of their friends were long gone, the children of their friends started to die.

01:21:23.980 That's for them kind of what's going on. We're losing now the second generation. What's coming 1.00

01:21:30.720 next? You're absolutely right. I'm thinking about it for immortalists. So immortalists, they have 12

01:21:37.980 billion years. We don't know, but one of the estimates, 12 billion years for this planet. I'm a

01:21:44.660 little tired from some of the shit now. Really 12 billion years, January 6th, wars and stuff. It's 0.99

01:21:55.860 true. But what I try to tell you in a different words is that those centenarians have been adapting

01:22:03.240 not only because what got them to be centenarians, but because of physiology of aging that allowed them

01:22:11.820 to harbor and maybe to put in compartments lots of their stuff. Yeah, lots of grief.

01:22:19.340 And it's not that you come to centenarians and they want to tell you, first of all, about the

01:22:23.480 spouse that they lost. That's not how they are talking to you. They're talking to you about where

01:22:28.620 they went and what they do. And they're going to a concert and their grandchild is coming to visit

01:22:33.820 them. At least the centenarians in my study. So it's not all centenarians in the world.

01:22:38.920 Look, I think living to a hundred with preserved health span would be a wonderful thing. And even

01:22:45.100 if it meant you lost a spouse, a child, it means you probably have three generations of people you

01:22:51.360 know. One of the questions I love asking my patients is how many of your great-grandparents

01:22:57.700 can you name? So we all have eight great-grandparents and I say, tell me the names of them.

01:23:02.200 I've only had one patient who could tell me two of them. Most people can't tell me a thing about

01:23:09.880 one. I can't. I didn't even meet two of my grandparents. They died so young. So if you

01:23:17.060 think about that for a moment, and the reason I ask my patients this, by the way, it has nothing to do

01:23:21.260 with living longer. It's more about living better. It's basically, it's a way to remind everyone,

01:23:27.060 myself included, that we're not that important. Like my point being is my great-grandchildren will

01:23:32.680 never know who I am. Maybe in the age where they have video, they'll see a video of me or something,

01:23:37.380 but like, I'm not going to be an important part of their life. Meaning the only people who I matter

01:23:41.780 to are very narrow and close to me now. So let's not lose sight of that. But these centenarians have

01:23:47.560 a gift, which is their great-grandchildren will know them. And when you use this example, when you can

01:23:54.360 go to concerts with your great-grandchildren, that's amazing. When you could take a vacation

01:24:00.100 with your great-grandchild, and you're not only able to give them money for college, but you go

01:24:04.920 to their graduation. Think about the implication of how much of their life you've been a part of.

01:24:09.580 So the flip side of everything I just said is, I've never really met somebody who's dying at the age

01:24:14.680 of 75, and that's a phenotype I've seen an awful lot in my training, who didn't wish to have another

01:24:20.940 year of life if it could be had at a higher quality. Which now brings this whole discussion

01:24:26.060 full circle. 9,999 out of 10,000 of us will not live to 100, directionally speaking. But if we want

01:24:35.280 to live an extra year or five years, what is the most important lesson we can take away from the

01:24:41.340 centenarian that we can actually do something about? So first, I want to start with what you mentioned

01:24:47.980 before one of my darkest days in research, when Jay Leno in The Tonight Show said, there's those

01:24:56.080 people at Einstein, and they said, the secret for longevity is don't exercise, don't, you know, be

01:25:02.260 obese. And you know what he said? If you die, you don't care anyhow. And that was the wrong lesson from

01:25:09.740 the centenarian study. If you're going to be centenarian, maybe it's not important. By the way, it could be

01:25:15.480 important for centenarians. I mean, this woman that I have that smoked for 90 years and died at 110. I

01:25:22.200 just wonder, wouldn't she be the next Madame Clement without smoking? So the lesson for most of us is

01:25:29.500 still exercise and nutrition, whatever it means to everyone and everything else that you give. That's

01:25:36.780 the lesson. And it's not the lesson from centenarians. The lesson from centenarians is that there are

01:25:42.400 longevity genes that could be translated into drugs. And I believe that they could afford years

01:25:50.260 of healthspin, however we want to define that. And that's really what I'm trying to say that is

01:25:57.160 not an emotional part, but a clinical part.

01:26:00.740 Would you agree with my takeaway from this cohort? Because the single most important lesson I

01:26:06.420 glean from everything we've said, in addition to lots that we haven't said that you and I have

01:26:11.200 talked about elsewhere, or that my own work has pointed me to based on my study of this problem,

01:26:16.840 their superpower is simply delaying the onset of bad things. Bad things just happen to them 20 to 25

01:26:24.080 years later. It's not that they don't get heart attacks. They just don't get them when they're 65.

01:26:29.080 It's not that they don't get cancer. They just don't get it when they're 60. It's not that they don't

01:26:34.560 even get dementia. They just don't get it when they're 75. And the last time I looked at the

01:26:41.180 distribution of death for centenarians, it was shockingly similar to that of non-centenarians

01:26:47.680 with a couple of differences. They tended to have a little more atherosclerosis, a little more heart

01:26:52.520 attacks, a little less Alzheimer's disease, and I think a little bit more pneumonia. But directionally,

01:26:58.840 they had the same actuarial table of death as people dying in their 80s. It was just a time

01:27:06.540 shift. In fact, I reviewed the paper from Germany where they looked at pathology. Okay, it's a

01:27:13.980 pathological. They looked at a thousand centenarians that over the years died in their homes. And they're

01:27:21.920 right, because in the hospitals, we kill them in other ways, right? So they died in their homes versus

01:27:28.040 thousands. I'm not sure about the numbers of other people that died at their homes. Basically,

01:27:33.700 the paper was funny because the title was like, there's nothing special about the centenarians.

01:27:38.500 They're dying for the same thing. But 30 years later, okay, they missed the point. It was like a negative

01:27:46.720 study. But you're right. They're kind of dying from the same thing much later on. So you can look at it

01:27:55.720 about what you said, the resiliency that got them there, the resiliency for anything that detect them

01:28:02.520 to get them there, or the fact that their aging was slow. And so what's the takeaway for us? To me,

01:28:10.020 the takeaway for us as physicians or people who want to have an extra five years of life or 10 years of

01:28:15.780 life, even if we can't have an extra 20, is nothing matters more than prevention of chronic disease.

01:28:21.240 And by the way, you don't get to prevent it once you have your heart attack. Secondary prevention is not

01:28:26.780 prevention. We're talking ultra, ultra, ultra primary prevention. And if health span is something

01:28:35.380 that the medical system hasn't been poised to teach, ultra primary prevention is also something that we

01:28:42.420 haven't really been prepared to teach. So let's pivot to another topic that we visited last time, but I think

01:28:51.540 so much has happened in the interim, which is metformin. I probably get an equal number of questions near

01:29:00.280 about the following three things. So the frequency of this would be several times a week, a patient or a friend

01:29:08.680 is asking me about metformin, rapamycin, or some combination of NR, NAD, or NMN. Somehow those three things

01:29:19.200 seem to rise to the level of everybody's curiosity when it comes to Giro protection. Let's talk a little bit about

01:29:27.320 all of them, but let's focus on metformin. So for folks coming to this who have heard of metformin, because I think

01:29:33.900 at this point many people have, but don't know much more than that. Can you give a brief background of

01:29:39.520 what metformin is, how long it's been around, what it's historically been used for, and of course

01:29:44.680 we'll talk a little bit more about why we're going to talk about it. Oh, so boring. I also get those

01:29:49.740 questions. What's the distribution, by the way? What do you mostly get asked about while we're on that

01:29:55.360 subject? So the most common question is, I hear you're doing a metformin study. Can I volunteer?

01:30:02.540 Can you assure me I'm not on placebo? Yeah, I was just about to say, are you sure you want to volunteer?

01:30:08.320 Okay, yeah. By the way, I'll tell you something really interesting. Because I'm leading this study,

01:30:14.620 I'm not pushing metformin. Again, the difference between advocate and scientist. I'm doing this study,

01:30:21.400 okay? I believe in that, but I'm doing this study. So I cannot tell you that it's good. So I'm not going

01:30:27.200 to prescribe to anyone. Even to my friends, I'm not going to prescribe. But I have a method that works

01:30:34.260 100% of the time. There are two papers that I wrote, one I think in 2016 about metformin and TAME

01:30:42.780 and the clinical data, and one in 2020 about actually the mechanisms of action that was in,

01:30:50.340 both of them were in cell metabolism, but the mechanism of action, the fact that metformin

01:30:54.780 actually hits all the hallmarks of aging, miraculously. I'll explain it later. And I'm

01:31:01.140 saying to those people, send it to your doctors and say, ask the doctor, what do you think about

01:31:08.500 me taking metformin? And there are one of two things that always happen. One could be that the

01:31:15.260 doctor reads the paper and said, oh shit, I should be on metformin myself. Of course, I'll give 0.99

01:31:20.900 metformin. Or the other is the doctor said, I'm busy, but you know, metformin is a safe drug and

01:31:28.040 stuff. Let me just give them metformin. And that's the end of that. And that's 100% success. Everybody

01:31:33.820 can get metformin. Now I know that I'll get emails asking for those papers. I'll put it in your box.

01:31:40.760 We will absolutely link to those papers in the show notes. Yeah.

01:31:43.180 But I think what people don't realize is that metformin is an extract of the French lilac.

01:31:50.740 Some people say it's nutraceutical. It is modified and it is a drug.

01:31:54.980 The three best drugs in the world are nutraceuticals, rapamycin, statins, metformin. There you go.

01:32:00.980 Right. But you need prescriptions. But also look, one of the modification was fenformin. It's an early

01:32:07.100 form of metformin that kind of was a little too potent, too potent on the mitochondria and

01:32:15.140 caused the trouble. But we're over that for 80 years. But metformin was used initially to treat

01:32:21.680 the flu and malaria and inflammatory diseases. And it's at that time that people noticed that people

01:32:30.880 with type 2 diabetes have lower glucose when they took metformin. So the whole metformin moved to

01:32:36.680 diabetes. And that was about the 1950s or 60s, right?

01:32:40.740 Right. 1950s and 60s. Interestingly, in the United States, it was approved only in 1993 or so. I came

01:32:51.520 as a fellow to Ralph DeFronzo at Yale. And my mission was to describe the mechanisms of action of metformin

01:33:01.480 in humans. I did insulin clamps. And I'm the first that described that metformin decreased hepatic

01:33:07.880 loose production rather than increase insulin sensitivity in the periphery. It's serendipitous. I didn't

01:33:13.780 think about it as aging. I came to Ralph DeFronzo and I did aging in another way to insulin resistant and

01:33:20.180 aging, but not metformin. But the point is there are billions, billions years of use. Remember, every patient

01:33:27.560 takes it takes it for years. And there are people on metformin who are 90 years old. I actually know

01:33:32.120 them. So there are billions use of year. And I cannot think of a drug with better safety record.

01:33:40.040 If there are any side effects to metformin, they happen usually in the first week of use. And that's

01:33:46.040 even if you didn't take them the way you should, which is small doses with food. And when I say with

01:33:52.360 food after the first bite, when your stomach is full, and then usually there is no side effects,

01:33:58.320 but three to 5% of people and maybe more with elderly have diarrhea that doesn't stop. And we

01:34:05.900 basically stop metformin in those people to stop diarrhea. By the way, in TAME study, we're going to

01:34:13.720 follow those patients that were sensitive to metformin and see if there's something unique about

01:34:20.080 them. That's one of the things we hope to do. Well, we're definitely going to talk about TAME. 1.00

01:34:25.080 So let's pick up the story 10 years ago. All of a sudden, people start to notice things. People start

01:34:32.800 to notice that diabetics who take metformin, when compared to diabetics who don't take metformin,

01:34:40.420 do better. And when I say do better, I mean they have lower mortality from all causes and lower

01:34:47.640 mortality from very specific causes. And if you look at it in some extremes, if you compare the

01:34:54.640 patients taking metformin to the patients taking insulin, the difference could be 50-80% difference

01:35:01.520 in mortality. Is that a fair assessment of where curiosity began to peak that this wasn't just a

01:35:08.160 good diabetes drug, but it might have some protection against aging? Yes, absolutely. We should just know

01:35:15.040 that's the base of gerotherapeutic. It has to be a drug that has effect beyond its disease. Whether

01:35:22.760 it's on other diseases, whether it's on overall mortality, not a disease-specific mortality,

01:35:28.880 that's when you have to think this drug is a gerotherapeutics. And that's exactly where metformin

01:35:36.140 fell. I would just add one thing. Look, some of those studies are association studies,

01:35:42.860 but some of the studies were clinical studies. The Diabetes Prevention Program, the DPP,

01:35:50.600 is an NIH-funded study where metformin was one arm. There was also an arm of lifestyle changes.

01:35:58.300 Both of them were preventing diabetes in about 30%. And by the way, the study was stopped early because

01:36:05.300 it was significant after four years, although the study was funded for five years.

01:36:10.560 Now, I do want to talk about that study near, and some of this is, of course, you and I have

01:36:15.600 already talked about, but I think it's just good for people to be able to hear our discussions on

01:36:19.340 this. One of the challenges of these studies, and I think this is frankly true of every single study

01:36:24.500 I've looked at for metformin, is when you squint your eye at the study, there's always a problem.

01:36:30.340 And in the case of the DPP, the problem is the patients in the metformin arm could easily fall out

01:36:37.020 of that arm and therefore not be counted. So if they progressed to a need for medication beyond

01:36:42.880 metformin, or if they couldn't tolerate metformin, or if they were not compliant with metformin,

01:36:48.300 they were no longer counted. And this is true not just of the DPP, but of course, when you compare

01:36:53.500 the metformin patients to the other patients who were taking other diabetes drugs, you always have

01:37:00.800 this potential confounder, which is you are disproportionately selecting the healthiest people,

01:37:07.940 which are the people who, A, can be compliant with medications, which might say much more about their

01:37:15.280 behaviors outside of medications, whose diabetes is maybe just mild enough that it's always kept at

01:37:21.960 bay with metformin, never requiring other medications, including insulin. So am I correct in saying

01:37:30.000 we don't, to date, and TAME we will talk about in a moment, but we don't, to date, have any clean

01:37:36.280 example of a study that demonstrates metformin's giroprotection in humans? First of all, you're

01:37:44.100 right. You can say it on many studies, but it's certainly true with diabetes. Diabetes is a problem

01:37:50.460 because it's a progressive disease no matter what. And you can go back to the data and show whether the

01:37:57.320 people were metformin from the beginning versus other have done better and whether it stops. And

01:38:03.120 there's a lot of things, but first of all, you're right. You're also right about, we had this discussion

01:38:08.840 about the mortality data and you pointed out that maybe the control were not controlled enough for

01:38:16.200 getting them out for some reason. You watch them all the time. You're absolutely right. Even the clinical

01:38:22.960 studies are not perfect studies, but there are still enough of clinical studies or small studies

01:38:30.260 that gives you the confidence. For example, there are two studies on people with mild cognitive

01:38:37.900 impairment that were treated with metformin, one for half a year and one for one year, and some of the

01:38:43.580 outcomes have changed. And there is no difference in how they were treated. In other words, they didn't

01:38:49.620 have diabetes. So it's not that they switched to other medication. So there are lots of examples

01:38:55.460 like that. But you are right. If this was compelling on its own, you could argue we wouldn't have to have

01:39:03.440 TAME even. And really, TAME is not that there were not studies for each one of the diseases, but there was

01:39:11.260 no studies to be agnostic of the diseases. We don't care. Look, we're targeting aging. We don't care

01:39:19.380 what disease you have. And we don't care which disease you're going to get. If you're obese and

01:39:25.000 your mother is diabetes, you're going to get diabetic next. We have to think in geroscience that

01:39:30.960 aging is going to drive your next disease. And therefore, it's the cluster that is going to

01:39:37.240 count. And we're counting the clusters. By the way, we had this discussion about mortality.

01:39:43.660 Mortality, you get a point for mortality, just like you get a point for a disease.

01:39:49.220 And the important thing of the cluster is that we cannot do TAME study. Imagine we do a TAME study, 0.98

01:39:57.720 and in two years, cardiovascular disease comes up as significant.

01:40:04.460 As in significant reduction of cardiovascular mortality, yeah.

01:40:08.120 And the FDA will say, hey, we have to stop this study because we cannot go with the study when with

01:40:13.440 placebo. And it will ruin it for us. So the whole problem of the statistics of TAME is to make sure 0.71

01:40:21.600 that we're not getting to any significance in any disease, just to trance. By the way, the one way to stop

01:40:28.460 the study is if mortality is significant. That will trigger a stoppage of the study. Otherwise, what will

01:40:36.080 stop the study is the integral approach of the cluster of disease.

01:40:41.480 So statistically, your biggest mistake here is overpowering the study for mortality, and therefore,

01:40:49.200 appropriately powering the study for subsets of mortality. Yes?

01:40:54.000 Subsets of mortality? You mean diseases?

01:40:56.200 Disease-specific mortality is what I mean. Yes.

01:40:58.500 Yeah. Well, let's talk a little bit about TAME. This is something you and I spoke about, 0.89

01:41:03.000 God, two years ago. So this is a study that is going to look at people who do not have type 2 diabetes?

01:41:12.280 Right. It's an exclusion criteria.

01:41:14.680 And they are over the age of 65?

01:41:17.740 65 to 79.

01:41:19.840 Okay. So 65 to 79, no type 2 diabetes. Any other exclusion criteria?

01:41:25.160 Yes. But it's not important for our discussion. So if you have cancer in the last year,

01:41:31.260 things like that, you know, very specific.

01:41:33.680 Yeah. Okay. But again, these people are going to be taking statins. They're going to be

01:41:38.520 taking medication for blood pressure. Some of them will be overweight. Some of them will be normal

01:41:43.060 weight, presumably.

01:41:44.560 Right. The point here, though, is that, look, we don't want to recruit a bunch of centenarians

01:41:50.100 to be in our study.

01:41:52.120 Yeah. You don't want to recruit future centenarians here. It's not going to be,

01:41:54.960 you're not going to get an answer.

01:41:56.180 So they have to have something. For example, walking speed less than 6 meter per second

01:42:01.580 is an inclusion criteria. You have to have something that shows you age.

01:42:06.460 I see. So you don't want exceptionally fit people in this group.

01:42:09.920 Right.

01:42:10.260 You know, it's almost like you want to take the patient population that Predamed started with, 0.97

01:42:15.780 because this was a primary prevention study for cardiovascular disease that found a statistical

01:42:22.180 significant difference in under five years. I think it was four and a half years. They expected

01:42:26.060 to go for seven. You're really looking for that type of population. You really do want people who

01:42:30.740 are going to, I mean, it's morbid to say this, but you're looking for people whose risk of death in

01:42:34.640 the next five years is high enough that you're going to move the needle. Now, the risk of that is

01:42:39.960 you get a negative study, which means there is no difference in all cause mortality or even

01:42:46.560 disease specific mortality. And the counter argument might be you started too late. That's

01:42:51.540 like applying the brakes on a car that's driving towards a cliff when it's only 20 feet from the

01:42:57.180 cliff. Should this be a longer study where you start this at people when they're 50 and your five

01:43:03.820 year mortality expectation is very low? Again, I'm not saying this can be done because that's a very

01:43:08.460 expensive study, but it is a risk here, correct? There are two arguments here. First of all, we

01:43:14.160 needed to do, look, to start a study at 50 where you have to show mortality is a 20 year study. We

01:43:22.520 cannot afford it. So we needed to start it when people are starting to accumulate disease in order

01:43:28.600 for us to have lots of events. Our hypothesis is that the aging part of your biology doesn't stop

01:43:37.940 working when you had your first disease. It's still going to get your next disease. If we think

01:43:43.440 biologically like that, we should be able to intervene as we do with animals quite late also.

01:43:50.800 That's one thing. The second thing is there are metformin studies which included elderly people. For

01:43:57.520 example, the DPP, the DPP, by the way, got 20% funding from the NIA in order to include 20%

01:44:07.940 of the subjects over the age of 65, which they didn't. They had 20% over the age of 60, but

01:44:15.460 there's still people over the age of 65. And their results were similar in prevention diabetes to

01:44:22.560 younger people. It's an example, but we have several other studies that tells us that metformin will

01:44:29.640 still target aging even if you started it. It's not that the first disease cancelled.

01:44:34.760 Yeah. It just means that you have a lower period to apply the brakes on. And it suggests that if

01:44:42.060 TAME shows a reduction in all-cause mortality in a subset of people so old, in quotes, when I say so

01:44:49.460 old, meaning in five years is really what I'm saying, it would suggest biologically that there

01:44:55.320 would be a benefit to starting sooner. But of course, then the question goes back to our original

01:44:59.740 discussion. How soon? So Peter, you said an endpoint of mortality in TAME. There could be only trend of

01:45:07.020 mortality. So I'm sorry, I misunderstood. What is the primary outcome of TAME? It's the cluster of 0.99

01:45:13.340 bunch of cardiovascular, cancer, cognitive, and mortality. It's the cluster. You get one point for

01:45:21.480 each one. I see. So let's define them again. So you get one point if you die. Right. You get one point if

01:45:28.620 you have a major adverse cardiac event. One point if you have a cancer occurrence or recurrence.

01:45:33.960 Except skin cancer, right? Okay. And MCI or dementia. Got it. Do you have any points for health

01:45:40.620 span outcomes such as frailty, falling, breaking hips? This is in the 70 million program. Let me explain.

01:45:49.640 We would like to start longitudinal study where we capture a lot of the other health span issues,

01:45:55.420 hospitalization and function and depression and all that. We are still powered to do it at the end

01:46:03.360 of the study or as the study goes on. We probably wouldn't have money to do it longitudinally through

01:46:10.760 the whole study, but we have power at the end of the study to see if there are difference in

01:46:15.220 frailty, et cetera. Just a statistical question. Why is mortality given equal

01:46:21.060 weighting to disease occurrence? Is there a reason that mortality wouldn't be three points to one point

01:46:27.840 for MCI, mace, and cancer? Well, it was questioned for each one of them. And we decided just that we

01:46:36.940 cannot rationalize that. Everyone is an outcome. We basically said those are the outcomes. We don't know

01:46:44.660 which one you're going. You're going to get an outcome. You're going to get the point for that.

01:46:48.980 By the way, it's the time until, of course, you're moving health span.

01:46:53.660 Yeah. Now, earlier you spoke about how animal models are not really great for Alzheimer's disease,

01:47:02.280 cancer, and cardiovascular disease, which I think any listener of this podcast is well aware of those

01:47:07.880 limitations. But you note that, look, animal models are pretty darn good for aging given how conserved

01:47:14.440 it is. And yet one of the challenges metformin has had is in animal models. Now, Rich Miller was on this

01:47:21.240 podcast a while ago. We spoke about the ITP where metformin was successful in the ITP when combined with

01:47:28.840 rapamycin, but alone was not. Again, I think the ITP is a very rigorous type of experiment.

01:47:36.620 I know you've probably thought a lot about this. What explanation do you think exists for why

01:47:43.060 metformin did not succeed in isolation in the ITP study, which I'll just take for a moment to

01:47:48.440 explain to people the ITP. If you haven't listened to that podcast with Rich Miller, you should go and

01:47:52.620 do so. We'll link to the appropriate section here. But these are studies that are conducted using a

01:47:59.160 particularly good model of mouse that is kind of less troubled by the usual difficulties my studies

01:48:05.680 have. It's also done independently at three separate laboratories concurrently. And then the

01:48:11.000 final thing is they look at all-cause mortality. So tell us why you think that we could be misled by

01:48:17.660 the ITP. So let me make a big picture statement now. It is possible that some of the drugs that had

01:48:26.640 mild effect in ITP will have much bigger effect in humans. And the other, maybe rapamycin is going to

01:48:34.480 be not as effective in humans as it is in animals. We just have to accept it. Rich Miller, if it's not

01:48:41.580 true in mice, it's not true. Well, which I don't think Rich says, by the way. I've never heard Rich say

01:48:46.900 that. Rich just says, this is what it is in this model. I'm kidding. He started saying that initially,

01:48:52.240 and he took it back, of course. I think one of the problems with animals is the dosage that they're

01:48:59.060 using. The dosage in animal where, if you look back, has 0.1% in a solution to 1%. 1% is deadly

01:49:09.600 because, listen, metformin, after all, is a weak cyanide. It binds to a complex in the mitochondria.

01:49:17.900 It probably affects complex 1 and also complex 3. It's not totally clear, but if you give too much,

01:49:24.740 there is a trade-off. So you have to give right. And I think that the studies in animals, except,

01:49:32.160 by the way, mice, it's not true for nematodes. And there are several other fish. There are several

01:49:38.020 other animals that live longer and healthier with metformin. But I don't think that 0.1%

01:49:44.140 is really the appropriate dosage. What do you think is the... So you think...

01:49:49.040 I don't know. I don't know. There's no dosage response.

01:49:52.600 So you're saying 1% is clearly problematic.

01:49:55.840 Yeah. I think 0.2% maybe will be better if we want to optimize that.

01:50:01.140 There's a paper that gets a lot of attention. I think it's 2013, the Rafael de Cabo paper,

01:50:06.300 which gets touted as, oh, look at metformin. I got to be honest with you. If you look at that paper,

01:50:12.520 I don't know how the editors let the title of that paper slide with the actual data.

01:50:17.620 Let's call a spade a spade. You had two groups of mice, one of them getting 0.1%,

01:50:22.980 one of them getting 1%. The group getting 1% were assassinated, to your point. I mean,

01:50:29.600 these animals were killed by the toxicity of metformin. The group getting 0.1% lived a staggering

01:50:37.540 4% longer. I mean, it was basically a null trial that was touted as the definitive animal study for

01:50:45.800 why metformin works. It's like an Onion article when you dissociate the title from the actual data.

01:50:52.100 Look, the average, when I took and make a table with all the longevity data,

01:50:58.100 the effect of metformin across studies were between 7% and 10%. Not a huge effect.

01:51:03.620 Well, that's pretty good. 7% to 10% in animal studies, if it's consistent, would be pretty

01:51:08.740 good. But the nice thing with metformin, the effects on healthspan is much bigger.

01:51:13.680 But what are the healthspan effects besides the metabolic effects?

01:51:17.780 Prevention of cancers. Isn't that effectively already captured in lifespan?

01:51:22.900 Rich Miller will give you a better thing. I mean, the problem for us with animals,

01:51:28.060 by the way, it's the problem with centenarians also. They die with cancers.

01:51:32.700 We don't know if they die from cancers.

01:51:35.920 Right. It's like the prostate cancer issue.

01:51:37.960 Right. As much as we say in humans, it's cardiovascular disease, cancer is probably

01:51:43.300 the leading disease for death for animals. And it's true. The other point, which I think

01:51:49.220 I'm trying to get to the bottom of that, but when you gave metformin to animals in the ITP,

01:51:55.320 in two centers, they had 10%, you know, one was 11 and one was 9% increase in longevity. But Rich

01:52:04.380 Miller's point was minus 2%. And I just wonder if they're all actually delivering the same dose

01:52:11.980 and what's going on. And so it wasn't significant, but it actually was in two centers, there was a 10%

01:52:18.760 effect.

01:52:19.320 Yeah. There's always an issue that could be methodologic. Do you think that the animal studies

01:52:24.820 in metformin are largely irrelevant once we have TAME underway? One of the reasons that we rely so 1.00

01:52:31.920 heavily on all of the animal data for rapamycin and why the ITP, which has studied rapamycin five

01:52:38.560 ways to Sunday, starting at late in life, starting at early in life, high dose, low dose, pulsing it,

01:52:44.040 not pulsing it, continuous. I mean, every way you study it, rapamycin works. And we have to look at

01:52:51.100 that and really pay attention to it because I don't think we're going to get the human trial of

01:52:56.120 rapamycin. I don't know that we're going to get the TAME equivalent of rapamycin. And if we are,

01:52:59.980 it's not going to be for some time. But now that we're moving to a human clinical trial of metformin,

01:53:06.620 should we even care about this question in mice? Does it matter if it's 0.1% or 0.2%?

01:53:11.820 No. What I'm saying is that the preliminary data from humans overall, as you said, there are

01:53:18.400 problems everywhere, but it's the same story. It's a 20-30% effect on each study, no matter how you

01:53:24.780 look at it. It's a really very impressive studies and there's no better studies that came with

01:53:30.320 different results. So I think once TAME is there, I don't know that we need to go into animal studies 0.69

01:53:38.440 and discover more about metformin at this time. Will TAME allow you on its current budget,

01:53:46.140 which I think is 50 million, is the bare bones budget. Sounds crazy, right? On the $50 million

01:53:52.600 budget, how much will you be able to look at mitochondrial function and omics that are associated

01:54:01.400 with other deeper markers that go beyond the hard outcomes that feed into your primary outcome?

01:54:09.460 Well, the NIA has given us a grant that is now delayed to do basically the biomarker part of TAME.

01:54:18.960 So even with this 50 million and with the NIH money, with AFAR, we're going to store lots of plasma,

01:54:29.180 DNA, and other resources like cells and other things that will then be open for omics studies. And we

01:54:37.660 wrote a beautiful grant about that. In part, it was preliminary data from our centenarian studies on

01:54:44.660 proteomics.

01:54:45.620 By the way, I just realized I used the word omics. Do you want to tell people what the full suite of

01:54:51.100 omics means so that people listening to this know what we're talking about?

01:54:54.860 Yeah. And it's really back to something that I said before when I talk about, if I said my study,

01:55:01.280 I mean our study. And when I talk about teams, I'm talking about teams that have computational

01:55:07.000 capacity. I talked about Zheng Dongdeng that did a lot of my study about Yuxin Su that is doing

01:55:12.660 functional genetics. I'm talking about Sophia Millman, who are doing other studies. We're a

01:55:18.140 really big team. And the reason to have a big team is very simple. All I knew when I started my

01:55:24.740 training was insulin. I knew insulin. I know insulin signaling. Now I'm losing billions of data

01:55:32.020 sets that are under my teeth. If you think of genetics only, whole exome sequencing for 3,000

01:55:37.600 people. Or my proteomics, which is 5,000 protein for 1,000 people. It's all big data. And in order

01:55:46.560 to do something with the big data, you have to ask the right questions. Because what we noticed,

01:55:53.040 the first paper that came out of the UK Biobank for Aging said that longevity is all about

01:56:00.400 assortative mating. Now, yes, if you're a smoker, you may marry a smoker. If you're obese, you will 0.98

01:56:06.820 marry obese. If you're poor, you will marry poor. It all has consequence of health. But we knew that 0.99

01:56:12.500 without the UK Biobank, we knew that obesity is a risk for diabetes without the Biobank too. So you

01:56:21.580 have to ask the right question. And it's really only asking the right question that you get the

01:56:27.400 answer. So the question that is very important for us, there are two questions. What are the

01:56:34.320 biomarkers for aging? How can we do a test at 50 years old and know if we're 40 or we're 60? If we're

01:56:41.080 40, we skip colonoscopy, okay? If we're 60, we have to do something about it already. And the second part,

01:56:48.760 it's more important for me is biomarkers that change with treatment. We want to make sure that

01:56:54.800 when we try all those treatments that we have in two, three months and answer, are they likely to

01:57:00.760 work? And then maybe we can get to phase two and phase three trial, but we need something more

01:57:05.360 immediate. And that's what TAME will try to provide. Now, in my study, we did this proteomic,

01:57:12.960 and it was really incredible for many reasons. And I'll give you a tidbit, but we have those

01:57:21.060 thousand people and 5,000 proteins. And we asked what changed between the ages 65 and 95. This didn't

01:57:28.840 include our centenarians. And the answer is a lot of proteins are changing, but this is, I think,

01:57:34.280 the most important part. A lot of the proteins that we're capturing, by the way, number one is IGF

01:57:40.000 related proteins. Number one that comes up even in the proteomic, not only in the genomics.

01:57:46.260 But then a lot of what you see is breakdown. You see breakdown of collagen. You see degranulation of

01:57:52.820 thrombocytes. You see breakdown of extracellular metrics, lots of things like that. And at first I

01:57:59.260 said, okay, that doesn't tell me anything until I thought, no matter what we do, we have to stop the

01:58:05.260 breakdown. This is probably going to be the best marker for any treatment. You just stop the

01:58:11.780 breakdown. It's funny. That seems so obvious, doesn't it? This gets back to what we talked

01:58:17.100 about earlier. You want to really know what health span is? It's not some nonsense definition

01:58:22.000 that the NIA gives us about freedom from disability and disease. Who cares? Who cares if you have cancer,

01:58:29.160 but you're living in remission with it, and you can climb three flights of stairs with a bag of

01:58:34.140 groceries in your hands. That's living. It's, yeah, when your collagen is breaking down,

01:58:40.160 your knees aren't working. I mean, this is the essence of what we're talking about.

01:58:46.300 I think the proteomic story is very interesting, and I hope it gets its appropriate funding.

01:58:51.960 By the way, the other part of proteomic, there are a few other aspects that are important. The proteins

01:58:57.740 that are 10 to the minus 80 significance, the two top ones, when you express them in animal,

01:59:04.280 they live long. In other words, they're protective protein. When you get those proteins, you don't know

01:59:10.820 which are protective and which are causing problems. And that's a challenge because we are saying,

01:59:17.920 let's bring it to normal. No, you don't want to lower the protective mechanism.

01:59:21.600 The third thing that is really cool is the proteome of females is much more stable. In other words,

01:59:31.000 it's only half of the proteins are significantly changing in women than in men between those ages.

01:59:38.400 So you'd need to look at actually different biomarkers for women and men. And by the way,

01:59:45.760 this was Eureka moment for all of us. From ITP and on, female and males have different biology as far

01:59:52.840 as aging. Some of the gerotherapeutics work for both, like metformin and rapamycin, maybe not equally,

01:59:59.440 but work for both. And some are not. They're totally sex-specific.

02:00:04.200 Yeah. Look at the 17-beta estradiol. I mean, that's a remarkable story.

02:00:08.560 And then the last thing that is interesting, those 1,000 people that I had, 500 are what we called opus.

02:00:15.580 They are offspring of parents with usual survival, no longevity in the family. And one are opel,

02:00:21.400 offspring of parents with exceptional longevity. The offspring of parents with exceptional longevity

02:00:26.420 have half of the biomarkers of the control group because they're younger. They'll get those

02:00:33.500 biomarkers later. So look, I don't know that methylation is going to be the best biomarker

02:00:41.560 for treatment. I think methylations are complicated. They're kind of stable. They are going up,

02:00:51.300 they're going down, but they're kind of stable. I think proteomic is going to be a better biomarker.

02:00:57.380 I think metabolomic is very complicated because it so depends on how you establish the sample.

02:01:06.140 If somebody was fasting less or more, if their insulin level was less or more, you're done.

02:01:12.100 Yeah. That's sort of my problem with these biomarkers. And you know, I mean,

02:01:16.360 you've probably heard me rail on these biological clocks. I've never seen a worse biomarker than a

02:01:22.260 biological clock. These are so easy to manipulate and game. In fact, I need to do this just to

02:01:28.540 demonstrate it, but I'm too lazy. Is get five copies of a biologic clock and do five different

02:01:35.420 self-experiments. And I promise you I can change my biologic age by 20 years. When fasting glucose

02:01:41.320 and vitamin D level factor into a biologic clock, I'm sorry, that's useless. It might be valuable at

02:01:49.040 the population level. It is as useful as a warm bucket of hamster vomit at the individual level.

02:01:55.260 So we have to have things that can't just change on a day-to-day basis. You're fasting glucose.

02:02:01.520 The difference between 95 and 105 has everything to do with the meal you had last night. How much

02:02:07.280 cortisol was coursing through your veins. If your water heater broke that night, if you got startled

02:02:11.900 in the morning before you got your blood draw. I mean, I'm just amazed near at the attention. I'm

02:02:17.460 sorry. I'm going off on this tangent, but I'm amazed at the attention that is being given to these

02:02:21.880 clocks and the entire cottage industry of businesses that are spinning out this type of,

02:02:28.340 I think this is complete buffoonery. So Peter, I have more positive outlook of life,

02:02:34.900 maybe because I'm so much older than you. So much older, yes. But at least it's good for the economy.

02:02:42.740 Yeah, exactly. Hey, Theranos was good for the economy until it wasn't, right? The other thing is

02:02:49.040 even with methylation, and I thought methylation was very interesting until the twin astronaut brother

02:02:54.800 experiment. Then I stopped thinking it was very interesting. Remember you had the twins, right?

02:02:59.840 The one astronaut went to the ISS for a year. His twin brother stayed down here. You then looked

02:03:05.000 at methylation clocks of them after a year, and they were vastly different. And then three days later,

02:03:11.300 after the twin brother was back on earth for three days, they repeated the test and it was right

02:03:15.200 back to his twin brother. I mean, come on, how biologically relevant can this be when it changes like

02:03:22.420 that? Right. Have you had Morgan Levine? You know about her. I do, of course. And I haven't had Steve

02:03:28.420 Horvath on as well. I've communicated with him a little bit via email, and I would certainly love

02:03:32.900 the opportunity to speak with him about this. But there's probably something interesting there. I'm

02:03:37.020 just having a hard time seeing that. Because again, what's the purpose of doing this? We have to

02:03:40.980 never lose sight of the purpose of biomarkers of aging. It's not so that you have bragging rights

02:03:45.820 that, oh, I just took this test and it says I'm 37 when I'm 50. No, the only reason you do this test

02:03:51.860 is if it can guide therapeutics. The only reason you do this test is if it helps you determine

02:03:56.840 if I'm doing this thing, is it making me better or worse? And my fear is I haven't seen a single

02:04:03.560 test that is validated at a level that would give me even a modicum of confidence in that. Have you?

02:04:10.680 No, no. I mean, again, I'm saying this in a very deliberately provocative way because I'm

02:04:14.840 waiting for you to say, oh, no, no, Peter, you need to look at this test. No. So I'm telling you,

02:04:19.340 I don't want to be a spokesperson. All I'm saying is methylation, because it's not mechanistic,

02:04:25.440 I don't think it's going to help us when we come to gerotherapeutics, even if they're examples.

02:04:31.120 I don't know what the examples mean, just like the example you gave. I don't know what it means.

02:04:36.320 Morgan Levine and Steve Horvath, by the way, they're really good and they're responsible and

02:04:40.480 they understand exactly what they did and what they didn't do. And Morgan Levine has a mechanistic

02:04:47.460 way of looking at epigenetics. In other words, remind me what the inputs are to the Levine clock.

02:04:54.440 I'm not going to get the difference between the clock I want to talk about. The way they're moving.

02:04:59.720 Okay.

02:05:00.440 So one thing that Morgan Levine is doing is she's saying, well, if methylation has a function,

02:05:06.060 then I want to cause methylation and see that there's change in the expression of those genes and

02:05:12.380 that those genes are something that we know are relevant to aging, et cetera, et cetera, and then

02:05:18.040 build a clock that is mechanistic. Because the only way this clock was built was to measure

02:05:24.200 chronological age.

02:05:26.020 Will you guys be using any of the biologic clocks through TAME as well?

02:05:30.320 So what's going to happen with TAME is there are two things. We are going to have a committee that 0.64

02:05:36.440 with time and we didn't decide we're going to do the right epigenetic test. And by the way,

02:05:42.340 we're going to do epigenetic, not the clock epigenetics, but enough epigenetic data on a genome

02:05:48.920 wide in order to then discover which are the ones that have changed with treatment that way.

02:05:56.980 But it would also be good, Nir, to have a clock running in parallel, wouldn't it? Because if you can

02:06:02.220 do this longitudinally with multiple data points, you sample people every three to six months

02:06:07.120 with multiple different versions of these clocks, and then you can look at how well those clocks

02:06:13.420 predicted the actual outcomes you will see over five to seven years. That strikes me as a very

02:06:18.460 valuable way to get at something that no one is doing today. Because in other words, take advantage

02:06:23.760 of the fact that you have two things going for you that nobody has today, which is a very large

02:06:30.620 longitudinal data set with hard outcomes, and an intervention that is relevant and interesting to

02:06:37.140 study. Right. So let me explain again. If we're doing the methylation scan, we can obtain any clock

02:06:45.140 from them. That's the idea. Yes, that's my point. I mean, you'll have the bio. I'm just saying,

02:06:49.560 just make sure you're collecting all of the data that goes into multiple clocks in a concurrent manner,

02:06:55.580 not just at a time A and time B. Right. The second point is the NIA is going to have RFA

02:07:02.380 for people who want to have sub-projects in relation to what we collected. Those RFAs can be in centers.

02:07:09.740 For example, there's a center that wants to look at skin aging. If 250 people are enough for a project,

02:07:16.360 then you have a center. If you need 500, you'll take two centers. In other words, there'll be a process

02:07:21.980 by which application will be reviewed, and people can come in and look at a variety of things that

02:07:28.680 they're interested in. Can I, if it's not too late, make some crazy suggestions? And I don't know if

02:07:34.120 this changes your IRB, but are you going to be doing any bone marrow sampling, even at time beginning

02:07:39.500 and end? Because I'll tell you, one of the other things I'd really, really love to see with a study of

02:07:45.860 this magnitude is, are you impacting immune function, specifically memory, T and B cell function?

02:07:53.560 Obviously, this is a topic that's very interesting and germane to COVID at the moment, but I think it

02:07:58.320 goes well beyond COVID. And it would be very interesting to know if metformin is having any

02:08:03.580 impact on immune function, specifically memory function, because it doesn't just play an important

02:08:09.740 role in infections. It plays a very important role in cancer.

02:08:12.480 So look, the answer is mundane. We're taking elderly people into a study, we're not going 1.00

02:08:18.560 to torture them. We're not going to have any excuse for them to leave us. It's the same with

02:08:23.920 our centenarians. By the way, what we're offering the offspring, we're offering MRI of their brain,

02:08:29.260 we're offering them coronary CT, we're offering things, but we're not doing biopsies. That's in

02:08:36.260 another study.

02:08:37.680 Are you doing brain MRI and tame?

02:08:39.820 No.

02:08:40.360 Just because of cost?

02:08:41.240 Yeah. Look, we are powered enough to do things eventually, because it's not going to happen

02:08:49.260 before and after. We're not going to have everything from before and after. This is only the things

02:08:54.660 that we have to measure for the outcomes.

02:08:57.980 What about exercise function? One of the things I want to talk about before we leave metformin

02:09:02.500 is the impact metformin may have as a negative impact on cardiorespiratory fitness.

02:09:08.040 Right. So no, the answer is no. We're not going to have anything like that.

02:09:13.460 Because of cost?

02:09:14.740 No, because of how much time can you get an elderly to spend time with you every three 0.90

02:09:23.300 months when really what you want to make sure is that they're taking their drugs. That's why.

02:09:29.980 How many centers? You said 3,500 subjects?

02:09:32.560 Yeah. Well, we have currently, before we get okay to expand a bit, we have 14 centers and about

02:09:39.420 250 people in each centers.

02:09:42.680 All in the U.S.?

02:09:43.840 All in the U.S. 0.98

02:09:45.020 I don't know. I think that would be potentially a lost opportunity. Not that you have to do

02:09:49.940 CPET testing every three months, but would doing it every two years be an issue? At a minimum,

02:09:57.680 I would look at fasting lactate levels or resting lactate levels.

02:10:02.060 Everybody has suggestion for TAME. 0.64

02:10:04.420 I know. Everybody has their pet idea for you. Yeah, of course. Yeah.

02:10:07.660 Right. And by no way I'm saying it's not important. I'm just saying, what are we practically

02:10:13.600 doing? Because at the end, we want the FDA. Let me state it again. I'm doing TAME, not because I

02:10:20.240 don't believe in metformin because we need to have a target that's similar to aging. That's the reason.

02:10:28.340 So let's talk a little bit about this other issue, which is, I've talked with you about my

02:10:32.700 experience, right? So I took metformin for eight years or something like that. But three years ago,

02:10:38.540 when I really began checking and constantly monitoring my lactate levels, both in and out

02:10:45.240 of exercise, it became clear to me that my lactate levels were too high. So my fasting lactate level

02:10:52.300 was typically above one millimole. It was between one and two millimole. And as a result of that,

02:10:58.280 my perceived mitochondrial efficiency was lower. And there's ample data to suggest that fasting

02:11:05.540 lactate level, now, of course, this is not necessarily in people taking metformin, but if you

02:11:09.720 take lactate levels in people at rest, there's a high association between what that tells you about

02:11:16.180 their general health. So the less healthy an individual is, the higher their lactate level is.

02:11:21.040 So for me, I just said, well, boy, there sure looks like a lot of compelling reasons why metformin

02:11:25.740 could be beneficial. But if I really stopped to look at it, none of the cohorts of people in whom

02:11:32.180 we would infer that look anything like me. These are not people who are exercising constantly,

02:11:38.780 doing all of these other things. So that was the decision for me three years ago. I'm not going

02:11:42.860 to take metformin anymore until I have better data. So now we have a couple of studies that

02:11:47.340 have looked at the impact of metformin on cardiorespiratory fitness, and we see that it is

02:11:52.040 indeed impaired. And then we have studies that look at the impact on metformin of strength training,

02:11:58.640 and we see a mixed response. We see that it does not appear to impact strength gains. It only

02:12:04.960 appears to impact hypertrophy. The good news is we know that strength matters more than hypertrophy

02:12:10.920 in longevity. I had a whole AMA on that, which you probably heard. So I think it was pretty clear

02:12:16.180 that we can say strength matters more. So what do you make about these potential limitations with

02:12:21.660 respect to metformin on cardiorespiratory fitness? And what do you think it says about people who

02:12:25.920 exercise a lot? First of all, I think one point that is very important is when we go from a drug that

02:12:35.920 has certain capabilities to personalized medicine, and all of a sudden it becomes about the person.

02:12:44.160 And there's nothing I can say about what you're observing except that lactate is one of the biomarkers

02:12:50.760 of giving metformin in every patient. In 1987, when I did the metformin study, lactate went from below

02:12:58.460 one to above one in everyone who took metformin. I hope in TAME you are measuring lactate if for no

02:13:04.780 other reason than to determine compliance. Yeah, absolutely. Absolutely. By the way,

02:13:09.140 there is a better test. It's called GDF-15. It's one of those peptides that goes up with aging. It's a

02:13:15.000 protective peptide. It goes by three and a half fold in people who take metformin. So there are other

02:13:20.160 ways to look at it. But this is the point with exercise. And let me make the most important point.

02:13:26.680 We can talk about muscle, but remember, just like with exercise itself, exercise is not only about

02:13:31.940 muscle. It improves brain function. It decreases cancer. Metformin has many other effects. So let's

02:13:39.160 remember that metformin has other effects than its effects on the muscle. As far as the muscle is

02:13:45.020 concerned, what we've done with Charlotte Peterson, which you reviewed two papers of her during the

02:13:51.400 time, we took the biopsies and look at the transcript of the people who were on exercise and metformin

02:13:59.280 versus exercise only, right? That's the only thing we had. And the reason we did that was because in

02:14:07.320 supplement four of her first paper, she showed that the strength of the muscle didn't change,

02:14:15.280 which means that per gram muscle, you did better. And we tried to understand what is the biology.

02:14:22.240 And it became very clear. Look, in order for muscle to grow, it needs to activate mTOR. So mTOR is good

02:14:30.260 for muscle and for muscle growth. It's not good for aging, but it's good for muscle growth. A huge

02:14:36.520 transcripts of mTOR has been decreased by metformin, which was not news for us. On the other hand,

02:14:42.680 there are other transcripts that had to do with inflammation and autophagy and oxidative markers

02:14:50.260 that were in people with metformin and not with exercise. So there were trade-offs. You get less

02:14:57.340 muscle, but the muscle is healthier. Is healthier? Can we really say that? Or can we just say is as strong?

02:15:04.380 It's biologically younger or as strong? Yeah, whatever. They're trade-offs.

02:15:10.460 But how do we say it's biologically younger?

02:15:12.720 Because the change in transcript that we saw is from an old transcript to young transcript.

02:15:19.580 So in our studies, we have biopsies from people who are old and young. And with every treatment,

02:15:25.180 and we've done it for resveratrol, and we've done it for metformin, and we've done it for

02:15:29.420 acarbos and for exercise, and now we're doing it for fasting, we have young and old. And we see

02:15:35.480 if in old with intervention, the transcript goes back to young. And it does.

02:15:42.180 So the exercising people without metformin versus exercising with metformin,

02:15:47.500 the with metformin group had a maintenance of the transcriptome, or it actually declined

02:15:55.260 relative to the other group. No, it's transcript-specific. It's which groups of transcript

02:16:02.240 changed, and did they change either with decrease or increase, or did they change to resemble more

02:16:08.940 young than old? And this was done on the patients in the master's trial?

02:16:13.440 Yes. And what was the duration of that trial?

02:16:16.300 16 weeks, I believe. So what do we think we can learn in four months with respect to the

02:16:22.580 transcript? How much does it apply from what you've learned over a much longer period of time

02:16:27.820 in a non-intervention setting? First of all, the answer is what we learn in four months. We learn

02:16:32.920 what it is in four months, but I think it's a long period of time. It's also a dynamic period of time in

02:16:40.280 this experiment because they were building muscle throughout time. So it is what it is in this

02:16:46.400 experiment. But in this experiment, the transcripts were younger by metformin, and the mTOR was increased

02:16:53.720 by exercise. And this was the trade-offs. And what about in terms of the cardiorespiratory

02:16:59.120 fitness differences? I mean, I guess what I'm getting at is the most plausible explanation here

02:17:03.900 for a blunting of cardiorespiratory fitness would be what you said earlier, which is metformin is a

02:17:10.980 mitochondrial inhibitor. It's a weak one at the doses we take it, but it nevertheless is. So it

02:17:17.200 seems to me that that's the most obvious explanation. That's the place I'd be looking

02:17:21.620 for the fire, given where the smoke is. So do we think that the net benefits of that are probably

02:17:28.740 positive in some people, but in others, for example, those who exercise a lot, it might not be

02:17:36.320 beneficial because they're getting so many of those other benefits of exercise, as you point out?

02:17:40.520 Yeah, look, I'm taking metformin and exercise, but I have a different way I'm thinking about it. For

02:17:46.540 me, we're very personalized. And this is back to the issue of when do we start metformin? Or what is

02:17:54.200 the biological age of those people who exercise and take metformin? And I would say that if you're

02:18:00.260 young or biologically young, I don't think you should take metformin when you exercise at this level.

02:18:06.120 Maybe, Peter, you've been doing it for years. You're probably biologically much younger than

02:18:11.240 most people. I don't know if metformin is for your age and for what you're doing for your health.

02:18:17.860 I can tell you that with metformin and fasting, my exercise capacity has increased significantly. I'm not

02:18:23.700 measuring lactate. I'm not exercising the way you are. So I think those are good discussion. We'll find

02:18:29.960 out eventually who can or who cannot, but we have to make sure that we don't generalize where we're

02:18:38.960 being so specific. Yeah, no, indeed. Again, I said at the outset, I get asked about metformin

02:18:44.620 constantly. And the only people who are asking me for whom it really matters are the patients. I don't

02:18:49.600 particularly care what someone asks me on social media. But it has been a change in my practice over

02:18:54.100 the past few years where I'm really reserving metformin only for people in whom I see an

02:18:59.620 otherwise obvious indication, such as even a trace of insulin resistance, hyperinsulinemia that is not

02:19:07.800 otherwise treated with the right amount of exercise, nutritional changes, sleep, and things like that.

02:19:13.700 So it'll be interesting to see what TAME does and how it can change that practice. But I also worry

02:19:20.240 that it won't fully answer the question for our patient population, because the TAME patient 1.00

02:19:26.560 population is not a very healthy population by definition. If they can't walk a certain speed,

02:19:32.280 they're 65 to 80 years old. You're basically selecting people who we expect to have a bad

02:19:38.060 outcome in five years. So it's a very important question for the world at large. I think it's

02:19:43.580 important that people listening to this understand we may not get the answer we want for them,

02:19:47.100 for the healthy 40-year-old person. Absolutely. Remember, all I want with TAIL is an FDA indication

02:19:53.720 for aging. That's all I want. What do you think that opens the floodgates for? How does that change

02:19:59.480 things now? Assuming that we get there and that we now have an FDA indication for aging, what follows?

02:20:05.720 First of all, pharmaceuticals will jump in. And they wouldn't make many mistakes because I think

02:20:11.460 the biomarkers will be there from TAME and others. There'll be more biomarkers, better biomarkers,

02:20:17.100 so that they could do testing two and three months and find out if their drug is working and then do

02:20:23.480 a phase-free trial. Now, phase-free trial, think of it. For any diabetes treatment, you need 12,000

02:20:30.500 people. For TAME, you need 3,000 people. The study is going to be much, much cheaper.

02:20:37.520 We think. This is still a hypothesis that being tested, right? Is 3,000 the right number? Yeah.

02:20:42.380 If this composite outcome works. Yeah, if it works. Look, the 3,000 is a question of when and not if for

02:20:51.980 me. How many years will it take us? And I hope that we'll actually do the 3,500 and not lose a lot.

02:20:59.700 Also, the question is who's stopping the study? That's another thing. So in other words, that's

02:21:05.020 administrative questions. But the reason we do that is to get an indication that for the FDA is not

02:21:13.200 making aging a disease, which is another big topic. But for us, it's OK, you can call it whatever we

02:21:20.640 want. But if there's a drug that targets variety of age-related disease and mortality, that's for us

02:21:26.740 a gerotherapeutics. And this is what we want to get. All the other questions, when to use, whom to use,

02:21:32.880 you're absolutely right. And the one thing we don't want is to kill anyone on the way to success.

02:21:40.600 We're almost out of time. But one thing I want to just touch on before we go, we've talked a little

02:21:45.140 bit about rapamycin, a lot about metformin. We didn't talk about the other one you and I probably

02:21:49.540 get asked about a lot, which is the NAD precursors. Where are you in your thinking on the efficacy of

02:21:56.120 these as gyroprotective agents? They're in a different class because they're not pharmaceutical,

02:22:00.000 they're nutraceutical. How is your appraisal of the data in that space?

02:22:04.840 As a biologist, I have a problem understanding it. On one hand, you give it to animals and the

02:22:13.480 animals are doing better. Not in the ITP, but not in the single most important animal study. They

02:22:20.480 didn't do better. Yeah. But when you ask people, OK, follow the drug, tell me where it goes,

02:22:27.260 where can you measure it? You cannot measure it anywhere and you cannot measure derivative of

02:22:34.160 that. So I don't know where is the biology. So Joe Bauer, do you know Joe Bauer from UPM?

02:22:40.920 He's a really good NAD biologist. He basically thought the NAD goes, we swallow that. It goes to

02:22:49.420 our microbiome and our microbiome. Sorry, when he says NAD, you mean the NR or NMN?

02:22:55.420 Right. They go to the microbiome and the microbiome either transfer the NAD or does 0.78

02:23:04.020 something. The microbiome itself does something. There's indirect health benefit from a Duffer

02:23:08.900 system. And he actually did a really good study where he discovered that the NAD for the microbiome

02:23:15.820 comes not from the food, but from the gut walls. Everywhere you try to understand what's the

02:23:23.300 biology, I don't get it. And when I don't get the biology, I'm a little bit more worried of what

02:23:31.740 kind of a placebo is it. Actually, one of the studies I wanted to do, those short studies when

02:23:38.220 I get elderly, do placebo control crossover and do biopsies and look at the biology, I wanted to do

02:23:44.520 NAD. And then I said, I just have so little belief here that I'm now doing intermittent fasting

02:23:50.180 one, but uncomfortable with it. Now, saying that, I started taking NMN at one point. And what I noticed

02:23:58.460 is my REM sleep has improved a lot. And I stopped it and my REM sleep wasn't so good. I restarted it

02:24:08.160 and my REM didn't get better again. And then I tried, there's a Japanese company, I don't remember

02:24:16.140 the name off the top of my head, that has really the best preparation. David Sinclair tells me about

02:24:21.340 that, has the best preparation of NMN, $900 a month supply for the low dose, $1,800 for the high dose.

02:24:31.460 I got some of the supplement as a present.

02:24:34.380 More expensive than a PCSK9 inhibitor and rapamycin combined. Okay.

02:24:38.620 Right. And my REM sleep didn't return. So I'm currently not taking anyone. So what am I telling

02:24:46.880 you? I'm telling you, I'm just not convinced. It's not the whole truth. You never know the

02:24:52.500 whole truth, but I don't know enough of the truth to make anything about it. Again, if people want to

02:24:59.660 know, it's good for the economy. I'm not sure that the preparations out there, look, they're very

02:25:05.620 sensitive preparations. If they have a high time shelf, I think you lose a lot of it. I don't know

02:25:13.100 which to recommend. Yeah. I think it's always problematic when your rationale for taking a

02:25:19.040 supplement is that it's a mini stimulus program to the economy through the supplement company.

02:25:25.140 I think as a general rule, if we're going to talk about principles for trying to optimize your health,

02:25:30.520 I would put that very low on the list.

02:25:32.940 Peter, I did many jokes. Do I have to go over and tell you which was a joke and which is not? I don't

02:25:38.780 know. I know it's a joke. This is my joke right back at you. Yeah. Yeah, exactly. No, no. This is

02:25:43.980 my sarcasm coming right back to you. Supplement company stimulus program, not a high on my list

02:25:51.460 reason for it. Anyway, Nir, this has been great. Thank you again for the generosity of your time and

02:25:56.980 behind the scenes, people don't know how much I bug you and how much we interact on all matters that

02:26:02.680 pertain to this stuff. You're one of the few people that I know listens to every single podcast,

02:26:07.660 reads every single newsletter, and at least a quarter of the time sends me a note with something

02:26:13.720 remarkable in it that expands my thinking on a topic. So I want to thank you for that.

02:26:18.040 You're part of the people who spread the gospels, and I think it's so important. In fact,

02:26:23.520 you're one of the messiahs, really.

02:26:25.900 That's an awful stat. I don't want to be a messiah of anything.

02:26:28.440 And I want to come clean. I'm usually not there for the last half an hour of the podcast.

02:26:35.840 That's a shame because I always save the best thing for the last half an hour.

02:26:39.900 Okay. Next time I'll listen to the last half an hour. It's a pleasure, Peter. And anytime and good

02:26:48.420 luck. And I'm looking forward, not to this one, I'm not going to listen to, but to the next one.

02:26:52.860 Very well. Thank you, Nir.

02:26:54.880 Okay.

02:26:56.760 Thank you for listening to this week's episode of The Drive. If you're interested in diving deeper

02:27:01.020 into any topics we discuss, we've created a membership program that allows us to bring you

02:27:05.420 more in-depth, exclusive content without relying on paid ads. It's our goal to ensure members get

02:27:10.920 back much more than the price of the subscription. Now to that end, membership benefits include a bunch

02:27:16.420 of things. One, totally kick-ass, comprehensive podcast show notes that detail every topic, paper,

02:27:22.240 person, thing we discuss on each episode. The word on the street is nobody's show notes rival these.

02:27:28.240 Monthly AMA episodes or ask me anything episodes, hearing these episodes completely.

02:27:33.340 Access to our private podcast feed that allows you to hear everything without having to listen to

02:27:38.580 spiels like this. The Qualies, which are a super short podcast that we release every Tuesday

02:27:43.920 through Friday, highlighting the best questions, topics, and tactics discussed on previous episodes

02:27:48.660 of The Drive. This is a great way to catch up on previous episodes without having to go back and

02:27:53.700 necessarily listen to everyone. Steep discounts on products that I believe in, but for which I'm not

02:27:59.440 getting paid to endorse, and a whole bunch of other benefits that we continue to trickle in

02:28:03.940 as time goes on. If you want to learn more and access these member-only benefits, you can head

02:28:08.260 over to peteratiamd.com forward slash subscribe. You can find me on Twitter, Instagram, and Facebook,

02:28:15.600 all with the ID peteratiamd. You can also leave us a review on Apple Podcasts or whatever podcast

02:28:22.100 player you listen on. This podcast is for general informational purposes only and does not constitute

02:28:27.680 the practice of medicine, nursing, or other professional healthcare services, including the

02:28:32.580 giving of medical advice. No doctor-patient relationship is formed. The use of this information

02:28:38.120 and the materials linked to this podcast is at the user's own risk. The content on this podcast is not

02:28:44.660 intended to be a substitute for professional medical advice, diagnosis, or treatment. Users should not

02:28:51.100 disregard or delay in obtaining medical advice from any medical condition they have, and they should

02:28:57.260 seek the assistance of their healthcare professionals for any such conditions. Finally, I take conflicts

02:29:03.220 of interest very seriously. For all of my disclosures and the companies I invest in or advise, please

02:29:09.280 visit peteratiamd.com forward slash about where I keep an up-to-date and active list of such companies.

02:29:27.260 for all of my disclosures and the companies I invest in, please visit peteratiamd.com forward slash about where I keep an up-to-date.

The Peter Attia Drive - April 25, 2022

#204 - Centenarians, metformin, and longevity ｜ Nir Barzilai, M.D.

Episode Stats

Summary

Transcript