The Peter Attia Drive - #222 ‒ How nutrition impacts longevity ｜ Matt Kaeberlein, Ph.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

00:00:19.380 longevity into something accessible for everyone. Our goal is to provide the best content in

00:00:24.360 health and wellness, full stop, and we've assembled a great team of analysts to make

00:00:28.420 this happen. If you enjoy this podcast, we've created a membership program that brings you

00:00:32.680 far more in-depth content. If you want to take your knowledge of the space to the next level,

00:00:36.980 at the end of this episode, I'll explain what those benefits are. Or if you want to learn more now,

00:00:41.960 head over to peteratiyahmd.com forward slash subscribe. Now, without further delay,

00:00:47.760 here's today's episode. My guest this week is Matt Caberlin, who of course is a returning guest.

00:00:54.940 He's been a previous podcast guest a number of times, most recently joining me on AMA 35 back

00:00:59.980 in May of 2022. Matt is not only one of our most recurring guests, but he's also one of the people

00:01:05.100 I will consistently share emails with discussing various topics. Probably not a week goes by that

00:01:09.680 we're not sending each other a paper or something like that. And so when I found out when Matt was

00:01:13.580 going to be in Texas for a project, I figured let's sit down together in person and do one of

00:01:18.780 these things instead of remotely, which we normally do. In this episode, we really focus the conversation

00:01:23.460 around nutrition as it relates to aging and longevity. This really came out of a paper that

00:01:29.120 Matt wrote as a review article about a year ago, which I remember reading in draft, really appreciating

00:01:34.200 it and loved reading the final version of it. So even though nutrition science is not the topic I'm

00:01:39.640 most interested in talking about, given things I've mentioned in the past, which is sort of diets and

00:01:44.500 fads and the religion around that stuff, we tried to really make this as biochemical a discussion as

00:01:50.720 possible. So we obviously discuss Matt's recent review article, and we talk pretty deeply about

00:01:56.100 the literature on caloric restriction. We talk about epigenetic clocks, aging, and its effect on

00:02:00.960 DNA and cell reprogramming. We then focus around protein and aging. So this is the one macronutrient

00:02:07.660 that stands out, right? Carbohydrates and fats are really there for energy use. Protein is not.

00:02:13.620 We then get into this seeming dichotomy around protein and mTOR. You've obviously heard me talk a lot

00:02:19.020 about mTOR. We understand that a drug that inhibits mTOR, namely rapamycin, seems to produce a whole

00:02:25.700 bunch of wonderful effects. And yet protein, particularly an amino acid called leucine,

00:02:30.440 seem to really trigger mTOR. So how can those two things simultaneously be true if having muscle is

00:02:36.500 good, but taking rapamycin is probably good? We get into the importance of muscle mass, the RDA on

00:02:42.100 protein itself, IGF, growth hormone, and a lot more. I want to point something out here. This is a topic

00:02:48.280 for which we just don't have easy answers. And it's possible you're going to walk away from this

00:02:53.440 entire conversation with more questions than answers. My goal is that you come away from this

00:02:58.420 realizing that, yeah, there's quite a bit of uncertainty here, but I have a better way that

00:03:01.720 I can think about it. And I have a better sense of what questions to ask. Now, for those of you who

00:03:05.840 may not remember who Matt is, or maybe even didn't listen to any of our previous podcasts, let me just

00:03:10.300 give you a really brief reminder. Matt is a globally recognized leader in the basic biology of aging.

00:03:15.800 He's a professor of laboratory medicine and pathology and adjunct professor of genomic

00:03:20.560 sciences and an adjunct professor of oral health sciences at the University of Washington in

00:03:24.120 Seattle. His research interests are focused on the basic mechanisms of aging in order to facilitate

00:03:28.320 translational interventions that promote healthspan and promote a healthy way of life.

00:03:32.160 So without further delay, please enjoy my conversation with Matt Cable.

00:03:40.320 Matt, it's great to finally be able to do one of these in person with you. We've done a lot of

00:03:44.060 these remotely. We're taking advantage of the fact that you're in Texas filming a documentary about

00:03:48.540 aging, which is pretty awesome. So when we knew that this was going to happen, we said, well,

00:03:52.580 let's take advantage of you being here and let's come up with something that we both talk about so

00:03:56.720 much over email, which is to say, I don't think a week goes by that we aren't exchanging an email

00:04:02.160 about some aspect of the relationship or the inner space between nutrition and longevity.

00:04:07.900 Does that speak to our ignorance? Does that speak to the ubiquity of such content? I don't know.

00:04:11.920 What does that say about us? It's an area that a lot of people are

00:04:14.640 really interested in and it certainly intersects with popular culture. So having been in the aging

00:04:20.540 field for a long time, I certainly recognize how complicated that biology is. And I think the biology

00:04:26.380 of nutrition is equally complicated. And when you get at the interface of those two, it's really hard,

00:04:33.200 I think, sometimes to draw a definitive conclusion. So a new paper will come out and you usually read the

00:04:38.160 papers before I do and you're like, hey, what do you think about this? And then, you know,

00:04:41.660 we throw it back and forth. It's hard sometimes to get to concrete answers. So certainly we'll try to

00:04:46.540 do that today. But I also think this will be a little bit of a theme that there are many things we

00:04:50.980 don't understand yet about optimal nutrition and how that intersects with optimal health span.

00:04:56.880 You and I have spent so much time on the podcast speaking about the molecules. Of course,

00:05:02.100 our favorite being rapamycin, but all sorts of them, right? We recently talked about NMNNR,

00:05:06.020 NAD. We've talked about metformin. And it's easier almost to ask the questions from the standpoint of

00:05:12.700 gyroprotective molecules because the intervention is much cleaner.

00:05:15.520 Yeah, absolutely.

00:05:16.360 Like, are you taking this drug? Yes or no? And of course, what's interesting about that,

00:05:20.940 and I think it speaks to what we're going to talk about today, think about the one drug among those

00:05:26.420 that stands out, which is rapamycin. Even within that, just I think yesterday or two days ago,

00:05:31.440 you and me and David Sabatini had a back and forth about timing of the dose, frequency within the

00:05:38.000 dosing schedule, the dose itself. I mean, even with a drug, it's still very complicated to say,

00:05:45.180 well, what about during this phase? Because the study I think we were talking about was looking

00:05:48.600 at mice and it was asking the question of early exposure of rapamycin later in life,

00:05:53.580 constant dosing, intermittent dosing. That's for a drug. And we're still struggling to piece it

00:05:58.520 together. Now imagine trying to ask that question of your food.

00:06:03.020 You know, we'll obviously talk a lot as well about the animal models and what they can tell us about

00:06:06.960 what might affect human aging. But the big piece that gets lost with the animal models on top of

00:06:11.860 all that complexity is the environment. You know, we keep these mice in a well-controlled environment,

00:06:17.180 usually relatively pathogen-free, and they live in that same environment their entire life.

00:06:22.460 Now you think about the human experience where our environment is extremely complicated. We're

00:06:26.600 constantly getting bombarded with all sorts of challenges and infectious agents. And our

00:06:32.240 environment changes dramatically throughout our lives. In fact, maybe this is something we want

00:06:37.200 to touch on. A lot of the epidemiological studies on optimal nutrition are from 20, 30, 40 years ago.

00:06:44.560 The average human environment is very different today than it was when those studies were done. And

00:06:49.140 how does that potentially change the interaction between nutrition and health outcomes? I think it's a

00:06:55.680 really interesting but challenging question to address to anybody's satisfaction, honestly.

00:07:01.820 Yeah, that's actually a great point. And I made a similar point on a totally different topic,

00:07:06.340 which was all of the studies that talk about cancer screening are very backwards-looking by definition,

00:07:12.520 right? You have to look at controlled trials that were done in the past. But the technology of

00:07:16.220 radiology is changing so much. Radiology is a very, you know, physics-based field of medicine.

00:07:21.640 And so when you read a study that talked about mammography for screening, you know,

00:07:26.580 it was a 15-year study, right? So it's a great study. Well, by definition, it was done based on 30

00:07:31.600 to 20-year-old technology that by the time the study has been completed, you have the follow-up data,

00:07:39.100 you write up the paper. It doesn't necessarily represent what's happening today. And that's a huge

00:07:44.340 challenge of evaluating that type of data.

00:07:46.000 And in people, because we age so slowly, there's really not a lot you can do about that if you want

00:07:52.900 to try to do correlative longitudinal studies of aging. Because people age so slowly, the people

00:07:59.200 who are in their 70s today were in their 30s 40 years ago. And so the environment that they were in

00:08:04.860 is probably quite different than the environment that 30-year-olds are in today. So there's not a

00:08:09.920 great way around that. I think the key is to recognize that limitation and be potentially

00:08:16.960 even more careful about assuming causation from correlation over many decades.

00:08:23.540 There's a bit of a mea culpa on the topic of nutrition, which is really my least favorite

00:08:27.840 topic, despite the fact that it keeps coming up on this podcast and it's unavoidable.

00:08:31.940 As I reflect back on my own understanding of this topic, the strength with which I held

00:08:38.740 convictions over the past more than decade, I would say I've gone in reverse, right? I have

00:08:44.740 looser and looser convictions as time goes on. And I view fewer and fewer things with certainty as time

00:08:51.160 goes on. When I think about this problem clinically, I have what I would consider to be an incredibly

00:08:57.620 simple framework, which is if I'm looking at a patient, I'm asking a question, are you over

00:09:02.520 nourished or under nourished? Are you under muscled or adequately muscled? So that's a two by two. And

00:09:09.680 then are you metabolically healthy or not? That's sort of my first order question. Now, one of those

00:09:15.880 spaces doesn't really have too many people in it. The adequately muscled, under nourished,

00:09:24.120 metabolically unhealthy bucket doesn't really exist. So these aren't people aren't uniform

00:09:27.600 distributed in those buckets, but it's a pretty good way to sort people. And you can't sort someone

00:09:34.480 by looking at them into that bucket, but by looking at them, doing some functional testing,

00:09:39.260 looking at their biomarkers, and that might include also doing things like a DEXA scan where you can

00:09:43.900 actually get some objective data. You can pretty quickly figure that out. And the reason we think

00:09:48.440 that's important is it helps us understand, do you need an energy deficit? Do you need an energy

00:09:54.920 surplus? What's your protein intake need to be to achieve that in combination with your calorie

00:10:00.980 needs? And the hardest of those to treat by far is over nutrition, under muscled. And unfortunately,

00:10:10.600 that's a very common phenotype. That's a lot of people these days. Yeah. I think as a general

00:10:14.900 approach, first order approach, that makes a ton of sense. You know, one of the things that that

00:10:19.620 allows you to recognize, right, is that the optimal strategy isn't, there's no one size fits all,

00:10:25.300 I guess would be the way I'd say it. Different people are going to have different needs nutritionally

00:10:30.100 and what works really well for one person may not work at all for another person. And so I think

00:10:34.600 looking at that level allows you to not have to try to say everybody should be doing X. That is pretty

00:10:42.380 similar to the way I think about it. Obviously, I don't practice medicine and I try not to make

00:10:46.800 recommendations for what people should do. But in my own life, that's generally the way that I

00:10:51.160 try to approach it as well. And I hope I'm doing okay. You haven't tested me yet. So you can't tell

00:10:56.440 me which bucket I'm in. But I think I'm doing okay for my age with my nutritional strategies. And the

00:11:01.360 other thing that I sort of have realized similar to what you were saying before is that, you know,

00:11:05.340 it's an ongoing learning process. And so I think it's really important that we be willing to

00:11:09.880 change our beliefs about nutrition and other aspects of health as more data comes in. So I

00:11:18.020 think if you take that strategy, then you can be open to the possibility that what you believed

00:11:22.820 10 years ago might not have been exactly right. And maybe we need to tweak it a little bit.

00:11:27.700 I'll be honest, I have real trust problems with nutritionists. You know, in part, it stems from I

00:11:31.780 remember very vividly when I was, I think it was probably in my early 20s, I read one of these diet

00:11:37.100 guru books. This was, I'm going to date myself, but this was, you know, early 90s, I guess.

00:11:42.940 The theme back then was, you could eat anything you wanted, as long as you cut out the fat,

00:11:47.480 you could have this really high, simple carbohydrate diet, just keep it low fat, and you know,

00:11:52.600 you'll be fine. And we now know that's exactly wrong. I can't help but look at a lot of what people,

00:11:58.920 what I would put sort of on the fad diet side, the diet gurus, what they're saying today,

00:12:03.660 how do we know 10 years from now, we're not going to look back on that. And again,

00:12:07.560 be like, that just makes no sense. I think some of us today can look at some of what's out there and

00:12:11.840 say, that just makes no sense. But again, this gets back to what I was saying before. It's not

00:12:16.500 that I would say nutrition science is across the board, low quality. I think they're actually really

00:12:21.680 good scientists doing really good work in this area. It's just a really hard problem. And I do think

00:12:27.000 to some extent, the biology of aging, and the biology of nutrition do share that these are extremely

00:12:32.580 complicated biological systems, we're trying to understand in the context of this changing

00:12:40.080 environment over time. So I don't blame the scientists, I just think we have to be really

00:12:44.660 careful to recognize what the limitations are, and not draw really strong conclusions like,

00:12:51.460 everybody should eat, you know, a low protein diet. That's kind of one of the fads that are out

00:12:56.220 there today. That's a mistake to recommend across the board, nutritional strategies for everyone.

00:13:01.720 I guess the last thing, sorry, I'm talking a long time here. But I guess the last thing that what

00:13:05.380 you said makes me think of as well, and I think this is really important, because people lose sight

00:13:09.400 of this is exactly what you said, if you can be somewhere close to optimal nutritional intake,

00:13:16.700 just say total calories, regardless of composition, body composition is somewhere close to where it

00:13:22.340 should be. That's a big chunk of what you need to give yourself the best chance of being healthy

00:13:28.140 going forward. You don't have to optimize every single thing. And I know you're all into

00:13:33.240 optimization. And I respect that about you. I think if you can do that, that's great. But you

00:13:37.100 don't have to to get most of the benefits. And so I think starting from that big picture perspective

00:13:41.220 allows you to get most people most of the way there. And then when they're most of the way there,

00:13:47.200 you can focus on how do we get that last 10, 20, 30%, whatever it is.

00:13:51.700 I couldn't agree with you more, Matt. And I would argue, and I do argue now in a very different way

00:13:57.220 from where I used to be a few years ago. There are most things in my life where I don't like the

00:14:02.560 80-20 principle. My good friend, Tim Ferriss, he's the king of this. He's the king of how can I get 80%

00:14:09.360 of the learning with 20% of the time? And I've never seen anybody who can do it like Tim. Like the guy

00:14:15.820 can learn a language in a month. He can be 80% proficient in a language in a month. I'm the

00:14:21.760 opposite. I'm the guy who loves the tail. I love the asymptote. I love the perfection of something.

00:14:28.820 I would say in nutrition, that is exactly not where my interest lies. I agree that you can just get 80%

00:14:36.360 of this right by focusing on exactly what we've talked about. And the details, the complete

00:14:42.340 optimization are not worth it. And it's instead better to put that effort into exercise. That's

00:14:49.980 where I think if you're going to really go down the rabbit hole and put more of your mental energy,

00:14:55.440 more of your time, and more of your focus into something, you have far more of an ROI on the

00:15:00.900 exercise front than eking out incremental value on the nutrition front. I've joked about this before.

00:15:07.140 Other guests on the podcast, Lane Norton and I have had riffs on this back and forth.

00:15:10.200 The people who sit there on Twitter, which I realize is not a representative sampling of the

00:15:16.180 world. It's simply an annoying vocal group of people who will waste endless hours debating the

00:15:23.480 finer points of their dietary pet peeves who can't do 10 pull-ups is amazing. There should be a rule

00:15:30.440 that says if you can't deadlift twice your body weight and do 15 pull-ups, you shouldn't be allowed

00:15:36.700 to pontificate endlessly about the finer points of nutrition.

00:15:40.980 We can talk to Elon about that. Maybe that can be a new rule.

00:15:45.620 I think we've established nutrition matters here. But I think at the same time,

00:15:50.320 David Allison said it once to me, it's amazing how little we know about this subject matter.

00:15:55.020 Kind of rehashing what we've said. We know that too much and too little are bad. And for most of our

00:16:00.740 existence, we were worried about the too little problem. The too much problem has become a

00:16:05.780 relatively recent phenomenon. And they're bad in different ways. Acutely, chronically, they have

00:16:10.280 different limitations. We know that certain things are toxic, acutely or chronically. Not a lot we

00:16:16.800 know. I mean, with definitive clarity, there's not a lot we know beyond those things. One thing that

00:16:23.140 seems to be true is, at least from the animal literature, caloric restriction seems to reproducibly

00:16:32.040 improve lifespan. Let's kind of talk about how that came to be as an understanding.

00:16:37.100 This area of research is actually quite old.

00:16:40.060 It's like 100 years?

00:16:40.840 Yeah. The first experiments were published in the early to mid-1930s, which means they were probably

00:16:46.220 started in the 1920s. So almost 100 years ago, people were going down this line of thinking of

00:16:53.040 asking, you know, what is the effect of significant restriction of calories on the aging process in

00:17:00.160 mammals? So the early studies were all done in rats. If I remember correctly, these studies were

00:17:04.780 originally designed from a developmental perspective. So they were really thinking about

00:17:09.120 malnutrition and its effects on development. And as a byproduct, made the observation that yes,

00:17:15.720 when you restrict calories in a rat early on in life, they have a smaller body size. But then if you let

00:17:22.040 them live out their entire lives, this is in the laboratory, and I think that's really important

00:17:26.040 to keep in mind, they live 40%, 50% longer. So we're talking really significant increases in

00:17:32.860 lifespan. And then the other thing that was appreciated pretty quickly was, not only are

00:17:36.820 they living longer, but they seem to be healthier as they're living longer. So this concept of health

00:17:42.020 span and the period of life that is spent in good health, free from disease and disability,

00:17:47.120 it seemed as if caloric restriction was not only increasing lifespan, but also extending health

00:17:52.540 span. That led to a large body of literature since then, studying the effect of caloric restriction in

00:17:59.600 not just rodents, rats and mice, but also all sorts of simpler organisms, invertebrates like fruit

00:18:05.880 flies and C. elegans and yeast. And the common theme seems to be that, again, starting from laboratory

00:18:11.660 conditions, if you restrict nutrients by a whole variety of different methods, you can increase

00:18:18.720 lifespan and apparently increase health span proportionally, at least proportionally. So

00:18:24.160 there's a lot of nuance there, a lot that we can dive into and to unpack. But I think that's generally

00:18:28.540 the take-home, is that over and over and over again across the evolutionary distance we're talking

00:18:34.000 about is much, much greater than the evolutionary distance between rodents and humans. So over a very

00:18:40.400 wide evolutionary distance in pretty much every organism where it's ever been studied, you can

00:18:46.140 find evidence that caloric restriction slows aging. Again, there are cases where that didn't happen,

00:18:52.900 where lifespan wasn't extended, where lifespan was shortened. Maybe we want to talk about this at some

00:18:56.900 point. The interaction between genetics and environment and caloric restriction. But in general,

00:19:01.520 the take-home message is caloric restriction can slow aging in laboratory animals pretty much

00:19:07.460 everywhere where it's been studied. The one question that some people have is whether that's true in

00:19:12.300 non-human primates.

00:19:13.200 I was going to say, before we get to NIA Wisconsin, which is perhaps the single greatest experiment that's

00:19:19.840 ever been done to test this hypothesis, both in terms of its duration, level of control, and proximity

00:19:26.420 to our genome. Let's spend a moment on that. Before we do, any things that come up from the rodent

00:19:33.860 studies that are worth talking about? So for example, one of the things that I think is always

00:19:38.220 important to point out is there's a very particular death that tends to fall on laboratory mice. If you

00:19:44.660 look at the death bars for humans, there's much more heterogeneity, but the leading cause is

00:19:49.520 atherosclerosis. Now that's true in the United States. It's true across the globe. When you mix in

00:19:55.300 develop and undevelop, it doesn't matter. Laboratory mice aren't that way. They die of pretty much one

00:19:59.860 thing and one thing alone. And that is... Actually, it's euthanasia, but I know where you're

00:20:03.440 going. Cancer, right? So certainly every old mouse at time of death will have cancer. And again,

00:20:13.000 because of the way animal studies are done, usually you have defined endpoints where when a mouse

00:20:17.920 reaches that endpoint, they have to be euthanized. But the expectation is if they hadn't been euthanized,

00:20:22.760 they would have died from the cancer. So I think you're absolutely right.

00:20:24.820 They're not dying from atherosclerotic.

00:20:26.700 That's right. When you look at their arteries, they're not littered with plaques the way ours

00:20:30.600 are. At least the commonly used inbred mouse strains, that is definitely true for. There

00:20:35.820 are, this is maybe getting in the weeds a little bit, but there are certainly mouse strains that

00:20:39.820 have been designed either transgenically or through selection to develop other pathologies that will

00:20:45.460 shorten their lifespan. But if you let a typical mouse strain in the lab live out its natural life,

00:20:51.940 it will have a very high tumor burden at the end of life. And most likely, I guess I should know this.

00:20:57.780 I don't know exactly. I'm guessing 80% of the animals would die from cancer. So it's different

00:21:02.500 from humans in that way. And I actually think this is a legitimate criticism to some extent of the

00:21:07.320 caloric, the interpretation of the caloric restriction literature that is, could it be the

00:21:12.120 case that really what caloric restriction is doing is preventing cancer. And that's why you see these big

00:21:18.100 increases in lifespan. And I think that's really difficult to definitively answer one way or the

00:21:23.600 other. What I would say is mice do develop functional declines in every tissue and organ

00:21:30.280 as they age, very much like people do. So a person may die from cardiovascular disease,

00:21:35.520 but at the same time, if they're in their 80s, their kidney isn't functioning as well. Their heart

00:21:39.600 isn't functioning as well. Their brain probably isn't functioning as well. So mice show all of those

00:21:43.920 same declines in function with age and caloric restriction seems to delay or outright prevent

00:21:51.740 those declines as well. So yeah, maybe the lifespan effect is primarily due to cancer, but caloric

00:21:57.580 restriction is having an effect apparently on the underlying biological aging process in all sorts of

00:22:03.660 different ways. And I really like the functional measures. A lot of people in the field these days

00:22:07.760 are really enamored with the aging clocks, epigenetic clocks, biochemical markers. I think those are all

00:22:13.160 useful and important. But from my perspective, what really gets my attention is if somebody shows

00:22:18.300 that the heart is still functioning like a young heart or the immune system is still functioning.

00:22:22.040 Yeah. I wasn't planning to go down that rabbit hole, but since you brought it up,

00:22:25.280 can you convince me of the utility of the clocks absent the type of data that would actually demonstrate

00:22:32.240 longitudinally their benefit, which to my knowledge, we really don't have yet?

00:22:36.300 I would say a couple of things on that. I think we need to be precise in what we mean when we talk

00:22:41.040 about the clocks because there's lots of flavors of clocks. Most people these days, if you just say

00:22:45.820 aging clock, what they really mean are the epigenetic clocks that are showing the characteristic

00:22:50.540 changes in the epigenome, the epigenetic marks that are seen with age. Again, in every organism where

00:22:56.640 it's really been studied, you do see these characteristic changes in the epigenome with age.

00:23:01.660 And so I would say one place where their utility is clear, at least to me, is as a chronological

00:23:07.040 measure. Now you might ask, okay, why would I ever want to use an epigenetic clock to tell my

00:23:11.940 chronological age? I know how old I am, but forensics, for example, might be a place where

00:23:16.600 that's useful. Their crime has been committed. They want to know with some level of precision,

00:23:20.300 how old the perpetrator is. You could use an epigenetic clock for that reason. In my world,

00:23:25.260 as part of the dog aging project, there are many dogs that are rescued. An owner might want to know

00:23:29.880 their age. So I think that is a real use and clearly the clocks will work for that. I think really what

00:23:34.020 you're asking though is, can I convince you that the epigenetic clocks and potentially other types

00:23:39.240 of clocks are actually measuring biological aging? And that's a harder, in my mind, that's a harder

00:23:45.980 thing to prove. And personally, I have no interest in convincing you of that because I'm not convinced.

00:23:51.460 So I think this is an area where the field is in flux a little bit. And there are certainly

00:23:56.000 scientists who I respect a lot in the field who believe at their core that these epigenetic clocks

00:24:02.840 tell us about biological aging or can be used to tell us about biological aging.

00:24:08.200 Then there are people like me who want to see the proof. And I think the proof is really

00:24:12.340 being able to show at an individual level, that could be in a mouse, could be in a person,

00:24:18.400 could be in a dog. At an individual level, you can predict someone's biological age at some point

00:24:24.720 in their life and with some level of precision, predict what's going to happen in the future.

00:24:30.380 What are their future health outcomes? How long are they going to live? Nobody has done that yet.

00:24:35.960 What they've done comes close, I guess. So what has been done is to look at longitudinal studies

00:24:41.660 in people where we have samples from people 10, 20, 30 years ago, measure the epigenetic profiles of

00:24:50.200 those people 10, 20, 30 years ago, and ask how well does that correlate with mortality outcomes,

00:24:56.060 for example, in the future. And they do work to some extent. I think people will debate how well

00:25:02.800 they work. Are they any better than other markers you could look at in predicting mortality? I think

00:25:08.080 that's unclear, but there is some correlation there. So it really depends to some extent maybe

00:25:12.780 on how skeptical you are. I'm a skeptic by nature and I want to actually see the proof. I guess the last

00:25:17.680 thing I would say about this, I'm talking mostly about the epigenetic clocks. Maybe it's worth talking

00:25:21.740 about other types of clocks that people can make. The other thing I want to caution people on though

00:25:27.100 is assuming that the epigenetic clocks are the only important thing about aging. There is again,

00:25:32.720 a small number of very vocal and popular people in the field who talk as if changing the epigenome is

00:25:42.660 going to change everything about aging. We have no data to support that. I just have to say it,

00:25:48.000 that is not true at this point. We have no data to support it. What we know about the biology of

00:25:53.660 aging is that epigenetic changes are one of, depending on how you categorize things, you know,

00:25:59.760 eight or nine or 10 molecular processes that seem to contribute, that the field has reached consensus

00:26:05.120 on. It's only one of those things. Is it possible it is sort of in a hierarchy, the most important and

00:26:10.980 drives a lot of those other changes? Yes, that's possible. We don't have any data to support it. So

00:26:16.120 this idea that reversing the epigenome is reversing aging is at best an exaggeration,

00:26:24.940 at worst, an outright lie. I mean, it's just not true.

00:26:29.260 What a set of experiments technology-wise would you need to be able to do to even test that hypothesis,

00:26:34.080 say, in a mouse?

00:26:34.780 We're close. Well, maybe close. I guess I should qualify that a little bit. Conceptually,

00:26:39.320 we're close. So there have been these factors called the Yamanaka factors that can reprogram the

00:26:45.720 epigenome. So this has been done in cells. So if you take cells in culture, in a laboratory,

00:26:50.700 and you passage them many, many times, you can see changes in the epigenome, just like you might see

00:26:56.220 changes in the epigenome in an animal, in tissues. And you can put these reprogramming factors into

00:27:01.800 the cells and turn them on. Now there are four Yamanaka factors?

00:27:04.860 There are four Yamanaka factors, and people are trying different cocktails, adding some other stuff in,

00:27:09.860 taking some stuff out. But yes, there are the four classic Yamanaka factors. And what those factors do

00:27:15.440 is they basically wipe clean the epigenetic changes that have happened over time. And also,

00:27:22.840 what's amazing is that they restore those cells back to a, if you take it far enough, back to a

00:27:27.960 pluripotent state. So essentially, you get virgin new cells that could differentiate into any cell type

00:27:34.240 in the body. So this has been known for many years. What is relatively more recent over the last

00:27:39.360 eight or nine years are people are trying to express these reprogramming factors in an animal.

00:27:46.580 So instead of doing it in cells in the laboratory, do it in an animal. And I think the most compelling

00:27:50.760 work is work in a premature aging model of mice. So it's called a progeroid model, where they're very

00:27:57.020 short-lived, they're very sick. But these reprogramming factors can extend lifespan by, I don't remember what

00:28:02.880 the exact numbers are, but a significant amount, maybe 40, 50%.

00:28:05.820 Well, which seems like a lot, except you have to recognize these mice live maybe 25% of the length

00:28:11.580 of a normal mouse, right? So they're very sick. But there are impressive changes that happen that

00:28:16.680 are consistent with the idea that you fixed or made something better. So the experiment to do

00:28:21.320 would be to express these reprogramming factors in an old mouse and make that mouse young again.

00:28:27.060 And this is where I think the exaggeration, I'll use the nice word, has gotten ahead of the actual

00:28:32.980 data. So what has been done is showing that in one or two, maybe three tissues, you can see an

00:28:41.240 improvement in function. The most impressive, I think, is work from David Sinclair's lab, where they

00:28:45.400 use this optic degeneration models. So degeneration of the eye showed that they could reverse that with

00:28:51.660 these reprogramming factors, and then tried to do the same thing in an old mouse. You know, the data was

00:28:56.300 mixed, but I think pretty compelling that you could, to some extent, regenerate the optic nerve in an

00:29:00.980 old mouse. So that's certainly impressive, exciting. But nobody has ever taken an old mouse and turned

00:29:07.520 it into a young mouse. So when people start talking about reversing aging, that implies that you have

00:29:13.600 taken an old animal or person, and to some extent, biologically made them young again, that hasn't

00:29:20.360 happened. So what I would say needs to happen to really convince me, there are two things. So I would

00:29:24.940 be convinced that this is useful, potentially therapeutically and important. I'm actually

00:29:30.260 already convinced it could be useful therapeutically. But I would become really excited if somebody could

00:29:34.540 do as good as rapamycin in a mouse. So I'm not asking for much, in my view. We know rapamycin can

00:29:39.860 extend lifespan 25% at least. Again, a dose hasn't been optimized, but 25%, let's stick with that.

00:29:47.140 And you can reverse functional declines in many tissues. So show me you can do that with

00:29:53.360 reprogramming, and I'll be excited. Nobody's done even that yet. Show me you can take a two

00:29:58.980 and a half year old mouse, make it look like a one year old mouse, and then it lives to be five

00:30:03.760 years old. I'll be really excited. Look, I'll be all on board. I might even come on your show and

00:30:07.980 apologize for saying that people were exaggerating, although they are exaggerating now. But I think the

00:30:13.520 enthusiasm has just gotten so far ahead of where the science is. Let's maybe help folks understand

00:30:19.380 what the Yamanaka factors are doing and how one can be sure that even if you fix the aging problem,

00:30:29.680 you don't create a new problem. So if the objective is, I want to take the DNA as I had it when I was

00:30:36.280 young. So when I was 20, this is what my DNA looked like. Now that I'm 50, it looks different.

00:30:43.600 It has literally these methyl groups that are sitting directly on the cysteine residues,

00:30:50.960 like literally on my DNA. Okay, we want to take those off. Maybe?

00:30:55.400 First of all, it's important to understand why that's even a problem.

00:30:58.360 Why is my 50-year-old crappy DNA not as good as my 20-year-old DNA?

00:31:04.140 So again, this is taking a step back to sort of basic biology. So the DNA is where all the

00:31:09.480 information is. But then that DNA has to get turned into RNA. That's called transcription

00:31:14.740 or gene expression. We'll just call it gene expression. And then that RNA has to get turned

00:31:18.420 into protein. And in general, it's the protein that does the work. So what these epigenetic

00:31:22.620 changes, the methyl groups that you were talking about, do primarily, we think, is affect expression

00:31:28.720 of the genes. So basically what you're seeing with aging, we think, is a shift in the epigenome

00:31:35.980 that leads to certain genes being expressed that shouldn't be and certain genes not being

00:31:41.140 expressed that should be. And I think there's a little bit of a debate about which is more

00:31:44.700 important right now, but it probably doesn't really matter, right? So the idea is you're

00:31:48.220 getting things turned on and turned off inappropriately as we get older. So there's a loss of regulation,

00:31:53.360 which probably contributes to a loss of homeostasis. And homeostasis is, I think, a really useful

00:31:58.240 way to think about aging. If you're healthy, your body is generally in homeostasis.

00:32:03.700 And what happens as we get older is it becomes harder and harder for our body to maintain

00:32:08.340 homeostasis. When you get out of homeostasis, if your defense mechanisms are working right,

00:32:13.080 you can get back in. So you get COVID, for example, your immune system works, you're out

00:32:17.100 of homeostasis, but you come back in and then you're okay again. I think as we get older,

00:32:20.820 it gets harder to come back into homeostasis. And that's why we start to see pathology and mortality.

00:32:25.540 So let me differentiate two states of pathology. My five-year-old son was on his scooter

00:32:31.860 two weeks ago, going down the steepest hill in the world, which I had no idea how I didn't see

00:32:36.700 that he was about to do that, like face planted. And when he came up, all I could think is how

00:32:42.500 quickly can we get to the hospital? I mean, it was a bloodbath. I'm not making this up, Matt.

00:32:47.580 Six days later, there was one little tiny scar. Eight or nine days later, you would have had no idea

00:32:55.520 this kid ripped his face off on pavement. He's five. I get a cut. It's like nine months until

00:33:02.800 the scar is gone. So there's a very clear distinction between a five-year-old's DNA and

00:33:08.280 a 50-year-old's DNA in terms of how he can literally make new proteins that are better than

00:33:14.080 my proteins. Let me stop you there just for a second, because I think this is actually the

00:33:17.320 crux of the question. You said it's a difference in your DNA. Well, I'm asking. I think what I'm trying

00:33:21.060 to get at is that's a clear case of the protein that he makes is better than my protein. He's

00:33:26.220 making much better protein. Certainly functions better. I guess what I was getting at, though,

00:33:29.900 is the one question I think that's really important here is there can be changes to the DNA to the

00:33:34.320 sequence, right? So the sequence of the DNA is the information. Those are called mutations,

00:33:38.440 and those accumulate as we age. And that's honestly what drives a lot of cancer. So we've known this for

00:33:43.380 a long time. The epigenetic changes are sort of on top of this. Yeah, and while it more regulates

00:33:48.100 this expression, I'm wondering how much that factors into the example I just gave.

00:33:52.280 It's a good question. I'm sure it does to some extent. Absolutely.

00:33:54.900 Like what else explains why his collagen is so much better than mine? What are the other

00:33:58.300 factors that go into that? I mean, I think there are probably many reasons why healing,

00:34:03.320 our ability to heal, declines with age. I actually, again, we've talked about this before,

00:34:06.880 I think inflammation is a huge driver of our loss of ability to recover as we get older. So,

00:34:13.280 you know, all sorts of things go wrong if you have a high level of sterile inflammation in your

00:34:17.860 body, including the ability of stem cells to function. And a lot of injuries require stem

00:34:22.480 cells to function to build back what's been broken. So it's complicated, I guess I would say,

00:34:27.220 but the question is... Yeah, it could be that I have more senescent cells and more senescent cell

00:34:30.440 factors that are impairing the ability of cells to heal.

00:34:34.280 Just to throw a wrench in that, there's actually a body of thought that senescent cells actually

00:34:37.920 promote wound healing. Again, this is where the biology is so complicated. But I think the crux of

00:34:42.500 the question we started from is, if you only fix the epigenome, do you fix everything?

00:34:46.860 How do you know you fix all these things?

00:34:47.980 Yeah, do you fix everything? And nobody knows, I think is the fair answer. I would be shocked if

00:34:54.160 that was the case, that epigenetic changes drive all of aging. But it's possible. I think we have

00:35:00.160 to be open to that idea that epigenetic changes sit on top of or upstream of the other hallmarks

00:35:06.660 of aging. First of all, let me say one thing. It won't fix everything. You will not fix mutations

00:35:11.200 by fixing the epigenome. The question is, do mutations, do they happen with enough frequency

00:35:18.000 to be a major contributor to functional declines that go along with aging? Certainly cancer you

00:35:25.080 can point to. Well, cancer for sure. But let's now talk about something else, which is near and dear

00:35:29.040 to your heart, no pun intended, but ejection fraction. Again, because you study dogs, not only is

00:35:34.420 cancer a big problem, but so is heart failure. So now we're dealing with a muscle, a set of cells

00:35:41.100 that really aren't being turned over the way skin is. So when we think about the example of my son,

00:35:45.980 when you think about your gut epithelium being sloughed off when you get sick, when you think

00:35:49.500 about your fingernails in your hair, boy, it's really easy to think about those things as rapidly

00:35:53.960 being turned over. But neurons, cardiac myocytes, these things don't get turned over a whole heck of a

00:35:59.460 lot. So what is it about reprogramming that we think is going to fix an aging neuron or an aging

00:36:08.340 cardiac myocyte? This is an area where the biology of what's really happening, at least to my knowledge,

00:36:14.480 is so poorly understood that I think the real answer is we don't completely know. I'm going to

00:36:18.740 give a very simplistic answer, which is that what people are trying to do is not reprogram all the way

00:36:26.540 back to the pluripotent state. So it's called partial reprogramming. It should be pretty dangerous.

00:36:30.420 Well, that's what I was going to say. If you're a single-celled organism, no problem going back to

00:36:33.920 the pluripotent state. You can then start over. In a complicated animal, if we reprogram you back to

00:36:40.620 the pluripotent state, that's not going to end well. No. Right? So I think the idea is to go back

00:36:45.040 far enough that you restore the epigenome to its pristine state, young state, and then hope that when you do

00:36:54.520 that, you restore gene expression to where it's supposed to be. Maybe one way to think about it

00:37:00.180 is you restore the homeostatic mechanisms to a more youthful state where then the homeostatic

00:37:06.680 mechanisms that all of our cells have can basically clean up the rest of the mess. Because we know as

00:37:11.660 we get older, for example, we all accumulate damaged mitochondria. Changing the epigenome, which is the

00:37:17.120 nuclear genome, isn't going to fix anything that's wrong with your mitochondria directly. But maybe by

00:37:22.460 fixing the epigenome, you restore the homeostatic mechanisms that then maintain mitochondria in a

00:37:28.960 healthy state, and you can fix the damage to the mitochondria. So that's the concept. And again,

00:37:35.060 I would say the evidence is suggestive that if you do it just right, you can improve function in at

00:37:42.900 least some aged tissue organs by partial reprogramming. I've yet to see anything that

00:37:48.620 convinces me that anybody has made an old heart into a young heart in an old animal with partial

00:37:54.740 reprogramming in the heart. But you can improve function. I would also say the same thing's true

00:37:58.440 with rapamycin, right? I would not argue. We see that short-term treatment with rapamycin in mice

00:38:04.200 makes an old heart function functionally, to some extent, more like a young heart. I would never argue

00:38:09.940 that we have taken that heart and now it's young. It's just in an old body. We don't know that,

00:38:14.340 and that's hard to prove. You can see some evidence that it should be possible with partial reprogramming

00:38:19.580 to do that. And the question is, will it work everywhere? Will it work in some tissues and

00:38:24.900 organs and not in others? We don't really know. So let's just say 10, 20 years from now, people

00:38:30.540 have figured out a lot of the complexity, starting to move these things into the clinic. Maybe we will

00:38:35.460 see really large effects on lifespan and healthspan in mice. What I've yet to hear anybody give a

00:38:42.460 convincing explanation of is how you do that in the brain. Because so much of who we are and what we

00:38:48.200 are comes from our experiences and our memories. And so how do you ensure that you can reprogram

00:38:55.420 somebody's brain in a way that isn't going to change that? And I just think that's going to be

00:39:00.000 a really hard problem to overcome. But maybe somebody will figure it out. There are tons of

00:39:04.440 really smart people working in this area, lots of resources going into this area. So I think it's

00:39:08.960 exciting. Again, my big concern is that we don't mislead people into thinking that we're close to

00:39:16.540 reversing aging. And I think it's a problem from the perspective of the general public. I think it's

00:39:20.540 a problem from the perspective of the scientific communities. Other scientists look at that and

00:39:24.000 they're like, this is snake oil. This is just not true.

00:39:26.700 My concern with it is actually in terms of the impact it has on people, which are, hey, this is

00:39:32.900 awesome. This thing's going to get worked out. I can sort of do what I want because in 10 years,

00:39:36.700 they're going to reprogram me. And my view on that is even if that is true, or even if you have a high

00:39:44.660 degree of confidence that that is true, how would you not hedge? You know, again, hedging is such an

00:39:50.360 important part of how companies manage risk. So the difference between good companies and bad

00:39:55.820 companies when it comes to risk management is everything. That's why some companies do really

00:40:00.400 well in economic downturns and others don't. It's basically about risk management. And a very important

00:40:05.280 part of risk management is indeed hedging. So if we think of ourselves each as little companies,

00:40:11.000 you know, you're the CEO of Matt Co. I'm the CEO of Pete Co. I can't think of a more important

00:40:17.140 asset within my company to manage than my own life. Do I have enough money? Yeah. You know,

00:40:22.720 do I have enough fun? Yeah. Those are all important assets, but existing would be the number one asset.

00:40:28.520 And to not take a risk management approach of hedging to that is insane. And yet what I see is

00:40:34.620 so many grand promises of this stuff and nobody's sort of paying attention to what they eat or how

00:40:40.500 much exercise they do, because I don't need to, this is going to be worked out. So the thing that I

00:40:44.660 always find amazing is some of the most vocal advocates for this stuff don't have an ounce of

00:40:48.680 muscle on them. You know, they're overweight or whatever, like they don't look healthy. And I'm like,

00:40:53.060 guys, you can do both. You can believe that in 10 years, we're going to fix this problem,

00:40:57.520 but you could still actually care about your health.

00:41:00.520 No, I think that's a really important point. And having, again, been in this field for a long time

00:41:05.340 now, I think you can just look back over the last 20, 30 years and look at predictions people made on

00:41:10.720 how fast these things were going to come along and get into the clinic. And none of that has happened.

00:41:16.160 So I totally agree with you. Also being in the center of it, I take a view of, again,

00:41:21.700 pretty strong skepticism when people say this is going to happen in 10, 15 years.

00:41:26.580 I honestly have not appreciated that there are maybe a lot of people out there looking at what

00:41:31.880 they read in the New York Times or on CNN and thinking to themselves, oh, I don't have to worry

00:41:36.900 about this. This is going to get worked out. So my advice would be don't expect major changes in

00:41:44.220 treatments to improve lifespan and healthspan in the next 20 years. And that doesn't mean I'm not

00:41:50.020 optimistic. I think there are opportunities there. It would not surprise me if we do see

00:41:54.940 some of these things get into the clinic, but I certainly wouldn't expect it because there are

00:41:59.040 so many barriers that we don't yet appreciate. There are lots of barriers just in moving something

00:42:03.960 through the clinical trial process. I think the reprogramming stuff is a perfect example.

00:42:07.840 So you actually alluded to this earlier. Are there potential side effects? Absolutely.

00:42:12.000 You push it too far, you reprogram too far, you're gone. We know that certain types of cancers

00:42:17.940 are a side effect of this partial reprogramming in mice. Again, it doesn't mean it can't be worked

00:42:22.400 out, but there are really reasons I think to be concerned that this is going to be hard

00:42:26.720 to implement therapeutically. The other thing I would say, even if those things can be worked out,

00:42:32.420 the FDA is going to be extremely skeptical of this kind of approach. So as people move these through

00:42:38.140 the clinical trial process, they are going to have to show with really rock solid, compelling data

00:42:43.780 that reprogramming strategies are not going to cause significant side effects. So I think it's

00:42:48.100 a long road before we have reprogramming strategies that get into the clinic. Maybe somebody will

00:42:54.380 identify a small molecule that can do some of this. And I know people are working on that. Maybe

00:42:58.520 that'll be an easier path. But for now, I think it's going to take a while. That's the best case

00:43:02.700 scenario. That's if we really can partially, I'm going to say partially reverse aging, reverse

00:43:07.260 aspects of aging. It's still going to be a long road.

00:43:09.960 And I wonder if the first wins are going to be things like what David Sinclair has done,

00:43:14.420 where you've got one very niche application. I think another one that would be amazing would

00:43:18.940 be osteoarthritis. If you could figure out a way to regenerate human cartilage without joint

00:43:24.120 replacements, those are huge wins that seem at least a little more feasible. But again, I agree

00:43:30.940 with you. I think this stuff takes four times as long and costs four times as much as we think.

00:43:36.640 You and I are, I mean, honestly, we're pretty lucky because we know about a lot of this stuff.

00:43:40.680 We actually can start practicing some of this stuff like rapamycin before it gets out there,

00:43:45.360 right? Again, I'm not recommending anybody take rapamycin necessarily without talking to your

00:43:50.200 physician first. But we know this stuff and we have at least a pretty good idea of the relative

00:43:55.200 risk reward. But before it gets out to where it hits the mainstream from a clinical perspective,

00:44:01.460 it's a really long path. I totally agree with what you said, though, about specific indications where

00:44:06.840 you can target it very precisely, hopefully, and where there's no other solution currently. I think

00:44:13.560 those are opportunities. That's exactly the strategy that people have tried to take with

00:44:17.280 senolytics, that these molecules that will clear senescent cells. And even that's been hard. I mean,

00:44:22.340 Unity is the sort of largest company in this space and their first clinical trial for osteoarthritis

00:44:27.600 failed. So now they're looking at the eye because it's a nice indication where for some of these eye

00:44:33.380 diseases, there isn't any solution. And you can, in principle, target it quite precisely to the eye.

00:44:39.260 So yeah, I think that is exactly the strategy that people will be taking. And hopefully it'll be

00:44:43.600 successful. I want this stuff to work. I just try to be a realist at the same time.

00:44:48.420 The way I would kind of describe this to people is if you want to bring it back to a financial

00:44:51.660 analogy, it's a lottery ticket. And so if your entire financial planning system is based on

00:44:57.940 winning the lottery, the odds that you're going to win are pretty low. Instead, if you're going to

00:45:03.540 play the lottery, play it in the context of an otherwise great saving and investing strategy.

00:45:09.080 I guess the other thing I would add to that is, and this is what we talked about before,

00:45:12.440 you don't have to do everything right. Get 80% of the way there, which nutritionally I don't think

00:45:17.300 is, I mean, for some people it's very challenging, but I think most people could do that.

00:45:20.540 But exercise, you don't have to optimize your physical activity. Do something and that'll

00:45:26.200 get you most of the way there. So yeah, I totally agree.

00:45:28.240 Yeah, the exercise curve, which we've covered a lot in previous podcasts, you get most of

00:45:33.240 the benefit. I would say literally 50% of the benefit based on at least the so-so epidemiologic

00:45:39.540 data, about 50% of the full benefit of exercise is captured going from nothing to about 15 met

00:45:48.320 hours per week. You know, that would be 15 mats times one hour would be one way to get there.

00:45:53.760 But in reality, no one who's that unfit is going to do 15 mats, but that would be like

00:45:57.320 three hours a week of five mats to put that in perspective. And five mats is like a very,

00:46:03.560 very brisk walk or a slow jog, something to that effect. So you get a sense of like 15 met hours per

00:46:09.400 week. By extension, I do about a hundred met hours per week of exercise. I think of everything in terms

00:46:14.580 of met hours. But the point is that you can get, depending on the study, 30 to 50% of the benefit

00:46:20.420 going from being completely sedentary to 15 met hours per week is pretty amazing.

00:46:25.780 Which is a big benefit, right? And again, it's sort of remarkable that that information isn't

00:46:30.020 out there. And for the most people in the general public don't know that. I don't know what the

00:46:34.840 solution is. I think you're obviously doing a great public service by trying to get that

00:46:38.760 information out there. But it's unfortunate because I think, again, most people understood how much

00:46:43.520 benefit they could get from just getting out and moving a little bit. Maybe a lot, maybe three

00:46:49.180 hours a week is a lot for some people. But the magnitude of the benefit compared to the effort

00:46:54.200 that you put in, I think most people just don't know that. And it's unfortunate.

00:46:58.020 Let's go back to the CR stuff. So what do we know about the effect of CR in the laboratory animals

00:47:04.840 on the immune system?

00:47:06.020 So it's a little bit complicated. First of all, laboratory animals in the laboratory are kept in

00:47:12.380 what's called a specific pathogen-free environment. So that doesn't mean there's no pathogens,

00:47:16.380 but it's a relatively low pathogen environment where they are not obligated to really use their

00:47:22.660 immune systems against all the challenges that we would face in the real world. So one question

00:47:27.880 has come up. Are animals that are on calorie restriction immune compromised? And again, I think

00:47:33.400 the data is a little bit mixed. There have been studies where people have done pathogen challenges

00:47:38.320 on CR animals and they respond better. At least the old animals respond better than age-matched

00:47:43.280 ad libitum fed control. So ad libitum just means eat as much as you want. But then for certain types

00:47:48.080 of challenges, caloric restriction clearly causes a deficit.

00:47:51.700 Yeah, the sepsis experiments are pretty clear. With the CR animals compared to controls, when you

00:47:56.300 induce sepsis in them, the CR animals die much more quickly.

00:47:59.640 And so, of course, the obvious implication of that is that maybe CR would impair immune function in

00:48:05.080 people and lead to higher risk of all sorts of infectious diseases. And this gets additionally

00:48:10.540 complicated though by the question of optimal CR with optimal nutrition. So you might sometimes

00:48:15.600 just see this CRON, C-R-O-N, right? Caloric restriction with optimal nutrition or CRAN,

00:48:20.860 caloric restriction with adequate nutrition. That can be done in a mouse. We can control all of that.

00:48:25.920 So we make sure that they get all the micronutrients and vitamins that they need

00:48:29.200 when they're on this CR diet. When you move into the real world and people start practicing caloric

00:48:35.820 restriction, that all goes out the window. If I wanted to do caloric restriction off the top of

00:48:40.180 my head, I wouldn't even know what to do to make sure that I'm getting optimal nutrition.

00:48:44.440 And so in that state where you are CR without optimal nutrition, I think that's where I really

00:48:50.200 become worried about the side effects, particularly as you raised immune deficits, because you may not

00:48:56.500 be getting the nutrient value or the specific micronutrients and vitamins that you need

00:49:02.600 to maintain a functioning immune system. Sure, you may affect some aspects of the biology of aging in a

00:49:09.160 way that you're aging biologically more slowly. That doesn't matter if you get influenza and die.

00:49:15.240 So again, I think that's an additional complication that comes into play. When we start talking about,

00:49:19.660 we haven't talked about all the other anti-aging nutritional strategies. When we start talking about

00:49:24.440 recommending those nutritional strategies to the general public, based solely on mouse studies,

00:49:31.580 I get really concerned because of this environmental complexity that humans live in.

00:49:38.140 And we haven't even talked about the genetic complexity, right? So there's all sorts of things

00:49:41.160 that are just different about laboratory animals compared to people living in the real world.

00:49:46.880 And then what can we say about frailty, sarcopenia, as it changes in an animal in a CR environment,

00:49:54.280 and can that be extrapolated also? It's pretty clear, I think, that much like rapamycin,

00:50:00.100 most functional measures of aging seem to be preserved in calorically restricted animals,

00:50:05.060 including measures of frailty and measures of sarcopenia. The same thing, again, is true with

00:50:09.500 rapamycin. This actually surprised a lot of people when the first studies were done because the

00:50:14.120 expectation was, because mTOR plays such a big role in muscle synthesis, that if you inhibit mTOR with

00:50:20.700 rapamycin or caloric restriction, which is a potent inhibitor of mTOR, that you would actually see

00:50:25.700 accelerated sarcopenia. And that just isn't the observation in laboratory animals. Again, we have

00:50:30.460 to be careful not to extrapolate to people, but it doesn't seem to be the case that you lose muscle

00:50:35.680 mass and function in the way that people would define sarcopenia. I think the important complication

00:50:41.380 here is that all of the caloric restriction studies that I'm aware of, when they look at muscle

00:50:46.260 function, normalized to body weight. And the calorically restricted mice weigh substantially

00:50:51.320 less than the ad libitum fed mice. Usually, I think it's on the order of 30, 35% less.

00:50:56.540 So it's usually grip strength normalized to weight.

00:50:58.960 Right. So what you're actually seeing is that the calorically restricted mice have maintained muscle

00:51:07.040 function proportionate to their body weight. And I don't know the answer to this, but it's something

00:51:12.320 that I thought of when we were talking about this show. Let's just say you did that in a person.

00:51:17.240 You would be able to answer this. I'm sure you've got a 60-year-old person who needs to lose 30% of

00:51:22.400 their body weight. But of course, you want to maintain their muscle mass, their muscle function.

00:51:26.960 Would you view it as a good thing or a bad thing if they lost 30% of their body weight and 30% of their

00:51:34.900 strength?

00:51:35.220 I don't think we would. And I don't think we would view it as a good thing. If you're telling me that

00:51:39.300 someone needs to lose 30% of their body weight, presumably their body composition isn't great to

00:51:43.460 begin with. So no, I think you would view that as maybe a better thing than where they started,

00:51:49.840 but not optimal either. Optimal might be you would lose 30% of your body weight, but it would

00:51:56.520 disproportionately be adipose tissue, and you might only lose 10% of your strength or none at all,

00:52:01.440 depending on the change in lean body mass.

00:52:03.220 This is just a complication of the CR studies. It's hard for me sometimes. It takes me 20,

00:52:08.180 30 minutes of trying to dig through the paper to really figure out what normalization did they do

00:52:13.460 to look at metabolic rate or muscle mass or lean mass or fat mass or muscle function. But usually,

00:52:19.780 these studies will be normalized to body weight. This actually comes up also in some of the

00:52:24.700 intermittent fasting studies where the question sometimes in these studies is, are they

00:52:29.600 isocaloric or are they calorically restricted when they're put on intermittent fasting?

00:52:34.240 And people will claim they're isocaloric, but the mice lose weight. And what they really are is

00:52:40.120 isocaloric when normalized body weight, right? So they're really calorically restricted, but you have

00:52:45.040 to kind of dig to get how the normalizations were done to really understand.

00:52:50.260 When we think about what we know in humans, you know, there was a study that looked at the

00:52:54.320 difference in bone mineral density in people who underwent equal amounts of weight loss,

00:53:00.000 one driven by a caloric restriction strategy, one driven by an exercise-driven strategy.

00:53:05.720 And the exercise-driven weight loss group did not experience a reduction in BMD, but the CR group did.

00:53:11.580 Yeah.

00:53:12.180 So, you know, that's interesting. That's yet another thing that makes you think there's a little more

00:53:17.180 nuance to this, which is not to say CR from a weight loss perspective isn't valuable,

00:53:22.540 but it begs the question, is CR the right tool for longevity? Once you've achieved optimal weight,

00:53:29.380 is additional CR beneficial?

00:53:31.640 Well, that makes the assumption we know what optimal weight is. I mean, I think that's kind

00:53:34.620 of the crux of the question, right? We're asking, does CR impact longevity positively? We know if you

00:53:41.080 go on CR, you're going to lose weight. So if the answer to that is yes, then by definition,

00:53:44.780 optimal weight is lower than what we think, right?

00:53:47.440 Well, in humans though.

00:53:48.620 I would say we still don't really know what optimal weight is. So again, this, I think,

00:53:53.860 just reflects the challenges in coming to definitive answers. And the way I think about

00:53:59.140 it more so is what are the downsides potentially to caloric restriction? And if we don't know that

00:54:06.180 caloric restriction has big benefits in terms of health span and perhaps lifespan, what are the

00:54:12.920 downsides? And do those downsides outweigh the uncertainty we have about whether caloric

00:54:18.480 restriction is beneficial? And unfortunately, I think this is something that not very many people

00:54:23.640 in this field pay attention to. We all expect if you do a clinical trial of a drug, you're going to

00:54:29.480 report adverse events and you're going to look at side effects. Very rarely do people think about

00:54:34.500 that before they write a book recommending that people should do diet X. Even in the clinical trials,

00:54:40.140 some of the nutritional clinical trials, they don't really carefully monitor adverse events.

00:54:45.880 It's a bias in the way we think about interventions. We feel like nutritional interventions

00:54:51.220 are by their very nature safe. And certainly for extreme nutritional interventions, that's clearly

00:54:56.960 not true. So I think we should be thinking about what are the risks associated with significant

00:55:01.540 caloric restriction in people as a therapeutic strategy.

00:55:05.840 So let's talk about the experiment to end all experiments with respect to caloric restriction,

00:55:09.440 which is the very famous one we alluded to earlier at the University of Wisconsin and the

00:55:14.040 NIA. I've read this study a thousand times. If I can get the details right once, I'll be happy. 0.99

00:55:19.660 But between the two of us, I hope we can do this. You had two groups of animals, one at the

00:55:24.160 University of Wisconsin and one directly in Bethesda, Maryland. This was obviously a huge NIH funded

00:55:28.440 effort. It ran for a couple of decades given the lifespan of rhesus monkeys. The Wisconsin animals

00:55:35.580 were fed, the controls and the treatment CR animals were fed a very processed diet. At least after the

00:55:44.040 fact, the investigators there suggested they wanted to more mimic a standard American diet. Of note,

00:55:49.860 I recall the amount of sugar, pure sucrose in their diet was 28.5% of total calories. So a high quality

00:55:58.580 diet facetiously, the CR animals, the calorically restricted animals were fed 25% of what the control

00:56:06.460 animals were fed. And in that experiment, we found a benefit to caloric restriction. The CR animals

00:56:13.300 outlived the control animals.

00:56:15.800 And they had fewer age-related diseases. So I think if you go back to that original 2009 paper, you know,

00:56:21.220 the lifespan effect is compelling and it looks real. But what again is really indicative of that it might be

00:56:27.740 having an effect on biological aging is that they saw reduced rates of cancer. Again, not surprisingly, as we

00:56:33.680 talked about in mice, but also heart disease and metabolic disease. So it's consistent with the idea that in

00:56:40.900 that cohort of monkeys, again, given what you mentioned about the dietary composition, caloric restriction was in

00:56:47.940 fact having a beneficial impact on the aging process.

00:56:52.300 And those animals all came in at about the same age.

00:56:55.440 So that was sort of an apples to apples comparison.

00:56:59.300 Now we go down the road to Bethesda. We have a totally different experiment in a way.

00:57:03.540 I don't know how much of this was deliberate and how much of it was not.

00:57:06.540 The diets were different. So that's maybe a good contrast. These animals were actually fed the closest

00:57:12.740 diet that could mimic their real diet. It didn't have any, you know, sugar in it really. I think it was

00:57:18.040 like about 3% sucrose. It was almost kind of like a vegetarian, pescatarian sort of diet.

00:57:24.320 Fish was the dominant source of protein. You know, it was a high quality diet relative to the

00:57:28.800 Wisconsin animals.

00:57:29.740 Higher quality for sure.

00:57:30.660 The complicating factor here was the animals didn't come in at all the same age. So you had

00:57:38.600 some animals that came in young, some animals that came in old. The net result of the study was there

00:57:44.160 was no difference. The CR animals did not outlive. And so while the Wisconsin study was first published

00:57:49.660 in 2009 and it said CR works, the 2012 publication for NIA said CR doesn't work. At least that's the

00:57:57.420 lay press interpretation of it. So how do you kind of reconcile these findings?

00:58:01.640 One thing to add to that is the NIA study at Bethesda, in their paper at least, they did show some

00:58:08.440 evidence for improvements in at least some healthspan metrics. So if you read that paper closely, I think

00:58:13.880 what they're really saying is CR didn't extend lifespan, but it did have what appear to be some

00:58:18.780 beneficial effects on healthspan metrics. So it wasn't a complete failure in that sense. I mean,

00:58:23.800 I think it's interesting because since then, I remember when the 2011 paper came out, the Wisconsin

00:58:29.120 people were pretty upset, understandably so, I think. Since then, they've had sort of a reconciliation

00:58:33.540 paper and where they try to figure out what does it mean that we got these different results.

00:58:38.480 And I think their conclusion, which certainly is plausible, is that a lot of it comes down to the

00:58:45.040 difference in diets. And if you look at the actual body weights of the animals and how much food they

00:58:50.040 ate, not just the composition, but actually how much they ate, you know, you could make an argument

00:58:55.020 that the Bethesda monkeys were somewhat slightly calorically restricted.

00:59:00.220 The controls.

00:59:00.940 The controls, yeah, sorry.

00:59:01.940 Yes. The controls at Bethesda ate less than the controls in Wisconsin. And that would have

00:59:08.580 narrowed the gap between them and the treatment. And so then I think, as you also alluded to,

00:59:13.060 the fact that the Bethesda study was a little bit less controlled for age of onset. I don't remember

00:59:19.740 the details exactly. There were also some genetic differences in there. So there's a combination of

00:59:23.940 factors that make it a little bit difficult to conclude that it all is about the diet. The monkeys

00:59:30.440 in the Bethesda study came in at different ages. There was at least a hint, I think, that the

00:59:35.040 monkeys that came in at older ages, started CR at older ages, maybe got somewhat of a benefit,

00:59:41.020 whereas the ones that started early didn't get any benefit. So it's complicated to interpret. And

00:59:45.600 it's interesting because we see this a lot of times in the basic biology of aging, basic science

00:59:51.500 studies, where different labs will get different results in what seems to be the same exact experiment.

00:59:57.680 And then you start to dig into it. And yeah, there's all these differences in the way it was

01:00:01.860 done. It's really hard to know which of those differences contributed to the different outcomes.

01:00:07.000 In this particular case, because it was a 30, 40-year experiment, we're never going to find out.

01:00:11.860 That can't be done again.

01:00:12.680 Yeah, it just won't be repeated, both because of how long it takes and also because

01:00:16.360 the view on primate research, these are rhesus macaques, the view on primate research publicly has

01:00:23.360 changed. I just don't think we'll ever see that experiment done again.

01:00:26.380 My gut feeling is that the Wisconsin study, to some extent, probably does mirror what is closer

01:00:35.140 to a typical American situation, at least these days. I do not believe that they started with

01:00:41.780 that intention. But where we're at today, it probably is relatively as close as you can get

01:00:47.260 for a controlled laboratory study. The question, though, in my mind is, between these two studies,

01:00:52.420 do they suggest that caloric restriction slows aging? And let's just start relative to the

01:00:58.060 typical American diet. Somebody is moderately obese and they're eating terrible. Is it caloric

01:01:03.240 restriction or is it just returning to what you would call like an optimal body weight,

01:01:09.260 optimal body mass? And I don't think we know the answer. From these studies, you can't draw many

01:01:14.660 conclusions. I think the one thing you can do, and Roz Anderson, who's still at Wisconsin,

01:01:19.220 has really, I think, been a leader in this, is you can study the molecular signatures of caloric

01:01:26.700 restriction in the monkeys and ask, does it look similar to the molecular signatures of caloric

01:01:33.000 restriction in rodents? And you might ask, well, why would you do that? It seems obvious. But again,

01:01:37.640 a lot of the questions that people have around caloric restriction studies in mice is,

01:01:41.940 will it work the same way in people? And obviously, rhesus macaques are much closer

01:01:48.180 evolutionarily to people than mice are. So if you see the same molecular changes, it's suggestive that

01:01:54.960 caloric restriction is having the same molecular changes in people, certainly in primates. And in fact,

01:02:00.440 that seems to be the case. A lot of what we see in terms of, you know, changes in mTOR signaling

01:02:05.860 and mitochondrial function and other metabolic pathways is, in fact, shared between mice

01:02:11.880 and monkeys. That is one important outcome from these studies that we can definitely say is

01:02:16.820 rock solid. I tend to believe that the pretty dramatic declines in age-related disease seen in

01:02:24.400 the Wisconsin studies are telling us something. But again, is it just telling us that not being obese

01:02:30.240 reduces your risk for a lot of these diseases? We kind of already know that from the human literature.

01:02:35.380 Exactly. The other thing that isn't entirely clear, given that the NIA study didn't find a

01:02:41.840 difference, is we don't know how much of this was the CR versus the DR, the dietary restriction.

01:02:48.380 In other words, what the Wisconsin experiment suggests is, if you have an awful diet, reducing

01:02:55.280 the amount of awful food you eat is a good thing. Right.

01:02:59.240 What the NIA experiment doesn't tell us is the contrapositive. It doesn't suggest that if you

01:03:05.420 have a good diet, eating less of that will help you live longer. It might, but it isn't definitive.

01:03:11.040 So we don't know if the Wisconsin animals lived longer simply because they lost weight or because

01:03:16.460 they lost weight and they were eating less processed food.

01:03:20.360 Right. And I think the other thing to add to that is the NIA monkeys, which were eating, 0.98

01:03:25.160 you know, what we'll call a superior diet to the Wisconsin monkeys, also ate less than the 0.89

01:03:29.460 Wisconsin monkeys. In total, yeah.

01:03:31.180 So in other words, if you ate more of a good diet, would that be detrimental? We also don't know that.

01:03:36.520 It's an interesting question, actually, and it's too bad we don't know the answer to that. But I think

01:03:40.720 if they had been body weight matched or caloric consumption matched, that would have been an

01:03:47.100 interesting comparison to be able to see are there differences there.

01:03:50.640 And the other thing that just kind of gets off into weeds that we don't need to necessarily go into

01:03:55.060 is I don't really have a great understanding of even how we differ from the rhesus monkeys.

01:04:01.560 So, you know, I recently read Herman Ponsner's book. I don't know. Have you read it, by the way?

01:04:05.500 No.

01:04:05.660 So he kind of goes into the ecology and evolution of humans as a species and how different we are,

01:04:11.320 even from our closest evolutionary cousins. And one of the fundamental differences are incredible

01:04:17.940 capacity to store excess energy. So our metabolic rates, you know, he documents this through lots of

01:04:24.520 assessments of doubly labeled water on not just ourselves, but also hunter gatherers that are

01:04:30.440 still around today. And then of course, all the primates is we're really kind of unique in our

01:04:35.980 energy expenditure. Our energy needs are far greater than anything else. And people like that would argue,

01:04:42.700 hey, that was kind of an advantage that we took to allow our development, including our brain

01:04:48.600 development. So there's kind of a reason we're at the top of the food chain, which is we have a much

01:04:55.140 greater brain. And the price we pay for that is higher energy expenditure. And the price we pay for

01:05:01.840 that is we better be able to store energy because we will have a much harder time tolerating a low

01:05:08.520 energy environment. And so he talks about how even when you put these animals in captivity and you overfeed

01:05:14.560 them, they're not getting that much fatter. They're actually putting on lean mass. You know, I think

01:05:19.620 what you could argue, and he doesn't talk about this, but knowing what we know about human biology, you

01:05:24.160 might argue that they're still getting metabolically sick. Just as humans, when you're overfed, the real

01:05:29.620 metabolic sickness comes not with the inflation of your subcutaneous fat. It's when that spills out into the

01:05:36.680 viscera, into the liver, into the peripancreatic space, into the perinephric space, into the pericardial

01:05:42.820 space. It's that fat that escapes the normal depot of sub-Q fat that is truly inflammatory and truly

01:05:50.000 metabolically disturbing. So I throw all that in there just to say, like, it's just one more

01:05:54.220 confounding variable that makes it difficult to compare us even to an organism as complex as a

01:05:59.540 rhesus monkey. People certainly have made that criticism of the caloric restriction literature

01:06:04.660 writ large, not even taking into account the monkey studies, but the mouse studies, right? That there

01:06:09.600 are all sorts of differences between people and mice, and the metabolic state that people have

01:06:16.980 evolved to fill is just completely different. Having said that, you're absolutely right that

01:06:21.780 even mice in the laboratory as they get older will show metabolic syndrome, right? You will see

01:06:26.260 many of the same changes, insulin resistance, for example, that you see in people.

01:06:32.440 And do you see it absent the adiposity? Can you see it?

01:06:35.380 Mice gain adiposity with age two. They do, in fact, become obese with age. Again, on a pretty

01:06:41.500 crappy diet, right? Well, I don't know if it's crappy or not, the standard mouse diet. I don't

01:06:45.700 remember what the number is you made, but in the Wisconsin study, right, a significant fraction

01:06:49.800 of the control-fed monkeys develop diabetes.

01:06:53.820 Yes. I want to say, like, a quarter of the controls were pre-diabetic by the end of the study.

01:06:58.760 Again, which probably speaks to, even though they weren't overweight, when you get 28.5% of your

01:07:04.320 calories from sugar, it's probably going to impair your metabolism.

01:07:08.180 The other point that's maybe worth at least just mentioning here, because I hear people

01:07:11.700 talk about how certain diets are better for humans because it more mimics what we evolved to eat.

01:07:19.720 I don't know whether that's true or not. You could argue both sides of that. I don't see any

01:07:22.780 particularly compelling reason to think that that was the optimal longevity diet that,

01:07:27.480 you know, humans ate 100,000 years ago.

01:07:29.600 That argument is illogical on several fronts. The first is, and I don't know who coined this

01:07:35.400 phrase, but it's so ubiquitous that it's obvious. Like, by necessity, we had to be opportunistic

01:07:40.940 omnivores to even suggest that our hunter-gatherer forefathers were sitting around pontificating

01:07:47.980 about what they were and were not going to eat.

01:07:50.000 What they should eat, yeah. 1.00

01:07:50.640 It's just the dumbest thing I've ever heard. I mean, I don't think people are actually arguing 0.97

01:07:54.280 that. But my point is, the argument becomes so nonsensical when you realize our evolution

01:07:59.820 necessitated the most flexibility from a nutritional standpoint.

01:08:04.620 Yes.

01:08:05.200 And therefore, we ate anything and everything. And I think because we never probably existed

01:08:13.700 in an environment where food abundance was so great that we could reach the level of overnutrition,

01:08:20.840 it gave us even more flexibility with what we could eat.

01:08:24.480 Is that maybe part of the reason why humans seem to be fairly robust towards eating really,

01:08:32.820 really crappy diets? Obviously, we have an obesity epidemic and all of that stuff happening. But people

01:08:39.620 seem to be able to tolerate a wide variety of different diets, some of which are pretty darn

01:08:45.580 bad for them for many, many years before you start to really see the significant consequences.

01:08:51.500 And it may be that metabolic flexibility.

01:08:53.100 I was going to make a totally different point that's almost orthogonal to that, which is

01:08:56.720 you can make a case that people can survive in really remarkable health with diets that look

01:09:01.560 nothing like one another. In other words, you can look at somebody eating a really well-formulated,

01:09:07.000 strict vegan diet where they're not getting any animal protein, which clearly our ancestors all had

01:09:13.240 animal protein whenever they could. They're often a little protein malnourished, but they're very

01:09:18.900 healthy. And similarly, look at the opposite end of that spectrum. You can look at somebody on a

01:09:22.240 ketogenic diet. The only thing they would have in common between that other person is probably a lot

01:09:26.540 of leafy vegetables. But other than that, it's a much higher fat, higher protein diet. They can be

01:09:31.160 very healthy. That to me speaks to the resilience of our genome in terms of its interaction with

01:09:38.540 nutrition.

01:09:39.540 And that's sort of where I started, which is that there's no reason to think that the

01:09:43.100 ancestral diet is best. There's no reason to think that. But the other thing that I was thinking

01:09:49.220 about when I started down this path is that like many other things, our, as a species, our dietary

01:09:56.080 options and the typical diet is evolving rapidly now. The quality of the food, the stuff that's in it,

01:10:03.560 the preservatives is dramatically different than it was 50 years ago, both in caloric content and

01:10:09.320 nutritional content and taste and taste. Right. Absolutely. Which contributes to why a lot of

01:10:14.560 people want to eat more. So high calorie, really good tasting food that's often cheap. But the

01:10:19.900 environment that we evolved into obviously is completely different than it is today. But our

01:10:25.300 environment is changing at an accelerating pace, I think. And that makes it really, again, complicated

01:10:31.580 to try to get into the minutiae of what is optimal. Maybe we should be thinking about what's good

01:10:36.500 enough first, right? Because I think it's going to be really hard. And again, this is where I struggle

01:10:41.680 with the data that comes from epidemiological studies of people 20 years ago. The environment,

01:10:48.320 the food quality is just very different for most people today than it was even 20 years ago.

01:10:52.660 Well, this is where the grandmother test comes in. And this is where when I watch like the extremists

01:10:57.420 on both sides argue, I say two things. The first is, look, there are really good and really bad ways to do

01:11:04.560 your respective diet. I don't want to hear somebody tell me that everybody on a vegan diet is doing

01:11:08.800 well because I watched a lot of those kids in college and they literally were going to kill

01:11:13.280 themselves eating ramen noodles and crackers and cookies all day. So you can be vegan and eat pure

01:11:18.300 garbage. You could be keto and eat pure garbage. The second thing I would say is if you're eating those 1.00

01:11:23.180 diets well, and I'm being a little subjective when I say well, you're all shopping on the outer part of

01:11:28.800 the perimeter of the grocery store. It doesn't matter if you're carnivore, vegan, keto, low carb,

01:11:35.200 paleo, whatever. If you're doing those diets in the way that they were at least thought to exist,

01:11:41.160 you aren't going down any aisles of the grocery store. And that's kind of this grandmother test.

01:11:46.320 Like if your great grandmother didn't recognize what you're eating, it doesn't mean it's not good.

01:11:52.400 I don't want to say that a protein bar is not a good thing to eat. You just have to acknowledge

01:11:56.800 there's a little more risk there. Eating a carrot is inherently less risky than eating a protein bar

01:12:03.540 with 14 ingredients in it. That's just a fact. I think this is what you're getting at. Just a

01:12:07.880 little bit of a humility around what is known, what is not known. And as we push the envelope

01:12:14.220 of convenience, of nutrient density, of economics, price, shareability, portability, right? The ability

01:12:21.580 to preserve things. We're going to take some risks. Yep. I think that's exactly right.

01:12:25.380 But let's talk about more broadly, a paper you wrote, how long has it been? Two years?

01:12:30.540 We probably wrote it longer than two years ago. I think it came out at the end of 2021.

01:12:35.260 Oh, okay. Okay. Yeah. So it's fairly recent.

01:12:36.960 So talk about the impetus for that paper, which I thought was a great paper and we should discuss it

01:12:40.680 in some detail. Yeah. So I was asked by one of the editors at Science to write a review,

01:12:46.240 I think on mTOR actually. And like, well, lots of people have written reviews on mTOR. I've been

01:12:50.200 thinking a lot about caloric restriction and particularly other nutritional strategies that

01:12:56.200 people have been studying in the field, like ketogenic diet, protein restriction, time-restricted

01:13:02.880 feeding, intermittent fasting. And what do we actually know about those diets and their effects

01:13:08.220 on aging? Because I was of the, before I started to really dive into it, and this isn't something

01:13:12.600 that my lab researches directly. So we've previously done work on caloric restriction in invertebrates

01:13:19.300 and C. elegans, but we never really have done a lot of dietary interventions in mice. You know,

01:13:24.060 before I dove into the literature, I had this impression that all of these diets were similar

01:13:30.400 in some ways and had maybe comparable effects on lifespan. At least that's the way it gets portrayed

01:13:36.520 if you read some of these reviews. And I don't even like to call them reviews because I don't think

01:13:41.340 honestly, much of what gets into the literature as review articles are actually reviews. It's more

01:13:46.520 one person's opinion piece on their specific thing that they study, which is unfortunate.

01:13:52.260 But if you read most of the reviews on caloric restriction and other dietary interventions,

01:13:56.740 they're very one-sided. They usually have phrases like, fasting is known to have all of these

01:14:02.420 fantastic benefits, slows aging in every place where it's been looked at. And you can see that for all

01:14:07.520 these different dietary strategies. So I proposed to the editor that, you know, maybe we should do

01:14:12.360 a critical review of this space and think about what do we know? What do we don't know? Are they

01:14:17.920 equivalent? And to the extent possible, can we gain any insights into whether or not these nutritional

01:14:25.340 strategies, whether there's evidence that they have an impact on the aging process in people? So

01:14:30.000 that's kind of where we started. And I knew it was an ambitious thing to tackle when I said it.

01:14:36.480 And I'm not sure I really appreciated exactly how challenging that was going to be because it's a

01:14:41.820 huge area of literature. And it turns out, maybe not shockingly, that there are many more questions

01:14:48.860 than there are answers when you really dive into it. So what was your process?

01:14:52.680 The first step was, and I should say I had a fantastic set of co-authors, all, you know,

01:14:56.800 really great early career scientists who really helped me with this and did a lot of the legwork.

01:15:01.860 I just want to mention them by name.

01:15:03.040 Please do.

01:15:03.460 So Alessandro Bito, who was a postdoc with me, Mitchell Lee, who was a former graduate student

01:15:08.740 with me, and Crystal Hill, who's at the Pennington Biomedical Research Institute. And she works on FGF

01:15:14.740 21 and protein restriction. So those three were co-authors on this paper with me, all just really

01:15:19.940 fantastic early career scientists. So we started by asking ourselves, okay, what are the different

01:15:24.920 popular dietary interventions that people have claimed have an effect on aging? And we came up with,

01:15:30.920 I don't know, six or seven, the ones I've already mentioned. So there's true caloric restriction,

01:15:35.100 which is pretty straightforward. That really just means limiting the overall caloric intake that an

01:15:41.200 animal gets by somewhere between 20 towards the low end. And the most I've ever seen is 65% of calories.

01:15:48.060 And you were doing this in animals and humans?

01:15:50.160 We were mostly focusing on mice. We narrowed it pretty quickly when we realized the scope of

01:15:54.800 what we had undertaken. So we could have tried to do it in, you know, fruit flies and worms and all that

01:16:00.260 stuff. We said, let's start with mice, see what's known, and then try to look into humans and ask,

01:16:07.260 are there parallels? So caloric restriction, pretty straightforward. We actually don't go very deep

01:16:12.100 into caloric restriction because that literature is huge. And other people I think have done a pretty

01:16:16.000 good job of reviewing true caloric restriction. But there are some points there that we probably

01:16:20.800 want to touch on that are important. And then there are variants of caloric restriction,

01:16:24.340 which include intermittent fasting, time-restricted feeding.

01:16:28.980 How did you differentiate those two? I have a definition, but I want to make sure yours is

01:16:32.600 clear.

01:16:32.860 So in mice, well, so first of all, the first differentiator we need to put across all of

01:16:37.320 these things, is it isocaloric or is it a flavor of caloric restriction? Because it turns out,

01:16:43.960 I would say the vast majority of studies in mice of all of the things that we're going to talk about

01:16:48.600 are flavors of caloric restriction. And what I mean by that is the experimental group ate less

01:16:53.620 calories than the control group.

01:16:54.920 So it's time-restricted feeding, but it's really caloric restriction in a narrower window.

01:16:59.400 Intermittent caloric restriction, maybe is the way you want to think of it. And there's actually

01:17:02.700 some nuance there that we can get to. So how am I differentiating between time-restricted feeding

01:17:06.780 and intermittent fasting? I would say, to my view, the easiest differentiator is time-restricted

01:17:12.300 feeding is limiting the number of hours in any 24-hour period that the animal or person eats.

01:17:18.920 And there are obviously, you're aware of this, there are flavors of time-restricted feeding in

01:17:22.500 people where the window can be anywhere from 12 to 6, sometimes even more extreme than that. But you

01:17:27.540 limit the hours per day that the animal or the person eats. Intermittent fasting, I would put in

01:17:34.140 a 24-hour or more fast. That's a reasonable definition.

01:17:37.100 That's actually the definition I use. An intermittent fast is a fast that occurs at a frequency of greater

01:17:41.860 than once a day.

01:17:42.720 Right. Exactly. The other thing I would say, though, is that time-restricted feeding gets

01:17:46.280 even more complicated than that because there's evidence that it's not only about how big the

01:17:51.340 window is, but where in the day the window is. And that's actually one of the things that came out of

01:17:55.420 our review of the literature is there is this clear connection between how much we eat and when we eat

01:18:03.120 that ties into circadian rhythms. And that circadian biology, even since this review came out,

01:18:08.560 there have been papers that have come out that reemphasize the importance of when we eat and

01:18:15.080 what we eat. I don't think it's either. I think it's both that suggests that that's probably going

01:18:19.140 to be significant in terms of the consequences of the long-term health effects.

01:18:23.020 All right. I'm hoping I'm going to remember to come back to that, but let's keep going.

01:18:26.100 So then there's what people call fasting mimicking diets, which are diets that I've been engineered to

01:18:32.840 some extent to induce the same metabolic changes as caloric restriction, usually very low sugar,

01:18:38.240 relatively low protein, high fat, but also very low calorie. So that clearly goes in the bucket of

01:18:44.080 a flavor of caloric restriction. There's ketogenic diets is another one. And then there's protein

01:18:49.760 restriction.

01:18:50.620 So isocaloric protein restriction.

01:18:52.420 Well, both. So again, you really have to look, you have to take each paper one by one

01:18:55.920 and figure out, is it isocaloric or isn't it? And that's in some cases simply not possible

01:19:01.880 because the data is just not there, but you have to look closely. So there are examples of both.

01:19:06.420 I guess one way to think about it is, is it ad lib or not? In other words,

01:19:09.800 an ad lib ketogenic diet might end up restricting energy, but non-deliberately.

01:19:14.740 That's one way to think about it, but I don't know that that answers the question of whether the

01:19:18.480 benefit comes from caloric restriction or not, right? So that's a complication, but I agree with you.

01:19:22.480 That is, it's different. We don't think about this much in mice, but certainly in people,

01:19:27.260 it's true. If you are not ad lib, there are psychological consequences to not eating when

01:19:33.160 you want, to being hungry all the time. Good, bad, indifferent, but those have biological

01:19:37.560 consequences as well. So they are different. Absolutely.

01:19:40.980 Let's go back to the circadian one. I want to kind of get the insights there.

01:19:44.860 So first of all, let's talk about what you know in mice, and then let's figure out if there's

01:19:49.100 any extrapolation.

01:19:49.900 So when we wrote the paper, there wasn't much on this. I mean, people were thinking about it,

01:19:54.060 particularly in the context of time-restricted feeding, that there might be differences in

01:19:59.080 the window of time-restricted feeding for in humans, right? Early in the day, late in the day.

01:20:03.300 There's been a couple of papers that have come out since we wrote the review in mice that I think

01:20:08.220 make a pretty compelling case that the lifespan benefit from, say, a 30% caloric restriction diet

01:20:16.000 is a combination of when the animals are eating and how much they're eating. Most of the benefit

01:20:23.960 seems to come from the calories. So let's just say, this may not be exactly right, but I think

01:20:29.240 it's close. Let's just say that you get a 30% lifespan extension from 30% caloric restriction,

01:20:34.160 that the two-thirds of that benefit comes from the calories. But one-third of the benefit actually

01:20:39.700 comes from the fact that those mice eat all their food in a short window and are fasted essentially

01:20:45.340 the rest of that 24-hour period. And if you force them, and I say force because if you give a

01:20:50.420 calorically-restricted mouse its food, it's going to eat it right away. So if you force them to eat

01:20:54.980 little bits throughout the day, you lose a portion of that lifespan benefit, which is really interesting.

01:21:01.420 Now, a mouse eating in an hour and then going 23 hours without food, what would we even compare

01:21:09.020 that to in a human? I don't know. I really don't feel comfortable even speculating. So the first

01:21:13.340 simplistic approach would be to say, well, a mouse lives about three years, a human lives about-

01:21:18.140 I was thinking more of like, how long does a mouse take before it dies from starvation?

01:21:22.060 Yeah. So that's where I was going to go next. I think that length of lifespan is not the approach

01:21:25.840 you take when it comes to metabolism. So I would say that, and this is total back of the envelope

01:21:30.700 calculation. Maybe it's like a one to four ratio. So a one-day mouse fast might be a four or five-day

01:21:37.060 fast in people, but that's not even perfectly true because a mouse will go into ketosis relatively

01:21:43.300 quickly within 24 hours. And a human can go into ketosis that quickly.

01:21:47.860 Depending on their incoming diet.

01:21:48.840 Yeah, exactly. It's not a perfect equivalency, but maybe one to four or five. I hope I'm not

01:21:53.140 saying something totally stupid here, but I think that's probably pretty close. So again, 0.69

01:21:57.080 it's very different, potentially, these kinds of studies in mice. The other thing that I think

01:22:00.660 most people don't appreciate unless they've actually done these caloric restriction experiments is that

01:22:05.700 if you go back to the classic experiments of Rick Weindruck and Roy Walford, those mice are fed a

01:22:11.620 calorically restricted diet. They're also fed three times a week. So they are in fact-

01:22:16.020 It's insane. It's like they're basically doing a two-week fast between their meals.

01:22:20.300 Yeah. And so what you see, even in 24 hours in a fasted mouse, is you see pretty dramatic

01:22:24.700 reductions in organ size. The mice are being fed three times a week. They're going through this

01:22:29.260 reduction in organ size and then this really rapid hypertrophy. And you can see that decrease in

01:22:34.860 organ size and then rapid increase even on some of the fasting mimicking diet work that Walter Longo

01:22:39.880 has done.

01:22:40.280 Has anybody done their reverse experiment where you try to actually mimic the way humans eat and you

01:22:45.300 take two groups of mice and the controls are fed, whatever, 100% of the nutrient, but they're fed

01:22:52.360 every two hours over the course of the day. And the CR group are given 70% of that, but they're fed at

01:22:59.940 the same time intervals constantly throughout the day. In other words, you make it purely a calorie thing

01:23:04.380 and you really take out the fasting except when they're sleeping.

01:23:07.220 At least one of these two studies that I was referring to did that.

01:23:10.960 Oh, so that's how they were able to identify that two-thirds of the benefit came from the

01:23:14.820 reduction in calories and a third of it came from the additional fast.

01:23:19.540 Right, exactly. So, and in my mind, I think this is really important because this is one of the

01:23:24.540 points that we made in our review is if you look at the vast majority of the literature around

01:23:30.100 intermittent fasting and time-restricted feeding and fasting mimicking diets,

01:23:34.180 they're calorically restricted. So, there's a fasting period and a caloric restriction component.

01:23:41.080 And none of the prior studies really, really teased that out in a way that allowed us to

01:23:47.000 have an understanding of how much is calories and how much is fasting and maybe how much is

01:23:54.240 when you're fasted. That's still, I think, is an open question.

01:23:57.960 What else can we say about early feeding versus late feeding?

01:24:00.920 You mean early in life?

01:24:02.660 No, early in day versus late in day.

01:24:04.740 Yeah, I mean, this is an area I'll admit I'm not an expert in. So, I don't honestly

01:24:08.960 have an opinion about which is better. And again, this is where I think mice are not going to be

01:24:16.160 a good model for humans. Those studies need to be done in people.

01:24:20.080 Some have suggested that an early feeding window versus a late feeding window produces

01:24:25.220 better pairing of our insulin sensitivity to our nutrient arrival, right?

01:24:29.600 I think that makes sense. Most people would agree that particularly if you're eating something

01:24:34.960 that causes your blood sugar to spike, that doing that right before you go to bed,

01:24:38.380 probably suboptimal, right? So, I think that maybe that can explain most of that observation

01:24:43.880 that has been made that if you're going to do a time-restricted feeding, it might be better

01:24:47.160 earlier somewhere, at least not right before bedtime, I guess I would say.

01:24:51.000 These kinds of questions are really complicated in humans because you could ask,

01:24:55.800 what benefit are we looking at? So, if you're looking at overnight blood glucose levels,

01:25:01.960 it makes perfect sense. If you're looking at sleep quality, maybe it's going to be different

01:25:05.820 or maybe it's going to be different in different people. If you're looking at other biomarkers,

01:25:10.660 again, it could be different. So, in my mind, at least, maybe you have a different opinion on this.

01:25:14.940 In my mind, at least, it's not even really clear how we evaluate what is better and what is

01:25:20.900 suboptimal. It may depend on what your endpoint is, what you're actually interested in optimizing.

01:25:26.880 Clinically, we see in people who wear CGM that early feeding produces an overall lower

01:25:33.440 average glucose for sure because even if you get the same spike, like if you're doing the same meal

01:25:39.520 early in the day versus late in the day, there's something about how long it takes to come down at

01:25:44.240 night versus in the morning. Now, that could be you're more insulin sensitive in the morning and

01:25:48.320 therefore, it comes down quicker. It could be something to do with pairing sleep with the

01:25:52.400 nutrition that is tweaking this and that there's a feedback loop where the excess glucose creates a

01:25:58.320 little more cortisol. You get a little more hepatic glucose up, but I don't really know if that makes

01:26:01.860 sense. I mean, I've heard people argue that, but at the same time, you theoretically should have the

01:26:05.280 lowest cortisol at night anyway, so that really shouldn't be an issue. I don't really know what it is

01:26:09.080 other than just to say I've observed it empirically. You know, it generally doesn't produce a great

01:26:13.040 quality of sleep, but to me, this starts to get into – which I want to hear more about,

01:26:17.640 but this gets into the minutiae. At some point, you just got to focus more on other things, but

01:26:22.200 I want to go down this rabbit hole just for the sake of completeness.

01:26:25.160 Yeah, sure. To some extent, that's almost where we ended up. Let me give the big picture answer

01:26:29.880 for why I think this is important. So, I think these nutritional intervention studies in mice

01:26:35.560 are very powerful for dissecting the biological mechanisms that underlie the effects that they have,

01:26:43.260 and some of these diets clearly have effects on aging. I'm very, very hesitant to suggest that

01:26:50.660 people should adopt any of these diets based on the rodent literature where it's at today,

01:26:55.660 and I think there are a whole variety of reasons for that, but that's kind of where I ended up. I

01:26:59.620 think they're super useful for understanding the biology. I'm really not sure that they're going to

01:27:04.380 work the same way in humans. What did you learn about the protein restriction in the ketogenic diet

01:27:08.560 mechanistically in the mice? The ketogenic diet studies, there have really only been two that I'm

01:27:12.840 aware of that looked at lifespan and healthspan in mice. They were slightly different, but in mice,

01:27:18.600 you have to go to really, really low sugar to actually get the mice to go into ketosis.

01:27:23.700 Essentially, 1% or less carbohydrate diets. So, again, that's a difference from people.

01:27:29.560 One of the studies that fed a ketogenic diet lifelong saw no effect on lifespan, but they did an

01:27:34.500 intermittent ketogenic diet. I don't remember the exact protocol, but it was something like every

01:27:39.060 other day or maybe once every three days or something. And there, there was about, I think,

01:27:42.780 a 15% increase in lifespan. And I'm sorry, what did they do on the other days, the animals?

01:27:48.020 Regular diet. Oh, interesting. Wow. Yeah. So, it was just back and forth between the control diet and

01:27:51.880 the ketogenic diet. And that didn't result in caloric restriction? That's the thing. The mice were

01:27:55.920 calorically restricted. So, in some ways, it's an intermittent caloric restriction. And this is what I would

01:28:00.740 say. It's also interesting because the fasting mimicking diet papers are intermittent ketogenic

01:28:05.220 diets. Maybe that's one thing to agree on is that intermittent ketogenic diets in mice can increase

01:28:14.980 lifespan and seem to have benefits for healthspan. The effects aren't huge. That's the other take-home,

01:28:20.120 I would say, from our study. There are two nutritional interventions that relatively consistently give

01:28:25.680 big effects on lifespan. One is caloric restriction and one is protein restriction.

01:28:31.540 Caloric restriction, the most extreme study that I've seen is 65% restriction. And that gave about

01:28:38.720 a 65% increase in lifespan. So, these are big, big effect sizes. Wow. I wasn't aware of that.

01:28:44.740 Yeah. That's this Weindrich and Walford paper. And when did they start that and how long did that

01:28:48.800 restriction last? That's a good question. I don't remember. It was probably six or nine months. I think

01:28:53.740 most of their studies were early onset caloric restriction. This study was really interesting

01:28:59.260 because they did a graded response from 90%, 80%, 60%, 50%, 40% of ad lib. You get essentially a

01:29:09.800 graded response in lifespan and it's roughly linear. So, 90% animals?

01:29:14.920 No, but they didn't go that far. They didn't go beyond 60 or 65%. And I also think this is an

01:29:19.700 interesting study because I don't think you could do that study today because the animal care wouldn't

01:29:24.880 allow you. Yeah. This gets back to an element that we don't think about enough, which is what

01:29:28.900 do those mice feel like? Like think about how angry those mice would have been on a third of their

01:29:35.320 normal caloric intake. You know, again, I haven't done these kind of mouse caloric restriction studies

01:29:40.080 myself. I've obviously talked to a lot of people who did. I think to really appreciate that,

01:29:44.220 you've got to probably be in the animal room seeing them. Certainly activity goes up quite

01:29:50.320 dramatically. And that's one of the remarkable things about caloric restriction in mice is that

01:29:53.860 they are more active throughout life than ad lib fed mice are. And maybe it's a sort of foraging

01:30:00.280 response, evolutionarily selected foraging response, but they are definitely, you give them a running

01:30:04.240 wheel and they'll just run and run and run and run. Yeah. There are behavioral changes for sure in

01:30:09.820 mice that are calorically restricted. And this is actually one of my real concerns about caloric

01:30:15.360 restriction in people. First of all, we should be realistic and recognize you're never going to get

01:30:20.340 a significant fraction of the population to calorically restrict. It's hard enough to get

01:30:25.260 people to calorically restrict down to a healthy weight. To get them to go 30% beyond that, it's just

01:30:29.840 not going to happen. But of the people I know, I mean, being in this field, I know people who have done

01:30:36.220 every possible anti-aging intervention you could imagine. And of the people I know, and I know a lot

01:30:41.820 of people who've dabbled in various forms of caloric restriction, certainly true caloric restriction has

01:30:47.080 real psychological consequences. And I really would be concerned. I have been concerned for some of the

01:30:54.080 people I know who've done this. If we started trying to do this in the general public, there's social

01:30:58.960 isolation that you get when you're calorically restricting, but then there's the biological changes in

01:31:03.980 the brain and you're hungry all the time. We often don't appreciate those aspects of some of these

01:31:09.560 nutritional interventions. But in the mice, it's hard to know what their psychological consequences

01:31:14.260 are. And what do we know about caloric restriction later in life in the mice versus earlier? The sort

01:31:19.240 of traditional thinking is you have a window in which you can do it early and beyond that, it's not

01:31:24.080 as effective. I think we're going to talk about some data that counter that. And then of course,

01:31:27.620 you have the NIA experiment we talked about earlier. In the monkeys. In the monkeys, where the early

01:31:32.780 fast didn't improve longevity, the late fast appears to have, although that was sort of a

01:31:38.520 subgroup analysis. Hard to draw causation there. What I would say about the mice is that for a long

01:31:43.680 time, the dogma was that caloric restriction didn't work if you started it past, I don't know,

01:31:49.140 15 months of age, which is maybe the mouse equivalent of a 40, 50-year-old person. So most of the early

01:31:54.920 caloric restriction studies were done, like I said, starting sometimes pre-development. The early rat

01:32:00.320 studies were pre-development and then sometimes, you know, six, nine months of age. When I first

01:32:04.720 started in the field, that's kind of what I was told. Like this is a settled question. More recent

01:32:09.320 studies that have been done in some ways more carefully, different diets, certainly. If you do

01:32:14.560 a graded onset of caloric restriction, in other words, don't go right from ad lib to 40% restriction

01:32:20.620 the next day. If you do sort of a graded onset, you can get lifespan benefits from caloric restriction

01:32:26.800 20, 22 months of age. So whether it's as good as starting early, I think the consensus is still

01:32:34.740 that the answer is no. You're never going to get the same magnitude of benefit from caloric restriction

01:32:41.080 starting late as you do starting early. But that could be wrong. So I would say that's the consensus,

01:32:46.880 but I don't think we know for sure whether it's possible if you did it just right, that you could get

01:32:52.700 most or all of the benefits from starting late in life. So Matt, on this topic of CR in mice,

01:32:58.400 again, the dogma has generally been, and I've been victim of this just blindly assuming it to be the

01:33:03.700 case, that CR in mice only works early in life. How applicable is that to humans? I don't know.

01:33:09.700 But a listener of the podcast actually pointed out that, in fact, there are some data that try to get

01:33:14.180 at this question. So there's this Han study 2019, which we'll link to, that looked at 800 female

01:33:20.540 mice. Now, this is a pretty elegant experiment. So for the first three months, they ran these mice

01:33:25.500 out on an ad libitum diet. And then at three months, they were split, randomized to, I believe,

01:33:31.600 a 40% calorie restriction versus ad lib. They ran that out until 24 months. And then each of those

01:33:38.400 groups was further split ad lib versus continued on. So you had one group that was, everybody's the

01:33:44.520 same until three months, one group that spent the rest of their life on dietary restriction,

01:33:48.020 one group that spent the rest of their life ad lib, and then you had the middle groups.

01:33:52.980 21 months, calorie restricted, then to ad lib, 21 months ad lib to then calorie restricted.

01:33:58.400 Okay. So the ends of this were not interesting, meaning the ad lib group lived the shortest.

01:34:03.560 We were looking at the figure earlier today, 1,200 days, roughly.

01:34:06.920 Maximum lifespan.

01:34:07.700 Yeah. Maximum lifespan. That's right. Median lifespan would have been,

01:34:10.780 looking at the graph, about 900 days.

01:34:12.360 Which is pretty good.

01:34:13.020 Yeah. So I was going to say, how does that stack up with what we talked about on the last podcast

01:34:16.900 about length of life of control?

01:34:18.660 That's a reasonable lifespan for control. I think if I remember correctly, this was also done not

01:34:22.640 in C57 Black 6, but in a little bit longer lived hybrid strength.

01:34:26.100 That's right. F1 hybrid.

01:34:27.180 So it's reasonable lifespan.

01:34:28.500 Okay. Looking at the all CR all day, mice looks like they had a maximum lifespan of just below,

01:34:35.900 call it 1,400 and change, with a median that I'm going to say was about 1,150.

01:34:40.980 Good lifespan extension.

01:34:41.820 Okay. So now what's interesting is the middle groups, which is really true. So I'm going to

01:34:45.820 just give you my little iPad so you can look at that table, which will link to this.

01:34:48.480 Okay. I got it right here.

01:34:49.320 You got it right here?

01:34:50.540 I remember the take-home message.

01:34:51.720 You do. Okay. Yeah, yeah, yeah. So what happened to the two middle groups?

01:34:54.760 One thing I would say is I think this is a pretty early onset of CR.

01:34:58.320 It really is. Three months.

01:34:59.620 Yeah. This gets back to what I was talking about before, that it seems likely from the early studies

01:35:04.300 that were done in rats, where they got some of these really, really large effects,

01:35:06.920 that some of the benefits of CR come from actually being restricted during development itself.

01:35:12.380 So I think that's useful to put into context. So then the big question here is what happens

01:35:16.580 if you start caloric restriction late in life? Or what this study did that I'm not really aware

01:35:21.960 of anybody doing previously is kind of the flip. It's almost like a crossover.

01:35:25.780 That's right. Yeah, it totally is.

01:35:26.820 So in this case, when they started CR late in life, there is a significant but not huge effect.

01:35:33.540 Like the magnitude of the lifespan extension is much less than in the mice that were on

01:35:39.320 CR from three months of age. That makes sense. That fits with what else is in the literature.

01:35:45.120 There were earlier studies. I think Steve Spindler did one, not too many years, maybe four or five

01:35:49.520 years before this one that did sort of a similar sort of approach starting around 15 months of age.

01:35:54.120 They saw a significant but not as large benefit from starting late in life. So that seems to be

01:36:00.260 the consensus. The thing that's really interesting here is, you know, what happens if you're CR'd for

01:36:06.140 an earlier period in life, and then back on AL? Do you lose the benefit? And it seems like the answer

01:36:13.040 is no. Those animals actually were longer lived than the mice that went on CR late in life. You could

01:36:19.960 ask some questions about, is it about the total amount of your life that you're restricted? Is it

01:36:25.360 about when you go on and when you come off? And I think in mice, this is still an open question. We

01:36:30.360 don't really know what the mechanisms are. Although the early in life mice had a longer median. The

01:36:38.260 median life expectancy was- The ones that were on CR and then switched to ad libitum. Yes, that's right.

01:36:42.740 They lived a little bit longer, but the bigger difference was the median life expectancy was higher

01:36:47.760 than the flip. Yes. Although I think we have a little bit difference in definitions. I tend to

01:36:52.160 think first about median. You seem to think first about maximum. But yeah, I mean, I think what you're

01:36:56.100 saying is right. The median lifespan is quite different between those two groups. It is the

01:37:01.440 difference. The maximum is very trivial. That's right. The real question here is, well, aside from

01:37:07.020 what does this mean for humans, which I would say we can't draw too many conclusions from humans from

01:37:11.160 this, but what is the underlying mechanism? And is it really just about the total amount of time that

01:37:16.220 you've been on CR? Or is it an interaction with how old you are, the developmental process,

01:37:22.360 and then what happens at the end of life, which is mostly the degenerative process and when you go on

01:37:27.440 CR? One thing that's worth adding to this too, is it's an interesting comparison to what we know

01:37:32.580 about mTOR and rapamycin. So with rapamycin, the data are pretty clear that you can start rapamycin

01:37:39.180 certainly well into middle age and maybe even into very old age and get most of the benefit. So

01:37:45.700 if you compare the curve here where they started the mice on CR at 22 or 24 months, whatever it is,

01:37:51.000 the effect is pretty small compared to CR. With rapamycin, you get almost exactly the same

01:37:56.360 benefit starting at 20, 22, 24 months as you do starting early in life. So that might tell us

01:38:02.840 that there's a difference, right? There clearly is a difference.

01:38:05.340 There's a different mechanism potentially as well. It could be that rapa is doing something different,

01:38:08.980 or it's a different dose effect relative to it.

01:38:11.340 Exactly. So it's an open question exactly why it's different, but it seems to be different.

01:38:16.160 Yeah, I'm really glad you brought that up because we talked about that with Rich Miller on his

01:38:19.480 podcast, which was a fortuitous accident, basically because they couldn't get the formulation of

01:38:25.340 rapamycin in that first ITP study.

01:38:27.540 One of my favorite stories in science, yeah.

01:38:29.320 Yeah, tell people that story.

01:38:30.320 So take a step back. The NIA started this program called the Interventions Testing Program. It must

01:38:36.980 have been the early 90s. And the idea here was, maybe it was early 2000s, sorry. Dating myself

01:38:41.600 again, losing my decade. So early 2000s. And the idea here was, I think, really smart. The idea was

01:38:47.660 that we could create a tool where the scientific community could nominate interventions for lifespan

01:38:55.360 testing in mice. And it was set up so that it would be done in triplicate, three sites. There still are

01:39:00.040 three sites for the ITP. So anybody in the community can nominate any intervention. There's a selection

01:39:06.440 committee that selects them every year. And if an intervention is selected, then the Intervention

01:39:11.280 Testing Program sites start the cohorts of mice on that intervention, you know, in whatever year it

01:39:17.100 was selected for. So sometime back in the early 2000s, Dave Sharp nominated rapamycin. Some ways he was

01:39:23.780 ahead of his time because I think when he nominated rapamycin, it was even before the first

01:39:28.640 invertebrate studies on mTOR and rapamycin. It was right around the same time they were being

01:39:32.320 published. So he, I think, was thinking about it from a cancer perspective primarily. In any case,

01:39:38.220 he nominated rapamycin. It got selected. It went into the cohort. And they typically test five or six

01:39:45.680 interventions or drugs each year. So they have a huge number of animals at each of these three sites

01:39:51.380 that are destined for these interventions to be tested in. And rapamycin was one of them.

01:39:56.380 Randy Strong, who's one of the PIs on the ITP, who's also got a strong biochemistry background,

01:40:02.260 I think recognized pretty quickly that the rapamycin wasn't stable in the food.

01:40:06.680 We could actually come back to this if you want to, because this is relevant for people as well.

01:40:09.960 So, and it gets broken down in the pH of the gut. So basically if they just put the powder in the

01:40:15.700 food, there's no bioavailability. It doesn't get taken up by the mice. And so they recognized that

01:40:20.640 right when they were supposed to start the experiment. And, you know, of course they were like,

01:40:24.700 crap, what do we do? We could just not test rapamycin. And I don't know if it was Randy or 0.97

01:40:29.600 who. Somebody said, well, I think I can figure out a way to stabilize the rapamycin, put it in the

01:40:35.460 food so that we can give it to the mice and we can do the lifespan experiment. I think what they

01:40:39.040 didn't recognize was that it was going to take 18 months or so to figure this out. So once they

01:40:43.900 finally developed what they call E-RAPA, encapsulated rapamycin, it's basically designed

01:40:49.120 so that it won't break down in the gastric pH. Once they developed that, they were now 18 months

01:40:54.180 into this lifespan experiment. Before this, I think everybody, myself included in the field, thought

01:40:58.900 you had to start early in life or you weren't going to get much of a benefit. There was really

01:41:02.660 almost no chance a drug was going to increase lifespan starting that late in life. But fortunately,

01:41:08.040 they went ahead with the experiment starting at 20 months of age. And what they found was that they

01:41:14.740 got this robust lifespan extension from starting with rapamycin treatment at 20 months of age.

01:41:19.920 And just to give some context, that's about the mouse equivalent of a 60 or 65-year-old person.

01:41:25.720 And I love the experiment. I love the outcome, obviously, because first of all, nobody thought

01:41:30.500 it was going to work except maybe Rich Miller. I'll give Rich credit. Maybe he thought it was going

01:41:34.400 to work. And it was really the first time anybody had convincingly shown that you could start

01:41:39.760 intervention in middle age in a mouse and get robust lifespan extension. And for me,

01:41:47.020 honestly, I reviewed that paper. And when I first time I saw that result, I'm like, this changes

01:41:51.200 everything. We actually have a chance for translational geroscience because you might be

01:41:56.120 able to intervene late in the aging process and have significant impact. That was 2009 when that paper

01:42:01.700 came out. So in the 13 years since then, the whole paradigm in the field has changed. Most people who

01:42:06.860 are studying interventions today are studying things that they test for efficacy late in life

01:42:13.240 because that's what we need to do in people. So it was a super important result for the field for that

01:42:18.600 reason. And it all came about by an accident. Nobody would have designed that study that way

01:42:23.260 beforehand.

01:42:24.980 Now, you were going to make a point about the bioavailability around.

01:42:27.780 So this is something that's only recently come across my radar, but I've heard several results now that

01:42:32.280 convinced me that it's true. So, you know, I mentioned the reason why they had to make this

01:42:36.840 E-Rapa is because rapamycin isn't stable at the gastric pH of mice. Same thing seems to be true

01:42:42.160 in people. So there are people who are getting their rapamycin from the rapimmune, which is the

01:42:47.180 brand name generic or the brand name serolimus comes in these triangle-shaped pills. There are also

01:42:52.220 people who are getting it from compounding pharmacies. And I've heard of several cases now

01:42:56.060 where the bioavailability is much lower in the compounded rapamycin in a capsule than in the

01:43:03.380 actual rapamycin.

01:43:04.160 The triangle, the white and yellow triangle.

01:43:06.640 Exactly. So it's just something for people to be aware of. And I don't think most physicians are

01:43:10.600 aware of it. I don't think most compounding pharmacies are aware of it. We've never had

01:43:14.140 it compounded. So we've only prescribed serolimus or rapamycin. You know, it's not a cheap drug. So I

01:43:19.800 can understand why there's a desire to compound it because it's, I don't know, it's got to be like

01:43:23.760 five, six bucks a milligram.

01:43:25.260 Yeah, I think that's about right.

01:43:26.620 That's very interesting. Yeah. So Matt, obviously one of the other things that came out of that review

01:43:30.520 article in the animal stuff was, as you said, the protein restriction. And I think of all the

01:43:36.680 topics in nutrition, this is the one I'm most interested in. I really don't care that much

01:43:41.680 about fat and carbs. Don't tell anybody, but I care an awful lot about protein. You know, in fact,

01:43:46.400 when you came over today, you probably saw me chasing down what was left of a protein shake. And

01:43:51.000 I think I was mentioning to you or to my wife, that's the only part of nutrition that is kind of,

01:43:56.440 I don't want to say a chore, but it's a very deliberate part of how I go about the day,

01:44:01.320 which is I really have to think about it. And the reason is I'm trying to eat a gram of protein

01:44:06.900 per pound of body weight spread out into four buckets. There's reasonable evidence to suggest

01:44:13.220 that if you consume too much protein in one sitting, and it's typically more than about 0.25

01:44:18.880 grams per pound is the general thinking, you're going to end up oxidizing some of that protein.

01:44:24.720 So it's not that it's harmful. It's just that you're not getting the amino acids you need for

01:44:29.620 muscle protein synthesis, which is of course our objective. So that means I'm kind of walking

01:44:33.840 around trying to get 40 grams here, 40 grams there, 40 grams here, 40 grams there. And truthfully,

01:44:39.600 that's not trivial if you're not willing to consume a whole bunch of crap with it,

01:44:44.660 if you're really just trying to focus on the protein quality. So look, the RDA says I'm crazy.

01:44:49.580 The recommended daily allowance of protein is 0.8 grams per kilogram, which is less than half

01:44:57.440 of what I would consume. And by the way, it's not just that I'm making up the amount that I'm

01:45:01.140 consuming. I'm doing it on the basis of other data that suggests that this is the amount of

01:45:05.000 protein consumption you need for optimal muscle protein synthesis. So where does this disconnect?

01:45:10.120 First of all, we can talk about the rodent studies, which is in the biology of aging. I think

01:45:14.860 the RDA question, that's a different question. It's my understanding that that actually was

01:45:20.040 developed to be protein balance for 95% of the population when sedentary. What that means,

01:45:28.360 first of all, that's a minimum amount, not necessarily the optimal amount. And it probably

01:45:32.700 very much depends on lifestyle. And lean body mass to begin with, even though it's sort of normalized to

01:45:38.720 it. And the reason why I bring this up is I think there's a lot, again, a lot of confusion

01:45:42.200 among the general public about what the RDA means. And it's not necessarily a bad thing to

01:45:47.740 be above the RDA in some areas, maybe a lot of areas. So I think that's just worth expanding on

01:45:52.540 just a little bit. I sort of jokingly think of the RDA for protein as what you need to

01:45:56.200 not waste away and wither up and die. Right. So you're not losing muscle mass.

01:46:00.680 So then the question of what is the relationship between protein and aging, I think is a really

01:46:04.980 important one. And it's gotten a lot of attention in the field. And like I think a lot of other

01:46:09.660 things, there's a lack of clarity about what we actually know and what we should be recommending

01:46:14.220 to people. So let's take a step back and start with the animal studies, the mouse studies.

01:46:19.320 I think there it's pretty clear that you can extend lifespan through protein restriction.

01:46:26.220 And there are actually a couple of flavors of protein restriction. You can restrict all protein

01:46:31.080 down to some percentage, some low percentage, or you can restrict specific amino acids, particularly

01:46:36.700 branch chains, tryptophan, methionine, or branch chain amino acids are the ones that have been

01:46:41.780 studied. And again, I make that distinction because it's not really clear that the mechanisms are the

01:46:47.200 same across these different flavors of protein restriction. The common mechanism that does seem

01:46:53.660 to potentially underlie all of these forms of protein restriction is inhibition of mTOR. And again,

01:46:59.140 that's partly why this becomes complicated, especially when we start talking about extrapolation to

01:47:03.440 human. You and I both recognize that inhibition of mTOR can have beneficial effects in the context of

01:47:10.620 aging and healthspan, certainly in mice, almost certainly in people, I would say. And protein is

01:47:15.800 an activator of mTOR. And we know a fair amount about the biochemistry of that, that particularly

01:47:20.620 branch chain amino acids can directly activate mTOR through cestrins, and that's sort of all worked out.

01:47:27.740 And so it seems intuitive that protein restriction would be beneficial by turning down mTOR. It seems

01:47:33.660 counterintuitive that what you were just talking about would be beneficial because you might be

01:47:37.780 hyperactivating mTOR. So we can dive into that. That's the simplest possible mechanism I can think

01:47:42.980 of for why protein restriction, especially branch chain amino acid restriction, would be having an impact

01:47:50.380 on lifespan and healthspan in mice. The other player that seems to be important, particularly in

01:47:56.020 total protein restriction, is a protein called FGF21, fibroblast growth factor 21, that is secreted in

01:48:04.780 response to a low-protein diet and then has effects on liver metabolism and also inhibition of mTOR

01:48:11.960 reduction of IGF-1. So that seems to be required for the lifespan extension that is seen from protein

01:48:18.360 restriction in mice, potentially partially upstream of mTOR and liver metabolism. The interesting thing

01:48:24.600 there is FGF21 overexpression by itself has also been reported to be sufficient to extend lifespan in

01:48:31.620 mice. So it kind of fits that that could be part of the story. So the question, one question is, is

01:48:37.660 protein restriction always beneficial in mice and can we separate it from caloric restriction? This is where

01:48:44.420 you really have to look closely at the studies and determine, did the mice on protein restriction

01:48:49.740 eat less, eat the same amount, and eat more? And it's interesting because you can actually find

01:48:54.580 examples of all of those. And honestly, I don't really understand why that's the case, except it's

01:48:59.000 something about the different compositions of the diet. What does seem to be the case is that when you

01:49:04.000 restrict for certain amino acids, if you're deficient for methionine, for example, or tryptophan, the mice

01:49:10.760 absolutely will eat more and they don't gain weight and they do seem to live a little bit longer.

01:49:15.120 So that could be a somewhat distinct mechanism there that we don't really understand.

01:49:20.340 What was the most compelling evidence you saw when you tried to tease apart

01:49:24.220 the relationship between protein and total intake?

01:49:27.540 I think the branched chain amino acid and methionine restriction studies are pretty clear

01:49:32.100 that those animals are consuming more calories than certainly if you matched a weight than the

01:49:38.340 ad libitum mice and they're living longer.

01:49:40.500 And what do we think is the route or mechanism through which methionine exerts this effect?

01:49:45.680 That's still really being worked out. There are lots of mechanisms that have been proposed. I

01:49:49.440 suspect mTOR plays a role. Methylation, methyl donors are important for a bunch of different

01:49:54.940 epigenetic modifications. So there may be a role there going back to the epigenome that we talked

01:49:59.500 about. Methionine is the first amino acid in every protein. So there could be effects on protein

01:50:04.860 synthesis. There's evidence linking methionine restriction to sulfur amino acid. Biology,

01:50:10.280 which has been implicated in aging. So it's hard to know, and maybe it's not one thing.

01:50:14.960 And those all sound like potentially just a substrate reduction problem, right? Like

01:50:18.520 less sulfur cross-bridging, less protein synthesis.

01:50:22.100 You know, if you look back in the literature in the invertebrate, inhibition of protein synthesis

01:50:26.440 in some cases is enough to extend lifespan. And of course, mTOR is a primary regulator of protein

01:50:32.660 synthesis. So when you inhibit mTOR, you can also inhibit protein synthesis. That's part of the

01:50:37.820 challenge here is this network is so interconnected that when you tweak one part of it, you have

01:50:42.760 effects throughout the network. And it's really hard to know which of those effects are causal.

01:50:47.180 So let's talk about time course. When you consume a protein-rich meal, do we have a sense of how long

01:50:53.620 mTOR is being activated in response to that set of amino acids?

01:50:58.560 I'm sure somebody does. I don't know the answer to that. Almost certainly, it's going to depend on

01:51:05.380 what you eat in combination with the protein, when you eat, how active you are.

01:51:10.340 I remember talking to David Sabatini about this through the lens of BCAA drinks. If you're going

01:51:16.240 to pound branched-chain amino acids during a workout because you want as much anabolic signal as

01:51:22.320 possible. And this is a couple of years ago, so maybe things have changed. But based on that

01:51:26.720 work, I think Bobby Sutton had done the work in his lab, if I'm getting his name right, was it Bobby

01:51:30.920 Sutton was the guy who did that science paper that looked at the leucine sensor on mTOR?

01:51:35.920 The answer was it didn't stay on long at all. Free amino acids were so short in their ability to turn

01:51:43.900 on mTOR that unless you had an intravenous drip of this stuff, it was going to be very difficult. So much

01:51:50.680 so that the idea of using BCAA analogs to treat sarcopenia was going to require drugs that could

01:51:58.940 stay on much longer. Is that kind of within your frame of thinking?

01:52:03.080 I think so. And I think it also makes sense in a biological context. I mean, cells and tissues,

01:52:09.080 you know, again, this gets back to the whole homeostasis concept. Cells and tissues have evolved

01:52:14.100 to maintain metabolites, and amino acids are metabolites, right? They're involved in many

01:52:18.220 different metabolic reactions within certain levels. And there are all sorts of mechanisms

01:52:22.580 to ensure that if a metabolite gets outside of that range, that we soak it up, we do something

01:52:28.940 else with it. So I think it makes sense that you're probably not going to have a persistent increase

01:52:35.400 in branched-chain amino acids far outside the normal range. What I would say, though, is that slightly

01:52:40.880 elevated branched-chain amino acids chronically can have big effects on the sort of downstream

01:52:47.240 processes. And there are some inborn diseases of childhood where you have elevated levels of

01:52:51.940 branched-chain amino acids. We know that there are consequences to even having somewhat modest

01:52:57.740 increases in mTOR, hyperactivation of mTOR signaling chronically. So again, I think the context really

01:53:03.940 matters. But yes, it's my intuition that it's probably hard to get very large, persistent increases

01:53:12.040 in mTOR simply from taking a branched-chain amino acid supplement. It doesn't mean it couldn't have

01:53:17.900 some effect on muscle building right after a workout, but I suspect it's hard to have long-term

01:53:23.360 persistent effects on mTOR.

01:53:24.540 I mean, the anabolic data suggests it's not necessary. It's just, again, muscle protein synthesis

01:53:29.540 window is open long enough that simply delivering a great source of whey protein in the hours after a

01:53:37.060 workout seems sufficient to not restrict muscle potential growth.

01:53:41.300 I think the other thing, though, that is also important to appreciate, and this is true with

01:53:45.220 rapamycin as well, I think a lot of people get confused about this, is it's not only about

01:53:49.100 how high mTOR gets turned on or how low it gets turned down, it's also about where that happens.

01:53:56.840 People for a long time thought that rapamycin would cause muscle loss. We don't see that. I mean,

01:54:03.160 we just don't see it in mice, and we don't see it in people, and I think it's probably because

01:54:07.360 I'm guessing you're not seeing it in dogs.

01:54:09.240 We have not seen anything to suggest that in dogs, yeah. I'm guessing that has as much to do

01:54:13.800 with how much, maybe more to do with where mTOR is being affected than how much we're inhibiting mTOR

01:54:19.360 or when we're inhibiting mTOR, and so I think the same thing is probably...

01:54:21.740 And do we know where the selectivity of rapamycin is? I mean, is it more selective in

01:54:25.560 hepatocytes? Is it more selective in adipose tissue? I mean...

01:54:29.500 I don't know of any good studies that have really carefully looked at this. There have

01:54:33.840 been a few studies in mice that tried to look at tissue mTOR signaling in the context of rapamycin.

01:54:39.640 It's a very technically challenging problem.

01:54:41.560 Well, and this is what I was just going to say. It gets even more complicated because

01:54:44.720 even in a mouse, where you can essentially control almost everything, what the mice are eating

01:54:49.780 and when they last ate has, if anything, as big, maybe bigger effect on mTOR signaling than rapamycin.

01:54:57.540 There have been, like I said, a couple studies that looked at this, and I'm not sure...

01:55:00.400 And they got different answers, and I'm not sure who to believe because I don't think either was

01:55:04.240 wrong. The only way I could imagine doing this is you have to be able to do subtractive studies

01:55:08.180 where you have to be able to do it in the context of a whole bunch of different diets first,

01:55:12.060 get kind of a baseline that you then pull out of potentially what you're seeing. But I mean,

01:55:16.240 it gets... It's complicated. And again, that's why I often will gravitate back towards

01:55:21.060 what are the functional consequences we can actually measure? Sure, I get it. You think that treating a

01:55:26.900 mouse with rapamycin is going to cause sarcopenia? Let's do the experiment and find out. The answer

01:55:31.060 is no. It doesn't. Right? So that tells us it's at least not as simple as we thought it was going to

01:55:35.340 be. Now, what about the flip side of that is more protein versus less protein activating mTOR in a way

01:55:42.240 that is counterproductive? I think it can. I think there are probably certainly cases where it can.

01:55:47.900 I don't know that anybody has really carefully done that study in mice. There was a study... It's a

01:55:53.120 really interesting study by Steve Simpson and colleagues where they did this nutritional geometry

01:55:57.820 work where they basically looked at different compositions of carbohydrates, fats, and proteins.

01:56:03.160 In Australia? Yeah, exactly. And, you know, looked at... I don't remember how many diets. There's a whole

01:56:07.660 range of diets, different compositions of the three macronutrients. Tried to control for caloric intake,

01:56:12.780 which is hard, as you can imagine, but I think they did a pretty good job. And then asked, what does it look

01:56:17.440 like in terms of metabolism, energy expenditure, lifespan? So the lifespan studies, I think, are pretty clear

01:56:23.200 that most of the diets where the mice lived the longest were towards the low end in protein. But there

01:56:30.820 were some things that I think called into question exactly what was going on there, because it wasn't the

01:56:36.340 case that the mice that were energetic, the diets that were energetically lowest gave the longest lifespan,

01:56:41.940 as you might expect from caloric restriction. And the diet that actually gave the absolute

01:56:46.240 longest lifespan had like, I don't know, it's like a 40% protein in it. So the way I interpret that is

01:56:51.240 that there are many ways to get to longevity. And how calorie restricted was that? They were not

01:56:56.660 calorically restricted at all. So you're saying that a diet that was ad lib with 40% protein had the best

01:57:02.800 outcome? The best absolute lifespan, yes. How do we even reconcile this body of literature? My view is

01:57:09.300 there are probably multiple paths to longevity. And we really don't understand the inner relationships

01:57:16.940 of these macronutrients in the diet with enough sophistication to get beyond sort of broad, general

01:57:24.040 predictions. And again, this is an area where I believe, like I can't prove it, but my intuition from

01:57:30.200 the data that I've seen and just my observations of people is that in humans, this relationship between

01:57:36.440 protein and health during aging is probably very different than it is in mice. I think mice are

01:57:42.260 able to tolerate a very low protein diet without some of the consequences that we see in people.

01:57:48.380 That's my intuition. I don't know that that's true, but that's my intuition.

01:57:51.340 I mean, it's my intuition well as well, because clinically what we see in what I call the death

01:57:57.940 bars, the death bars is our internal nomenclature for how people die. We just constantly look at death

01:58:03.220 bars and we double click and double click and double click all the way to try to tease out

01:58:07.480 everything that is reducing lifespan and health span. And the problems that occur in humans when

01:58:13.960 they are under muscled are insane. And it ranges from the metabolic consequences of being under muscled.

01:58:21.360 Our muscles are a sink for glucose. They are the single most important sink we have for glucose

01:58:28.400 and our ability to tolerate glucose and maintain glucose homeostasis in the presence of larger,

01:58:34.560 more metabolically healthy muscles is the difference between having diabetes and not having diabetes.

01:58:39.960 Furthermore, when you think about sarcopenia and when you think about osteoporosis, which again,

01:58:46.240 I just don't think we're talking about how these things impact animals. Like we don't study any animal,

01:58:51.140 including primates in a setting where sarcopenia and osteoporosis are problematic.

01:58:55.740 And yet I would ask anyone to consider the entire population that they know over the age of 75.

01:59:03.780 And I would ask you take every person that is alive today that's over 75 and tell me how many of them

01:59:10.380 are not suffering at least some consequence of one or both of those phenomenon. And if somebody did that

01:59:17.100 analysis, I would be shocked if we didn't find at least 80% of people over the age of 75 are

01:59:24.300 experiencing this. And if you look at the activity, just monitor the activity level. Once they hit 75,

01:59:30.640 they fall off a cliff. So muscle mass dramatically plummets, activity levels dramatically plummet,

01:59:36.500 difficult to say which one's feeding which, but there's no question that something is happening

01:59:41.140 to our species at about the age of 75 that is a structural problem. And none of this other stuff

01:59:47.200 matters if that sucks. I don't care if I live to 100 and don't have cancer if I'm an invalid for the 0.73

01:59:55.140 last 25 years and I can't play with my grandkids and throw a ball. For me personally, I'm not saying

02:00:01.440 that's a, that's not a view that everyone should take in the world. I'm just telling you that's my

02:00:04.860 view. I mean, I think that's absolutely correct. I guess the question, and I think this is still where

02:00:09.920 some of the confusion comes from, is how important is dietary protein in that maintenance of muscle

02:00:16.680 or loss of muscle in people who are going to go the wrong direction? I think the data is that it is

02:00:21.880 quite important. There are lots of studies that have compared the RDA versus the double RDA standard,

02:00:29.660 and it's a significant difference. Protein makes a very big difference following, obviously,

02:00:36.280 the training that is necessary to stimulate muscle protein synthesis. So I think those have to be

02:00:40.920 coupled to some extent. Absolutely. I believe there are data, and I hate when I have to say

02:00:45.640 this because I just, I'm going to say something and it's going to be wrong and 20 people are going

02:00:49.180 to respond. It's okay, I do it all the time. Don't worry about it. In anticipation of the fact that

02:00:52.220 there are data that I've read and I don't have the memory I once had, I believe there are data that

02:00:57.120 show just the protein difference alone can make some difference, but it's not nearly the difference you

02:01:02.960 get when you pair it with hypertrophy training. That's my recollection as well.

02:01:06.640 Which brings us to the interesting question then, why is it that there is a camp? And in my field,

02:01:13.260 it's a pretty vocal camp in the aging field that would argue that low protein is the best

02:01:20.600 nutritional strategy for aging and health span in people. And this gets back to the point I kind of

02:01:28.660 started with, which is that you can find the answer you want for almost any question in this area that

02:01:34.420 intersects in nutrition and aging. There will be a study that will fit your belief. So I think you

02:01:39.340 really have to be careful, or I try at least, to take a global view and try to understand what is

02:01:44.680 the totality of the data say. But there are epidemiological studies and one in the field

02:01:50.720 most people will point to when they go to humans and they talk about low protein. The study that Walter

02:01:56.760 Longo was the senior author on and Morgan Levine was the first author on where they looked at

02:02:01.500 protein consumption and all-cause mortality as a function of age in people. There were some

02:02:08.220 studies in, I think they had some yeast studies in there as well, maybe some cell culture studies.

02:02:12.220 The take-home message was that low protein is beneficial up to about 65 years of age. And

02:02:20.000 then once you get above 65 years of age, it kind of flips and people who ate a higher protein diet

02:02:25.560 have lower all-cause mortality. I should be clear, when I say beneficial, we're talking specifically

02:02:30.100 about all-cause mortality. Which at the end of the day is a very important metric.

02:02:36.020 Sure, you want to be alive.

02:02:37.040 Yeah, it's not the most important metric necessarily. You could argue it's equally important to the health

02:02:41.900 span metrics. Okay, so let's make sure people understand what that means. That means below

02:02:46.260 the age of 65, the epidemiologic data in this study suggested people eating less protein had

02:02:53.040 lower mortality in all-cause mortality. And above 65, you saw that reverse.

02:02:57.520 That's right.

02:02:58.000 Now, did that paper make any attempt to quantify the net impact on mortality? Because the very

02:03:03.720 misleading thing about an assessment like that is when you look at mortality adjusted by population

02:03:10.380 before the age of 65, it's relatively low. Above the age of 65, it goes up very non-linearly.

02:03:19.640 So when we do our death bar analysis, it's like, this is the death per 100,000 people.

02:03:26.740 If you're 40, 50, 60, 70, 80, 90, like, you know what I mean? It just becomes insane.

02:03:33.120 So you could argue through that analysis, you're much better off with a high-protein strategy,

02:03:38.940 even if it's throughout life, because the absolute reduction in mortality would unquestionably be

02:03:44.420 lower as a result of the benefit you would have later in life.

02:03:48.700 I absolutely agree conceptually with what you said. The impact of a change in mortality late in life is

02:03:54.060 going to usually swamp the impact, certainly swamp the same impact on mortality early in life.

02:03:59.660 I think the question here is, what are the relative effects? They did model this a little

02:04:05.220 bit, and it is in their model. I couldn't get the data. I can't evaluate exactly what they did.

02:04:11.400 But in their model, the relative risk crossed somewhere in the 60s, right? In other words,

02:04:18.540 your total mortality benefit was lower eating a high-protein diet. I think it was starting

02:04:25.000 somewhere in the 60s. And that actually surprised me, because for exactly the reason you said,

02:04:29.020 the relative impact of the high-protein diet early in life would have to be an order of magnitude

02:04:35.100 greater than the relative impact of the... So I'm sorry, say what they're finding was again

02:04:39.280 at the age of 60? I don't remember the exact number. It's in the paper. You can see the curves

02:04:43.180 crossed. It was much later than I thought it would be, given that 65 was the point that they kind of

02:04:47.340 pick. So I would have thought maybe in your 50s. So I actually tried to do my own modeling of this

02:04:52.220 off of the data that I could find on relative risk for low and high-protein. Again, what you define

02:04:58.000 low, what you define high, you know, they're... And you're trying to ask the question, when should

02:05:01.880 you switch the diet? Or maybe more formally, at what age does the risk equal out? Yeah, what's the

02:05:07.560 crossover? Yeah. And what did you find? So mine was closer to like 50. That's the point where once you

02:05:13.440 get past 50, the benefit of a high-protein diet on mortality seems to outweigh any detriment that you

02:05:20.100 would get from starting earlier. So that's odd to me because whether it's 50 or 60, Matt,

02:05:24.360 it's a benefit on mortality, which is really where more of the argument is. There can't be any benefit

02:05:30.920 on healthspan. From low-protein, you mean? No, from high-protein. Early in life or late? Why? Why

02:05:35.820 can't there be a benefit? Oh, late in life, I'm saying. Why not? Well, I'm saying like if you're

02:05:39.680 protein-restricted late in life... Low-protein has no benefit on healthspan. Yeah, yeah, yeah. So I would agree

02:05:44.900 with you intuitively. I'll exclude special cases. So I'm not talking about people who have renal

02:05:49.800 insufficiency for whom they have to restrict. I agree with you conceptually. The only thing that

02:05:53.780 makes me hesitate a little bit is I've just seen, like I was talking about the mouse rapamycin

02:05:58.200 experiments where everybody who knew anything about muscle said that if you gave a mouse rapamycin

02:06:03.260 throughout life, it was going to get sarcopenia. And that just didn't happen. No, but I'm saying we

02:06:07.240 have clinical data that suggests that when people over the age of 65 are protein deficient versus

02:06:12.780 protein significant, there's a huge difference in muscle mass. Which we know is going to be

02:06:17.780 associated with frailty and poor outcome. I would totally agree with that. I don't know.

02:06:22.660 Do we have controlled studies where people were eating low-protein and doing resistance training

02:06:26.720 late in life? There are nuance here that could complicate things. But I think in general,

02:06:30.260 you're probably right. I think the other area where this gets

02:06:32.200 very complicated is the, I don't want to say by necessity, but just by convention, we use IGF-1

02:06:39.740 as a biomarker for protein intake. It's certainly associated with protein intake, but you want to tell

02:06:45.860 people what IGF-1 is, where it comes from, and what it's a proxy for?

02:06:49.600 So IGF-1 is insulin-like growth factor one. It's a hormone that's in the growth hormone pathway. So

02:06:55.460 you can think of as a growth-promoting hormone. It's part of this central pathway that promotes

02:07:00.660 growth in many, many different tissues. So if you have high growth hormone levels,

02:07:04.780 you'll have high IGF-1 levels and high mTOR. This is a part of the mTOR pathway as well, upstream of mTOR.

02:07:10.260 The reason why people have been really interested in IGF-1 in the field of aging biology, it comes

02:07:17.640 from studies, again, in the very simple laboratory model systems. So the most famous and one of the

02:07:24.280 first genes that was shown to clearly from a mechanistic perspective affect aging is it comes

02:07:29.940 from Cynthia Kenyon and even Tom Johnson a little bit before her, which is the insulin-like receptor

02:07:35.580 in C. elegans called DAF-2. And Cynthia published a classic paper showing that if you make a mutation

02:07:40.700 in DAF-2, it could double the lifespan of worms, and they seem to be healthier about twice as long.

02:07:46.160 And what that mutation does is it turns down signaling through this pathway. Now, a little

02:07:51.920 bit more complicated in worms because it's called the insulin IGF-1-like signaling pathway. So it's not

02:07:57.480 identical. There's one path in worms that kind of takes the place of both IGF-1 signaling and

02:08:02.520 insulin signaling, but you can kind of think of them as equivalent. And then there are a whole

02:08:05.740 bunch of studies in mice for mostly mutations in the growth hormone upstream signaling upstream of

02:08:13.860 IGF-1 that lead to increased lifespan. So there are- So this means GH does not activate the production

02:08:20.520 of more IGF-1. That's right. So you have through a variety of mechanisms- You have high GH, low IGF-1

02:08:25.580 animals. Well, low GH signaling. But they probably are high in IGF-1. Oftentimes it's the receptor that's

02:08:30.600 mutated. That's right. So those animals tend to be very long-lived. They rival caloric restriction

02:08:36.000 in terms of the magnitude of lifespan extension. And there are several different mutations in that

02:08:40.500 pathway. The mutations in IGF-1, I guess I should know the current state of that literature a little

02:08:45.700 bit better. It's complicated. And there have been some controversies in the field about the different

02:08:50.980 mutations that directly affect IGF-1 itself and the effects on lifespan. So I'm not going to wade into

02:08:58.740 that because I think it still hasn't been resolved. But there's no question that mutations that reduce

02:09:04.140 growth hormone signaling in mice extend lifespan. Now, it's important to understand though that with

02:09:11.180 one exception, those studies are all cases where the animals are growth hormone signaling deficient

02:09:17.260 through development. So they are very, very small animals. And then they have constitutively low

02:09:24.400 levels of signaling through that pathway for the rest of their life. There's one study that I think

02:09:29.340 it used a monoclonal antibody to the IGF-1 receptor in mice. This is from Nir Barzilai and Hasi Cohen,

02:09:36.240 where they treated mice with this antibody late in life. And they got, you know, a reasonably sized

02:09:40.640 lifespan extension. I think it was, I don't know, 14, 15% median lifespan.

02:09:45.820 That was an antibody that did not penetrate the CNS, if I recall.

02:09:49.460 I think so.

02:09:49.840 I remember Nir talking about this and saying, you would get all the benefits of IGF in the brain

02:09:54.980 without the benefits of IGF in the, or without the potential harm of IGF in the periphery.

02:09:59.360 Another complication, right? Where the effects of IGF in the brain might be fundamental on,

02:10:03.900 for health span and cognitive function, might be fundamentally different than high IGF-1 in the

02:10:08.420 periphery. So that study, I think, is the best evidence in mice that you can get some benefit

02:10:13.740 specifically from reducing IGF-1 signaling in middle age.

02:10:17.900 And this is such an important question I get asked all the time. I have a lot of patients

02:10:21.420 that are asking to be put on growth hormone. We just don't do it. The reason is, I just am not

02:10:28.060 comfortable with, I don't see enough data in humans to suggest that it's necessarily safe.

02:10:34.700 Conversely, I don't really see evidence to suggest it's not. This is sort of the weird thing with growth

02:10:39.220 hormone. Like if you buy hook, line, and sinker, the argument that more growth hormone equals more IGF

02:10:44.380 equals more mortality. And you look at how much growth hormone is being used. I mean,

02:10:49.460 it is hands down the most abused drug in sports. It's first, second, third. Nobody's even within

02:10:55.740 the zip code. And this is going back 35, maybe 40 years, probably to the early 80s. Where are the

02:11:02.140 bodies? There need to be more bodies. So I'm stuck with, like, I don't see where the bodies are.

02:11:08.420 But at the same time, it's still a bit of a leap for me. And I don't have the luxury of rapamycin data

02:11:14.400 where I can at least point to all of the humans who have taken rapamycin for 23 years. And we know

02:11:21.100 what that looks like. And then even though it's not for gyro protection, and then all of the mechanistic 0.98

02:11:26.360 stuff that is consistently pointing the right way. So there's going to be some patient of mine

02:11:30.080 listening to this saying, Peter, you almost talked me into taking growth hormone based on your

02:11:34.240 discussion. And it's, no, I can't. It's funny. I even took it for a week after my shoulder surgery.

02:11:39.940 I had sort of looked at some literature using GH and anabolic steroids to help with recovery.

02:11:45.880 And it could have been true, true, and unrelated. But I felt the worst I've ever felt after a week of

02:11:51.460 growth hormone and nandrolone. And I was like, yeah, I'm done. Now, again, I think it was, I happen to be

02:11:57.580 sick as well. But my blood pressure went up. My blood sugar went up. I felt like crap. I couldn't sleep.

02:12:02.520 Again, a lot of confounding factors, shoulder surgery and a nasty virus. So it could all be

02:12:08.520 irrelevant. So first of all, obviously, I've never given growth hormone to anyone. I've never taken

02:12:12.820 growth hormone. I'm not an expert in the human application of growth hormone. But I've certainly

02:12:17.280 tried to follow that literature. Because based on the mouse studies, you would have predicted,

02:12:22.580 right, that growth hormone therapy should be bad. Should be the most toxic therapy you could give a

02:12:27.000 human. Yeah, certainly should cause increased risk for a bunch of different diseases, including cancer.

02:12:31.240 It's mostly cancer. And my understanding of the literature here is that, like you said, it's not

02:12:36.960 clear that there are significant benefits, particularly for strength. I think there's some

02:12:40.600 evidence that muscle mass may increase, but strength doesn't. But it's also not clear that

02:12:44.440 there's any real detriment, that there's any significant risk, which is a little bit surprising.

02:12:49.120 Yeah, it is surprising. And I do have a couple of patients who have taken it. Usually other doctors

02:12:55.740 were prescribing it or, you know, they came in under the care of somebody else. And they all seem to

02:13:00.700 claim they feel infinitely better on it. There may be something to that. It might be that in 20 years,

02:13:06.660 we have enough data to say, you know what, by the time you're 60, you should just be on a slow amount

02:13:11.960 of growth hormone for all of these reasons. I'd love to see somebody do this study. Because it's a very

02:13:18.840 important question to be asked. And I also think we have enough data to suggest that such a study is not

02:13:24.280 unethical. In other words, we don't have an abundance of data. In fact, we have a paucity of

02:13:29.820 data suggesting it's harm, that it would justify ethically doing a study like this. That's sort of

02:13:35.320 a hope I would have, because I really find this to be one of the most confusing questions in this

02:13:39.880 space. I agree. And again, this is sort of why I personally have settled around the idea for now,

02:13:45.360 at least, that IGF-1 particularly is probably not that informative in people, particularly, you know,

02:13:53.220 once you get past 50 years. 50 years is arbitrary, but that's kind of where I would put the number.

02:13:57.840 Obviously, again, IGF-1 itself is complicated, because you don't really know what that means in

02:14:03.300 terms of IGF-1 signaling and downstream activity. Yeah, important, I guess, for people to understand

02:14:07.820 that. Just like testosterone is mostly bound to sexism binding globulin, there's only a small

02:14:13.640 amount of testosterone that's free. It's the same with IGF-1. It has these IGF-BPs,

02:14:18.380 these are binding proteins that bind most of it. And therefore, total IGF is not really

02:14:25.220 completely informative as to what's happening, even in terms of the quantity that's there for

02:14:29.200 signaling, because it's not the unbound portion of it. So, some people look at things like IGF-2

02:14:35.380 IGF-BP ratio. The bigger that number is, in theory, the more IGF signaling you would have. But,

02:14:41.660 you know, this gets to now when you look at sort of the epidemiologic curves,

02:14:45.420 which on the x-axis would show in, you know, deciles or quartiles or whatever buckets,

02:14:51.720 IGF levels rising. And then on the y-axis would show you mortality. I've never seen one of those

02:14:57.780 curves that just goes up. Sometimes they're U-shaped. Sometimes they're downsloped. Sometimes

02:15:02.660 they're flat. And it depends on the indication. But the story seems much more complicated than IGF is

02:15:09.360 bad. You know, going back to the Dean paper that we were talking about, again, it's an important paper.

02:15:14.240 It's a well-done paper. You really have to recognize that population you're looking in

02:15:20.140 might make a big difference as well. If you're talking about a population of people where 30%

02:15:25.240 of them are obese, some high percentage have metabolic disease or diabetes, having high IGF-1

02:15:31.500 in that context might be very different than somebody who is appropriate rate, exercising,

02:15:39.020 eating a high-protein diet, right? And again, those kinds of things don't typically come out

02:15:42.860 in these epidemiological studies. The other thing I'll say is today, I went and tried to look

02:15:47.380 through the literature and see what other studies have shown that same relationship. And they're all

02:15:52.840 over the place. You can find studies that really don't show, or protein consumption particularly,

02:15:57.040 you can find studies epidemiological that really don't show any downside to eating a high-protein diet

02:16:02.500 in people. It's hard for me to draw too much confidence that high-protein is significantly

02:16:08.980 detrimental when you're younger than 50. And I feel pretty confident that a higher, at least certainly 1.00

02:16:16.660 higher than the RDA level of dietary protein intake when you're above 50 is beneficial, particularly if

02:16:24.220 you're exercising. I mean, that's where I would be a little bit concerned. If you've got somebody who's

02:16:29.280 overweight, obese, diabetic... Sedentary.

02:16:33.060 So high-calorie plus high-protein could be problematic.

02:16:36.400 Totally agree. And by the way, I frankly think a lot of the epidemiology is tainted by that.

02:16:42.440 It's high-protein in the context of high-calorie.

02:16:44.860 Exactly. The other thing that I think is also potentially interesting to think about in human

02:16:48.820 are these people who have mutations in the growth hormone pathway. So this is now maybe more akin to

02:16:55.360 these mouse models where they have low-growth hormone signaling from development, even in

02:17:00.040 utero, potentially. They go through their entire lives. A couple of studies. Again, Walter Longo,

02:17:04.160 obviously prolific in this area, had a study in little people of Ecuador, right? There have been

02:17:08.720 several studies, but the most...

02:17:10.360 Leron-dwarf syndrome.

02:17:11.480 Yeah, that's right. The Leron syndrome. Yeah. The most famous study is one that was published

02:17:15.200 in Science where they looked at lifespan and age-related health outcomes in the people with

02:17:22.140 low-growth hormone signaling versus controls in their same environment.

02:17:26.400 Environment. Yeah. It's a really fascinating study. The interesting things are there's no

02:17:30.880 difference in lifespan, but the people with low levels of growth hormone signaling, the reduction in

02:17:36.380 cancer risk is profound. I don't remember the exact numbers, but I think it was zero. There was one

02:17:42.900 person in their cohort who developed a cancer. I don't remember what it was, and she was treated,

02:17:46.760 and then she lived the rest of her life. But none of them died from cancer. And the rate of diabetes

02:17:51.660 was lower in the little people. But Ecuador, at least that part of Ecuador at that time, had a very low

02:17:58.460 diabetes rate to begin with, something 5%. So it's a little bit harder to say. But certainly cancer,

02:18:02.620 dramatic reduction in risk of cancer. So why didn't they live longer? And it's a little bit

02:18:08.340 ambiguous. They don't really say, but you know, they say that there is a higher, much higher rate of

02:18:13.180 alcoholism, liver failure, and accidents. This gets back to the social and psychological consequences

02:18:20.540 in humans that are just different than we have in mice. The growth hormone deficient mice aren't going

02:18:26.540 to be subject, well, they might be probably not subject to the same social pressures that somebody,

02:18:33.040 you know, has very low growth hormone signaling in people is subjected to, which may contribute to

02:18:38.100 other things later on, like alcoholism. So anyways, fascinating though, biology, which is consistent

02:18:43.440 with the idea, I think, that you can impact at least a subset of age-related biology by being

02:18:50.460 constitutively low in growth hormone through your entire life. You know, what would happen if you did

02:18:55.780 that in bursts, you know, like post-developmentally, just after puberty, say from your 20s and 30s,

02:19:01.360 who knows, right? We don't have any, there are no naturally occurring examples of that. I don't,

02:19:05.780 or very few that we could look at and actually evaluate.

02:19:09.480 By the way, do we have examples? Is there enough data to look at people with acromegaly during different

02:19:16.640 periods of their life to see if that's had the exact, do we see a higher incidence of cancer?

02:19:21.240 I don't know the answer to that. Those populations would be relatively small, but yeah, maybe,

02:19:26.280 maybe it's possible.

02:19:27.320 Yeah. It seems like, I imagine somebody's looked at that, the incidence of cancer in people with

02:19:31.440 adult onset acromegaly or something to that effect. The other thing I would say on the IGF thing before

02:19:35.960 we leave that is the interplay with insulin. And so high insulin, high IGF, low insulin, low IGF,

02:19:43.500 low insulin, high IGF. I mean, these are very different physiologic states. It's very difficult

02:19:48.780 to think that we're teasing those out when we look at broad swaths.

02:19:53.500 I think this just comes back to the fact that these, especially these epidemiological studies

02:19:57.360 are a mixture of normal people typically. And so the lifestyles most people are living are what

02:20:05.240 gets weighted in those types of analyses. And that may be very different as we talked about,

02:20:10.420 if you are normal weight, high protein, maybe high calorie, because you're extremely active.

02:20:15.060 Then if you're overweight, sedentary, and eating a calorie diet, I really think that's

02:20:20.400 underappreciated and probably really important. And thinking about the cancer risk, this is going

02:20:25.380 to be some pure speculation on my part. There's no question, I don't think, that high growth hormone

02:20:31.100 signaling and high IGF-1 signaling, everything else being equal in a person leads to a higher risk of

02:20:37.880 developing cancer.

02:20:39.200 You don't?

02:20:39.520 I don't. I think that's true.

02:20:40.860 Oh, you do think that's true. Okay.

02:20:41.940 I believe that that's true. Everything else being equal, of course, everything isn't going

02:20:45.440 to be equal. But if we just look at that one variable signaling through that pathway,

02:20:50.100 higher signaling, higher risk of cancer. So then if it's the case, which we could make an argument

02:20:55.400 that that doesn't seem to be the case, at least in certain populations of people that high growth

02:21:01.180 hormone signaling or treating with growth hormone dramatically increases the cancer incidence.

02:21:05.400 So why is that? Or in people who are...

02:21:06.940 And by the way, we should also differentiate between high causes it versus low removes it.

02:21:12.440 Just because we have a genetic example of where not having it creates a deficiency of cancer.

02:21:20.500 So going from sort of 100 to 30 decreases cancer doesn't mean going from 100 to 30 increases cancer.

02:21:27.680 100 to 130 increases cancer.

02:21:29.680 That's right.

02:21:30.040 I mean, we don't know. It could.

02:21:31.020 The word you use there is interesting because you said removes it. No, this isn't what you meant.

02:21:34.940 But this is, I think, something that is also important to appreciate. So to go from pre-initiation

02:21:40.980 of cancer to cancer to metastasis to, you know, somebody dying from it, there's steps that have

02:21:47.320 to happen there. And there are different defense mechanisms that act at each of those steps.

02:21:52.560 My guess is growth hormone and IGF-1 is primarily acting at the very early steps,

02:21:56.880 where we know that if you promote cell division, that that is a sort of a permissive early

02:22:02.220 environment for mutations to happen and cancers to get a foothold. In most cases, it seems to be

02:22:08.860 the case that those early cancers are detected and wiped out by our immune system. One of the

02:22:15.100 reasons why I think a lot of cancers become more prevalent as we get older is because the function

02:22:21.240 of the immune system to detect and clear those cancers declines. There's obviously other stuff going

02:22:26.000 on accumulation of senescent cells, which contributes to this process. But if you are,

02:22:31.000 say, I shouldn't even say this because I bother people about the biological clocks. Let's just say

02:22:35.180 though, theoretically, you're a 60-year-old person, but biologically, because you are exercising,

02:22:42.340 eating an appropriate diet, biologically, you're 40 years old. At least your immune system is

02:22:47.180 functioning like a 40-year-old. You might have a little bit higher IGF-1. You might have a little bit

02:22:52.020 higher of that early cancer risk, but you have a much lower total risk of developing cancer because

02:22:56.960 your immune system has a much better chance of catching it and getting rid of it. And those

02:23:00.880 are things we don't even think about. Well, Matt, I don't know that we settled anything today.

02:23:05.260 Pretty safe to say. We've probably, for the listener, created more questions than answers.

02:23:09.720 No, I'm sure we've done some good. It's a complicated question. And you know,

02:23:13.300 we actually did not dive into the genetic interaction with caloric restriction. So I mean,

02:23:17.100 I think the take-home there is that even in mice where we can control everything else, if you look

02:23:21.280 across genotypes, you get different results from the same diet and the effect of caloric restriction

02:23:26.760 on lifespan. So maybe we can't answer the big detailed questions. I guess the take-homes I would

02:23:32.880 have are, we've learned a ton from these nutritional studies in laboratory animals about the biological

02:23:38.420 mechanisms. We've learned a lot about which proteins and pathways are important. And that has led us to

02:23:45.280 things like rapamycin, which might be a more effective intervention in humans. So they have

02:23:51.760 value for that. The other take-home that we've talked about is you don't have to worry about

02:23:57.080 every little detail. Most people can get a big chunk of the way there by eating a relatively healthy

02:24:05.100 diet. Don't worry so much about how much protein, how much carbs, how much fat, eat good foods,

02:24:09.960 don't overeat, and be active. Exercise. I do worry a little bit that society does this, but scientists

02:24:16.440 do it sometimes too when we start really getting into the weeds and making recommendations to people

02:24:20.940 that we overthink things a little bit. Give people anxiety about, am I eating a low enough protein diet

02:24:25.720 or am I still in ketosis? I got to do my breath monitor every- Yeah, what should my fasting window

02:24:31.440 be? The questions are out there to what extent do any of these things have big benefits. I think you can

02:24:37.180 get most of the benefits without worrying about a lot of that. Yeah, I agree. Well, Matt, glad we

02:24:42.720 finally got to do one of these in person. Yeah, it's been a pleasure. Maybe the next one should be

02:24:45.240 in person as well. Absolutely. Thank you for listening to this week's episode of The Drive.

02:24:49.780 If you're interested in diving deeper into any topics we discuss, we've created a membership

02:24:54.040 program that allows us to bring you more in-depth exclusive content without relying on paid ads.

02:24:59.580 It's our goal to ensure members get back much more than the price of the subscription.

02:25:04.080 Now to that end, membership benefits include a bunch of things.

02:25:07.180 One, totally kick-ass comprehensive podcast show notes that detail every topic, paper,

02:25:12.540 person, thing we discuss on each episode. The word on the street is nobody's show notes rival these.

02:25:18.480 Monthly AMA episodes or Ask Me Anything episodes, hearing these episodes completely.

02:25:23.600 Access to our private podcast feed that allows you to hear everything without having to listen to

02:25:28.880 spiels like this. The Qualies, which are a super short podcast that we release every Tuesday

02:25:34.200 through Friday, highlighting the best questions, topics, and tactics discussed on previous episodes

02:25:38.940 of The Drive. This is a great way to catch up on previous episodes without having to go back and

02:25:43.980 necessarily listen to everyone. Steep discounts on products that I believe in, but for which I'm not

02:25:49.740 getting paid to endorse. And a whole bunch of other benefits that we continue to trickle in

02:25:54.240 as time goes on. If you want to learn more and access these member-only benefits, you can head

02:25:58.540 over to peteratiamd.com forward slash subscribe. You can find me on Twitter, Instagram, and Facebook

02:26:05.600 all with the ID peteratiamd. You can also leave us a review on Apple Podcasts or whatever podcast

02:26:12.400 player you listen on. This podcast is for general informational purposes only and does not constitute

02:26:17.960 the practice of medicine, nursing, or other professional healthcare services, including the giving of

02:26:24.240 advice. No doctor-patient relationship is formed. The use of this information and the materials linked

02:26:29.820 to this podcast is at the user's own risk. The content on this podcast is not intended to be a

02:26:35.860 substitute for professional medical advice, diagnosis, or treatment. Users should not disregard or delay in

02:26:43.180 obtaining medical advice from any medical condition they have, and they should seek the assistance of

02:26:48.560 their healthcare professionals for any such conditions. Finally, I take conflicts of interest

02:26:54.000 very seriously. For all of my disclosures and the companies I invest in or advise, please visit

02:27:00.000 peteratiamd.com forward slash about where I keep an up-to-date and active list of such companies.

02:27:24.000 Thank you.

02:27:31.000 Thank you.

02:27:35.000 Thank you.

The Peter Attia Drive - September 12, 2022

#222 ‒ How nutrition impacts longevity ｜ Matt Kaeberlein, Ph.D

Episode Stats

Summary

Transcript