The Peter Attia Drive - #175 - Matt Kaeberlein, Ph.D.： The biology of aging, rapamycin, and other interventions that target the aging process

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

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

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

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

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

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

00:00:36.840 at the end of this episode, I'll explain what those benefits are. Or if you want to learn more now,

00:00:41.740 head over to peteratiyahmd.com forward slash subscribe. Now, without further delay, here's

00:00:48.080 today's episode. My returning guest this week is Matt Caberlin. Matt is recognized globally for his

00:00:56.660 research on the basic biology of aging. And Matt is clearly for me on the short list of people who

00:01:03.580 I always reach out to when I have questions about aging. I consider him an amazing mentor and an

00:01:09.820 amazing scientist. Matt was a previous guest way back in episode number 10, circa mid 2018,

00:01:16.040 when we dove deep into his background and his interest in aging and the origins of the dog

00:01:21.480 aging project, which uses rapamycin to study companion dogs. In this episode, we pick up that

00:01:28.860 baton and go even deeper, but we also take an even broader look at aging, arguably the broadest

00:01:34.540 look I've taken in any podcast, because we really start from the nine hallmarks of aging and talk about

00:01:41.220 each of them. And then we tie it into a framework for how do you think about aging? Sort of a top-down

00:01:47.940 view, a disease-centric view, or a bottom-up view through the lens of biology. And we talk about

00:01:54.280 some of the pros and cons of each of these. We then pivot our discussion to talking about

00:01:59.760 rapamycin. Matt and I both speak very openly about our personal use of rapamycin and all the trials,

00:02:06.480 both in animals and humans that have led us to our conclusions. We then return to the dog aging

00:02:12.240 project where Matt provides an update on some of the exciting work that's being done there and some of

00:02:16.780 the exciting work that's being done elsewhere in his institution with respect to rapamycin trials

00:02:21.940 in humans and what they can be targeting in the short-term to better understand the impact of

00:02:28.320 these drugs and drugs like rapamycin, rapalog, so to speak, how they can impact humans in a long-term

00:02:34.900 way, even though we can't really study humans in a long-term way. We also discuss a relatively new

00:02:40.500 molecule that many of you may not have heard of called TORIN-2, which is like rapamycin,

00:02:46.400 another inhibitor of mTOR, but one that does so via a slightly different mechanism. It looks quite

00:02:51.820 promising. And we discuss what that may or may not imply. And we finish up our discussion with another

00:02:57.560 deep dive into sirtuins, NAD, NR, and NMN, the precursors to NAD. So I think if you have any interest

00:03:07.120 in the topic of aging, you're going to find this episode probably riveting, if not at least half as

00:03:12.740 riveting as I did, which is still to say very riveting. So without further delay, please enjoy

00:03:17.360 my conversation with Matt Caberlin. Hey Matt, awesome to have you back on the show. It's been a

00:03:28.100 while because you were one of the very, very initial and original guests on this podcast. So I've been

00:03:33.940 looking forward to having you back from almost the moment we finished our first discussion.

00:03:37.920 Yeah. Thanks, Peter. It's great to be back. Unfortunately, we can't see each other in person,

00:03:42.180 but this is the next best thing. Yeah. Yeah, I know. We certainly will and get together in person

00:03:47.120 soon, I'm sure. There are so many things I wanted to talk about. And, you know, I think last time we

00:03:51.900 spoke, we focused mostly on our mutual favorite drug, rapamycin, and our favorite pathway,

00:03:58.400 the mTOR pathway. And certainly for those listening, I don't want them to think we're not going to

00:04:03.500 touch on that today. We absolutely are. But I want to start even broader than that because

00:04:08.480 so often you and I are having these discussions about all things that pertain to aging. And I

00:04:14.560 find you to be one of the most thoughtful people across the topic. So I sort of want to start with

00:04:20.260 these broader questions about aging. A lot of people have different definitions of aging. And

00:04:25.260 truthfully, I'm not sure I even know how to define aging sometimes. It depends, I guess,

00:04:30.380 on the context in which I'm asked, right? I think if my five-year-old asks me about aging,

00:04:34.880 I would come up with one answer. And if I'm giving a keynote talk and somebody asks me about aging,

00:04:40.640 I'd have to give a different one. So what are some of the ways that you really describe aging?

00:04:46.240 That's a really good question. And I think I would probably say the same thing that you just said,

00:04:51.040 which is that I'm not sure I have a great answer and it changes depending on the situation.

00:04:56.900 I think in the past, I often gravitated towards sort of a molecular definition of aging. So

00:05:02.980 what are the types of damage that occur during aging? What are the consequences of that damage?

00:05:10.460 In some ways, a hallmarks of aging framework view, right? Where we know that things like

00:05:15.460 mitochondrial dysfunction, telomere shortening, cellular senescence, things like that happen during

00:05:21.560 aging at the cellular and molecular level and contribute to many of the functional declines

00:05:28.520 and diseases that go along with aging. And so, you know, given my training in biochemistry and molecular

00:05:34.200 biology, that sort of is the natural place where I go when I think about what is aging. I would say

00:05:41.320 over the last several years, I have developed, I think, a greater appreciation for a functional

00:05:48.320 definition of aging. And so I, you know, as I start to think more and more about translation of

00:05:54.720 interventions that seem to affect the biological aging process in laboratory animals outside of the

00:06:00.400 laboratory into the clinic, I spend a lot of time thinking about, well, what are some of the functional

00:06:05.440 changes that go along with aging? And we know that across every organ system in the body, we see functional

00:06:12.800 declines that go along with aging. And so I've started to think more about things like frailty

00:06:19.120 as an important component of defining aging from a biological perspective when we're talking about

00:06:26.640 having an impact on health and longevity. So I think it really depends on the context. I guess the other

00:06:32.320 thing I would say is both of those definitions and the way I almost exclusively think about aging, unless

00:06:37.600 somebody sort of forces me out of out of my box, is from a biological perspective. There are other aspects

00:06:44.080 of aging that intersect with biological aging, right, that are probably as important as what I think of as

00:06:50.960 fundamental biology of aging to quality of life. So social aspects of aging, for example, are extremely

00:06:57.920 important in people, especially, right, for quality of life as you get older. I don't tend to gravitate towards,

00:07:04.560 you know, that kind of a definition of aging, but I do recognize that it's important. So,

00:07:08.800 you know, I think I just, I'm naturally always thinking about the biology of aging.

00:07:14.080 And as I already said, I tend to focus on the molecular mechanisms that drive

00:07:19.600 the biological changes that go along with aging. So let's look at the sort of two extreme views within

00:07:25.520 the biologic and then try to figure out where disease fits in. So on the one hand, you referred to the

00:07:31.120 hallmarks of aging. And there's a very famous set of papers that have laid these out. And if the day

00:07:36.560 comes that I can actually recite all of them, I'll be very impressed, right? I can, I always have to

00:07:41.360 look it up. I think I can probably get seven. Yeah, yeah, yeah. We can do like, who can, who can name

00:07:45.600 the most hallmarks of aging? Right, right. So DNA damage is, is one of them. Cellular senescence is

00:07:52.080 another. Stem cell fatigue is another. Protein misfolding is another. That's four.

00:07:58.000 Telomere shortening is another. Mitochondrial dysfunction is another. My favorite, yep.

00:08:04.240 Did I mention nutrient sensing issues? Deregulated nutrient sensing, yeah.

00:08:09.040 Deregulated nutrient sensing. So that's my seven. What am I forgetting?

00:08:13.120 There's intracellular communication.

00:08:14.720 Bingo.

00:08:15.680 And what are we missing here? I'm going to be really ashamed if it's, if it's something that I study.

00:08:21.040 Yeah. So we got eight out of nine. That's not bad.

00:08:23.040 So you've got the, you've got those. And then at the other side, you talked about,

00:08:27.040 okay, let's talk phenotype. Let's talk about the outward expression within the organism. And

00:08:31.680 frailty is, I think, a fantastic example. Where does disease fit into this, right? Because with

00:08:37.600 aging comes more cancer. With aging comes more cardiovascular disease. With aging comes more

00:08:43.040 dementia. And you could argue that those are functional. Maybe less so cancer, but certainly

00:08:49.680 cardiovascular disease and dementia are very functional forms of decline. But of course,

00:08:54.640 at their root, they have a very cellular component and a very strong set of cellular contributions.

00:09:01.520 Do you think of disease as basically the bridge or the link between these fundamental cellular

00:09:06.480 declines and the ultimate phenotypic declines?

00:09:09.520 No, I would not say they're a bridge. So I think that I actually, I actually personally tend to think

00:09:16.800 almost the other way around, where I think that the functional declines that often precede overt disease

00:09:23.840 or clinical diagnosis of disease are probably as important or more important from a quality of life

00:09:30.800 perspective as we get older. And, you know, this may simply reflect the fact that I'm getting older.

00:09:35.280 And, you know, I've noticed some of these functional declines in myself, right? And I think these

00:09:41.120 functional declines happen, as I already said, in every organ, every tissue. We don't always recognize

00:09:47.120 them as such, unless we sit down and think about and try to list out all of the ways that we have

00:09:52.480 changed as we get older. And those often happen far before you get diagnosed with any age-related

00:09:58.000 disease, right? So I'm pretty fortunate. I turned 50 in February. I don't have any age-related

00:10:03.040 disease, at least that I've been diagnosed with. Yet I have a multitude of functional declines,

00:10:07.840 which fortunately don't impact me severely, but that I recognize I'm not, I'm functionally impaired

00:10:13.440 to where I was 25 years ago. I think we, you know, it's just a fact. Any 50-year-old is. And so I kind

00:10:19.520 of think of those as the first things that you can observe that happen during aging, often way before

00:10:25.920 you get an age-related disease. The other thing I would say about disease, so there's actually two things,

00:10:30.160 two points I want to make. One is, I think it depends on the disease, but we really don't have

00:10:35.200 a good understanding of when the pathology of the disease is no longer normative aging. And what I

00:10:43.280 mean by that is, you know, we've got some understanding of the molecular cellular mechanisms that drive

00:10:48.080 biological aging, that contribute in some way to our risk of developing Alzheimer's disease,

00:10:53.680 cardiovascular disease, all of the age-related diseases that are major causes of death and

00:10:59.200 disability. But in most of those cases, there comes a point where the pathology of the disease

00:11:05.440 is not necessarily at a molecular mechanistic level, an extension of aging biology. It becomes

00:11:12.480 something different. And I think that's really important to recognize because one of the implications

00:11:17.040 of that is that an intervention that affects biological aging, let's just say rapamycin,

00:11:22.080 we can discuss whether rapamycin really affects biological aging in people. I think that's still

00:11:26.160 a little bit of an open question. Let's just say for the sake of argument that it does,

00:11:30.000 it's not clear that that same intervention is going to be effective once a pathology progresses

00:11:36.320 to the point that it's not the same mechanism anymore. So I think that that's a really important

00:11:40.880 point that sometimes gets lost in this discussion of aging and disease.

00:11:44.320 So let's actually double click on that using a disease. So pick one of the big three, atherosclerosis,

00:11:52.080 cancer, dementia. I think cancer is a really good example, right? So we know mTOR, which,

00:11:57.600 and I'll go back to rapamycin in part because, you know, it's again, something I think a lot about,

00:12:01.920 but it actually, it's a really good, I think, example in this specific case, because we know that mTOR,

00:12:08.320 which is the target of rapamycin, right? The protein that rapamycin inhibits plays this fundamental

00:12:14.400 role in regulating cell division and cell cycle, right? So if you inhibit in a non-cancerous cell,

00:12:22.160 if you inhibit mTOR enough, you will stop the cell cycle. The cell will stop dividing, right? But there

00:12:28.160 are mutations that can happen that lead to cancer that cause the cell to no longer pay attention to

00:12:34.000 the mTOR break, right? And so once that's happened, if that's the type of cancer you have

00:12:38.960 that no longer responds to mTOR inhibition, rapamycin won't do anything to cell cycle in that,

00:12:45.120 in that case. So that's a really, I think, specific example that you can point to. There are,

00:12:49.680 you know, sort of an infinite number of other examples that we could use, but that's a really

00:12:53.600 nice one because there, rapamycin will be quite effective at preventing cancer before that mutation

00:12:59.600 happens. But after that mutation happens and the cell's not responding to rapamycin,

00:13:03.920 anymore because it doesn't sense the mTOR break, it's completely ineffective, right?

00:13:10.240 So that, I think, is a case where the mechanisms have changed. The mechanisms that are important

00:13:16.400 for preventing cancer before that mutation occurred are different from the mechanisms that might

00:13:22.560 deal with that cancer after that mutation has occurred.

00:13:26.240 Yeah, it's funny. This is a little off topic, but I've often contemplated this question in the context

00:13:31.360 of nutrition. Because in as much as there's an optimal nutrition to prevent a condition,

00:13:37.920 it might not be the same as the optimal nutritional strategy to treat the disease once it's present.

00:13:44.320 An example of that in an extreme sense might be a ketogenic diet. I happen to believe a ketogenic diet

00:13:49.440 is probably the best treatment for someone with type 2 diabetes. Because of course, type 2 diabetes,

00:13:54.960 by its very definition is a carbohydrate intolerance disorder. So once a person has it, you pull out

00:14:00.240 the carbohydrates completely and you let them heal, right? You basically let them recover and regain

00:14:06.000 their ability. And again, we've seen that people who have been on a ketogenic diet for a long enough

00:14:12.160 period of time can resume some amount of carbohydrate consumption provided their other factors are changing,

00:14:17.840 such as exercise. Does that mean one needs to be on a ketogenic diet to prevent diabetes? No,

00:14:22.960 I don't think so. So it's a little bit of the same idea, though it's still something that's unclear.

00:14:29.040 One thing I want to go back to on the disease front is, and I believe it was Cynthia Kenyon who

00:14:34.000 spoke about this once. I think I read it in a paper. Something to the effect of using a disease-based

00:14:40.400 definition for aging is, she didn't use the word tautology, but she effectively said it is a bit of a

00:14:46.360 tautology because at what point is a disease a disease? It's only a disease when some people have it and

00:14:52.480 some people don't. If everybody has cancer by a certain age, then it's normative aging to your

00:14:58.880 point. It's no longer a disease. And then we get into, well, what does it mean that some, like,

00:15:03.840 why do, you know, 0.0004% of the population live to be 100? They've managed to not succumb to a disease

00:15:10.640 by the age of 100. And what does that tell us about their normative aging versus everybody else?

00:15:16.960 All of this is to say, I literally still don't think I understand what aging is,

00:15:21.840 which is unfortunate given my line of work. We have to just accept that it's extremely

00:15:26.820 complicated, right? And so you're never, I shouldn't say never. I don't think I will ever

00:15:31.420 understand aging fully. And I don't think the field will, at least in any timeframe that I can

00:15:37.880 expect to experience, right? But I also believe that we don't have to understand it fully to be able

00:15:46.080 to have an impact on the biology of aging through interventions. And that's kind of where I'm at.

00:15:50.920 I feel like I've got a conceptual flavor for what aging is. And I have some information about what

00:15:59.920 the molecular mechanisms are. And it's enough information that I can come up with rational

00:16:05.040 approaches to target those mechanisms with the prediction that those approaches should have an

00:16:10.280 impact on health and longevity as animals and people get older. And then we have to test those

00:16:16.200 predictions. That's kind of the way I think about it. I do want to come back to one point, though,

00:16:20.760 which I also think is often underappreciated in this relationship between disease and the biology of

00:16:27.420 aging. Sometimes people get into this debate about whether or not the biology of aging causes

00:16:35.120 diseases of aging, right? So does the biology of aging cause Alzheimer's disease, cancer,

00:16:41.460 cardiovascular disease, right? People get in debates about that. And I personally think the

00:16:46.860 data are pretty good that what I think of as the biology of aging, the molecular mechanisms,

00:16:51.960 the hallmarks, whatever you want to call them, contribute in a causal way to your risk of developing

00:16:56.640 diseases. But I also think it doesn't matter. And this, I think, is really important.

00:17:02.360 From the perspective of what is the best strategy to keep people healthier longer,

00:17:07.660 it just doesn't matter whether aging causes disease or it creates a permissive physiological

00:17:15.980 state for disease. You can't argue that biological age is the single greatest risk factor for every

00:17:24.920 major cause of death and disability in developed countries. That is just a fact. And whether or not

00:17:30.760 biological age causes those diseases or creates a physiological state that allows those diseases to

00:17:37.920 manifest themselves, it doesn't matter from the perspective of what is the most effective way

00:17:42.880 to prevent those diseases. And I think that's where this debate is counterproductive, right?

00:17:48.280 Should we call aging a disease? Does aging cause disease? I think that those are not the right

00:17:54.640 questions to be asking in my view.

00:17:56.600 Yeah. Let me see if you would agree with my assessment. I think you would, but I'll tell

00:18:02.180 you how I think about this problem clinically. So let's use atherosclerosis as an example. And I

00:18:07.940 want to highlight what you just said in case the person watching this or listening to this missed it

00:18:13.260 in any way, shape, or form. When I used to give talks, I'd always lead with a question like this.

00:18:18.760 What is the greatest risk factor for atherosclerosis? Hand will go up. Yes. Smoking. Nope. Hand will go up.

00:18:26.600 High blood pressure. Nope. Hand will go up. If it's in an especially erudite audience. ApoB or LDL-C.

00:18:33.660 Nope. And they'll just keep rattling. Inflammation. Nope. Nope. Nope. Nope. Nope. Nope. Nope. The number one

00:18:40.300 thing is age. Hands down. You take a 70-year-old person who has everything perfect about them and you

00:18:49.500 compare them to a 20-year-old train wreck who has not a single thing that is in their favor. There is

00:18:55.720 no comparison about their 10-year mortality prediction. The 70-year-old is in hands down a

00:19:01.820 worse shape. So you can't undo that. Now, here's how I think about this problem clinically.

00:19:08.540 Atherosclerosis is a great example because it's the disease that we understand the most of the big

00:19:12.440 three, right? Which is not to say we understand it completely, but we have a far better understanding of

00:19:16.880 what its drivers are and how to prevent it than we do cancer and Alzheimer's disease.

00:19:22.220 Lowering ApoB pharmacologically, nutritionally, et cetera, is arguably the most important strategy

00:19:29.340 you have to reduce it along with probably improving metabolic health. So those two things, right? So

00:19:34.240 regulating glucose, insulin, lowering ApoB, all of these things can be done through lifestyle,

00:19:39.360 through drugs, et cetera, can dramatically reduce a person's risk of atherosclerosis.

00:19:43.980 None of those things are necessarily directly targeting the nine hallmarks of aging. I think

00:19:49.400 indirectly they certainly do. But when you give somebody a PCSK9 inhibitor, which specifically

00:19:56.200 targets a protein that allows the body to clear more ApoB particles out of circulation, it is by no

00:20:03.260 means targeting one of those nine pillars, but it's having a measurable impact on reducing their risk of

00:20:09.840 disease. And in the end, that's the part that I think is hard for some people to understand within

00:20:15.320 the aging community is that you can still target metrics of a disease specifically without going

00:20:22.140 after a hallmark. I don't disagree with you that I think that it's a concept that sometimes is

00:20:27.520 underappreciated, right? Especially within the aging community. I also think it makes sense,

00:20:32.640 right? So even if the hallmarks, we focus so much on the hallmarks because it's easy,

00:20:37.100 right? I think there are some reasons why the hallmarks are incomplete, but let's just keep using

00:20:41.020 that term sort of as a surrogate for the molecular mechanisms of aging, right? If we accept that the

00:20:47.300 hallmarks are at some level preceding the damage, whatever that is, that's causing the disease,

00:20:54.740 it makes sense that you could intervene sort of, if we think about it in an upstream, downstream

00:20:59.160 perspective, right? The hallmarks being upstream, the disease being ultimately downstream. It makes sense

00:21:04.040 that you could intervene at the level of the hallmarks. It also makes sense that you could

00:21:06.980 intervene in whatever the bridge is that's connecting the hallmarks to the disease. You

00:21:12.940 wouldn't necessarily impact the biology of aging at all. And this gets back to what I was talking about

00:21:17.740 before, which for many diseases of aging, there comes a point where the pathology of the disease

00:21:23.800 is not normative aging anymore. And so you can quite successfully treat or cure a disease of aging in

00:21:31.800 an individual without impacting the biology of aging. In fact, I would argue that's almost exclusively

00:21:36.800 what is done clinically, right? Is to try to either alleviate the symptoms of the disease or cure the

00:21:44.640 disease. There's starting to be more on the preventative side, but I still think that lags far behind,

00:21:49.920 you know, waiting until people are sick and then trying to do something about their disease. But you don't

00:21:54.040 have to impact the biology of aging to be successful at any of those things. I would just argue that

00:22:00.880 impacting the biology of aging is going to be a much more effective and efficient approach

00:22:05.900 from sort of the overall health perspective. I think the other point that's obvious, but is

00:22:10.700 important to make, you can be quite effective at treating an age-related disease without actually

00:22:16.060 targeting the hallmarks of the biology of aging. But I think an important point to make is that if you

00:22:21.680 don't actually confront the biology of aging, that you're really only impacting that one disease.

00:22:27.680 Say it's cancer, you can cure somebody's cancer, you can take the tumor out, right? And they can go on

00:22:32.400 and live a normal life. But you haven't done anything to the biology of aging. If you don't confront the

00:22:38.640 biology of aging, right, you still have all of these other functional declines and diseases of aging that

00:22:44.500 are increasing essentially exponentially as you're getting older. And so the effect on health and longevity is

00:22:51.300 quite small. In fact, I think J.L. Shansky was the person who originally kind of did the math here,

00:22:56.320 right? And it's really very striking. So if you consider a typical 50-year-old woman in the United

00:23:01.800 States and you say, what would happen if we had a cure to all cancers, right? Every type of cancer,

00:23:09.640 we had a cure today and we implemented that. Life expectancy for a typical 50-year-old woman in the

00:23:16.480 United States would go up about three years. That's it. So we've won the war on cancer. We get

00:23:22.720 a plus three on life expectancy. It's about the same for curing heart disease. If you cure both of

00:23:29.220 those diseases, it's roughly additive. It's, I think, about seven years. So the impact on life

00:23:34.540 expectancy is actually, from curing disease, is actually quite small. If you compare that to the

00:23:40.660 impact of targeting biological aging, which, again, I think we have to be honest, right? This is

00:23:45.960 hypothetical in humans at this point. So all we can really do is extrapolate from what is done in

00:23:51.220 laboratory animals. But it's pretty routine now. There are, I don't know, maybe a dozen, 15 different

00:23:58.260 interventions that can increase lifespan, slow aging in mice by between 15 and 30%. If we just extrapolate

00:24:06.140 that to the human condition, the impact on life expectancy is about 20 years with the added value

00:24:12.760 that because you've sort of slowed all of the functional declines of aging simultaneously,

00:24:17.740 those extra couple of decades are spent in relatively good health, right? So I think it's just

00:24:23.300 important to appreciate the potential difference in health and life expectancy from targeting aging as

00:24:32.260 opposed to what we do right now, which is trying to cure individual diseases. And I've sort of adopted

00:24:38.060 the term, you know, I think of that as 20th century medicine, that the disease first approach, or I would

00:24:43.540 even say 19th century medicine. We've been doing this for a long time. And I contrast that to what I think

00:24:48.700 of as 21st century medicine, which is approaching health and longevity from the perspective of the biology

00:24:55.160 of aging. And we're not there yet, right? But I do believe that this is happening. There is a transition

00:25:02.360 occurring, a slow appreciation among the broader clinical community of the potential of this kind

00:25:08.580 of approach. And I do believe that we will get there in this century. So it is my hope and expectation

00:25:14.200 that 21st century medicine will really become targeting the biology of aging to enhance health and

00:25:20.900 longevity, hopefully by much, much greater amount than, than we're currently able to do.

00:25:27.440 Yeah. So I agree with a lot of that. The one thing I would throw a little bit of a question at is,

00:25:32.940 and I'd love to have Jay on the podcast because I don't know that I fully agree with his analysis.

00:25:37.700 I agree with the spirit of his analysis. So I completely agree that disease-based whack-a-mole

00:25:43.980 is not going to be nearly as effective as targeting the basic biology of aging. So completely agree with,

00:25:50.580 with your points. I don't know that I agree with the magnitude, right? The de minimis magnitude of it,

00:25:56.880 because I really think that Jay's analysis is focused on an independent look at each disease.

00:26:03.780 Whereas in reality, if you eliminated cardiovascular disease, by definition, you have reduced inflammation

00:26:13.140 significantly. You've reduced the burden of microvascular disease significantly.

00:26:17.360 Those are going to play into other diseases. So his analysis is very actuarial, but it's not

00:26:25.240 actually very biological. That's my opinion. Yeah. I think that's fair. I think you could make

00:26:29.760 the case, at least for cancer, that it may be closer to just what the straight math would suggest.

00:26:36.760 Now, with all of that said, I couldn't agree more with the macro point here, which is delaying chronic

00:26:44.640 disease is probably not going to be as effective as targeting something at the foundation. Now,

00:26:51.480 that said, let's take a look at NIR's data. So when you look at Nir Barzilai's work with centenarians,

00:26:57.880 the overwhelming statement here is they get chronic diseases later. They don't live longer

00:27:06.920 with them. And I remember really being struck by that. That would not have been my null hypothesis

00:27:13.920 going into a study of Nir's work and Tom Pearl's as well. Between Nir and Tom, you really have

00:27:19.300 the greatest assessment of the centenarians and their siblings. And you realize that they don't seem

00:27:26.180 to have magical protection from a disease once they get it. When they get cancer, they're just

00:27:30.640 as hosed as the rest of us. When they get heart disease, they're just as bad off as the rest of

00:27:35.100 us. But they get a phase shift. They get a 20-year, 20, sometimes 30-year phase shift in when they're

00:27:42.200 going to get the disease. So how does that factor into my thinking clinically? My factoring into that

00:27:48.420 clinically is the sooner you begin prevention, the more you can mimic the centenarian. What we don't

00:27:55.420 fully understand is molecularly, why is that the case? We've identified some of their genes.

00:28:02.100 We know that they're going to be more likely to have ApoE2 versus ApoE3 or ApoE4. They're going to

00:28:09.240 be less likely to have ApoC3, high regulation versus low. I mean, you have a lot of genes that produce

00:28:17.120 phenotypes that are favorable. But when you bring it back to our nine hallmarks, by the way, the ninth

00:28:22.920 hallmark is epigenetic. Of course. Oh my God.

00:28:26.040 It's the epigenetic modulation. Now, the other point, I guess, that we're both well aware of,

00:28:33.400 but maybe for the listener, is how difficult it is to study aging in humans targeting the basic biology

00:28:41.060 because you don't get to do what our clinical trials apparatus is set up to do, which is pick a

00:28:50.740 disease, a very clear endpoint, and target it. If your endpoint is nutrient sensing, that's a tough

00:28:57.700 clinical trial. Yeah, I agree. And I think we should definitely come back to that point because

00:29:02.900 it's clearly a major focus of the field right now is how do we start to test some of these interventions

00:29:09.560 that we know work in laboratory animals in the clinic? That's a challenge. I want to come back to

00:29:14.380 your comment about centenarians, though, because I think this, again, it's a conceptual area where

00:29:19.400 there's a lot of, I think, misunderstanding both among people in the field and also lay people who

00:29:26.580 pay attention to the field. So you're right, I think, that the bulk of the evidence supports the

00:29:32.340 idea that centenarians do not live longer with multiple age-related diseases, but they have

00:29:39.900 genetic risk factors or genetic variants that put them at lower risk, at least for the major killers.

00:29:46.600 So there is a genetic component to being a centenarian. It's not huge, but it's significant,

00:29:51.320 probably somewhere, I think, around 25, 30 percent. And they tend to not have the high-risk

00:29:56.880 genetic variants for Alzheimer's disease, certain types of cancer, heart disease, things like that.

00:30:03.400 And then there's a luck or environment or something else we don't understand, right,

00:30:07.240 that comes into play. But people often look at that observation and make the assumption that we

00:30:14.700 don't see variants in things like mTOR or sirtuins, not that there aren't variants that haven't been

00:30:20.920 talked about, but that have strong effects. So I think the question becomes, why don't we see

00:30:25.280 variants in SIRT6 or mTOR or pick your favorite longevity gene, FOXO, right, that cause people to live

00:30:33.560 to be 180 years old? Because we can do that magnitude of lifespan extension in a mouse or a

00:30:39.220 C. elegans. And so they go from that. So that's an observation. We haven't found those variants yet.

00:30:44.980 Although FOXO is one of the genes.

00:30:47.240 But not strong effect variants, right?

00:30:48.560 No, it's not. You're right. It's a very weak effect, but it's clearly a higher amount of...

00:30:52.020 Yeah, as is 36, right? And as is mTOR. There's evidence for all of those things.

00:30:55.940 You can find genetic evidence, but the effects are relatively small.

00:30:59.340 But then the interpretation that people make is the reason we don't see those variants is because

00:31:04.140 humans are fundamentally different from mice or whatever your favorite laboratory organism is,

00:31:09.940 and none of those things will work the same way in people. Now, I can't prove that they will work

00:31:13.940 the same way yet for aging in people, but that's not a logical sort of interpretation, right? I would

00:31:21.300 argue the reason you don't see strong effect variants in mTOR, for example, in people is because

00:31:28.460 we know that strong effect variants in mTOR are incompatible with life and development,

00:31:33.620 even in mice, right? You make an mTOR knockout mouse, it's dead.

00:31:39.060 So I think the reason why you don't see these strong effect variants in people is because there

00:31:45.360 is such a strong selective pressure. Either they don't make it through gestation, they don't make

00:31:50.460 it through development, or they're sterile. All of those are incompatible with evolutionary success.

00:31:55.620 And that's my hypothesis for why we don't see these very strong effect genetic variants in people

00:32:01.780 that lead to 160, 170 year lifespans. Again, it's a hypothesis, right? It's hard to know what the

00:32:09.780 explanation is, but it certainly fits with what we see in laboratory animals. Any of these mutants in

00:32:16.420 mice that have 30, 40% lifespan extension, I mean, they are all significantly defective in,

00:32:24.620 from an evolutionary perspective, right? They would not be selectively advantageous mutations. And so

00:32:31.120 I think that's at least a consistent explanation for why we don't see these, what we would think of

00:32:35.880 as slower aging variants pop up in people. The other point that's important to make as well,

00:32:42.020 though, is that that doesn't mean that we can't intervene in these pathways to have an impact on

00:32:46.700 health and longevity. And for me, this has been probably the most exciting aspect and development

00:32:52.820 in the field over the last probably 15 years. You know, if you asked me 15 years ago, whether I

00:32:58.920 thought that it would be easy to slow aging in an old mouse, I would have said, no, I would have said,

00:33:04.920 you probably have to start at six months of age and treat them all the way through life to get the

00:33:10.140 benefits. And that was honestly, largely based on caloric restriction, right? Which that seems to be the

00:33:14.380 case for caloric restriction. You know, what has emerged now is there are five or six or seven,

00:33:19.320 or maybe even more different interventions that can be initiated in middle age or even in late age

00:33:25.520 in mice. And you actually go, it's not only that you slow down the declines, you actually reverse

00:33:31.760 the declines, right? And again, I come back to rapamycin because in pretty much every tissue where

00:33:35.940 this has been looked at, you take an old mouse, you give it rapamycin, functionally it's younger

00:33:40.280 in that tissue or that organ and lifespan is extended. And by the way, to your point, Matt, I was having

00:33:45.540 dinner with a friend last night who asked me, why don't we start giving rapamycin to children? And I said,

00:33:51.640 look, I think, I think rapamycin is, is the most important, you know, gyroprotective agent out there 1.00

00:33:57.140 today, but you actually, you wouldn't want to give it to a developing child, right? So he said, if you had

00:34:03.520 to guess when would be the right age to start, I said, I have no honest clue, but it wouldn't be before

00:34:07.900 25, right? It just wouldn't be before about the age of 25, which speaks to your point, any genetic

00:34:14.700 manipulation, or in this case, naturally acquired genetic variation that mimicked rapamycin would

00:34:20.920 probably not be selective because you wouldn't be able to turn it on and off the way you would need

00:34:25.760 to with the drug. That's right. Yeah, that's right. And I mean, I think that's a really interesting

00:34:29.880 question about rapamycin and other interventions as more and more interventions start to, you know, get to

00:34:37.180 the same sort of level of confidence that we have about rapamycin. When is the optimal sort of

00:34:42.960 treatment period, right? To get the biggest benefits. And I think it's not something that

00:34:48.020 often gets talked about in that conversation, but it's really a, it's a balancing act, right?

00:34:52.580 There, any intervention is going to have some risk associated with it, right? The risk can be low,

00:34:57.640 the risk can be high, but there's always some risk of side effects. And so the question of optimal

00:35:02.660 treatment is a balance between the beneficial effects that you get and the risk of detrimental

00:35:10.840 effects, side effects. And so I think we still don't, even with rapamycin, we still don't have a

00:35:17.100 great feel for what that risk looks like, right? And that's in part because we haven't had long-term

00:35:25.480 controlled clinical trials at multiple doses, multiple strategies of intermittent versus continuous

00:35:31.760 treatment, things like that. So we just don't know, right? And it's interesting in the case of

00:35:36.880 rapamycin because my impression is that, you know, there are a group of people, I think you and I both

00:35:42.180 know some of them, we might even be among them, who, you know, have self-experimented, right? And my

00:35:48.740 impression is that there has been a movement towards the idea that, you know, once a week dosing with

00:35:53.700 rapamycin is probably better than daily dosing with rapamycin for aging effects in people. And that,

00:36:00.040 you know, you might want to do it for three months and then you stop for six months and you do it

00:36:04.420 again for three months. I think that's where people are starting to coalesce around this idea.

00:36:10.480 There's not a lot of data to support that, right? It's a guess. So it's just interesting to sort of

00:36:15.960 see how that is evolving in the absence of long-term, large-scale controlled clinical trials.

00:36:23.080 And this is getting to the point that you raised earlier, which is it's really hard and expensive

00:36:27.900 to do long-term controlled clinical trials for aging in people. We'd really, I think still as a

00:36:36.680 field, don't have a great strategy for how to address that challenge of actually trying to answer

00:36:41.900 some of these questions in the way that, you know, at least traditionally we would want to see them

00:36:47.700 answered in a clinical setting.

00:36:48.880 I mean, I think there's an even bigger problem, Matt, which is I don't think there's a regulatory

00:36:53.580 appetite to even do what would be necessary because we have the regulatory appetite to say

00:36:59.100 you take an individual who is either at very high risk for a disease or already has said disease,

00:37:04.660 and we will go ahead and accept the risk of a clinical trial to assess it. But by definition,

00:37:10.680 if you truly want to understand how gyroprotective metformin or rapamycin or

00:37:16.800 canagaflazin or pick your favorite molecule is, you really need to be testing it in people before

00:37:22.820 they have a disease. So there's really two fundamental problems. This is one, we do not

00:37:27.580 have a regulatory environment that accepts that risk. So I would love to see the IRB submission that

00:37:33.460 says we're going to test rapamycin in healthy 50-year-olds whose 10-year risk of death is less than 1%.

00:37:39.020 And then secondly, because we can't logistically follow those people for the next 50 years,

00:37:45.420 which is what would be necessary to understand its true magnitude of gyroprotection, we would need

00:37:52.200 really good biomarkers of aging of which I will posit, and you may disagree, we have somewhere between

00:37:59.000 zero and epsilon of those. I mean, we don't have diddly squat in the way of meaningful biomarkers of

00:38:05.120 aging. By the way, I want to go back to your point on rapa. So as you do know, and I'm very happy to

00:38:10.200 talk about, I mean, I've been taking rapamycin for two and a half years now, and I am one of those

00:38:14.840 people who has coalesced around once weekly dosing. I have fiddled a little bit with my dose. I've

00:38:19.320 varied it from as little as five to as many as eight milligrams once a week. I've also done the cycling on

00:38:25.420 and the cycling off, but I am completely flying blind. Most of my dosing and frequency data comes from

00:38:33.320 Joan Manick and Lloyd Klickstein's 2014 paper with avarolimus, which of course is not the same as

00:38:39.880 rapamycin, but it's a pretty reasonable hand-drawn facsimile of it. And it was from that study that I

00:38:46.200 concluded that 20 milligrams versus five milligrams versus one milligram, your sweet spot was at about

00:38:52.780 five from a side effect standpoint and an efficacy standpoint, and the daily dose of one milligram

00:38:58.700 didn't seem to produce as good an effect as the five milligrams once a week. But how does that

00:39:05.540 stand in comparison to what we learned from all of the ITP studies where it's continuous

00:39:11.280 administration of rapamycin? And to your point, because we'll come back to these and talk about

00:39:15.640 them, remarkable results regardless of when it's initiated. Right. I don't know what to make of that.

00:39:22.340 I kind of agree with everything you said about why, you know, if you're going to pick

00:39:27.000 one regimen for rapamycin, that makes sense. And, and, and like you said, it's really based on

00:39:32.520 two clinical trials. I mean, really. And so it's a limited data set, but at least for immune function

00:39:37.700 in people, it seemed to work. We could spend hours talking about this because that's a whole nother

00:39:42.740 interesting and informative sort of experience was the failure of the phase three clinical trial

00:39:49.720 at Restore Bio and what happened there. And I don't think we actually, it's at least not in the,

00:39:54.180 it's not published yet. Well, let's, let's talk about this because it is really interesting.

00:39:57.900 So let's back up for a moment and give people the context. So when Joan was at Novartis and both Joan

00:40:03.640 and Lloyd were at Novartis in 2014, the Everolimus study was done. Do you want to just tell people

00:40:10.000 really briefly what that study looked at? So Everolimus is a derivative of rapamycin and

00:40:15.220 biochemically it has exactly the same mechanism. So it's, I, at least I conceptually think about them as

00:40:21.400 essentially the same molecule. So, so anything that we see with Everolimus we would expect to see

00:40:26.180 with rapamycin. And so that was a phase two clinical trial in healthy older adults to look

00:40:33.420 at whether or not a six week treatment with Everolimus could boost vaccine response. So the

00:40:40.580 ability of older adults to respond to an influenza vaccine. And that actually was based on one of my

00:40:46.820 favorite papers from Pan Zheng's lab at the university of Michigan, who showed that in mice,

00:40:51.660 I think it was four weeks of rapamycin. It was either four weeks or six weeks of rapamycin and

00:40:55.500 old mice boosted the ability of a flu vaccine to protect those mice against the lethal dose of

00:41:01.720 influenza. That if anybody hasn't, hasn't looked at that paper, I would just encourage you to go

00:41:06.880 check it out. I think it was a 2009 science signaling paper from Pan Zheng's lab. The title has

00:41:13.000 something about hematopoietic stem cells and rapamycin. We'll link to it in the show notes for

00:41:17.000 sure. Okay. Yeah. It's an amazing study, but my favorite experiment there was this experiment where

00:41:22.540 they took young and old mice and they gave them a influenza vaccine and then they waited two weeks

00:41:29.020 and then they gave them a leaf, what would be a lethal dose of influenza if they hadn't got the vaccine.

00:41:34.180 And in the old group, half the mice either got rapamycin or they got a control. So the first thing that

00:41:39.780 struck me about that experiment was if you look at just the difference between young and old mice

00:41:44.320 who got the vaccine, all of the young mice were protected. 100% vaccine response. I think it was

00:41:50.720 about 65%. So two thirds of the old mice did not have a sufficient response to the vaccine that they

00:41:58.320 were protected against influenza. I don't think the numbers are exactly the same in people, but there is

00:42:03.840 definitely that same trend where older adults tend to not respond to vaccines as robustly as younger

00:42:11.940 adults. Unless those old mice got rapamycin. If they got six weeks of rapamycin, they were 100%

00:42:19.880 effective at responding to the vaccine. So at least for that measure of immune function,

00:42:25.200 rapamycin fully restored the immune system back to that of a youthful immune system. So that was the

00:42:30.060 mouse data that sort of served as support for testing this in humans, in this clinical trial

00:42:36.160 that we're about to get to. And by the way, Matt, do you recall if the immune response in question,

00:42:41.960 i.e. the immune response that was better in the RAPA group versus the non-RAPA group of older mice,

00:42:48.300 was this a B cell response, a memory B cell response, or was it a T cell response?

00:42:53.500 I don't recall the data. I know they have a lot of data in there looking at the antibody titers and

00:42:58.620 different immune cell types and how they responded. As I said, the title of that paper and the model

00:43:04.980 that they proposed was that rapamycin was acting at the level of hematopoietic stem cells to rejuvenate

00:43:10.660 hematopoietic stem cell function in some way. I don't recall the details of the immunology. And

00:43:15.720 I've tried multiple times throughout my career to become fluent in immunologist speak and have failed

00:43:22.900 miserably every time. So that's a weakness of mine that will likely never be addressed adequately.

00:43:30.140 So I don't know the answer.

00:43:30.780 It's just got such amazing implications for what we're seeing with COVID vaccination,

00:43:35.340 because we are still in such a nascent stage of truly understanding. I mean,

00:43:39.700 we're seeing people who naturally acquired COVID, who within six months have no more circulating IgG.

00:43:47.760 And we have no clue if they still have immune function. For example, do they still have memory

00:43:54.620 B cells in their bone marrow that when challenged will make antibodies? Do they still have memory T

00:44:01.020 cells that will show up and immediately respond? And so we're actually involved in a study that's

00:44:06.720 looking at that. But my intuition is you can probably still have an immune response or an immune

00:44:13.500 response that's ready to go absent circulating IgG. It'd be interesting to see where RAPA plays.

00:44:19.420 Yeah. The one thing I will say about rapamycin and everolimus is that it seems like it's not

00:44:25.180 one simple mechanism at play. So Joan, for example, has published data that one of the effects of

00:44:33.960 everolimus and another drug called RTB-101, which is also an mTOR inhibitor, is to boost antiviral gene

00:44:40.140 expression, right? And we know also that rapamycin, at least in mice and almost certainly in people,

00:44:47.320 tamps down on the sort of chronic sterile inflammation that goes along with aging through

00:44:52.820 mechanisms that still are being worked out. So I think it's probably multiple things that are going

00:44:57.720 on that can impact immune function in different ways when you treat with rapamycin, especially in

00:45:04.140 the context of an old animal or an old person. So it's probably not going to be one mechanism is my

00:45:10.740 guess. So based on this study, so Joan and colleagues take a group of, I believe, 65-year-olds.

00:45:17.740 I think they're, you know, 320 of them. So divide it into four groups of 80. You've got a placebo group,

00:45:24.700 one milligram a week group, 20 milligrams once a week and five milligrams once a week. Yeah?

00:45:30.340 That's right. Yeah. And so they did a six, I think it was a six-week treatment period.

00:45:35.400 And then they also waited, I think, two weeks. And then they gave them a flu vaccine. And what they

00:45:40.760 showed, I mean, obviously it's people, so you can't do the mouse experiment and then give those

00:45:44.760 people a lethal dose of influenza. So they were sort of, you know, stuck with looking at things like

00:45:49.500 antibody titers and to try to get an assessment of, did the people who got everolimus respond better

00:45:57.220 to the vaccine than the people who got the vehicle control. And it looked like they did. And it was,

00:46:01.860 you know, you can argue about the strength of the data. I think it was pretty clear that they had

00:46:06.580 at least at the five-mig once a week, and I think the one-mig daily, I can't remember,

00:46:12.760 maybe it was the 20-mig once a week, but the five-mig once a week looked pretty convincing 0.78

00:46:16.260 that they had a better response to the vaccine. I think the other thing that's really important about

00:46:21.100 that paper is that the incidence of side effects was really hardly different at all between the

00:46:27.860 different everolimus groups and the placebo, and particularly the five-mig once a week.

00:46:33.360 There really was no significant increase in adverse events in that group, which gets back to how we kind

00:46:38.880 of started on this tangent, which is that that's part of the reason why, you know, I think people are

00:46:44.300 moving towards the idea that you can still get efficacy from once weekly dosing and that the

00:46:51.160 side effects are reduced at once weekly dosing to essentially zero or close to zero. That was the

00:46:57.460 first phase two trial. And before we leave this trial, Matt, I think it is important for someone

00:47:02.260 listening to understand why people like you and I and Dave Sabatini and all the people who live in

00:47:08.680 mTOR land, I'm not putting myself in the same category as you guys, but you know what I mean,

00:47:12.500 I'm an mTOR fanboy. Why was that such an interesting and landmark paper? Well, I think it's because up

00:47:18.300 until that point, the only human application for rapamycin was immune suppression. Rapamycin is a

00:47:25.720 drug I have known about forever because as a surgical resident, I was giving rapamycin to kidney

00:47:31.560 transplant patients, heart transplant patients, and liver transplant patients, along with a cocktail of

00:47:36.560 cyclosporine, prednisone, mmf, all of these other really nasty drugs. And it was one part of that

00:47:44.440 in their ability to suppress the immune system. So I think prior to that paper, you have the ITP that

00:47:52.080 the first ITP that came out in 09 suggesting this is a remarkable tool for longevity, at least in the

00:47:59.560 ITP mice. But then on the other hand, you're saying, well, but it can't really work in humans because

00:48:04.400 it's going to be really horrible to the immune system. So you had this sort of dialectical

00:48:09.620 dilemma, which all of a sudden it became, well, maybe that's not the case. So how do you reconcile

00:48:16.360 the data that we saw in Joan's paper with the fact that rapamycin is, at least in theory,

00:48:24.900 an immune suppressor as it pertains to organ transplantation?

00:48:28.500 Yeah. And I think, you know, you're absolutely right. That was one important aspect of that paper.

00:48:33.480 And I unfortunately think that the old view, which in my room is the wrong view, that rapamycin

00:48:38.980 is an immunosuppressant still is prevalent in the clinical community. I think most clinicians

00:48:45.340 still think of rapamycin and rapalogs as immune suppressants, because that's often how they're used

00:48:50.940 clinically to prevent organ transplant rejection. So I think the simple answer is it's all about dose,

00:48:55.700 dummy. I mean, we know every drug has a dose response, right? And that you can get different

00:49:00.860 effects, different outcomes, different side effects, depending on the dose. And so, you know,

00:49:07.100 when you back off on the dose of rapamycin or everolimus in the context of an aged physiology,

00:49:14.240 so there's, you know, there's multiple things going on here in that study. They used lower doses,

00:49:19.360 they tested once weekly dosing, and these are old people who show a functional decline. You're not 0.99

00:49:24.700 going to see a functional improvement in vaccine response if you did this experiment in young people,

00:49:29.980 right? You're only going to see it in the context of older people where they already have a functional 1.00

00:49:36.420 deficit. And that's, again, it's an important conceptual point that I think often gets lost

00:49:41.060 in these discussions. The other thing to appreciate about how rapamycin has been used, you know,

00:49:46.060 traditionally and how it was first approved is these are people who had an organ transplant,

00:49:51.220 taking high doses of the drug. And also, I think always, I'm not an organ transplant physician,

00:49:56.800 but my impression is these people are always taking other immune suppressants in combination

00:50:02.320 with mTOR inhibitors. So in that context, yes, it does seem to be the case that mTOR inhibitors

00:50:09.700 can help reduce the chance of organ rejection in transplant patients. I don't, I could be wrong.

00:50:17.700 I think there's, I think there's preclinical data. I don't know of any strong clinical data

00:50:21.680 that by itself, rapamycin, even at higher doses has substantial immune suppressing effects. It

00:50:29.320 might, it wouldn't shock me if it does at high doses, but I don't know that that's really ever

00:50:33.180 been, been shown clearly in healthy people, you know, taking higher doses of the drug.

00:50:38.240 Yeah, that's interesting. I don't think I've ever gone back and looked at the FDA approval process in

00:50:42.800 99 because obviously it was approved, approved for that use. And I think that's the only use that

00:50:49.360 it's been approved for. So I wonder what trials led to that approval, you know, whatever trials

00:50:54.960 would have taken place in the mid nineties. Yeah. Another point that I think is worth making

00:50:59.040 is, you know, we use, and I I'm as guilty of this as anybody else. We use sort of broad terms when

00:51:04.600 we're talking about the immune system suppression, you know, lower function. We say the immune system

00:51:09.820 doesn't function as well in older people as compared to younger people, which is true. But I think when 0.94

00:51:15.420 we say it doesn't function as well, we tend to think of as it's functioning at a lower level,

00:51:20.340 right? There's just less activation, which actually isn't necessarily the case in old mice,

00:51:26.160 in old people, there's a lot of immune activation that shouldn't be happening, right? There's a lot

00:51:30.800 of sterile inflammation that's occurring. So it's not necessarily that the immune system is

00:51:36.280 functioning less, it's functioning inappropriately. And, and I think there's a ton of evidence that,

00:51:42.680 that many of the benefits that we see in mice from rapamycin occur because it tamps down on that

00:51:50.820 sort of sterile inflammation that goes along with aging, right? The inappropriate activation of the

00:51:56.880 immune system. So in that sense, you could think of it as an immune suppressant, but, but through

00:52:01.540 mechanisms that I don't understand at all, and I don't think anybody really understands, it seems to

00:52:06.860 be more effective at targeting the bad part of the aberrant immune response in an aged physiology,

00:52:13.960 which might be the reason why you're able to then have more of the good part. And here I'm, you know,

00:52:19.840 clearly showing my ignorance of immunology because I'm referring to it as bad and good, right? But I

00:52:25.060 think in some ways that's, it's, it's useful to kind of think about it at that level because,

00:52:30.000 because it's easier to, to appreciate that it's not, you know, not immune function isn't intrinsically good

00:52:35.800 or bad, right? It, it is. And there are different types of immune responses. And if you get the,

00:52:41.640 the wrong type of the immune response at the wrong time, that's bad. If you get the right type of

00:52:47.760 immune response when you need it, that's good, right? And, and, and I think in, in the context of

00:52:52.620 aging, we just see a lot more of the bad and probably a decline in function of the good. And, and I don't

00:52:58.140 know if rapamycin is affecting both of those, but I think it's definitely affecting the bad and bringing

00:53:02.880 it back down, which might just be enough to allow the good to come back up where it's supposed to.

00:53:08.320 Yeah. Which, you know, look, I think that's a great point to make. And it's, it's very similar

00:53:11.600 with reactive oxygen species, right? I mean, you have to have them. They are vital signaling molecules.

00:53:17.080 And yet if they run amok, they cause damage. And so, yeah, a lot of biology is, is in the Goldilocks

00:53:23.200 framework of not too much, not too little. Can I just throw one more thing out there as well,

00:53:28.360 which is, and this is, this is, again, I'll admit if you'd asked me 15 years ago, like how important

00:53:33.240 did I think that, you know, chronic inflammation and immune function would be in aging? I probably

00:53:37.860 would have said, you know, I don't know, but my guess is it's not going to be the major player,

00:53:41.220 right? Because, you know, my background came from working in yeast and worms, which, you know,

00:53:45.620 worms sort of have an immune system, but it's not really right. I would have pointed to things like

00:53:50.020 translation and autophagy and mitochondria, all of which are clearly important for affecting

00:53:55.760 inflammation and immune function. But I, I wasn't a big, I wasn't really bullish on

00:53:59.740 inflammation at that point. And I have become very bullish on inflammation as, you know,

00:54:05.240 critically important for many of the functional declines and diseases of aging that we see in

00:54:10.600 people. The point that I want to make here is that maybe I shouldn't say every, all of the

00:54:15.780 interventions that I know about that, that I'm enthusiastic about translationally seem to hit

00:54:22.360 inflammation in mice and they seem to tamp down on the chronic sterile inflammatory signaling that we

00:54:29.740 see go along with aging, which makes sense. It's encouraging that they all seem to have this shared

00:54:35.800 mechanism. But the other, the flip side of that is people talk a lot about their, people are very

00:54:40.620 excited now about, you know, senolytics or reprogramming now is the new, the new senolytics,

00:54:45.580 right? It's the thing everybody's excited about. I'm not sure that those are fundamentally

00:54:49.400 different from rapamycin in terms of the way that they're working. And, and I think we,

00:54:55.300 obviously we need to find out because, you know, it would be nice to know whether combinations of

00:55:00.240 these interventions are going to do better than one alone. But, but to me that the underlying theme

00:55:05.380 that seems to be similar about all of these things that work in mice is if you look in tissues of

00:55:10.500 aged mice at inflammatory cytokines, P16, P21, markers of senescence, they seem to be tamped down

00:55:18.960 by all of these, these interventions, which might explain the functional improvements that we see

00:55:25.680 from using these interventions in aged mice. I think that's actually a really interesting point.

00:55:30.740 And I would kind of say I'm in the same boat. I have become probably more convinced at the

00:55:39.660 importance of inflammation, certainly in Alzheimer's disease and certainly in atherosclerosis,

00:55:45.160 because I've seen enough people who either develop these conditions without otherwise clear

00:55:53.980 reasons for it and vice versa, right? People who do have other risk factors and don't go on to develop,

00:55:59.940 but have at least have a demonstrably low measured amount of inflammation. Again, the question becomes,

00:56:07.080 is there a direct target of inflammation or do you simply reduce inflammation by targeting these other

00:56:12.060 mechanisms, which we'll, we'll get to? Let's go back to the Everolimus progression. You were just

00:56:19.640 about to talk about RTB 101. So what, what is that and how is that the next stage in the evolution of

00:56:26.120 this Rapalog? Right. So there was the first phase two trial was, was done. They hit their endpoint. It

00:56:32.240 looked like it worked and they didn't see any, I mean, this was a phase two trial. So, so the important

00:56:36.140 thing was they didn't really see any substantial adverse effects as well, right? So then for the

00:56:41.340 next trial, they added in another drug called RTB 101, which is one of these drugs that hits multiple

00:56:49.840 kinases in the cell. So mTOR is a kinase, which means that the biochemical activity is to, is to put

00:56:56.700 phosphate groups on other proteins, but you can think of it as a signaling sort of activity. mTOR

00:57:02.380 senses the environment and regulates the output based on this signaling function of being a

00:57:08.760 kinase. This drug, RTB 101 inhibits mTOR, the catalytic activity of mTOR. It also inhibits other

00:57:15.760 kinases. So it's talked about as sort of a dual kinase inhibitor, but in reality it hits multiple

00:57:20.480 kinases. So it's a dirtier drug. That's, that's maybe the way to, the point to make it. It's a dirtier

00:57:24.820 drug than rapamycin, which has a very specific biochemical effect on mTOR and mTORc1 specifically,

00:57:31.720 which we haven't dove into the mTORc1 versus mTORc2 situation yet.

00:57:36.760 And was the rationale for using a dirtier drug simply to create a new drug, to have a molecule

00:57:42.060 that is novel and therefore protectable by IP, or was it designed to be dirtier without thinking of it

00:57:48.920 as being dirtier, but thinking of it as being sort of pluripotent across several kinases?

00:57:54.940 I obviously wasn't involved in designing this trial, so I can't answer that question as to what

00:57:59.600 exactly their thought process was. I think it is true that RTB 101 had and has a longer existing

00:58:06.380 patent life than Everolimus. So there is that component, certainly, to moving a drug through

00:58:11.320 the approval process. And if you talk to Joan, she will tell you that they had unpublished data that

00:58:17.680 RTB 101, at the doses that they were using in this trial, appears to be specific for mTORc1,

00:58:25.520 like rapamycin. So I think their thought process, I mean, in reality, it was probably twofold, right?

00:58:31.320 The IP life was greater and there's a biological rationale for testing RTB 101 either alone or in 1.00

00:58:38.960 combination with Everolimus. And so that was really the design of the phase two trial. It was structurally

00:58:44.960 very similar to the first phase two trial, right? So you have older adults, same sort of age range.

00:58:50.480 I think they had to be 65. They couldn't have a pre-existing age-related condition, significant age-related

00:58:56.260 condition. They had a control placebo arm, Everolimus alone, RTB 101 alone, and both. And I'll be honest

00:59:03.740 with you, I don't remember the dosing on the RTB 101. And so they looked at vaccine response. And then they

00:59:09.760 also looked at, in this study, upper respiratory tract infections over the next season to look beyond

00:59:16.620 to just, you know, if there's an impact of the mTOR inhibitor on immune function, is it specific to

00:59:21.980 vaccines or does it look like it's broadly boosting immune function? You know, you can look at the data

00:59:27.420 and not be completely convinced, but certainly for the Everolimus plus RTB 101 and the RTB 101 alone group,

00:59:37.740 I think the Everolimus alone group in this case didn't reach statistical significance. But for the

00:59:42.840 combination and the RTB 101 alone group, they saw improved vaccine response. And I think what was

00:59:48.940 really striking was a lower risk of upper respiratory tract infection in the people who'd gotten the

00:59:54.860 mTOR inhibitor over the next season. So that suggested that not only is it boosting vaccine response,

01:00:01.340 but it's also broadly conferring protection against a variety of immune challenges in this older

01:00:07.700 population. And again, very, very little in the way of adverse events, which gets back to the point,

01:00:13.860 you know, you hear people talk about how mTOR inhibitors, you know, you could never use them

01:00:17.940 for aging in people because the side effects are so bad. And it, you know, it just gets frustrating

01:00:22.280 over and over and over again to say, go read the data, you know, go look at the data, right? That's

01:00:28.080 just not true at the doses that, that, you know, have been tested so far. That's just blatantly false,

01:00:33.720 right? There's no evidence actually for significant side effects from rapamycin monotherapy,

01:00:40.000 everolimus monotherapy, or RTB 101 at the doses that people are talking about using in this context.

01:00:47.120 Yeah. I mean, statins have far, far greater side effects than rapamycin. I mean, and it's not even

01:00:53.460 close. And I think statins are a very important type of drug, but to think how ubiquitous they are

01:01:00.140 and that we accept, Hey, 10% of people, I mean, I think the literature says 5%, but really clinically

01:01:05.620 10% of people have debilitating muscle aches from statins that, you know, render it impossible to

01:01:11.000 take them. People, some people have elevated liver function tests that are otherwise unexplained. I

01:01:15.040 mean, it is really frustrating for me to hear this, especially when people talk about how, well,

01:01:20.660 we could never really study rapamycin because, you know, it's just too unsafe. It's like, I,

01:01:25.320 I'm not sure what data you're pointing to, but when, when you have it, let me know.

01:01:29.280 Yeah, I agree. So anyways, so that was the second phase two trial, which was successful.

01:01:34.760 So then they went on to a phase three trial to hopefully get FDA approval for improving

01:01:40.480 vaccine response and immune function in, in elderly people. And in that phase three trial,

01:01:46.680 I'm trying to remember what they called that trial. I don't, I don't remember the trial,

01:01:49.500 but now what was rad zero zero one? What was that? That's the same drug.

01:01:54.480 That's what they were calling everolimus. Okay.

01:01:56.060 Yes. Yeah. And they had a name for the, all clinical trials have names. So they had a name.

01:02:00.420 I don't remember what it was, but, but in that phase three clinical trial, they only tested RTB 101.

01:02:05.700 So they, they took out the rapalog everolimus and only tested RTB 101.

01:02:12.600 So there was no rad zero zero one plus RTB 101.

01:02:16.260 That's right. No combination, no individual rapalog. That's right. And that clinical trial failed to hit

01:02:22.520 the end point. And it was, it was terminated halfway through. So they, this is my understanding.

01:02:26.900 They were going to do one group in the fall season and one group in the spring season or something like

01:02:33.040 that. And they got halfway through, they weren't hitting their end point. So they terminated the

01:02:37.240 trial early. Restore Bio was the company that was doing this clinical trial to try to,

01:02:43.340 to move RTB 101 through to approval. The board basically merged them with a, with a CAR-T

01:02:49.800 cell therapy cancer company, and they gave up on RTB 101 because of the failed clinical trial.

01:02:56.640 So, you know, why that clinical trial failed, I think still the data has not seen the light of day

01:03:03.320 yet. I don't think it's come out yet. So, so we don't know for sure what happened. I think,

01:03:08.460 you know, we can observe that the rapalog wasn't in there anymore, right? That is one difference

01:03:13.980 between the two successful trials and the one that failed. And I believe Joan has talked about this

01:03:19.680 publicly that, that, that they also in conversations with the FDA were required to change the end point

01:03:26.260 from laboratory confirmed infections. That was the end point, one of the end points in the phase two

01:03:31.660 trial to something else, which involved patient reported symptoms as the end point, and they

01:03:37.740 didn't hit it. And so I think that, I think my understanding from talking to Joan is that data

01:03:41.640 will come out at some point and, you know, we'll be able to really take a look and, and see, you know,

01:03:46.920 what the drug was doing and, and what it, what it wasn't doing in that third clinical trial.

01:03:52.220 My intuition is that it probably worked and they probably got screwed by being forced to change

01:03:58.720 the end point of that clinical trial. And if that's the case, then I think, you know, we really

01:04:04.080 do need to have a conversation around the way that FDA, I mean, we need to have this conversation

01:04:08.820 anyway, but the way that FDA is approaching clinical trials in the aging space, what needs to be shown

01:04:18.840 for an appropriate end point? And what is the acceptable level of risk when the, the goal of the

01:04:28.340 trial is to test whether or not an intervention is affecting a functional decline that goes along

01:04:34.380 with aging or aging itself, right? And that's a, that's a bigger question. I certainly have plenty

01:04:39.560 of thoughts around that. I know there are lots of people in the field who are thinking about this

01:04:44.140 and working on it. And some have talked to FDA. It's a real challenge. And I don't want to blame

01:04:48.600 anything on FDA, right? I think that they are, the people at FDA want to do their jobs to the best of

01:04:54.920 their ability. I think there are constraints around the way that FDA is required to work

01:05:00.040 and the culture that does not, in my view, appropriately evaluate risk reward. I think

01:05:09.680 that there is a culture of the risk has to be extremely low in people who are of normal health

01:05:17.320 status for their age. I actually am trying very hard. I still slip, but I'm trying very hard

01:05:22.500 not to use the word healthy when we talk about aging or older people. I'm relatively healthy

01:05:30.620 for my age. I would say I'm probably in the top 10% health wise for my age group. I am not as healthy

01:05:37.420 as I was 20 years ago. We already talked about this, right? So I'm of, you know, normal to upper

01:05:42.720 health status for my age, but I would not call a 65 year old, a typical 65 or 70 year old healthy.

01:05:49.460 They're not, they're functionally impaired. And we really need, we being, you know, the regulators,

01:05:55.900 society, policymakers really need, I think, to start taking a realistic look at what normal aging is.

01:06:05.820 It is a progressive chronic decline in function that at some point will lead to overt disease and

01:06:13.640 with 100% probability will lead to death, right? And so if, if we can intervene in that process to

01:06:21.300 slow it down or reverse it, there should be some level of risk that is acceptable for that potential

01:06:28.880 outcome. And I think that because we tend to think of 65, 70 year olds as healthy, as opposed to

01:06:35.040 functionally impaired, it makes it really hard to have a rational discussion around what the appropriate

01:06:39.460 level of risk is. Matt, I think, I think that is so astonishingly well said. And I've, I've had so

01:06:46.220 many discussions with near and to a lesser extent, Steve Ostad about the choice of metformin over

01:06:52.880 rapamycin in tame. And it always comes back to this point, which is the regulators will absolutely

01:06:59.720 positively not consider rapamycin, which of course comes back to this question, which is what are we

01:07:07.380 claiming to be studying and in whom? The other thing is the studies are being designed around

01:07:14.000 disease, right? Progression of disease or outcomes of disease. So this really comes back to a broader

01:07:20.520 theme, which is where are we on the spectrum of understanding aging in the way that you're defining it

01:07:30.280 and getting it further away from the discrete definition that involves a disease? Because until

01:07:37.260 we really get people to center around that, your very eloquent explanation of there being no such

01:07:44.960 thing as a healthy 70 year old, until we realize that at the regulatory level, right? At the policy

01:07:51.760 level, even at the scientific level, it's going to be very difficult to study the things that will

01:07:57.240 have an opportunity to give step function changes in longevity. Let me start by pushing back on,

01:08:03.680 on what you said about tame and metformin versus rapamycin, because I think that's a myth that

01:08:09.540 regulators would not allow you to do a clinical trial with rapamycin, right? We just talked about

01:08:15.640 three clinical trials that were done with mTOR inhibitors, two of which were with everolimus,

01:08:20.660 which is essentially rapamycin. So I think that is talked about as to why metformin was chosen for

01:08:26.600 tame. I don't think that is true. And I don't actually think that is the reason at all.

01:08:30.860 I've pushed, by the way. I've pushed. I needle near constantly about it.

01:08:35.880 I mean, honestly, I think there are good reasons why metformin makes sense to test in that context.

01:08:41.800 And there's data in people that support that. So I'm not trying to say they should have used

01:08:46.480 rapamycin. I think that people with expertise in the field can come to differing opinions as to what

01:08:52.220 the best shot on goal is, I guess.

01:08:55.140 You're just saying, let's just say we picked metformin because of these reasons, but not because

01:08:59.920 we thought that RAPA wouldn't be safe enough or the regulators themselves would decline it.

01:09:05.180 Yeah. I think they'd let you do the trial. Obviously, they would pay attention to

01:09:08.460 adverse events and there would be concerns around adverse events, but they would certainly let you

01:09:12.380 do the trial. And that's been proven, right? I mean, RestorBio did the trial. And it wasn't

01:09:17.680 because of adverse events that it got shut down. I think that's the other important point to make.

01:09:21.520 I think the regulators at FDA are doing the best job that they can within the constraints of how

01:09:27.060 they are required to regulate drugs. I have a real problem with the way we regulate drugs in

01:09:31.800 this country, but I don't necessarily blame it on the people at FDA. And I do think if you came to

01:09:37.140 them with a clinical trial where you had an endpoint that was quantitative and functional and related to

01:09:43.100 quality of life in people, they would let you do that clinical trial with rapamycin. I'm 100%

01:09:49.340 certain of that. The challenge is, I think the reason why this hasn't happened, one,

01:09:53.520 rapamycin is off patent. Nobody's going to make money off of it. And two, there is a misplaced

01:09:59.920 concern about side effects, which we've already talked about. It's just not reality that the risk

01:10:05.140 is significant at doses of rapamycin that would be tested. So I think the real challenge though,

01:10:11.160 is identifying the right endpoints for a clinical trial in aging. You're not going to do lifespan in

01:10:17.340 people. We can do it in dogs. You're not going to do it in people. I think we just have to accept that.

01:10:21.080 So what is the, I don't even want to say the right approach, because I don't think there's one,

01:10:25.160 but what are some approaches that one might consider if your goal is to move FDA towards

01:10:32.240 approving a drug in people of normal health status to prevent age-related functional declines in

01:10:39.720 disease, to target aging. That's what I mean when I say prevent age-related functional declines in

01:10:44.860 disease, to target aging. So if your goal is to get there, so what does that clinical trial look like?

01:10:49.620 So the TAME trial is a specific example of one strategy, which is to take people who are already 0.98

01:10:56.060 have one age-related disease and ask whether your intervention, metformin in this case, can delay the

01:11:02.020 onset of the second age-related disease. It's a comorbidity trial. That makes a lot of sense because

01:11:08.440 we know with pretty good precision how long that timeframe is on average, from the first age-related

01:11:13.700 disease to the second age-related disease. And so you can quantitatively assess, does your intervention

01:11:19.200 increase that length of time? That's the rationale behind TAME. In my personal view, the limitation of

01:11:26.340 that approach is it's not a true healthy aging study, right? It's not taking people who are of 1.00

01:11:33.420 normal health status for their age and asking whether the intervention can improve or extend the

01:11:40.940 healthy period of life, right? So it's a different design. But I totally get why TAME was designed

01:11:46.400 that way. And I like the design. I think there's a reasonable chance it'll work. But it's different

01:11:51.040 from the way I would think about a clinical trial in this space. What I would do is I would try to

01:11:56.720 identify the best single endpoint or set of endpoints that correspond to a significant functional deficit

01:12:06.500 that impacts quality of life in older people and assess whether or not my intervention improves

01:12:15.140 that. And optimally, that endpoint would have quantitative markers that you could measure to

01:12:20.780 show that you've improved it. So RestoreBio went with immune function and they went with vaccine

01:12:25.860 response initially. And then ultimately, I think in the phase three trial, they were also looking at

01:12:31.080 respiratory tract infections, right? Over the next season. That was their endpoint. It's quantitative.

01:12:36.680 The problem is it's really hard, right? It's a really noisy endpoint. And I'm not at all

01:12:41.520 criticizing them. In fact, I think, as you know, Joan Manik is one of my favorite people. She's a good

01:12:46.140 friend. I have an amazing amount of respect for her. I think they went for it. And for reasons that

01:12:51.820 were probably beyond their control, you know, that clinical trial failed. But I have so much respect for

01:12:57.220 what they tried to do. I'm just saying it's a tough clinical trial. It's a tough endpoint. It's noisy.

01:13:02.100 I think there are other endpoints that you could consider that might not be as noisy that you could

01:13:08.600 consider doing a clinical trial for. So one of my favorites at this moment is periodontal disease.

01:13:14.260 And that's because, you know, my lab has published that aged mice get periodontal disease,

01:13:19.320 that eight weeks of treatment with rapamycin reverses the clinically defining features of

01:13:24.420 periodontal disease in mice. We know that something like two-thirds of older adults have periodontal

01:13:30.440 disease or will get it. And those who have periodontal disease are at higher risk for

01:13:35.320 dementia, cardiovascular disease, diabetes. So it's connected in some way to other age-related

01:13:41.360 diseases.

01:13:42.520 And by the way, you know, a recent podcast of mine explored this topic and it may in fact

01:13:47.460 be causally related through that inflammatory axis, right? I mean, I think that that's probably

01:13:53.080 the strongest line of evidence connecting oral disease with systemic disease. It's through this

01:14:00.220 immune inflammatory pathway. So I think that's actually an elegant approach.

01:14:04.960 So the reason why I like periodontal disease, right, is the endpoints are extremely quantitative,

01:14:10.040 right? So what we looked at in the mice are gingival inflammation, bone around the teeth,

01:14:14.960 which can be measured, you know, crudely by pocket depth, more quantitatively by x-ray,

01:14:19.520 and microbiome. That's really the, at least in my understanding, I'm not a dentist,

01:14:25.060 but I've learned a lot about oral health over the last few years. That's my understanding of,

01:14:29.720 you know, if you have gingival inflammation, you've lost enough bone around the teeth,

01:14:35.100 you don't even have to have the dysregulation of the oral microbiome, but it always goes along with

01:14:38.720 it. You've got periodontal disease, right? So there are nice quantitative endpoints that can be

01:14:43.340 looked at in people. And it's extremely non-invasive, right? The way that you do this clinical

01:14:49.200 trial is you have people come in for a dental exam, right? Before and after treatment. So you've got

01:14:55.200 a shot at seeing changes in quantitative endpoints that we can not just delay, but actually reverse

01:15:03.500 the declines in mice in people, you know, in a reasonable timeframe. So you could easily imagine

01:15:09.880 something similar to the structure of the Restore Bio trial with rapamycin or everolimus or RTB-101,

01:15:15.980 or, you know, pick your favorite intervention where you treat people for eight weeks, three months.

01:15:22.140 You know, they have a dental exam before, a dental exam after, a dental exam six months later. And you

01:15:28.020 just look and see what was the impact of the intervention. So it's a pretty straightforward

01:15:32.600 clinical trial. And, you know, I hope we will get this off the ground. We're actually trying to

01:15:36.720 get some funding to do a clinical trial now. In humans or in dogs? In people. Yeah. In people.

01:15:41.680 And this is, this is really the person who deserves all the credit for this crystallizing in my mind is

01:15:46.640 a gentleman named Jonathan Ahn, who was a DDS PhD student. So, so he had already got his dental degree.

01:15:53.360 He did his PhD with me. And he, you know, before he came to my lab, he came to me one, one day.

01:15:58.260 I vividly remember this conversation because I had never thought about oral health. And he goes,

01:16:03.740 you know, people get periodontal disease when they get older in much the same way that people get

01:16:10.600 Alzheimer's disease or heart disease. If you look at the risk profile as a function of age,

01:16:15.760 it looks strikingly similar to these other age-related diseases. You know, what do you think

01:16:20.200 maybe the biology of aging is contributing? And, you know, in hindsight, it's like, oh yeah,

01:16:25.180 that makes a lot of sense. But I had never thought about that before. And so he came to my lab and

01:16:29.220 showed that we could do this in mice. We could, we could see age-related periodontal disease.

01:16:33.120 We could see bone loss around the teeth. We could see inflammation of the gums and that we could

01:16:37.220 reverse that with rapamycin. And so he's actually now a faculty member at the University of Washington.

01:16:42.000 And he's really the one who's trying to push this clinical trial forward. So,

01:16:45.120 so I'm sort of peripherally involved, but John is really the guy in this space. He's going to be a

01:16:49.640 rock star. And this would be a phase two?

01:16:51.320 Yes. Right. If it gets funded, I mean, we're, we're just submitting the grant now. So I don't know,

01:16:55.980 you know, it's very early, right? What would be the budget for this study?

01:16:59.480 So it's going to be a three-year grant. I, you know, I don't know. It's,

01:17:02.480 I think it's R01 size, a couple hundred thousand a year over three years.

01:17:06.000 But I want to go back to something. If money were no object,

01:17:09.840 what secondary endpoints would you add to that study? In other words, if you could power this

01:17:14.600 study to hit that as your primary, but also go after multiple secondaries, let's just throw in

01:17:20.900 immune function. There'd be no reason not to repeat what was done with RAD001. What else would you add

01:17:27.740 to that? Right. So if you look at the mouse data, I mean, I think the mouse data is, it's a reasonable

01:17:32.420 place to start. Again, obviously mice aren't people, but it's a reasonable place to start.

01:17:36.200 So where does rapamycin reverse functional declines associated with aging?

01:17:42.380 Hearing, right?

01:17:43.740 Yes.

01:17:44.100 Wasn't there that study that just came out? Yeah. It just came out two weeks ago.

01:17:47.680 Right.

01:17:48.060 It reversed age-related hearing loss, which I, that was, that got me very excited.

01:17:53.240 Yeah, I agree. So that, and that's again, a very easy and quantitative endpoint, right?

01:17:57.560 Wouldn't it be amazing if you could improve age-related hearing loss with rapamycin? And so

01:18:01.940 that's a no-brainer. Immune function, I mean, I agree. If money was no option, sure. But, but even

01:18:07.520 in the context of a, so let's say the periodontal disease clinical trial that I, that I talked about,

01:18:12.960 if you could also measure hearing in that clinical trial, right, you can, it's a, it's a big, big bang

01:18:18.680 for the buck, right? It costs you almost nothing and you get potentially another endpoint that you

01:18:22.880 could hit on. Where else? So muscle function, there's pretty good evidence that at least rapamycin

01:18:27.540 can prevent sarcopenia. I don't think there's a lot of data yet on improvements in, in muscle

01:18:32.300 function, but you could do things like, you know, grip strength, walk speed, things like

01:18:36.140 that. Cognitive function, that's hard. Again, that's kind of, I put that in the sort of the

01:18:40.940 same category as immune function as an endpoint. It's a tough endpoint, but it would certainly

01:18:46.220 be interesting to look at cognitive function in this elderly population of normal health status.

01:18:51.880 So some of these people are going to be on the road to dementia. They won't have dementia

01:18:56.080 yet because that's an entrance criteria to get into the trial. Heart function, it's another place

01:19:00.840 where we see reversal of, of age-related declines. This would be hard. So this is a different,

01:19:05.860 this is a different clinical trial, but also potentially cool reproductive function. You

01:19:10.240 wouldn't do that in that same patient population probably, but in mice, there's pretty good evidence

01:19:15.220 that you can reverse, at least in females, reproductive declines that go along with aging,

01:19:20.500 or at least delay them. Let's just pause on that for a second, Matt. That is staggering when you

01:19:26.560 consider where we are today from a standpoint of reproductive medicine. Women are having children

01:19:34.040 later and later, and I can't tell you the number of just patients in my practice, either male or female,

01:19:41.660 for whom this isn't a top of mind priority. It's simply unbelievable. And to think that there would

01:19:49.640 be a, you know, I never really had thought of this, honestly, because I don't think I paid attention to

01:19:53.740 that subset of the rapamycin literature. If you were to, and again, this is guessing, but how much of an

01:20:01.220 impact do you think you could have? So if you, if you had a woman who was 40 years old, she's obviously 0.72

01:20:07.020 still premenopausal, but her AMH is low and, or it's reasonable, but she's got significant

01:20:16.300 aneuploidy for the listener, meaning whenever her eggs are produced, they don't evenly divide into the

01:20:23.480 right number of chromosomes. So they don't show up with one of each chromosome. And that's an enormous

01:20:27.980 cause of infertility as a woman ages, is this aneuploidy. So they either omit a chromosome or 0.98

01:20:35.080 include two, and those almost universally lead to an early miscarriage. Some of them like trisomy

01:20:41.220 13 or 21 will make it. Do you have a sense of what type of magnitude of an effect this could have?

01:20:49.100 I think the honest answer is no. So I think there's a couple of things to say on this. One is there's

01:20:52.920 not a lot, there've been, I think two or three studies in mice looking at this, right? And so

01:20:57.260 there's not a lot of data on magnitude of effect, even in mice. So my guess is that in people,

01:21:05.280 so I don't even want to comment on magnitude of effects. I really don't know. I think the way

01:21:09.480 that you would design a clinical trial though, would be very similar to what we've talked about

01:21:14.580 with rapamycin already with the immune function or periodontal disease. You would take a woman 1.00

01:21:20.440 premenopausal, let's say early forties, a group of women and do eight weeks of treatment and then 0.95

01:21:28.700 a washout. And then you could look at your endpoints or if they are going through IVF,

01:21:34.120 for example, look and see whether or not you get an improvement in outcome using that functional

01:21:39.780 measure. Now, why the eight weeks? Do we really believe that that would be a sufficient amount of

01:21:44.380 time? I mean, are we just coming up with eight weeks because that's where we saw the vaccine response

01:21:49.160 in the Everolimus trial? Or do we think we, you know, if you were shooting for the moon and cost

01:21:54.680 were not an issue, what would be the downside of longer treatment? A couple of things. So one is

01:21:59.900 increased risk of side effects, right? I mean, I think that's the longer you go with any treatment,

01:22:03.620 the more risk there is for an adverse event, even though I think the risks are pretty low.

01:22:07.680 Here's a really interesting piece of data that we don't have with the immune function studies in mice

01:22:12.600 or people. In all of those studies, the treatment was stopped. And then there was a two week or so

01:22:18.060 period before the vaccine was given. So what was the rationale for that washout?

01:22:23.160 I think it's because people think and thought about rapamycin as an immune suppressant. So,

01:22:29.340 and this is where I was going with that. What we don't know is if they were still taking the drug,

01:22:33.540 would you have gotten the same effects for immune function? I don't know. So it's an important

01:22:38.000 question that I think we just don't know the answer to. But the other risk is that at least for some of

01:22:42.780 your functional measures, if it requires, let's say, hyperactivation of mTOR to get the response,

01:22:50.440 you might impair that by doing continued treatment with rapamycin. And at least all of the data that

01:22:55.500 I've seen, limited in people, extensive in mice, eight weeks is enough to give you essentially that

01:23:02.840 full benefit for whatever the functional output is that you're looking at. Maybe not for lifespan,

01:23:08.100 right? Maybe multiple eight week transient interventions would be better than one.

01:23:13.760 Only thing that I know of for data there is our study where we did three months

01:23:17.160 of rapamycin treatment between 20 and 23 months of age, and then let them go to the end of life.

01:23:22.380 The magnitude of effect was pretty similar to what the ITP saw. So it was reasonably close to

01:23:28.020 continuing treatment, but we don't know. But eight weeks in mice is, you know, what would that be in

01:23:33.760 humans, right? So here's the thing. Yeah, I know where you're going with that, right? If you were to

01:23:36.340 linearly extrapolate that, that would be a few years, right? In people. And I agree with that. 0.63

01:23:40.480 I think that the immune trials that we've talked about suggest that that might be long enough,

01:23:45.660 but I agree. We don't know. We don't know is the answer. I'll just say, so I haven't talked about,

01:23:50.500 you know, my experience with rapamycin and I, and I really don't talk about this publicly,

01:23:54.680 but I'll do it here. So I've tried eight to 10 week courses of rapamycin a couple of times.

01:23:59.580 The most recent time it was because this was probably spring of 2019. So I'm pretty active.

01:24:06.160 I play softball in the spring and I noticed I had a lot of shoulder pain. And by the end of the

01:24:11.300 season, it was to the point where I couldn't throw a football. Like I couldn't go play catch with my

01:24:15.440 son. Actually, that's one of the hardest personal sort of aging experiences I had was when we went

01:24:20.800 across the street to the park and I was going to play catch with my son and I couldn't do it

01:24:25.620 because my shoulder hurts so much and my right shoulder. And I thought I must have a rotator cuff

01:24:30.980 tear. I went in to see the specialist, finally got diagnosed with frozen shoulder, which is

01:24:36.340 inflammation of the shoulder capsule, which happens to people as they, some people as they get older,

01:24:41.200 extremely painful, completely limited my range of motion. And the doctor was like, well, I could

01:24:46.500 give you a shot of cortisol, but I don't really recommend it. That can, that can degrade the cartilage.

01:24:51.260 Really. There's not a lot you can do. Some people, it goes away after a year. Some people just,

01:24:55.980 you just learn to live with that. And I was pretty depressed by that diagnosis, right? I was like,

01:24:59.820 at least if it was a rotator cuff tear, I could get surgery, get it fixed. So I go home and I'm

01:25:03.940 sitting there and I'm thinking, and so I'm looking on the internet and I see, okay, it's an inflammation

01:25:06.940 of the shoulder capsule. I'm thinking to myself, what do I know that has anti-inflammatory effects,

01:25:12.540 you know, in the context of aging? Rapamycin. So I got some rapamycin and I did eight weeks. And I

01:25:18.840 mean, again, placebo effect is real, but this was so painful. I don't believe it was placebo effect.

01:25:24.380 Within two weeks, I had probably half my range of motion back. Within eight weeks,

01:25:28.400 I was back 195%. Were you just dosing once a week?

01:25:33.660 Once a week, eight migs once a week. Yeah. And I want to be careful because I want to say, 1.00

01:25:37.140 I'm not encouraging other people to go do this, but I am a true believer after that experience,

01:25:42.660 right? I don't think it's placebo effect. I don't see how it could be with how painful it was,

01:25:47.240 how real the limitations were on range of motion. So that goes to this eight week question,

01:25:52.180 at least for that indication, which I believe this was a real effect. Eight weeks was plenty and it

01:25:59.440 hasn't come back, right? Which actually kind of makes sense with frozen shoulder. I think when,

01:26:03.140 when people recover from it, it doesn't always come back. So I think that again, a lot of these

01:26:09.160 age-related conditions that are inflammatory driven, you can kind of reset that with an eight

01:26:17.660 week treatment and rapamycin. I suspect senolytics, if we had good senolytics, would do pretty much the

01:26:22.980 same thing. Do you think that rapamycin is itself a senolytic? I don't think it's a senolytic in the

01:26:28.640 sense that it, at least the classical way people have thought about senolytics where it kills the

01:26:32.520 senescent cells. I absolutely think it turns down the chronic inflammatory signaling that is driven by

01:26:39.060 P16, P21, NF-kappa B. Yes, we see that in mice, multiple tissues, no question about it. I still think

01:26:46.280 the senescence field is a little bit messy in the terminology. I think a lot of what people call

01:26:51.980 senescence isn't truly being derived from senescent cells, the way that we think about them.

01:26:59.440 It's P16, P21 mediated inflammatory cytokines, right? Doesn't necessarily have to come from senescent

01:27:07.320 cells. No question. Rapamycin shuts that off and it shuts it off within eight weeks, at least in mice,

01:27:13.100 in a lot of tissues. And do you think it's doing that independent of what it's doing at the mTOR or

01:27:17.820 do you think that it's doing that through mTOR? I would be shocked if it's not doing that through

01:27:22.980 mTOR. I don't know of any good evidence that rapamycin has off-target effects. Any activity?

01:27:27.900 Yeah. Okay. Yeah. So really it's these SASPs getting whacked that is how it would act through

01:27:35.160 via the senescent pathway as opposed to targeting a senescent cell directly. By the way, that's been my

01:27:40.760 reading of the literature. Would you agree with that? I think that's right. I think in some way

01:27:44.680 it shuts off those chronic inflammatory signals. The secretories. Yeah. Yeah. Yeah. And maybe other

01:27:51.140 stuff, right? I mean, we always look at the SASP because that's what we know. I mean, I think this

01:27:54.780 is natural in science. We look under the lamppost. We measure what we know to measure, right? It would

01:27:59.580 not shock me at all if there are other things that go along with the canonical SASP that senolytics or

01:28:07.640 rapamycin or caloric restriction. I think that's a big part of fasting. Fasting does the same thing,

01:28:12.280 right? Tamps down on that chronic inflammatory signaling, maybe through mTOR, maybe through

01:28:17.620 other mechanisms. So I think that what we know about is part of it. It wouldn't shock me if there

01:28:22.700 were things we don't know about yet that also are contributing. Yeah. But getting back to this eight

01:28:27.220 week, right? That's how we got started. So I tend to think based on my personal experience and the

01:28:32.220 little bit of data from these two clinical trials, that that's probably long enough for at least some

01:28:37.820 endpoints that are driven primarily by immune dysregulation and chronic inflammation.

01:28:44.240 I think two and a half years ago when I started, I was very strict about eight on, six off, eight on,

01:28:49.680 six off, eight on, six off. And I don't know, a little while ago, I just sort of said,

01:28:55.860 yeah, I'm just not coming off. And I wish I had a biomarker to point to. I wish I had some

01:29:03.560 way of measuring whether this is the right thing to do or if, you know, eight, four, eight, four,

01:29:09.820 eight, four. There's a symmetry to eight, five, because you'll get through exactly four cycles

01:29:14.600 a year. Maybe I do eight, five, eight, five, eight, five, eight, five. But it really frustrates me that

01:29:19.780 we don't have a biomarker for this. Yeah. Or aging in general.

01:29:24.020 Or aging in general. So let's talk a little bit about that. So what does that look like? I mean,

01:29:27.340 when I had Eileen White on the podcast, gosh, it's been maybe a year and a half now, we had a really

01:29:33.900 interesting discussion about why we don't even have biomarkers for autophagy. I mean, something that

01:29:38.760 is so important and we can't measure it. And this was important in the context of people who choose

01:29:46.160 to calorically restrict or fast. Is fasting for a day long enough to generate a meaningful amount of

01:29:53.520 autophagy in a human, in a mouse, it clearly is. But in a human, is it? No idea. It's two days,

01:29:58.400 three days, seven days. You know, seven days is almost assuredly enough. It's a big difference

01:30:02.900 between fasting for a day and fasting for seven days. Why don't we have biomarkers for that?

01:30:07.780 Why don't we have a biomarker that can assess nutrient sensing better? Why don't we have a,

01:30:12.600 you know, I mean, you could argue we have some biomarkers. We can measure telomere length,

01:30:15.940 but you know my feelings on this, Matt. I'm in the camp that thinks measuring telomere length is

01:30:20.460 not helpful at all for aging. And I think there's plenty of data to suggest that while telomere

01:30:25.400 length is a very important marker of cellular division, it really speaks very little about

01:30:31.260 the organism's state of aging. Despite the popularity of that biomarker, even the epigenetic

01:30:37.640 clocks, I don't find to be helpful. I find them to be far too, and I'd like you to push back on this

01:30:42.400 if you feel as much, but I've seen how easily they can be manipulated by short-term interventions

01:30:47.780 that don't seem biologically relevant. I'll start with the epigenetic clock,

01:30:51.460 because everybody, that's a big area of interest in the field.

01:30:54.080 Let's explain to people what that is. Let's start from the beginning. Assume people don't

01:30:57.420 know what an epigenetic clock is. The epigenetic clock refers to typically chemical marks on DNA

01:31:03.560 that regulate gene expression, whether or not, you know, the gene that is located at specific points

01:31:09.820 in your genome gets turned on or off. And what has been observed is that those marks change

01:31:15.580 with age in pretty much every organism where it's been studied, and that you can identify

01:31:20.600 patterns of change at specific locations in the genome. So specific changes in these chemical marks

01:31:27.380 with age that correlate very strongly with chronological age. And so that has led to the

01:31:35.660 idea that you can create clocks that look at specific changes in chemical marks in the DNA,

01:31:42.920 the genome, that are telling you something about how long that organism has been alive.

01:31:49.940 And then what sort of has emerged from that, there are two things that have emerged from that. One is

01:31:54.260 you may be able to use that chronological aging clock to find individuals whose marks don't fall

01:32:01.880 on the line that you would expect it to fall on based on their chronological age. In other words,

01:32:06.520 they have marks that make them look older or younger than their chronological age says that they are.

01:32:11.400 And so you would hypothesize that those individuals biologically, if those marks are really

01:32:17.460 reflecting biological age, might be aging more slowly or more quickly. And what's been shown is

01:32:22.640 that indeed, those individuals who tend to be off the line, depending on whether they seem to be aging

01:32:28.360 more slowly or more quickly, are at lower or higher risk for specific diseases. So that adds some level

01:32:35.020 of confidence that this epigenetic clock, chronological epigenetic aging clock is actually reflecting

01:32:42.500 biological age. And so the idea is maybe we can use that information to develop epigenetic clocks

01:32:49.220 that will, in a predictive way, tell you how old you are biologically. So you can get tests now,

01:32:55.920 there are plenty of companies now that are selling these things, where you can go buy your epigenetic

01:32:59.920 blood tests. Mostly this has been done in blood cells. That's one limitation to think about is

01:33:04.440 almost all of the literature in humans is developed on epigenetic clocks from blood. And it's still,

01:33:10.000 I think, a little bit of a question, even if this is reflecting biological age, it's the biological age

01:33:15.020 of your blood, which may or may not reflect the biological age of your entire body. But you can buy

01:33:19.740 tests now that based on your, you give them some of your blood, they will tell you your

01:33:23.960 epigenetic biological age or some number. They're looking at PBMC, I assume?

01:33:30.060 I don't know. Honestly, I'm not, I'm not involved in any of this stuff. So I don't,

01:33:34.520 I don't know exactly what, yes, most of the studies that have been published are PBMCs. I don't,

01:33:39.180 they may even have saliva tests now. I don't know, honestly, how these commercial companies are doing

01:33:43.380 it. I mean, some of the clocks I've seen where I've just immediately discounted them is

01:33:47.820 when some of their inputs are things like glucose level, vitamin D level, which are things that

01:33:52.980 vary so much from day to day. And by the way, are so easy to manipulate. Like you

01:33:56.660 can take a vitamin D supplement or not take a vitamin D supplement. You can, you know,

01:34:01.120 have a high cortisol spike one morning and your glucose is 110 versus have a good sleep the night

01:34:06.600 before and your glucose is 95. So something that's that malleable, I just don't think makes sense as

01:34:12.840 an ironclad marker of, of true biologic age. Yeah. Let me take a step back. So the epigenetic

01:34:18.920 clocks, right, are, are probably the one that people talk about the most and have gotten the most

01:34:22.800 traction in the field. And I guess I'm a little bit of a skeptic, but I mean, I believe these clocks,

01:34:28.640 I believe the data and I believe that the correlations are extremely strong. I, I'm a

01:34:33.580 little bit worried still that there are so many data points in the epigenome that you can find a

01:34:41.140 pattern that will fit anything you go looking for. So I'm a little bit worried about the dimensionality

01:34:46.500 of the data and, and whether or not it's pattern matching in some cases, rather than,

01:34:52.340 than it's really truly going to be a robust predictor of biological age. That probably

01:34:57.220 reflects what is admittedly my limited understanding of the mathematics behind a lot of the epigenetic

01:35:04.000 clocks that have, that have been built. So I don't view that as a strong criticism. It's just a personal

01:35:08.180 sort of concern that I, that I have, but I think these clocks are telling us something.

01:35:11.780 So you're basically saying without doing the complex mathematics to correct for so many,

01:35:17.580 the multiple looks that you can take at the data, you could be tricked. And I have not spent enough

01:35:22.560 time looking at that either. I would like to have Steve Horvath on the podcast at some point,

01:35:27.340 because I think Steve could speak to that probably better than anyone else.

01:35:30.240 And for sure. But the other point I wanted to make is what you alluded to is now what people

01:35:34.700 are doing is going beyond the epigenetic clocks to try to look at every possible thing you could

01:35:40.020 measure, sometimes combining that with the epigenetic clock to build these super clocks

01:35:44.020 or multi clocks or multi-omic clocks. Right. And I think there's huge power in that,

01:35:48.840 but it also increases that dimensionality problem that I just mentioned, because, you know,

01:35:53.840 all of a sudden now you've got, if you're doing omics stuff, you've got tens of thousands of

01:35:58.740 additional data points that you can measure and you can fit a pattern where a lot of this.

01:36:04.600 And I think, I think even Steve and other people who are in the epigenetic clock field would agree

01:36:10.200 with this, where a lot of this has yet to really mature is in getting us to biological explanations

01:36:16.800 for what the patterns are telling us, right? What genes are, are they that these, these marks are

01:36:23.900 located at and are those in any way causal for, you know, biological aging? So I think if you get to

01:36:31.360 the point where you can understand mechanism, it's going to be much more powerful. I also think

01:36:35.580 though, this gets to the fundamental challenge with biomarkers. And I think this is where you're

01:36:41.560 dissatisfied. We have a lot of biomarkers of aging. We just don't have any validated biomarkers of

01:36:48.060 aging, right? And this has been a problem, you know, since I was a graduate student, everybody's

01:36:52.320 wanted biomarkers of aging. The NIA had a huge program where they, they did all this funding to

01:36:57.840 identify biomarkers of aging. I think it was back in the eighties before my time, nineties, maybe you

01:37:02.600 can identify all sorts of things that correlate with age. How do you get to the point of convincing

01:37:07.740 yourself first and other people second, that these things are actually telling you something about

01:37:13.860 biological aging that can then be used to understand whether an intervention is working

01:37:21.900 first of all, at the population level, but ultimately where we want to get to

01:37:26.060 is at the individual level. So what we all want is a test that you can take and you can fast,

01:37:33.840 you can do your fasting regimen, you can do your rapamycin, you can take metformin, whatever,

01:37:38.800 and you can come back and find out, is it working from this set of biomarkers? And that's where we

01:37:44.700 want to get to. And we're not there yet. I think everyone would agree. I'm not sure when we're going

01:37:49.740 to get there. So who's the natural owner of getting there? Because, you know, I had this discussion

01:37:53.860 with Steve Ostad recently and he made the same point you did, which is look, the NIA tried to do

01:37:58.740 this a long time ago and tried, you know, validly, right? They put a lot of money into it. You could

01:38:04.060 make the case that the technology simply wasn't mature enough to do this. 30 years later, we have

01:38:10.820 a lot more tools at our disposal. You've got the entire world of omics at your disposable, plus you've

01:38:17.220 got machine learning, plus, plus, plus. Is there any reason this couldn't be done today? And if so,

01:38:24.240 this strikes me as a project that's almost too big for academics because it's too disjointed.

01:38:29.520 But at the same time, it's not a particularly interesting commercial problem to solve because

01:38:35.140 it's far too big an investment before you could get to why you would care about it, right? A commercial

01:38:40.720 problem is give me a drug. But I'm arguing you can't develop a drug really well without this.

01:38:48.520 So who's, like there's a bit of a cart and a horse thing, which is someone's got to pony up a lot of

01:38:54.880 money to develop the foundation of a pyramid that will ultimately become a great tool for drug discovery

01:39:02.320 and a much more streamlined manner in which we could do clinical trials around this.

01:39:06.800 Yeah. I think the answer to your question is it depends on whether you're talking about doing

01:39:12.080 this preclinically or clinically, right? I actually think this is a problem that is-

01:39:16.860 Don't you think it has to be done both?

01:39:18.360 Well, eventually, yes. It's a problem that can be solved today preclinically. Like there is no

01:39:23.580 real barrier to doing what you just said. So multi-omic analysis of aging in mice with interventions,

01:39:31.140 applying machine learning to identify patterns that predict the effect of interventions and

01:39:38.560 individual outcomes for longevity. You obviously have to think a little bit about, you know,

01:39:43.120 what can you measure? If you want to do this longitudinally, you can't kill the animals, right?

01:39:49.640 So you're sort of restricted to blood. So there are some practical aspects, but there's no

01:39:53.660 technical barrier to doing that now. Who should do it? Who might be doing it? I mean, I think this

01:40:01.640 would fall maybe in the realm of what Calico could do. They've got the resources, they've got the

01:40:06.520 expertise. Is there any evidence that Calico is interested in this type of a problem?

01:40:11.380 I think so. I don't know. I honestly don't know anything about the inner workings of Calico these days.

01:40:16.260 I think conceptually they are interested in multi-omic signatures of different aging processes.

01:40:23.380 I don't know if they've done this particular experiment. They certainly have the resources

01:40:27.440 and expertise to do it. They're not the only ones, but they're the first ones who come to mind. And

01:40:31.720 they sort of fit this space between true academia and industry, right? Where they're kind of this

01:40:38.540 interesting beast in the middle. So I think it could be done preclinically. And you could actually

01:40:43.240 then, let's just say you have this test, right? You get to the end of day, you say, okay, these are

01:40:48.300 the most predictive, I don't know, whatever, 24 things that give you 95% confidence on remaining

01:40:57.180 life, whatever your endpoint is. Then you get that test and then you show whether it works or not in

01:41:03.140 an independent study. And if it does, I'd be pretty convinced, right? If you can show me that you create

01:41:07.560 this test and then you go do a separate experiment and you can predict when the mice are six months old,

01:41:12.580 how long they're going to live at an individual level, I'm impressed. And if you can show that

01:41:17.980 this intervention, when you treat them, makes the signature go in the way that you think it should

01:41:21.980 go and you can predict they're going to live 30% longer, I'm even more impressed. I'll believe it

01:41:26.200 at that point. So that's not easy, but I think it's doable. I think we know enough now and we've

01:41:31.000 got enough things that we could measure that you could certainly build the test and then whether

01:41:35.980 it would work in the validation step or not, I don't know, but I think you could probably get it to

01:41:41.400 work. You can't take exactly that same approach to people. And this gets back to that, you know,

01:41:47.600 the same issue that we talked about with clinical trials, right? It takes a long time to do the

01:41:51.860 validation step and know that you have actually changed somebody's biological state so that as

01:41:58.500 they get older, they are at lower risk for disease and are likely to live some X percent longer.

01:42:05.180 So you're almost obligated to have some level, you have to have some level of faith in the test at

01:42:12.720 that point, right? And I don't know, it's going to be different for everybody. And I honestly don't

01:42:16.340 know what the regulatory step has to be before you could convince regulators that you can actually

01:42:23.620 go out and tell the general public that this test works. Although I will say there are already

01:42:29.720 people doing that and the regulators aren't doing anything about it as far as I can tell.

01:42:33.760 Let's come back to your dogs. We spoke about them at length three years ago when we sat down to talk

01:42:40.380 about rapamycin. But again, let's assume a clean slate and folks might not be familiar with some of

01:42:46.860 your work specifically around dogs and how working with companion dogs offers many advantages over

01:42:54.400 working in mice, beginning with some of the obvious, like they're far more genetically similar to us.

01:42:59.600 They live in our environment and they also seem to die of things that more closely replicate how we

01:43:06.240 die. They die of heart failure. They die of cancer, but not the same type of cancer that a mouse gets

01:43:11.720 that's almost genetically predetermined. Just to start from ground zero, we at the University of

01:43:17.260 Washington and Texas A&M and at several other institutions have a large project called the

01:43:22.060 Dog Aging Project. The goal of the Dog Aging Project is really to understand the biology of aging in

01:43:30.140 companion dogs or pet dogs. Some people, when I say companion, some people think I mean, you know,

01:43:35.160 like seeing eye dogs. Seeing eye dogs, yeah, yeah. Pet dogs. And there are really two aspects of this

01:43:41.540 project. One is a large scale longitudinal study of aging, completely observational. The goal there is

01:43:47.920 really just to understand what are the most important genetic and environmental factors

01:43:53.440 that influence healthy aging in dogs. And part of that, this is getting back to the conversation we

01:43:58.560 just had, is to measure as much as we can about those dogs every year as they go through their lives

01:44:05.780 in order to be able to identify patterns that are associated with health outcomes during aging,

01:44:12.020 lifespan, disease incidents. So in some respects, it's a similar approach as to what you would do if you

01:44:17.720 wanted to create biomarkers of aging, right? So that's the longitudinal study of aging. The second

01:44:22.720 goal, so one way to think about that is that's to try to understand aging in dogs. The second goal, and this

01:44:28.280 is really where I'm, you know, focused largely, is to do something about it. So can we slow or reverse

01:44:35.500 biological aging in pet dogs to increase healthy lifespan? So that ultimately, I hope, will be a series of

01:44:43.780 veterinary clinical trials to test interventions to figure out, can we slow aging, reduce disease,

01:44:52.340 and increase lifespan in pet dogs? The first clinical trial is with rapamycin. And so that's our first

01:44:58.560 shot on goal, but I hope it won't be the only one. And so you talked about, so why dogs? Why pet dogs in

01:45:05.220 particular? And I think you hit on, I think, most of the most important reasons, right? So they've got

01:45:10.680 this really interesting and powerful genetic architecture. We have a couple hundred purebred

01:45:16.620 breeds of dogs, which you can almost think of as inbred strains, right? And then on top of that,

01:45:21.520 we have this mixed breed population. And that's coupled with phenotypic diversity. So for almost any

01:45:27.600 trait that you think about, dogs are more diverse than people are even. Body size is a really easy one.

01:45:33.600 Everybody can just think about the difference between a Great Dane and a Chihuahua. So that combination

01:45:38.200 of unique genetic architecture with phenotypic diversity is really powerful for mapping

01:45:43.980 genotype onto specific traits. And lifespan is another case where you have this strong diversity. A Great

01:45:51.560 Dane will grow old and die, you know, maybe in eight to 10 years, whereas a Chihuahua often will live to be

01:45:58.300 16, 17 years old. So, you know, we're talking a hundred percent difference in lifespan. So that's

01:46:04.840 really powerful for mapping genotype onto lifespan. Dogs share the human environment is another big

01:46:11.640 one, right? And that's, that's for me, one of the most important because we cannot model that in the

01:46:17.540 laboratory. In fact, we do exactly the opposite. We really try to limit variation in environment to

01:46:23.640 extreme measures. Dogs share our environment with the exception of diet, share almost all aspects of

01:46:29.300 the, of the human environment. And so that's a bridge in some ways between laboratory studies and

01:46:35.280 human studies. And as I've already alluded to, they age more rapidly than we do. Their lifespan is

01:46:41.440 substantially shorter. We all are familiar with the idea of one human year is about seven dog years,

01:46:46.880 right? That's just another way of saying dogs age about seven times faster than people do.

01:46:50.940 And we can talk about that. It's interesting. If you actually look at the epigenetic clock,

01:46:54.820 it's not a linear seven time rate, but, but they, but I think it's close enough, right? It's close

01:46:59.840 enough. And as you suggested, they age very similarly to the way that people do. They get

01:47:05.200 essentially all of the same age-related diseases and they're all age-related and they show the same

01:47:10.940 functional declines that people do. So dogs get arthritis as they get older, right? You know,

01:47:15.660 and arthritis is, I guess it's a disease, but it's also starts as a functional decline,

01:47:19.320 anybody who's ever had an old dog, you will notice that your old dog doesn't move around as much,

01:47:24.520 doesn't walk as fast. So they are going through the same changes with aging at the functional level

01:47:30.980 that, that we are. Again, it just happens seven to 10 times more quickly. So they're a very powerful

01:47:35.980 animal in which to understand aging and test interventions for that reason. And you can do it

01:47:42.220 in a timeframe that's, that's, you know, feasible. That's again, I think, you know, we've talked about

01:47:47.240 the challenges with doing a clinical trial for lifespan in people. Even if we believe

01:47:52.340 metformin is going to extend lifespan in people, you're not going to do a clinical trial to prove

01:47:56.940 that. You can do that clinical trial in dogs. And so we've designed the rapamycin, the test of

01:48:02.620 rapamycin in aging dogs or triad. That's, that's what we call our clinical trial. We've designed triad

01:48:08.640 so that we'll be able, we'll be powered, statistically powered to detect a 15% change

01:48:13.960 in, in lifespan within a three-year window, right? So we can do a three-year clinical trial,

01:48:19.960 reasonable cohort sizes to see an effect on lifespan that's comparable to what rapamycin does in mice.

01:48:26.020 That's a pretty high bar, Matt. Are you at all worried you've underpowered that given that you

01:48:31.000 only have, and they're being administered the drug for the entire three years?

01:48:34.940 Yes, I'm worried that the study is underpowered. I will say from my experience now, we've done two

01:48:42.820 safety clinical trials, and this is our big clinical trial. So this is my third

01:48:46.500 veterinary clinical trial. I've learned that there are many reasons to be concerned

01:48:51.680 when you do a clinical trial. Clinical trials are a lot of guesswork. You have to take an educated

01:48:56.600 guess about a lot of different things. You can't test every dose. You can't test every duration.

01:49:01.720 You're, you know, you can't test an infinite number of study subjects. So I'm, I'm worried

01:49:08.420 about this clinical trial for many reasons. I think we've done the best that we can, given what we know

01:49:14.180 and given the constraints that we have to work under to give ourselves a reasonable shot. 15% might be a

01:49:21.160 high bar, but it's consistent with the lower end of what people see in mice. So the first study at the

01:49:27.580 low, low dose of rapamycin in mice from the ITP had a 14% effect, I think in females and a 9% effect

01:49:34.060 in males. Subsequent studies at higher doses had larger effects. So it's reasonable. It's a reasonable

01:49:40.380 place to start. Do you expect to see a sex difference in dogs? I don't know. I don't know. So people still

01:49:47.240 don't completely understand why female mice seems to respond to a given dose of rapamycin better than 0.68

01:49:55.240 male mice. So it's correlated with blood levels. So I think the simple idea is that, that male mice

01:50:00.840 either take up rapamycin less effectively or break it down more quickly. There's no evidence for that

01:50:06.520 in dogs. We will be measuring rapamycin levels in the dogs. So we will see if there is a difference

01:50:11.860 in blood concentration in females versus males from the limited data that we've got so far. We don't have

01:50:18.200 any evidence that that's the case in dogs. And I don't know of any evidence in people that that's the

01:50:23.080 case either. So that might be a mouse specific thing. I think there's this misperception that

01:50:28.180 rapamycin works better in female mice than in male mice. That's not true. At a given dose,

01:50:34.960 especially the lower doses. At a given plasma dose. We don't know if it's true, right?

01:50:39.200 Given dose in the food, at lower doses, females show a bigger lifespan extension. When you go to higher 0.61

01:50:45.520 doses, the males catch up. And if you push it too far, you can actually find a dose where female mice, 0.97

01:50:51.980 the lifespan extension starts to go back the other direction and male mice actually get a bigger

01:50:55.580 lifespan extension. So I think it's more about effective concentration than it is a male versus

01:51:01.400 female, truly sex specific response. In the dogs, you give it daily in their food? No, no. So it's once 0.84

01:51:08.400 a week. So this is a now, now, you know, guesswork, right? What's the best way to do it? Now we're back

01:51:13.420 into the alchemy. So we've tested three times a week and once a week, and we decided to go with

01:51:19.240 once a week for triad and, you know, it's a guess. And how many mg per kg are they getting? Is it one

01:51:25.620 dose? So it's 0.15 mg per kg once a week. And that was based on our observations from 0.05 mg per kg

01:51:35.120 three times a week. That's how we get, you know, 0.05 times three. So lifespan is our primary endpoint,

01:51:41.740 which is important because I think if, I think this is the first clinical trial that has lifespan

01:51:47.980 in a healthy or normal health status population as the endpoint. I will say also, it's funny because

01:51:54.680 I get pushed back from clinical people. You can't call this a clinical trial. It's just in dogs. It's

01:51:59.980 not in people. And I just simply respond that a veterinary clinic is a clinic. It's a veterinary

01:52:05.740 clinical trial, but it is a clinical trial. And I think this is the first clinical trial with

01:52:10.000 lifespan as the endpoint. I don't know if you can hear my dog. You want me to stop?

01:52:13.240 I can't. I can. No, no, it's funny. It's totally appropriate that as we're talking about dogs,

01:52:18.020 we can hear your dog. The environment. So the point I want to make though, is that lifespan is

01:52:21.900 our primary endpoint, but we are tracking multiple secondary endpoints to give us a picture of is

01:52:28.460 rapamycin broadly impacting the aging process. So we're looking every six months, the dogs get

01:52:34.020 echocardiograms to look at heart function. The dogs will be fitted with activity monitors

01:52:39.560 periodically to look at spontaneous activity. Every six months, the dogs will get cognitive

01:52:44.380 assessments to look at cognitive function. We're getting blood chemistry, serum metabolome,

01:52:49.840 fecal microbiome. And of course, we'll be tracking disease incidents as these dogs get older. So,

01:52:55.420 you know, over this three year window, I hope that even if we don't see that magnitude of lifespan

01:53:00.860 extension and we don't reach statistical significance, if there is a broad effect on multiple age

01:53:07.660 related outcomes, that we will be able to detect changes in other secondary endpoints.

01:53:14.380 What's the sample size? You're going to have two groups, placebo versus dose?

01:53:17.780 Right. Half placebo, half treatment. The intention is to randomize 350 dogs equally split between the

01:53:24.100 two groups. So 175 of each.

01:53:26.140 And do you know roughly if you had gone with a 10% effect size, how many that would have required?

01:53:31.940 How many more? I don't recall off the top of my head. I think probably around 500.

01:53:36.640 Yeah. In your previous work, you had already got a sense of what was happening in animals with heart

01:53:44.240 failure. Remind people a little bit about that if they don't remember the first episode.

01:53:49.220 Sure. So there's really good data. Three, at least three, maybe four now studies from different labs in

01:53:56.360 mice showing that if you look at age related declines in heart function, particularly left

01:54:02.460 ventricular function. So the several measures of left ventricular function decline with age,

01:54:07.840 that eight to 10 weeks of rapamycin is enough to reverse those changes and make the young heart by

01:54:13.260 echocardiography look functionally like a, make the old heart look like a young heart. There's also data

01:54:20.220 in mouse models of a few different types of heart failure. So dilated cardiomyopathy is work that we've

01:54:26.000 done in particular, where you can reverse dilated cardiomyopathy in mice with rapamycin treatment.

01:54:33.780 So I think there's really strong evidence in mice. In our first clinical trial, which was,

01:54:39.480 it was designed to really only be a safety trial. There were 24 dogs, 16 got rapamycin,

01:54:45.420 eight got the placebo, but we had the dogs also get echocardiograms before and after the treatment period.

01:54:51.420 And in that study, there were statistically significant improvements in two measures,

01:54:57.420 two of the three measures of left ventricular function by echocardiography in the dogs that got

01:55:02.780 rapamycin compared to the placebo. What was, I think to me, most interesting in that data was that

01:55:09.020 the improvements that we saw were exclusively found in the dogs that came in with lower function.

01:55:16.200 Now, one thing to note is none of these dogs had function so low that it would be clinically

01:55:21.660 diagnosed as heart failure, right? So it was normal age-related declines in function. And it was

01:55:28.160 exclusively the dogs with the lowest function that showed the improvements from rapamycin.

01:55:32.960 That is an interesting observation, which my gut feeling is real. Again, a small cohort. So I don't,

01:55:39.480 you know, want to, want to make it out to be stronger than it is, but it also is intuitive,

01:55:44.180 right? We know that individuals have individual trajectories of aging and develop a unique

01:55:51.620 spectrum of functional declines. And it makes sense that some of these interventions that are

01:55:56.980 restoring function would primarily be effective in people who've lost function. That's just intuitive.

01:56:03.360 So I think that's probably what we'll, we'll see, you know, in our long-term study in triad,

01:56:09.680 where we follow the dogs for three years is that there's going to be variation at baseline. And it

01:56:15.120 might be the case that, that those individuals that have the lowest function at baseline are the

01:56:19.600 ones that are going to see the biggest benefit. If there is a benefit at all.

01:56:23.780 Now in triad, will you be trying to create somewhat of a homogeneous sample in size? By that,

01:56:31.400 I mean the size of the dogs, or are you going to be completely heterogeneous with respect to

01:56:36.260 the size of the animal? And also what is the age of the animal? What are the exclusion criteria is

01:56:42.360 around the age on either end, low or high? Right. So age, they have to be at least seven

01:56:47.760 years old to come into the study. So that'd be middle-aged. And we do have a size range. So the

01:56:53.480 dogs have to be between 40 and a hundred pounds to be randomized into the study. So it's not for little

01:56:59.140 dogs. That's right. And the reason for that is not because we have any reason to think rapamycin

01:57:04.340 will work differently or better or worse than little dogs. It's because big dogs age faster

01:57:08.940 than small dogs. Let me come back to that in a minute because that might not be, that might be

01:57:13.900 an oversimplification, but they certainly live shorter than small dogs and develop many age-related

01:57:20.160 diseases and functional declines at an accelerated rate compared to small dogs. So we need a population

01:57:26.160 that's already middle-aged and that will be aging rapidly to be able to see potential benefits

01:57:32.360 from rapamycin in the timeframe of this study. And that was all factored into the power calculations,

01:57:36.740 the demography of dogs in that age and weight range. I said that it may not be completely

01:57:42.240 accurate to say that big dogs age more rapidly than small dogs because there's a growing body

01:57:47.440 of evidence, which we have some preliminary data in support of as well, that for brain aging and

01:57:53.540 cognitive function, that might not be the case. That the rate of cognitive decline in big dogs

01:58:00.300 looks, from a chronological sense, very, very similar to small dogs, even though the big dogs

01:58:07.720 are dying at an earlier age. They don't seem to show accelerated cognitive decline, which is

01:58:13.560 interesting. And I think there's a little bit of data in people, although obviously people don't show

01:58:18.500 the same diversity in body size that dogs do, so that's harder to see. It doesn't really seem to be

01:58:24.400 the case that accelerated aging due to increased body size is reflected in brain aging. And the

01:58:31.380 mechanisms there might turn out to be really interesting. So just an observation that I've

01:58:35.560 noticed recently and I think might be important. You're using once-a-week dosing. Explain to people,

01:58:42.720 because we just mentioned it briefly and then said we'd come back to it now. We're going to sort of

01:58:47.380 bifurcate mTOR into mTOR complex, one mTOR complex to give people a sense of how those function and why

01:58:55.460 it, on the one hand, is leading you to do what you're doing, me to do what I'm doing, and why in

01:59:02.240 some ways it makes it a little bit interesting why the ITP found what it did with daily dosing.

01:59:07.780 Let's tie all of that together, but first with an explanation of how mTOR works.

01:59:11.740 Sure. So mTOR, of course, is a protein. It's a kinase. We already talked about that.

01:59:17.060 But it acts in a complex with other partner proteins. And so the mTOR protein acts in at

01:59:25.020 least two, I think there's only two that we know of, two different complexes called mTOR complex one

01:59:30.860 or mTORC one and mTOR complex two or mTORC two. And the difference between those complexes is that

01:59:38.660 they have different partner proteins for mTOR. So there are a couple of things that are the same

01:59:43.440 across both complexes. And then there are a set of partner proteins that are unique.

01:59:48.460 And the two complexes do functionally different things in the cell. So mTOR complex one is largely

01:59:56.740 thought of as the mTOR complex that is most responsive to nutrient levels. So when nutrient signals

02:00:03.240 are low, that leads to lower mTOR complex one activity and mTOR complex one downstream is known

02:00:10.660 to regulate things like autophagy, mRNA translation, effects on metabolism. And mTOR complex two does

02:00:18.960 different things. And I think we know a lot about what mTOR complex one does. We know much less about

02:00:24.320 what mTOR complex two does, although people are studying that and learning more and more about what

02:00:30.120 mTOR complex two does. From the perspective of aging biology, people have focused almost exclusively

02:00:36.880 on mTOR complex one. There's a little bit of data in C. elegans on mTOR complex two affecting lifespan.

02:00:43.240 But outside of C. elegans, almost all of this, the data for rapamycin or mTOR as a regulator of aging

02:00:50.880 is thought to be mediated by inhibition or reduced signaling through mTOR complex one. So that's what people

02:00:58.040 almost always think about when they think about effects of mTOR on aging. And rapamycin as a drug

02:01:04.760 biochemically is a specific inhibitor of mTOR complex one. So the way rapamycin works is it actually,

02:01:11.960 it's a small molecule that binds to another protein called FKBP12 or FPR1 in yeast. And once rapamycin binds

02:01:20.520 to FKBP12, that complex of rapamycin with FKBP12 goes to mTOR complex one. And you could think of it

02:01:28.600 as sort of just messing it up, breaks it apart. So it inhibits mTOR complex one when rapamycin is

02:01:34.120 bind to FKBP12. So biochemically, rapamycin is an extremely clean drug. And that as far as I know,

02:01:40.680 and I haven't really seen any good data otherwise, there's no direct inhibitory effect of rapamycin on

02:01:46.520 anything other than mTOR complex one. What people have observed is that chronic long-term inhibition

02:01:55.160 of mTOR complex one can have feedback effects on mTOR complex two. And it's kind of confusing

02:02:02.120 because there's actually some data both directions that chronic inhibition of mTOR complex one can lead

02:02:08.360 to activation of mTOR complex two. In some context, mostly I think the data supports the idea that

02:02:13.720 chronic inhibition of mTOR complex one with rapamycin can lead to inhibition of mTOR complex two in the

02:02:20.040 long term. And that's definitely true in mice at higher doses of rapamycin. At lower doses of

02:02:25.800 rapamycin, it's not completely clear to me how much effect on mTOR complex two there is. So the reason

02:02:33.960 why that's interesting from the perspective of aging is I just said that almost all of the data for

02:02:38.840 lifespan at least is that it's inhibition of mTOR complex one that leads to lifespan extension.

02:02:44.760 Work from David Sabatini and Dudley Lamming when he was in David's lab led to the development of a

02:02:50.520 model, which I think is still the sort of preferred model. I will say upfront, I think it's at least

02:02:57.160 partly wrong, but it's the preferred model, which is that the side effects associated with rapamycin,

02:03:02.120 particularly the metabolic side effects associated with rapamycin are due to this chronic effect of

02:03:09.240 inhibiting mTOR complex two. So people will talk about rapamycin as if it induces something like

02:03:15.320 diabetes, a pseudo diabetes. And this is true in mice. There's evidence for it in humans as well, that

02:03:20.920 chronic long-term treatment with rapamycin leads to glucose intolerance. So if you give a mouse who's been

02:03:26.920 on rapamycin for a year, a glucose tolerance test, they will not clear that glucose as rapidly as a

02:03:34.520 mouse that never saw rapamycin. And the model is that that's due to the chronic effects of rapamycin

02:03:40.680 on mTOR complex two. Most of the evidence in support of that model comes from genetic experiments with

02:03:46.360 mTOR complex two deficient mice. So I have yet to see a really clean experiment showing that that's what

02:03:52.600 accounts for the rapamycin effects on glucose tolerance. I think it's a reasonable model.

02:03:57.480 In other words, that evidence comes from creating genetic mice where you manipulate mTORC1 and mTORC2

02:04:04.280 rather than experiments where you give the mice rapamycin. You know, I didn't ask Rich Miller this

02:04:09.160 question because I'm pretty sure the answer is they didn't do it. But in all of the ITPs where the

02:04:14.920 animals are getting rapamycin every single day, did they see impaired glucose tolerance despite longer life?

02:04:21.960 Yeah. My recollection is they did not look, but there have been other studies,

02:04:25.960 I think mostly in the C57 black 6J mouse strain. So that's a different strain than the ITP.

02:04:31.800 We don't need to get into it, but it's a different genetic background. But in the C57 background,

02:04:36.760 you do see the same changes in glucose tolerance test at the, I think at the 42 part per million for

02:04:43.640 sure rapamycin dose. So the higher ITP dose in older mice. So I think it's a real effect. I think you

02:04:49.960 see that effect with rapamycin. And like I said, there's evidence in organ transplant patients

02:04:53.880 for impaired glucose homeostasis as well. So I think it's a real effect. There's a couple of

02:04:58.360 things that I, and I actually, my intuition is it probably is due to mTORC2. I don't have any reason

02:05:03.560 to doubt that. I just don't think it's been shown cleanly that that's the mechanism. Where I differ

02:05:09.480 a little bit, certainly from Dudley, I'm not sure what David's view on this is right now. Where I differ

02:05:15.720 from Dudley's interpretation is that I am less convinced that these effects that we see for

02:05:21.000 glucose homeostasis are bad. I think there's at least as much likelihood that what this really

02:05:29.160 reflects is an underlying change in metabolism that might actually account for part of the beneficial

02:05:35.320 effects of rapamycin, where they shift away from primarily relying on glucose as the preferred carbon

02:05:42.200 source and switch over to fat metabolism and maybe even ketogenesis to some extent in that context.

02:05:48.840 So it's a different physiological metabolic state. In that context, when you challenge them with a

02:05:54.920 non-physiological amount of glucose, they don't respond the same way. So I don't know that it's

02:06:00.600 actually a defect in glucose homeostasis. I think it may reflect the test that's done. And in some ways,

02:06:08.760 it's an artifact of that test that you get a different response, which is in the context of,

02:06:13.880 you know, diabetes would be interpreted as a bad response. It might just reflect a different

02:06:18.920 underlying physiological state. And I haven't seen anybody really try to address that. And what makes

02:06:26.200 me believe that might be the case is we and others have seen that rapamycin treatment has pretty profound

02:06:32.920 effects on fat mobilization, fat metabolism, adipogenesis, and at higher doses, ketogenesis.

02:06:40.840 So it would not surprise me at all if that metabolic adaptation accounts for some of the beneficial

02:06:48.040 effects of rapamycin and is also leading to this apparent aberrant response to a glucose tolerance test.

02:06:55.160 Yeah, it's interesting. I've spoken with one other physician who uses rapamycin, although he uses it

02:07:00.600 very liberally in many of his patients. I do not. I think his patients are coming in much older and

02:07:05.720 much more metabolically deranged. And the results that he's seeing in an uncontrolled manner, meaning

02:07:12.920 you simply have no idea what the performance effect of rapamycin is. So when you give it to somebody who's

02:07:18.440 expecting to get better, they may go and change many other behaviors. But, you know, he's shared with me

02:07:24.600 some of his data and it's quite profound, right? So, you know, triglycerides falling from unhealthy

02:07:30.200 levels of 200 milligrams per deciliter to 70 milligrams per deciliter. And actually they're

02:07:36.360 seeing the opposite, right? They're seeing improved glucose homeostasis. Again, these are people starting

02:07:41.800 who are pre-diabetic and in some cases diabetic. And I think your point's a fair one about oral glucose

02:07:47.800 tolerance tests are very unnatural. And you can see a physiologic insulin resistance when you do them

02:07:54.360 on people who are either calorie restricted or carbohydrate restricted because that initial form

02:08:00.200 of muscle insulin resistance is actually a protective element there. So all of that said, you've decided

02:08:07.720 to go with once a week rather than daily, despite all of the other animal models that have shown great

02:08:14.120 success with daily dosing. Yeah. Is that a hedge? I don't know if I would define it as a hedge. I mean,

02:08:22.520 I think that, um, so the rationale there is both based on the human data, which again is limited,

02:08:28.760 right? But we've talked about the daily versus weekly dosing for immune function. And again,

02:08:34.760 we've also talked about why I think that readouts of immune function are probably telling us about the

02:08:40.440 underlying inflammatory state, which I think is a big part of what rapamycin is doing.

02:08:45.560 So that is suggestive that weekly dosing is at least as good as daily dosing. And it's also

02:08:51.480 suggestive that weekly dosing has fewer side effects. And because this is a, this is a challenge with human

02:08:57.560 clinical trials, as we've already talked about, it's also something you have to be, you know, absolutely

02:09:02.360 aware of when you're talking about doing a clinical trial in, in companion dogs, right? These are

02:09:07.080 people's pets. There is an extremely, rightly so, an extremely low tolerance for significant

02:09:12.840 side effects when we're talking about people's pets, right? I love my dog and, and I would be,

02:09:18.600 you know, devastated if I hurt anybody else's dog in this clinical trial. So you want to do everything

02:09:23.560 you can to reduce the likelihood of side effects. And there's reason to believe that once weekly dosing

02:09:29.880 is likely to have fewer side effects. And there's also the pragmatic aspect of we're asking owners

02:09:36.280 to give this medication to their dog in the context of a clinical trial. We expect that

02:09:42.920 there will be better compliance and fewer mistakes with once a week dosing versus daily dosing or three

02:09:50.280 times a week. So let's pivot a bit from RAPA to TORIN2. You recently, you and I have shared a couple

02:09:57.640 of emails on this topic. Tell people a little bit about what that is and why you're excited.

02:10:01.720 So TORIN2 is a different version of an mTOR inhibitor. So we just talked about mTOR complex

02:10:08.200 one and mTOR complex two and how rapamycin, at least biochemically, is a specific inhibitor of

02:10:14.120 mTOR complex one. TORIN2 is a, what's called a catalytic inhibitor or an ATP competitive inhibitor

02:10:20.360 of mTOR, which at least in theory... Versus what we call allosteric. Allosteric inhibitor,

02:10:25.720 which is rapamycin, meaning it interacts not through the catalytic site. Yes. At least in theory,

02:10:31.000 TORIN2 and TORIN1 and other catalytic inhibitors will equally inhibit both mTOR complex one

02:10:38.200 and mTOR complex two. And for reasons that I still am not sure about, as far as I know, nobody has

02:10:46.760 tested TORIN2, TORIN1, other catalytic inhibitors for effects on aging in mice.

02:10:54.680 Why has this not... Yeah, this seems like something that needs to be submitted to ITP tomorrow,

02:10:59.400 right? Or at least for the next cycle. Yeah. I don't know. I don't know why it

02:11:03.960 hasn't been done, to be honest with you. And maybe it has, and I just haven't seen the data. That

02:11:08.040 certainly is possible. But I don't think... I don't know of any data, not just for lifespan,

02:11:12.120 but for other functional measures of aging. I don't know of any data. I can tell you why I think

02:11:17.160 there's additional reason to test it beyond sort of the rationale that seems obvious. But I don't

02:11:23.160 know that I would say I'm excited about TORIN2. It seems like an obvious gap in knowledge and would

02:11:29.080 be actually a nice way to test the question we were talking about previously of TORIN1 versus TORIN2,

02:11:35.000 good, bad, all of that stuff. It seems like an important set of experiments to do. And as far as I

02:11:39.960 know, nobody's done it. So part of the reason, additional reason why I think it's an interesting

02:11:45.160 set of experiments to do is in addition to studying aging, my lab also works on mitochondrial dysfunction

02:11:51.160 and mitochondrial disease. And we have worked for many years in a mouse model of a childhood

02:11:57.560 mitochondrial disease called Lee syndrome. So this mouse is defective in complex one of the electron

02:12:03.960 transport chain of the mitochondria. And it is very short lived. It lives about 55,

02:12:08.920 60 days, and it develops many of the same molecular phenotypes and neurological phenotypes as kids who

02:12:17.480 have Lee syndrome, this childhood mitochondrial disease. And these children typically live how long?

02:12:23.080 It's variable onset, but it's anywhere from infants to eight, nine years old. Typically,

02:12:29.320 kids with Lee syndrome don't make it to be teenagers. It's a horrible, horrible disorder.

02:12:35.320 So we found out many years ago that rapamycin could roughly double or triple the survival of these

02:12:40.280 mice and basically prevent the neurodegeneration and brain lesions that are thought to limit

02:12:46.280 lifespan in both the mice and the kids with Lee syndrome.

02:12:49.400 And presumably, Matt, that's because when you knock out complex one, you're destroying oxidative

02:12:56.440 phosphorylation. And presumably, the neurons are going to be the most sensitive to that.

02:13:02.920 Yeah, it's an interesting question. We don't really know. I mean, I'll tell you what I'll tell

02:13:06.520 you the sort of current thinking in the field. So first thing is to recognize is this particular

02:13:12.200 component of complex one is an accessory stabilizing factor. So the mice are not 100%

02:13:18.360 deficient in complex one, they have a low level of complex one, I think you'd probably be dead if you

02:13:23.080 didn't have any complex one. And the same thing is true in the patients. So it's a deficiency in

02:13:27.720 complex one. And you're right. So one idea, there are actually several reasons why being deficient in

02:13:34.440 complex one could be a problem. One could be you just can't generate enough ATP and neurons or a

02:13:40.840 subset of neurons are particularly sensitive, right? Another could be that you're generating high levels

02:13:45.480 of reactive oxygen species. We know subsets of neurons are especially sensitive to that. And you could

02:13:51.320 come up. There's also an inflammatory component to this disease where we see a lot of neuroinflammation

02:13:56.200 in the brain regions where the lesions occur. And we don't quite understand what's causing that

02:14:01.880 inflammation. So there are multiple ways that this could be causing the symptoms of the disease.

02:14:08.200 An interesting body of work from Vam C. Muthu's lab and Isha Jain, who was in his lab and now has her own

02:14:14.280 lab, shows that hypoxia can also rescue these mice. And hypoxia actually rescues these mice to a greater

02:14:20.280 extent than rapamycin. So that at least is consistent with an oxidative stress model,

02:14:26.840 right? So one idea would be that when you're deficient in complex one, those neurons are not

02:14:32.200 using as much oxygen because they've switched over to non-oxidative glycolytic metabolism and

02:14:37.560 fermentation. So you have higher oxygen levels in the brain that leads to higher oxidative damage.

02:14:42.680 When you reduce oxygen through hypoxia, you've suppressed that. I don't have any reason to doubt that model.

02:14:48.600 By the way, have you ever tried the experiment where you give them excessive amounts of lactate?

02:14:54.280 Which neurons like lactate? And if you gave them lactate in theory, you would also come up with a

02:14:59.320 way to bypass the ETC. I mean, it'd be an interesting control or way to test that hypothesis.

02:15:06.040 We've never tried that. So what I can tell you is they have higher levels of lactate in their blood.

02:15:10.440 That's common in mitochondrial disease because they switch over to fermentation. But that doesn't mean

02:15:16.280 that your experiment wouldn't work. It's just an observation. And rapamycin suppresses that high

02:15:20.680 level of lactate, interestingly, as well as the accumulation of all the glycolytic

02:15:25.000 intermediates upstream of lactate. So anyways, oxidative stress, there's at least some reason

02:15:29.880 to believe that oxidative stress is important. Another phenotype that these mice get is a dramatic

02:15:35.960 loss of fat. So they are extremely lean and rapamycin suppresses that. So we think that there is

02:15:41.960 some, we know there's something going on with fat mobilization, adipogenesis, when we treat with

02:15:48.040 rapamycin that spares those mice from presumably using up all their fat. Presumably they're trying

02:15:54.680 to burn the fat in some way, and that's why they become so lean. So that's another observation.

02:16:00.040 So how do we get to mTORC2 and TORIN2? So we did a phosphoproteomic study with Judith Villain,

02:16:06.040 who's at the University of Washington in Genome Sciences, to try to just take a sort of a global

02:16:11.480 picture and ask, what is the effect of rapamycin on the proteome and the phosphoproteome in the

02:16:18.440 knockout mice compared to wild type and then with rapamycin treatment? There was a lot of interesting

02:16:23.240 stuff in there. This work is published, it was published in Nature Metabolism, I don't know,

02:16:27.720 several months ago. So if people are interested, they can look at that paper. There's a lot of

02:16:31.960 interesting stuff in that proteomic dataset, but one of the striking things that we saw

02:16:36.840 was that mTORC2 components were, at the protein level, decreased, which actually fits with the

02:16:45.000 idea that high-dose rapamycin chronically leads to inhibition of mTORC2. And associated with that was

02:16:51.640 an inhibition of protein kinase C, which is another kinase that's regulated by mTORC complex 2. So that

02:16:58.520 led us to hypothesize, and protein kinase C is known to regulate some aspects of inflammation.

02:17:04.280 So that led us to hypothesize that maybe some of the effects of rapamycin, at least in this mouse

02:17:09.160 model, are through protein kinase C and mTORC2, as opposed to what we had assumed was all mTORC1

02:17:17.400 related effects of rapamycin. So we tested a few drugs that are known inhibitors of protein kinase C,

02:17:25.320 and they rescued part of the lifespan of the mitochondrial disease mice. So they weren't as

02:17:30.520 good as rapamycin, but they did give significant increases in survival and delayed some of the

02:17:35.960 neurological symptoms that the mice experienced. So that is consistent with the idea that this

02:17:41.240 inhibition of mTORC2, inhibition of protein kinase C is part of the effect of rapamycin. So now we've

02:17:48.280 gone back and we've tested TORN2 in this mitochondrial disease model, seems to work as well as rapamycin,

02:17:56.200 which is interesting. No negative side effects that we can see from inhibiting TORC2 in that context.

02:18:03.080 And then we're also working in another mouse model of a severe metabolic disease. And I think I'm not

02:18:10.040 going to talk about it because this is a collaboration and it's not published yet. And so I don't want to give

02:18:14.200 away what the initial observation was really the observation of our collaborators. But let me just say

02:18:19.960 in another childhood metabolic disease, which is on the surface, completely unrelated to complex one of

02:18:28.440 the mitochondria, we see dramatic rescue from rapamycin and TORN2. So this makes me wonder, first of all,

02:18:37.400 I've thought for a long time that mitochondrial disease might be a good model for normal aging

02:18:43.960 in some respects. We know that, I mean, we talked about mitochondrial dysfunction as one of the

02:18:48.040 hallmarks of aging. So severe mitochondrial dysfunction, I would not argue as accelerated

02:18:52.680 aging, but I think interventions that are effective in a model of severe mitochondrial disease might also

02:18:58.760 be effective in the context of normative aging. Everything we've seen so far is consistent with that.

02:19:04.360 And it makes me wonder if TORN2 or other catalytic inhibitors of mTOR might be as effective as rapamycin

02:19:13.000 in the context of aging, and maybe even more effective in the context of some age-related

02:19:18.200 indications. So to me, it just seems like a gaping hole in the literature that really needs to be

02:19:23.720 explored and we'll learn from doing that sort of exploration. The other point that's interesting to

02:19:29.560 make is RTB-101, the drug we talked about previously in the RestorBio phase three clinical trial,

02:19:36.200 is a catalytic inhibitor of mTOR as well as other kinases. So it would fall at least biochemically

02:19:44.040 into the TORN2 class as opposed to the rapamycin class. How is TORN2 discovered or synthesized?

02:19:50.440 Was it synthesized directly to be a catalytic inhibitor? Was it discovered and modified?

02:19:55.560 I, you know, David Sabatini would be the person to ask. I should know the answer to that. My

02:19:59.880 recollection, so rapamycin is interesting, right? It was bound on rapanui, made by bacteria. I think

02:20:05.800 TORN2 came out of a more traditional sort of chemical screening as a mTOR inhibitor and then

02:20:12.360 was modified to be a more potent catalytic mTOR inhibitor. But I honestly don't know that literature,

02:20:17.800 recall it. And does it already have an IND? Is there already a molecule that's in a pipeline,

02:20:23.800 than an FDA pipeline? That's a good question. I have looked

02:20:26.600 a little bit and I don't know of anybody who's developing TORN2 or TORN1, which is another

02:20:32.040 catalytic inhibitor for FDA approval. And I've wondered about this. And honestly,

02:20:36.760 I just haven't had the bandwidth to really dig into it. Is it because people tried and there

02:20:40.840 were side effects? I don't know. But I don't know of anybody who's actively developing the TORNs,

02:20:45.800 at least, for FDA approval. RTB-101, again, we talked about that's being developed or has

02:20:52.440 attempted to be developed for FDA approval. There are other dual kinase inhibitors that are used

02:20:57.720 clinically that hit mTOR in addition to other, like a PI3 kinase, a different class of kinase.

02:21:03.480 So there are molecules that have similar types of activity, but I don't know of anybody who's really

02:21:09.640 trying hard to develop the more specific mTOR catalytic inhibitors. And again, I'm not sure why.

02:21:15.400 Let's change gears for a second and talk about NAD, NR, NMN, all these things. I've had David Sinclair

02:21:23.640 on the podcast a couple of times. He's very eloquently explained what sirtuins are, how they work,

02:21:28.280 why they require NAD. So for folks who want to get up to speed on that, you can do so in great depth.

02:21:34.200 Do you want to give the 30-second answer as to why sirtuins matter and why they need NAD?

02:21:39.560 Sure. But I would start by saying that I'm not sure that sirtuins are the only or most important 0.99

02:21:46.760 reason why NAD is important. Sirtuins are a class of NAD-dependent, mostly deacetylases. They also can

02:21:53.800 do some other activities. But basically, one way to think about it is that sirtuins take acetyl groups

02:22:00.920 off of other proteins. That's their activity. And that requires NAD. And it actually consumes NAD.

02:22:07.160 So NAD is a cofactor for many metabolic reactions where it gets converted between NAD, which is the

02:22:12.840 oxidized form of the coenzyme, and NADH, which is the reduced form. Many, many different metabolic

02:22:19.720 reactions use NAD in that way. Including the electron transport chain. Including the electron

02:22:24.280 transport chain and glycolysis and fermentation, yes. Sirtuins are fundamentally different in that they

02:22:30.440 use up NAD, right? And so NAD is required for their activity. NADH, the reduced form of NAD,

02:22:38.120 is actually an inhibitor of sirtuins. So the NAD to NADH ratio can be used as a proxy of likely sirtuin

02:22:45.800 activity. So sirtuins are important in the aging field. You know, this really goes back to my work

02:22:51.880 as a graduate student in yeast when I was a grad student with Lenny Guarenti. When we first showed

02:22:57.160 that you could overexpress the yeast sirtuin, which is called sirtu, that's where sirtuin comes from.

02:23:03.640 So the yeast protein is called sirtu. If you overexpress sirtu, you increase lifespan in yeast.

02:23:08.840 And since then, other people have shown that activation or overexpression of sirtuins in worms

02:23:14.760 or flies or mice can have interesting effects on aging.

02:23:19.080 You and David must have overlapped, right? You guys must have both been in Lenny's lab at the

02:23:23.480 same time. David and I overlapped in Lenny's lab, yeah. So he was a postdoc when I was a grad

02:23:26.520 student. And I got to give David, I mean, David and I have had our scientific disagreements over

02:23:30.200 the years. But I got to give David a ton of credit. As a postdoc, he mentored me in important ways.

02:23:37.080 And I think actually guided me to the project looking at sirtu, which is what I just talked

02:23:41.640 about when we overexpressed sirtu. So David was a very important early influence on my scientific

02:23:47.080 career. And the Guarenti lab at that time, you know, was full of really smart, I'm sure it's still

02:23:52.920 full of really smart people, but it was just a really great environment. It was a powerhouse.

02:23:56.200 Yeah. With lots of really fantastic scientists. So that was the first, and it was really Lenny's

02:24:02.680 lab that established sirtuins as important in aging in multiple model systems. So in mice,

02:24:09.640 David might disagree with us a little bit, but I think if you're being honest, right,

02:24:12.920 the evidence that sirtuins are potent regulators of lifespan in mice is mixed. It's not strong. There

02:24:20.760 are a couple of studies out of probably, you know, a dozen that have been published,

02:24:25.720 and there's probably two dozen that are unpublished where people saw no effects on

02:24:28.920 lifespan from manipulator or activating sirtuins. There are a couple of studies that show in one

02:24:34.360 case of a brain specific activation of one of the sirtuins called SIRT1 could slightly extend

02:24:39.880 lifespan. And another that, that overexpression of a different sirtuin, SIRT6 could slightly extend

02:24:45.640 lifespan, I think only in males, but nowhere near the reproducibility or magnitude of effect of other

02:24:51.800 things, including rapamycin. So the data on sirtuins is broad, but the absolute effects on

02:24:58.440 lifespan at least are, in my personal view, unconvincing. Like it hasn't been broadly

02:25:03.800 replicated like rapamycin has, and they're not big. What we do see with sirtuins is abundant evidence

02:25:10.200 that metabolic markers of health can be improved by activating sirtuins. And in a few other disease

02:25:16.680 specific models, pretty good evidence, heart disease in particular, some evidence for cognitive function

02:25:21.320 as well, improvements in age-related outcomes. So there's a lot of smoke there, but I think there's

02:25:26.520 a lot of confusion in the field about the relative strength of data for different interventions. And at

02:25:32.200 least in my view, there's really no comparison between the effects you get from inhibiting mTOR

02:25:37.800 and the effects at least so far that people have reported from activating sirtuins. It's like mTOR is

02:25:44.040 head and shoulders above sirtuins when it comes to magnitude of effect. I think it would be impossible

02:25:49.560 to dispute that. I don't think there could be any dispute of that. What's interesting is if I were to

02:25:55.080 tally up the number of questions I get per month about NR and NMN versus rapologues, the ratio is

02:26:05.800 the exact opposite to the magnitude. So if the effect size of rapamycin and the importance of mTOR is

02:26:11.240 10x that of sirtuins, it's flipped in the number of questions I get about it and just the pop culture

02:26:20.520 awareness of that. So let's put the marketing of that aside and talk about the chemistry of it for a

02:26:27.160 moment. Right. So sirtuins are NAD-dependent enzymes, right? So they need NAD to do their action.

02:26:33.960 And we've already talked about, in general, the model is that turning up sirtuins is a good thing, 1.00

02:26:39.560 right? That you're going to get, if you're going to get benefits in the context of aging,

02:26:43.400 that's going to happen from activating sirtuins. Again, that's probably a pretty massive

02:26:48.760 oversimplification because there are seven sirtuins, right? And they do different things

02:26:52.760 and different things in different tissues. But that's kind of where the field has gotten stuck.

02:26:56.360 The idea is that activating sirtuins is good. And so if you accept that and NAD is an activator of

02:27:03.960 sirtuins, then more NAD is good. And there's good evidence that NAD homeostasis becomes impaired with

02:27:10.920 aging. That the ratio of NAD to NADH, the oxidized to reduced form of NAD, in many tissues at least,

02:27:19.240 shifts towards more NADH and less NAD, right? And so that with age. And so the prediction would be

02:27:26.040 that you would have declining sirtuin activity due to that metabolic shift. And I will also note,

02:27:32.840 because I think this is important, that in mitochondrial disease, you see the same shift.

02:27:37.480 It's just much, much more pronounced. You see a shift towards NADH and less NAD. And that's exactly

02:27:45.480 what you see when mitochondria are less functional because the cells will switch over to glycolysis and

02:27:51.080 fermentation, right? Fermentation to lactate. And the whole reason why we ferment to lactate

02:27:56.600 is to restore NAD levels. Fermentation to lactate takes NADH and turns it back into NAD.

02:28:02.840 So this probably reflects an underlying metabolic defect, which could be mitochondrial in origin,

02:28:08.920 that leads to this shift towards the reduced form of NAD with age. So those two observations,

02:28:14.840 less NAD, bad. Sirtuin's good. The prediction is that if we could boost NAD, that would be good

02:28:22.520 because that would then restore sirtuin activity and have effects on aging. And so this led to the

02:28:28.520 development and popularization of these molecules called NAD precursors or NAD boosters. The two of

02:28:34.760 which get talked about the most are nicotinamide riboside, NR. That was kind of the first to gain

02:28:40.520 popularity. And then nicotinamide mononucleotide, NMN. Both of those are precursors of NAD that within

02:28:47.320 cells can be converted into NAD. And so there's a large body of literature in a variety of model

02:28:55.960 organisms showing that treatment with NR or NMN sometimes leads to benefits that are associated

02:29:05.240 with healthy aging. And in one study, lifespan extension in mice. I say sometimes because

02:29:13.240 there's also a large body of literature that doesn't reproduce those results. Some of it published,

02:29:18.040 a lot of it unpublished. Including the ITP. Including the ITP. That's right.

02:29:23.080 Which is what I think has to be considered the gold standard for at least mice data.

02:29:27.720 I think that's true. Although, you know, as we talked about before,

02:29:30.360 it's a different genetic background than C57 black six.

02:29:33.880 And I would argue it's a much better genetic background.

02:29:36.840 Well, I think you can make that argument and there are good reasons to believe that argument.

02:29:42.520 Nonetheless, I think it's important to note, right, that that could be why

02:29:45.800 it worked in one context and didn't work in another context. And we've struggled with this

02:29:50.520 as well in my lab. So, you know, it's been reported in this mitochondrial disease mouse

02:29:54.360 by a collaborator of ours who I trust their data, right, that NMN could increase lifespan in that mouse

02:30:00.680 model. We've tried multiple times with both NR and NMN in that mouse model and been unable to

02:30:05.880 to get these effects. So I think these drugs are tricky from a biological efficacy perspective to

02:30:13.640 there's something we don't understand about delivery or or uptake.

02:30:18.120 Do you think you're temperature stable?

02:30:19.880 Probably. Well, I mean, I don't know. I'm sure there are people who know the answer to that.

02:30:25.080 I've read conflicting things, right? I've read that, and I need to go back to the sources on it,

02:30:30.760 but I've read that at least through the lens in which they're provided as supplements,

02:30:35.400 the likelihood that by the time that thing arrives at your door, it still has the biologic activity

02:30:40.360 that it would have had in a refrigerated manner is low. Now, I don't know that the companies that

02:30:44.920 seldom recommend refrigerating them, but they might not recommend refrigerating them because then it

02:30:49.960 would imply that they're being shipped in an unrefrigerated manner, which sort of nullifies

02:30:53.320 the whole benefit. But I don't know if you've looked at any of that. We haven't. So definitely

02:30:58.520 in our mouse studies, we are careful to keep the food in the refrigerator until we put it in the

02:31:02.600 mouse cage. I suspect that there certainly might be some truth to the idea that the biological activity

02:31:07.800 goes off over time at room temperature. But I think that reflects a bigger problem with the NAD

02:31:13.880 precursor field, which is that there's a lot of controversy even among the two camps. So I mean,

02:31:20.040 it's sort of funny, right? Because there's an NR camp of researchers who really think NR is the tool

02:31:25.080 that we should use. And then there's an NMN camp. And they both say that the other camp, that their

02:31:30.280 molecule doesn't work because it's not biologically available or all sorts of reasons, right? So there's

02:31:36.280 a lot of lack of clarity around biological availability and efficacy with these molecules

02:31:43.480 in the preclinical literature, right? Where in theory, people should be able to exactly reproduce

02:31:49.480 the way that other people do the work and get the same results, right? There are lots of reasons why

02:31:53.960 scientific results don't get replicated. It's my impression that in the NAD precursor field,

02:31:58.600 that's a bigger problem than in some other areas. And I don't know the reasons for that. But all I can

02:32:03.560 say is we've experienced that in my lab as well. So I, you know, I've tried to stay on the fence here

02:32:08.760 because I think there's a ton of smoke, right? There's a ton of smoke with sirtuins. There's a 0.97

02:32:12.840 ton of smoke with NAD precursors that they can, if you do the experiment the right way, have positive

02:32:18.840 effects that look a lot like what we would expect for something that's impacting the aging process.

02:32:23.720 And as I said, I mean, it's funny because people who are in the field, I think, sometimes think of me

02:32:28.360 as this anti-sirtuin guy, which is absolutely not the truth. I'm the guy, I'm the guy who first showed

02:32:33.480 that you could overexpress a sirtuin and increase lifespan. If anybody's going to be pro-sirtuin,

02:32:37.800 it's me. I think the problem is that I've seen a lot of data that people have struggled to reproduce.

02:32:44.440 And I just honestly don't know how to interpret that. Whereas with rapamycin, it works for everybody.

02:32:49.160 It's robust and everybody gets the same result over and over and over again.

02:32:52.760 So I'm less enthusiastic, I would say, about sirtuins and NAD precursors as opposed to some

02:32:59.320 other interventions in the field. But I think there's a lot of data that suggests that these

02:33:05.400 molecules and that sirtuins are important for aging. I think what we haven't done yet is figured

02:33:10.760 out how to tweak the system in exactly the right way to get robust and reproducible experimental results

02:33:18.120 that I personally would feel comfortable moving forward with clinically.

02:33:21.880 And it might be that the answer is, you have to hit this system with two prongs. You have to

02:33:26.840 provide more of the precursor and you have to activate the sirtuin in the way that resveratrol

02:33:32.680 attempted to do, but wasn't doing, right? So maybe the answer is it's both.

02:33:37.240 Yeah, right. Sure. That's a possibility. Yeah. The other thing that I find weird about the NAD

02:33:41.880 precursor literature and the limited work that's been done clinically is often, I mean, in principle,

02:33:48.680 it should be trivial to determine whether or not you have boosted NAD levels, right? If you treat

02:33:55.240 somebody with an NAD precursor, we know how to measure NAD. That's not hard. So we know what the

02:34:00.440 biomarker is at least that far in this case. And oftentimes that's not done. So, so, you know,

02:34:05.640 if you're treating somebody with NR or NMN and you're not increasing NAD levels in the blood or in your

02:34:10.600 target tissue, that should tell you something important, I would think.

02:34:15.160 Well, and the other question is, is increasing it in the blood sufficient?

02:34:18.600 That's a different question, but it is an important question. I agree. That's why I said target tissue,

02:34:22.280 and we can't biopsy certain tissues, but we can biopsy some tissues. Yeah.

02:34:26.280 It's super messy. Look, I get asked about it at least once a week by patients, and I usually point

02:34:31.240 them to something I've written on the subject matter. But in the end, I say, look, I will say,

02:34:35.160 I think it's very safe. I really don't see a downside other than to your pocketbook of taking

02:34:40.600 NR or NMN. So they check the first box of any intervention, which is, is the downside sufficiently

02:34:47.800 low? And I think the answer is yes. I'm just having a hard time seeing upside.

02:34:52.360 Yeah. And I'll tell you, honestly, I would love to test NR or NMN or both in dogs for exactly that

02:34:58.840 reason, right? Because there is essentially no risk, and we could actually find out, does it

02:35:05.080 work? Three years from now, five years from now, I could tell you, do NAD precursors slow aging,

02:35:11.240 at least in pet dogs. And I think if I saw NR, slow aging, increased lifespan, and or improve

02:35:17.480 multiple functional measures of aging in dogs, I'd be much more bullish on taking NR myself.

02:35:22.440 I wouldn't prove it's going to work in people, but it gets you part way there and a pretty big part

02:35:26.120 of the way there. And I would put that on the short list of things I would like to test.

02:35:30.520 The fact is nobody's going to do the definitive clinical trial in people because they don't

02:35:34.920 have to, because they can sell that stuff to people now, right? They don't, they aren't required

02:35:39.080 to do the clinical trial to show that it works. Yep. Matt, you probably don't remember this,

02:35:44.200 but there was one night, oh, four, four, five years ago. Are you making comments about my

02:35:50.840 cognitive decline with aging? No. I just don't think you'd remember some random night of us having

02:35:56.780 dinner in New York, but it was about four years ago. Yes. Yeah. Okay. Wow. You do have the cognitive

02:36:02.520 function. So the three of us were having dinner at my favorite, not my favorite, but a decent Persian

02:36:08.600 place on the Upper East Side. And I don't know if I said so at the time, but I may have told you

02:36:14.600 after, it was after that dinner that I decided I got to do a podcast because the three of us had such

02:36:22.360 an enjoyable, she's at Einstein, isn't she? She was. Yeah. So she's actually moved to Columbia.

02:36:26.920 So now she's running a reproductive aging center at Columbia now. Apropos with our discussion today,

02:36:32.600 but we just had such an amazing discussion, which really means me asking the two of you guys

02:36:38.080 nonstop questions. And I remember thinking after that, God damn, why didn't I record this dinner? 0.86

02:36:44.160 Why didn't I have my phone sitting on the table to record it? And it was literally that moment 0.97

02:36:49.640 that made me realize like this happens the same time. Every time I go out with Sabatini,

02:36:54.480 I have the same feeling. Every time I go out with so-and-so, I have the same feeling.

02:36:58.260 I got to just do this podcast thing. So for anybody listening to this, who is appreciative

02:37:02.860 of the podcast, they owe you personally a great debt of gratitude for-

02:37:07.820 And Yuxin. I got to say, I'm sure it was Yuxin who made most of the really insightful comments

02:37:13.420 at that dinner. I think I had a few glasses of wine.

02:37:15.740 I think it was both of you guys, but it is always such an awesome time to be able to sit down with you.

02:37:22.420 I think as the listener appreciates here, the breadth of topics within the space of longevity

02:37:27.700 that you can cover is broad. So you're one of the few people who can go very wide and very deep.

02:37:33.720 And I think today's discussion demonstrated that. So thank you very much, Matt. Really,

02:37:37.720 as always, enjoyed this discussion.

02:37:39.700 Sure. Yeah. Anytime. I enjoy the discussion as much as you do, I think. So it's been a lot of fun.

02:37:45.280 All right. And best of luck with the triad study. I know a lot of people are probably more

02:37:49.480 interested in that than any of the human stuff, because the dog owners I know care far more about

02:37:55.540 what rapamycin can do for their dogs than they care about what it can do for them. So-

02:37:59.240 Absolutely.

02:38:00.080 Hopefully we'll have those results by the next time we speak.

02:38:02.720 Great. Thanks, Peter.

02:38:03.480 Thank you for listening to this week's episode of The Drive. If you're interested in diving deeper

02:38:08.180 into any topics we discuss, we've created a membership program that allows us to bring you

02:38:12.600 more in-depth exclusive content without relying on paid ads. It's our goal to ensure members get back

02:38:18.360 much more than the price of the subscription. Now to that end, membership benefits include a bunch

02:38:23.600 of things. One, totally kick-ass comprehensive podcast show notes that detail every topic,

02:38:29.020 paper, person, thing we discuss on each episode. The word on the street is nobody's show notes rival

02:38:34.360 these. Monthly AMA episodes or ask me anything episodes, hearing these episodes completely.

02:38:40.480 Access to our private podcast feed that allows you to hear everything without having to listen to

02:38:45.760 spiels like this. The Qualies, which are a super short podcast that we release every Tuesday

02:38:51.100 through Friday, highlighting the best questions, topics, and tactics discussed on previous episodes

02:38:55.820 of The Drive. This is a great way to catch up on previous episodes without having to go back and

02:39:00.880 necessarily listen to everyone. Steep discounts on products that I believe in, but for which I'm not

02:39:06.640 getting paid to endorse. And a whole bunch of other benefits that we continue to trickle in

02:39:11.120 as time goes on. If you want to learn more and access these member-only benefits,

02:39:15.100 you can head over to peteratiamd.com forward slash subscribe. You can find me on Twitter,

02:39:21.280 Instagram, and Facebook, all with the ID peteratiamd. You can also leave us a review on

02:39:27.020 Apple Podcasts or whatever podcast player you listen on. This podcast is for general informational

02:39:32.960 purposes only and does not constitute the practice of medicine, nursing, or other professional

02:39:37.400 healthcare services, including the giving of medical advice. No doctor-patient relationship

02:39:43.180 is formed. The use of this information and the materials linked to this podcast is at the user's

02:39:48.840 own risk. The content on this podcast is not intended to be a substitute for professional medical advice,

02:39:55.160 diagnosis, or treatment. Users should not disregard or delay in obtaining medical advice

02:40:01.280 from any medical condition they have, and they should seek the assistance of their healthcare

02:40:05.880 professionals for any such conditions. Finally, I take conflicts of interest very seriously. For all

02:40:12.460 of my disclosures and the companies I invest in or advise, please visit peteratiamd.com forward slash

02:40:19.880 about where I keep an up-to-date and active list of such companies.

02:40:35.880 Thank you.

The Peter Attia Drive - September 13, 2021

#175 - Matt Kaeberlein, Ph.D.： The biology of aging, rapamycin, and other interventions that target the aging process

Episode Stats

Summary

Transcript