#148 - Richard Miller, M.D., Ph.D.: The gold standard for testing longevity drugs: the Interventions Testing Program

Hey everyone. Welcome to the Drive Podcast. I'm your host, Peter Attia. This podcast, my website and my weekly newsletter all focus on the goal of translating the science of longevity into something accessible for everyone. Our goal is to provide the best content in health and wellness, full stop, and we've assembled a great team of analysts to make this happen. If you enjoy this podcast, we've created a membership program that brings you far more in depth content. If you want to take your knowledge of this space to the next level, at the end of this episode I'll explain what those benefits are. Or if you want to learn more now, head over to peterattiamd.com subscribe Now without further delay, here's today's episode. My guest this week is Richard Miller. Rich is a Professor of Pathology at the University of Michigan and the Director of Michigan's Paul F. Glenn center for Aging Research. He served on a variety of editorial and advisory positions on behalf of the American Federation for Aging Research afar and the National Institutes of Aging nia, which we talk about a little bit. He's also served as the Editor in Chief of Aging Cell. He's the recipient of the Nathan Schock Award and the Allied Signal Award, the Irving Wright Award and an award from the Glenn foundation, along with a number of other awards for Aging Research. Dr. Miller's research focuses on the problems of the basic biologies of aging, mostly in mice, but sometimes using other cell lines. Though we speak in this podcast almost exclusively about one of the amazing products of his life's work called the NIA funded Interventions testing programs, the ITPs. Now, over time you'll probably hear me or you have heard me talk about the ITPs on various podcasts, certainly with respect to the podcasts around rapamycin and metformin. The ITPS are a set of studies that are done concurrently in three labs, the Jackson Labs, Rich Miller's lab in Michigan, and Randy Strong's lab in UT San Antonio, we go into great detail about what the criteria are and how these are done, but suffice it to say that based on the types of mice that are used and the rigor with which these are done, statistically and otherwise, the ITPs represent effectively the gold standard of testing molecules for longevity in arguably the most important subset of mice we could study. And basically what we do in this podcast is go through the long list of molecules that have been tested, what the results were. A number of these are quite surprising. They're all very interesting, both in positive and and negative findings. I've been looking forward to speaking with Rich for about a year, but the reason we delayed this until now is I wanted to make sure that a couple of the findings that were in the pipeline were close enough to publication that we could speak about them confidently. And this podcast can now be released as soon as those manuscripts have been accepted. In particular, one around an SGLT2 inhibitor and another one around nicotinamide riboside, which is a very popular supplement that many people ask about as a precursor to NAD. So without further delay, please enjoy my conversation with Dr. Rich Miller. Rich, thank you so much for making time to chat today. I've been looking forward to this one for almost a year now, but we wanted to wait until some of these exciting results we're going to get to were far enough outside the pipeline that we could speak about them.

Good to chat with you, but I.

Think before we get to the most recent findings of the ITP and even explaining what an ITP is, I want to give people a bit of a sense of who you are and how you've played kind of an amazing role in the field of longevity, along with a couple of your colleagues. So where'd you grow up, by the way?

I was born in Philadelphia, and when we were five, my mom and dad moved me and my two brothers to the northern suburbs of Philadelphia, Cheltenham Township. I went to high school with Ben Netanyahu, of all people, who joined our soccer team as a sophomore. I was just born slightly too late. If I'd been born three years earlier, I would have had Reggie Jackson as a teammate.

Wow, interesting. Split your loyalties there between what city you're going to choose for sports. So when did you take an interest in science growing up?

When I was about five.

That's pretty unusual for a five year old, right?

Well, I really liked science. I was decent at school and I always liked getting good grades and stuff like that. And I decided early on that aging was bad for you. It made people sick and then die. I was kind of against that. And the best way to fight against that was to learn something about aging and to develop ways of slowing the aging process down. So you can't do that unless you're a scientist.

So you end up doing your MD and PhD degrees at Yale in the 1970s. At that time, I think you went off and did a postdoc at msk, is that correct? At Memorial Sloan Kettering.

Yes, that's exactly right. I actually did a postdoc for three years, first at Harvard, which did not work out at all. And then I was Rescued by a mentor at Sloan Kettering.

Tell me more about what you mean. That it didn't work out at all.

The first part, I don't want to be sued for slander, but let's just say that my postdoctoral mentor had some defects as a mentor, and I wasn't very productive, and he wasn't very happy with me, and I wasn't very happy with him. So as a beginning postdoc, you're always scared of offending the big boss, and you tend to put off the decision to leave as long as you possibly can. And eventually we both got fed up with one another, and so I moved to Sloan Kettering where things went better.

Did your field of study change when you made that switch?

It changed. Not at that point, but at the beginning of my postdoc, I knew I'd always wanted to work on aging. And the PhD project that I worked on involved somatic cell hybridization. I wanted to mix together two kinds of cells, one that could divide and one that couldn't divide and immortalize the one, the T lymphocyte, that couldn't divide. And that project went well. But the more I learned about the Hayflick hypothesis and all of that cell senescent stuff, the. The more convinced I was that it wasn't going to teach me anything. I wanted to learn about aging. And so I decided as a postdoc to become an immunologist. The lab I worked at Harvard was an immunology lab, and the one at Sloan Kettering was an immunology lab too. Then when I set up my own lab at Boston University, it was principally to study immunology and how aging modified the immune system. When I came to Michigan in 1990, I was still mostly an immunologist, focusing on aging and writing book chapters on how aging modified immunity, the like. Over the next decade or so, I switched gradually into focusing on aging more broadly, with immunology becoming only a minor interest. And now we do very little immunology. It's nearly all the kind of work that is focused on what is controlling the aging process and how you can use that to develop interventions.

Give folks a little bit of the history around Hayflick and the division of cells and some of these observations, these are kind of some of the foundational things in aging, right?

Unfortunately, yes. I don't know whether you want the 2 minute or the 20 minute or the 3 hour rant about this, but the phenomenon that Hayflick more or less discovered was that if you take normal human cells, which are not Cancer cells, and you let them grow in tissue culture, they're divide, but only a limited number of times, about 50 times, and then they stop. And if you listen to Hayflick, he then decided that was sort of like aging. He had sort of found a way to study aging in culture. Now, that's nuts. It's nothing like aging in the slightest. But he thought it might be like aging. So at that time, people were really excited by tissue culture. It was the brand new thing. Everybody wanted to work on tissue culture, and it seemed like a wonderful, exciting opportunity to be able to use tissue culture to study aging. So they bought this idea. And a whole generation of superb cell biologists spent their life investigating the Hayflick system under the illusion that they were studying aging parenthetically. About 15 or 20 years later, when I was just beginning to become well known, I was invited to give a talk at New York University, whose chair of pathology was Vittorio Defendi. And Defendi had been a colleague of Hayflick when Hayflick was making these discoveries. And Defendi told me that he and Hayflick had had lunch over and over again, and Hayflick had been a kind of a grump and had said over and over again, oh, my cells, they just won't grow. Oh, I tried them. I thawed them out only two weeks ago, and they already stopped growing. And Defendi, if you believe his story, which I do, told Hayflick, well, Len, maybe they're just getting old. And Hayflick, who, whose sense of humor is notoriously absent, did not understand he was being joshed. He thought it was a hypothesis, a scientific hypothesis. And after a while, he persuaded himself that it was his hypothesis and then that it was the correct hypothesis. So he's a persuasive fellow, and many people bought into the notion that the idea that when cells stop growing in culture, it's sort of like or caused by or something related to aging, too many people, in my view, accepted that as the truth of the matter. And then when it later became established that the limitation of growth, the Hayflick limit, was actually due to the shortening of telomeres, a very important finding and certainly true. People convinced themselves because they thought it was like aging, the telomeres were something to do with aging, too. And it was that way in which a whole new generation of people sort of bought into the Hayflick and the telomeres thing as a central cause of the aging process, despite all the evidence to the contrary.

Well, it's interesting you bring this up, of course, because at least at the time that we sit here recording this, it's only been about a week since a story came out of Israel, which I'm sure you've been emailed about as many times as I have. You haven't? Okay, well, consider yourself lucky because I've only been emailed it about 57 times, which is apparently the most rigorous study in the history of humans demonstrating the anti aging benefits of hyperbaric oxygen. So you see, rich, in this amazing study done in Israel, a group of volunteers were exposed to hyperbaric oxygen, and wouldn't you know it, their telomeres lengthened a bit. And so as a result of that, I've had. I don't know, I'd probably need scientific notation to count the number of people that have emailed me that story to tell me why are we not doing hyperbaric oxygen. You see, this is the fountain of youth. And I think I forwarded it to Joan Manock and Matt Kaeberline because the three of us have a grumpy old men women club where we groan on about the idea of until you have really good biomarkers for aging, it's really hard to study aging and we can't use telomere length as a biomarker of aging. But nobody really wants to listen to this.

I agree with all three of you. Yeah. And you know, it's funny, it's entirely possible that hyperbaric oxygen could well have health benefits under certain circumstances. Certainly I'm not endorsing it. There's no evidence that that statement is true, but it's not silly. There are certainly oxygen sensing circuits in multiple cells, which, at least in worms, can trigger anti aging programs, pro longevity programs. So it's not a dumb. And there's some really good labs pursuing it. The weakness, I haven't read the paper, but from what you've told me, the weakness is just the one you've cited. Aging and telomere length are not the same thing. And if you want to prove that hyperbaric conditions at suitable doses and it's suitable time intervals, et cetera, might be good for you, that's a very plausible idea. Well worth testing. It's just that you don't test it by measuring telomeres.

Yeah, I mean, it's interesting, obviously. I think that the telomere, I don't want to say story, I would say the telomere concept or notion as a biomarker, even though it's not really a biomarker, I think gets a lot of credibility based on the fact that a nobel prize was awarded for its elucidation. So it's interesting. I've heard people say look, and I think this is the most accurate thing I've heard stated, which is, look, the work that blackburn did is really amazing biology, and it's worthy of a nobel prize. But it doesn't really explain aging. And those two statements can coexist.

Oh, sure. But you've left out a third important statement. Telomere biology is a critical element in cancer biology. And the amazing work that several people, including blackburn, did to work out the telomere story is, I think, really a fundamental advance in our understanding of cancer in people. And it may also have an interesting sidelight in evolutionary biology. There are some species mice, for instance, with extremely long telomeres. They don't need much anti cancer defense at all. They're going to get eaten in six months. Vera gorbunova at rochester has done some lovely work on the ways in which different species with different body size and different lifespans differ in their ways of stopping the cancer process. Some relying, like people do, on the telomere alarm clock, where the telomeres get too short. This puts a whole lot of anti cancer defenses into play. And others, like mice, where this does not happen. The mice have very long telomeres. They don't need telomeres to tell them when their cells are getting cancerous. As a fundamental step in our understanding of cancer biology in people and the way in which anti cancer defenses evolve, it's great stuff. It's just the notion that it's sort of a shortcut for actually working on aging, in my view, does not hold up very well.

Yeah, well, I think I couldn't agree with you more. And it's certainly on the short list of frustrations I have in this space where people very quickly revert to, well, but look, but intervention x lengthened telomeres. Therefore, it's somehow a cure for longevity or a cure for aging. And despite all evidence to the contrary. So what prompted the move from BU to Michigan?

In 1990, my wife got a job as a professor of english at the university of michigan. And she said, would you like to move with me to michigan? Why don't you look for a job? So that worked out pretty well. I mean, I'm being a little coy. I had always wanted to move to michigan. Michigan was very strong university and very strong in aging research. And I had said to her, you know, she was coming up for tenure at harvard. Harvard had never tenured a woman in 372 years in English, they had never found a competent woman who could be an English professor. So we know they were not going to start with Patsy. And I said, hey, why don't you apply to Michigan? I'd love to move to Michigan. And it worked out for both of us.

Excellent. So what is it that brought to your mind this idea of the itp? This is partly your brainchild, right?

I was the midwife or something. So the National Aging Institute in its Division of Aging Biology at that time was headed by a visionary guy named Huber Warner. And Hubert thought it would be a great idea for the aging research community generally to have a program in which drugs were tried on mice or mice and worms, or mice, worms and flies and dogs or something. So he commissioned a committee of eight or 10 or 12 of us to sit down with him and Nancy Nadon, who was his deputy at the time. And we spent a day or two talking about what the NIA might do to develop a program in which interventions, potential interventions, were tested directly. And after the first night, I went into a room with Arlen Richardson, who at the time was at Texas and is now at Oklahoma. And Arlen and I put together what we viewed would be the best way to do this, focusing on mice, and we sold it to Huber. So Huber then asked me to write down a more formal plan, which he and I wrote up as an article for Mechanisms in Aging and Development. And when it was time for the NIA to formulate the rules for the program and to start saying, would you please everybody apply for this program? We have some money. The program that Arlen and I had sketched out and which Hubert helped to refine, and by then been through three or four versions, became the sort of foundation on which the program was developed. And then three grants were awarded. My own. David Harrison's at the Jackson Labs and Randy Strong at Texas. So the three of us, it took us a while, sort of get to the point where we all agreed as to what should be done and how it should be done. But the foundation of the program is one that Huber and Dorsen, which Arlen and I had helped to develop. By now, it's about 18 years ago.

This is around the time you wrote a paper extending life, scientific prospects and political obstacles. Right. That paper coincides almost with when you and Strong and Harrison kind of came up with what the principles of the ITP would be, right?

Yes. I mean, the two events were not really related in any cause and effect fashion, but it's about the same time I had been asked to give a talk at the law school at the University of Michigan. They were putting a seminar together, an all day symposium in the End of Life and the other people chatting. One was how to take care of old people, one was how to make decisions, triage decisions at the end of life. And I was the only biologist on the program. So I gave them the talk, which then became that article in the Millbank Quarterly that you've been mentioning.

Now, you make three big points in this article that I think are worth spending a minute on before we jump into the itp because I think it really sets an adequate stage. Right. So again, keeping in mind this is an article written about 18 years ago and it's important, I think, for people to realize that the world we live in today is much more open to the discussion we're going to have than it was, I think, 20 years ago, that's for sure. So you said, look, discussions of anti aging medicine aren't silly. The development of an anti aging strategy is starting to make headway and further work here would be worthwhile. That was basically the thesis of your paper.

Yes, I agree, and it's, I think, a really important point that deserves stress. It's still an evolving process. We're not there yet, but 18 years ago we certainly weren't even close to there. Around that time, the Gordon Research Conference on Aging, which is an invitation only conference about 120 more or less distinguished researchers. I got into a fight with my friend George Martin as to whether aging could be considered a single process. A drug might slow that process down, or whether it was a very complex spaghetti bowl filled with 30 or 40 or 50 different processes, each one of which had its own independent control. And George and I did not agree about that. So as a game, he asked the assembled conferees to vote and you could vote 0 to 10, 0 if you agree that there were too many processes ever to be controlled together, and 10 if you happen to take my position. The average score was about 2 or 3. Very few people at that time, even among the sophisticated, best educated professional aging researchers would check the box that said, yeah, there's a single aging process. And until you check that box, the notion that you could interrupt the process makes no sense. Because if there is no process, then interrupting it is a hopeless goal. I think you really sort of need to see it as a unitary process in order to devote much time to trying to slow that process down. It's confusing because there are dozens and hundreds of things that go wrong in old age, and learning about any one of them can be very productive. It's just that in addition to looking at any one of them individually, the one that leads to cataracts, the one that leads to muscle failure, etc. The concept that there's some underlying mechanism that speeds them up a lot in mice or a little bit in horses, or slows them down a lot in people or an awful lot in whales, Once you get that idea in your head, then suddenly it becomes permissible to think about ways to slow it down.

Now, there were a couple of things that we already quote, unquote, knew at this time that had to give you some hope. Obviously, there was the vast literature on caloric restriction. So almost without exception, and there are some very notable exceptions, but almost without exception, some form of caloric restriction. And we could get into all the details of different species and when it's applied during life and to what extent and wild types versus not. But there was clearly a path towards extending life with caloric restriction. And then, of course, we had the DAF mutants in C. Elegans, Cynthia Kenyon's work, which I think was probably early 90s, right. It was 1993.

1993.

And if I recall, her first. God, it's been a while since I've looked at this. Her first observation was with. Was it Daf16 was the first one, or was it Daf16?

The first paper, the critical paper, made two big discoveries. One was that mutants of Daf2 extended lifespan and that mutants of Daf16 blocked when combined.

Yeah. And now you could combine with CR and then get a better outcome. But yes, and DAF2 ended up being the analog of the IGF receptor. And I think DAF16 was basically an analog of FOXO3A. Right. Or 3B, I can't recall.

Yeah, it wasn't quite the receptor. It was the thing that the receptor turned on. But it's one step downstream.

One step downstream inside. Yeah. So explain to folks why that was relevant as a proof of principle, at least.

Well, first of all, it's just the point you were making. Those of us who knew the caloric restriction literature were screaming our heads off saying, see, look, you can slow the whole thing down. And no one quite bought into that. Cynthia's demonstration, which was also pioneered by other people like Tom Johnson and Gary Rufkin and some others at the same time, pointed out that a single gene mutation could extend longevity dramatically in a worm. The funny thing is Nature, early in 1993, published a very erudite, detailed six page paper by Linda Partridge and her colleagues proving that there could never be a single gene mutation that extended lifespan. Aging was too complicated, too many feedbacks are. It could never, ever be done.

And Linda's paper was a statement of any species.

It wasn't explicitly stated, but it was a theoretical proof that you could never have a single gene mutate that extended lifespan. Absolute mathematical proof, lots of equations. I couldn't follow that. It can't follow now. So about six months later, Cynthia's paper came out saying, hey, look, here's a gene that does that. So that same issue Linda was asked to comment on Linda Partridge, and she wrote a paper saying, well, of course C. Elegans is different. C. Elegans has dour mutants, they're very, very special. But there could never ever be in any other species a gene that does this. Trust me on that. So it took three more years for the second shoe to drop, so to speak, which is Andre Bartky's paper with Holly Brownborg and two other colleagues showing that the Ames dwarf mutation could extend lifespan by about 40% in a mouse. Once it was true for mice and true for worms. The idea that, you know, there might be something to this became defensible. One of my favorite conference stories, about a year or two after the Amesdorf paper came out, also in Nature, I was at a conference and Robin Holiday, an Australian scientist who was very well respected, gave a keynote lecture explaining why you could never have a gene that slowed the aging process and extended longevity. And I raised my hand to say, but you know, there are two papers here, one in worms, one in mice, saying that there are single gene mutations that extend longevity, and that's not compatible with your theory. And he said, yeah, but there could never be a single gene mutation that could extend longevity. And so we went through that loop about twice and then I went out to get a cup of coffee. It's a notion, a zombie notion that sort of was hard to kill. And many scientists still are distinctly uncomfortable with the notion. Not all of them anymore. I think our team is winning, but many people still are reluctant, even if they're sophisticated and know the literature, to accept the general principle that there are aging processes which can be slowed down if you consider them as a unitary group of changes with a common controlling factor.

And I always think that part of the debate is fixating on the wrong point. Right? I mean, whether or not it's a single gene or multiple genes is probably less the point, the bigger point, at least for Me as an outsider, is aging malleable? Yes or no? To me, that's the question that matters. And if prior to the 1990s, if the answer was believed, no, aging is not malleable, it is what it is, then it's a very uninteresting field to be in.

Yeah, no, I agree with that. I think you're right on point here. I would add one more, Philip, to that. The question everybody was asking and which I used to think was pertinent is what causes aging? What is the cause of aging? And now I don't think that anymore. I think the real question that we're all trying to answer or ought to be trying to answer is what is it that can slow aging? My own current sort of philosophical framework is that aging is caused by an awful lot of different things. Some cells die, some cells become mutated, and some tissue structures get cross linked or heavy metals accumulate in a key cell. All of those are really bad for you. And as they accumulate and accumulate and accumulate, you start to feel older and then more diseased. But the key point, I think is that there are biological processes that can postpone all of that stuff together, that can postpone it for five decades in people, or almost a year in a mouse or 25 years in a chimp. So I don't really care what caus aging. What I care about is what is the process that can postpone all the different aspects of aging. I think once you rephrase the question in that way, you're well on the way towards designing experimental paradigms to address the serious question, which is the coordination and eventually the postponement of these multiple aspects of the aging process.

I think it's worth restating that, Rich, because that might be one of the most important teaching points, I think for anybody who thinks about science, and certainly my mentor in the lab, I also studied immunology, would constantly make this point, which is if you don't ask the right question, you are guaranteed to flail. You know, you might get lucky, but you're, you're going to flail. Now, if you ask the right question, you may still fail, but you really, you know, it's the difference between starting on your own 20 yard line and the other guy's 20 yards.

Yeah, I agree, and I agree it's a really critical point. A lot of the time at like an aging meeting, somebody, usually someone fairly new to the field, will say, oh well, you don't understand what aging is. Let me explain to you what aging is. And everybody who's Heard this line of argument 40 or 50 or 60 times, will sigh. Sometimes we'll go out and get another cup of coffee and say, well, let's talk about that at the break or at the bar after the meeting. But they're wrong. The answer is you have to think about it. Just as you were saying, you have to think about these fundamental issues. What is the best way to frame this question? And it can be tricky. If the world is filled with people, including myself, who are convinced we're right, then you're surrounded by people who aren't right. They're wrong, but they're convinced they're right. And sort of talking them into viewing the matter from another framework takes a while.

Let's dive into the itp, the Interventions Testing Program, this amazing, amazing scientific framework. Think of it as a program. I don't know. How do you describe it to people? I sort of think of it as a gold standard by which we test single drug often, but sometimes not just single drug interventions in a robust, rigorous manner. And when things pass the ITP test, we take them really seriously. When things fail in an itp, even if they've succeeded elsewhere, we look at them very closely. So what's the word you would use to describe the ITP to somebody on the outside?

Well, I agree with most of what you've said, and I'm certainly very flattered to hear you describe it in that way, although I think you're being a little too optimistic and a little kind. So let me, let me sort of back. Go back over what you said. The goal of the itp, and we've tried our best to meet this goal, is to develop a fundamentally sound way of testing. One question, does this drug extend mouse lifespan? And we do our best by designing the program in ways so that our answers will be reliable and to the best we can, reproducible and believable. For example, we use an awful lot of mice because that gives us a lot of statistical power. We want to be sure that we will have enough mice present that if a particular drug extends lifespan by 8 or 10%, we'll pick it up almost all the time, 80% or 90% of the time. That's sort of an ambitious thing to do. To give you an example, Jay Olshansky and Bruce Carnes and their colleagues demonstrated years ago that if you abolish cancer in people, nobody ever gets cancer again. Human lifespan goes up by only 3%. So the drugs that we are interested in by designing our program are drugs that are designed to win at our Program, you've got to be at least three times better than a complete conquest of cancer.

I'd like to come back to Jay's methodology on that because I've never fully understood that calculation because I know he's also made the claim that if you completely abolish atherosclerosis, it's about a 3% bump. And if it's cancer and atherosclerosis, it's like a 7% bump.

Yeah, that's right.

And if you got rid of cancer, heart disease, stroke, and diabetes, it's about an 18% bump. So we'll revisit that. But let's go back to these principles, because there really are four remarkable principles of the itp, and that's what it is. That, for me personally, gets me excited. So you alluded to one already, which is from a statistical standpoint, you, power to 80% for an 8 to 10% detection rate. And to put that in English, it means you need an awful lot of animals.

Yes. Each year, each of the three sites will have 50 males and 50 females on each drug and also 100 male controls and 100 female controls. We double the number of controls because it gives us a lot better statistical power at relatively little cost. Then we pull all the data from all three sites together for our analysis.

And then you've alluded to now the second feature of the itp, which is they are run in parallel at three sites. So the Miller lab is doing this at Michigan, the Strong lab is doing this in San Antonio at ut, and Harrison is doing this at Jackson Labs, correct?

Yeah, that's right.

Okay, the next point that makes this pretty special is you are using genetically heterogeneous mice. These are not homogeneous. So explain to folks why using these genetically heterogeneous mice is an important feature of the itps.

Yeah, I will let me just say one more point about that second point you raised. Doing it at all three sites not only gives us power, but it also tells us that we're getting an effect that is reproducible at three sites. If, for instance, and we've certainly seen this, we get a drug that does just terrific at one of the sites, but not at all if the other two. Something may have gone wrong.

That's right. That's a methodologic issue right there. Potentially, yeah.

If you get a drug that works pretty darn well at all three sites, people have very good reason for believing it could work at their site, too. That sort of instantaneous reproducibility is a major reason for doing it. At three sites. The literature is filled with reports of a drug that did something good to a mouse at one site. And then no one wants to wait four years to test it out. Doing three tests at the same time gets us around that. Then the question you asked was why use the genetically heterogeneous mice? About 90% of the work in aging with mice and actually of medical research with mice uses a single inbred genotype where every mouse is the same, there's no variation from mouse to mouse in their genetics, and in addition, it's inbred. And everybody knows if you have a choice, you'd rather not be inbred. That's why we don't allow people to, to get married to their brothers and sisters, because it tends to produce weak offspring. And every one of the inbred strains has its own set of bizarre peculiarities. Many of them are blind by the time they're grown ups, most of them are deaf, et cetera, et cetera. So what we wanted to do in developing a genetically heterogeneous stock was avoid that. We did not want to trick ourselves into picking a drug that only worked on black six mice. And conversely, we did not want to miss a really good drug that just happened not to work on black six mice. So what we do is we obtain from the Jackson Laboratories two different kinds of mice which have represent four different grandparents, inbred grandparents. When we cross them together, just like when your mother and father produce offspring, each one of those children is genetically unique, but each one will share half of its genes with all the others at random. A random half. So all of the mice we've for the IDP, of which there are now more than 20,000, have that in common. No two mice are ever the same, but any two mice share half their genes. The other advantage of doing it this way is that the Jackson Labs sort of the gold standard in inbred stocks, they're black six this year and they're black six 10 years from now. And their black six 10 years ago are as close to the identical as science can guarantee. And what that means is that the four way cross mice we use, the HET3 mice, are the same year after year after year. We can make 100 or 1,000 or 10,000 of them with the same characteristics. Anyone else anywhere in the world that wants to make them can do so very inexpensively and be sure that their population characteristics, the average numbers of mice that have this gene or this pair of genes, or this triplet of genes will be the same as the ones that we are using. It gives a sort of reproducible heterogeneity, which I wish almost everyone would adopt. I think that's the way to do science. And people who were at the origins of the IDP agreed with me about that. So that's why the IDP does it that way.

Do you have a sense of what fraction of research is done in mice utilizing genetically heterogeneous mice in the fashion you described? Are we talking less than 10%?

Oh, yeah, much less than 10%. And the reason is that when people want to work in aging, they call the NIA the National Aging Institute, and they say, can I please have some old mice? And the NIA does not have genetically heterogeneous mice. They had them for I begged and pleaded and kicked and screamed, as did some colleagues. And they set up a colony of genetically heterogeneous mice. And after three or four years, they were not receiving many requests. Everybody was saying, I want black six mice. My scientific parents used black six mice. And everybody always uses black six mice. I need black six mice. So the nia, for lack of demand, folded up the colony. And now that, I think largely because the ITP people are beginning to realize the folly of using black six mice and the good reasons for using genetically heterogeneous mice. And the NIA keeps getting a sprinkling of flow calls every now and then saying, please, please, can you carry these genetically heterogeneous mice? But it's momentum. They have contracts with their mouse producers, and it would take at least Richard Hotas, the director of the nia, had two or three or four of his trusted associates to say, oh, yeah, we should do that. And so far they haven't.

All right, so there's one more feature of the ITP, which is that basically anybody can make a suggestion for a molecule, right?

Yeah.

So what was the first candidate molecule in terms of the first pool that was suggested?

I don't know. I have to go back and check my notes. And the reason I'm a little foggy on this is that actually we had a sort of a proto embryonic ITP at the University of Michigan. Only a single site for about two years. At that time, we had a Nathan Shock center funded by nia, and I devoted my chunk of the nation Shock center to testing four or five or six drugs over a period of two years. None of them worked. And then the ITP got formally funded. It became a multi institutional program. That first year, there were four suggestions that we accepted. One of them was Aspirin One of them was a molecule called ndga, which is norodihydroguaretic acid, which actually does work. It's worked three times in a row, although it works only in males. One of them was nitrofluorbaprofen, and I forgot the fourth. I'd have to check my notes, but those were the first four that were accepted for the initial round of the itp.

Now, aspirin's interesting because it initially showed that it worked in males, but this is one of the examples of something that didn't replicate after the itp. Is that correct?

Yeah. It was not actually a formal attempt at replication. The initial dose of aspirin was very low, that is 1/100th of the dose a person would take, even if adjusted for mouse body weight. And it gave only an sort of an 8 or 10% increase. And it was in males only.

Why the decision to use such a low dose?

The sponsor, that is the person who says, please use this drug, we take his or her suggestion very seriously. And the sponsor of this was a guy named Christian Van Levenberg, who's now at the University of Florida. And he gave a strong rationale for why he thought this dose was appropriate. And we accepted his suggestion. And it seemed to be a fairly good suggestion in the sense that it worked. But then several years later, we said, maybe it would have worked even better if we'd used a higher dose. Maybe it would have worked even better, maybe even in females, if we used a dose that approximates the sort of 83 milligrams a day thing that I used to take to prevent heart attacks that a lot of people take to prevent heart attacks. So we tried it at two doses higher than the one we had originally evaluated, and it did not extend longevity. So it was not an exact replication. It was a replication at higher dose. It succeeded the first time, sort of. It did not succeed the second time. And by that point we had enough things that really worked great that we weren't about to devote additional time and energy to trying intermediate doses or something like that. To give you a sense of the budgetary commitment, the total program budget each year in direct costs is about a million bucks. So if we test six drugs a year, something like that, $3 million divided by six, about a half a million buck investment for each drug accepted for an initial test into the program, crudely speaking. So we have enough money now, because the NIA has been extremely generous to us, that we can test five or six or seven drugs each year for the first time and also go back and recheck one or two of these a year. But we can't go much above that without running out of money. So if we were to decide to try aspirin at a few more doses, we'd be eliminating from our program some other drugs that, that have never been tested but look really promising. So we make that decision each year as a group with the help of five other scientists who are not part of the itp, but give us advice on what drugs to pick.

Is part of your decision to include a drug or not include a drug at all based on clinical data in humans?

Oh, yeah, sure. I mean, we know that mice are not humans and the things that work in mice might not work in humans and, and vice versa. But if someone says, see this drug shows potential for helping to solve, prevent cancer in humans or to prevent some aspects of neurological decline in humans, that's very strong argument as to why it might well be worth testing in mice. Testing a drug in humans with lifespan as an endpoint is almost hopelessly slow and almost hopelessly expensive. The ideal animal for making a test of that is the mouse. So whenever we have a suggestion, we consider a wide range of different kinds of evidence. And clinical evidence about a health effect in people can be given a very high weight. In addition, even if a drug is used in humans for something that's not arguably unrelated to the aging process, the fact that it's FDA approved also is a big selling point. Because if one of our drugs works in mice and it's already FDA approved, then the barriers towards evaluating it in human clinical situations become much easier to get over because of the safety tests that have already gone into obtaining FDA approval for some other indication, like Metformin, for instance.

Yeah, we'll come to Metformin because there's a few drugs on the ITP hit list. I want to really spend some time on Metformin as one of them, of course, rapamycin another. But I'm curious, like using aspirin as an example. If you were today going to study aspirin for the first time, would you be swayed by the fact that the evidence for use of aspirin to prevent cardiovascular disease is quite weak? Or would you chalk that up to, hey, maybe for that indication it is. But nobody's really evaluated all cause mortality over a long enough period of time in humans. We think it's still worth studying.

You're thinking about this in a, in a very good, sophisticated way. You could join our compound review committee. So all of these points can and do come up in the discussion of which drugs to give highest priority to the fact that something might or might not modulate longevity or all cause mortality in people is worth knowing. But there are lots of things that test interesting ideas in mice that may not be of primary relevance in the human clinical setting. Aspirin, for instance, is a famous anti inflammatory. And there are lots of good reasons which I'm sure you know, suggesting inflammation may go up as you get older and it's probably bad for you and it may contribute to lots of diseases. So it's a very strong argument for saying, let's test some anti inflammatory drugs in the mice. We're not testing aspirin in mice to prevent heart attacks. Mice don't get heart attacks. But the notion that it's a strong anti inflammatory and that the processes of inflammation contribute to aging in both species, that makes some sense. And that's a good reason to test it.

Yeah. Let's move to your second cohort here because this is where you sort of hit one out of the park. Right. So in Cohort 2 in 2005, started in 2005, Rapamycin is one of the candidate drugs. Who brought that to the committee?

Dave Sharp. Dave Sharp was a colleague of Randy Strong's at Texas. He's an expert on tor the target of rapamycin. He was aware of invertebrate data in worms and in flies saying that if you had a mutant that inhibited the function of the enzyme that led to longevity increases in these two invertebrate species. And so he said, look, I've got a drug that can inhibit the enzyme. Why don't you give the drug to mice? Rapamycin is that drug and it's actually safe enough that you can use it in certain high risk situations in people. So Dave Sharp suggested it and we tried it. Randy Strong noticed first of all that when you gave it to mice in the usual form, 90, 95% of it was eaten up in the stomach. So Randy and Dave and their collaborators then spent a year trying to successfully trying to reformulate their app rapamycin by coating it in a kind of a shell that would get it through the stomach into the small intestine where it would be absorbed. That's why for the first time around, instead of starting at four months, which was our goal, we could only start at 19 months of age or 20 months of age. It took us that long, took Randy that long to get the drug into a form which was effective when given by mouth. But yes, then that was the first drug that gave a very strong signal in both males and females. And it's still the only drug that we've tested so far which gives a very strong signal in females. We have two others that work repeatedly in females, but they're not as strong in terms of the size of the effect as rapamycin in females.

The other thing that's really noteworthy about this is what you just alluded to, which is based on the need to reformulate rapamycin to improve its bioavailability, you, without planning it this way effectively were giving rapamycin to 60 year old mice.

Yup, 20 month old mice, which are sort of roughly the equivalent of a 60 year old person.

So this is actually very interesting because the gold standard intervention for treating mice, which is caloric restriction, typically only works when initiated earlier in, in life. And so typically when you wait until a mouse is two years old to begin calorically restricting it, it's too little too late.

That's right. Richard Weindrich found that it works great and caloric restriction works great at sort of four or five months of age, 12 to 14 months of age. It still works, but not so great. And then Weindrich found that if you started as late as 19 or 20 months in his hand, it failed. Just as you were saying.

So how do you put the rapamycin finding in. So one way that you can think about this, I guess, is to think of it as, here's a 50 year old person, how many years do they have left in life? Right. And if you do nothing, we would say the average 50 year old person has 30 years left. Yes. Is that about the right number?

If she's a white woman, yes.

And a man wouldn't be too far off from that. Right?

Five years less, but yes, that's right. That's generally correct.

So what did the rapamycin data suggest? If they were to translate to humans, which is a big if.

The point you're making is exactly on point. I bet that it was a waste of time and money to give rapamycin to 20 month old mice to middle aged mice. I said this can't possibly work. Look, caloric restriction doesn't work at this age. Just as you were saying. And so it's a waste of time. And somebody said, yeah, but we've already got the mice, we've already paid for them, try it. And they were right, of course, and I was completely wrong. And that's a big surprise. That I think is the reason why the paper was attractive enough to be published in Nature. The notion is, and here this becomes hand waving because we don't have a molecular explanation. But the notion is that there are some processes really bad for you when you're a 20 month old mouse or 60, 65 year old person that are still not irreversible, that can be reversed and that that reversing can be dependent upon inhibition of TOR signals. So giving rapamycin to a 20 month old mouse extends longevity dramatically, just as dramatically as if you had started the rapamycin at a much younger age. That's a fundamental, I think, reformulation of how we thought aging worked. And independent of the notion that, say, you can take a drug and it can make you live a long time. The observation that some, though not all, of our drugs that work for us are fully effective when started late in life suggests two things. One is the theoretical thing. Some things are going on late in life which are still reversible and have a major effect on your health. And secondly, of course, if you have any friends who are 50 and 60 and 70 year old and they would like to take a drug that makes them live longer, that's fairly good news.

So let's explain a slight semantic point that's going to be interesting or I guess relevant as we kind of go through these. Can you explain to folks the difference between increasing median survival and maximum survival? Like how do you think about it? What moves the needle for you when you were thinking about something?

Yeah, I mean, just technically, median survival is the age at which half of them have died and half of them haven't died yet.

Okay.

Maximum survival is a folk notion. It's basically how long is the oldest mouse or person who's ever lived. And that's a value which with very limited statistical appeal because the oldest person in a group of 100 people is not going to live as long as the oldest person in a group of a million people, or 100 million people or 10 billion people. So the statistic that is probably intuitive and a best substitute for that is what is the age at which 90% of the people have lived? That we crudely refer to that as maximum lifespan, though it's really not maximum lifespan. The reason it's an important concept in terms of understanding how to interpret survival statistics and survival curves is that if you happen to have, let's say, a group of animals, many of which are dying, let's talk about in terms of human years, most of them are dying in their 30s and 40s and 50s or something, and you have something that extends Their median lifespan up to 60, you can see as a public health benefit. That's why a lot of people get immunized and don't want to have their kids smoke, etc. But it hasn't worked on the aging process. It hasn't in particular had any real effect on how much longer you're going to live when you're 70. But if you had a drug or a diet or some intervention that authentically slowed the aging process, then it would extend expectancy of additional healthy years of life, even for those who are already quite old, in their 50s, 60s, 70s and 80s. And that will modify the age of death at the 90th percentile. So drugs that only extend the median and don't affect the age of death of the longest lived 5 or 10%, they might be interesting in some ways, but they are considered less plausible as candidates for anti aging drugs. If you have a drug that's authentically slowing aging, one of the things you most want to see is that the very oldest animals in this drug treated group are living longer than the very oldest animals in the untreated control group.

And so when you look at the rapamycin data, they were very interesting. The median extension, which we've already said not as interesting, that's just saying what's the age at which half the animals have died. In the males went up 20%. In the females it went up 13%. These are huge numbers.

I think you may have the numbers backwards. In the females, the median goes up more than in the males.

Oh, I thought that was for the 90th percentile. It went up more, more in the.

Females for both of them. It's a technical artifact because when you give the same dose of rapamycin in chow, the blood levels in the females are three times higher than the males. So we don't know why.

I'd like to come back to that in a second. The numbers I've always remembered are actually not the median ones. I could be out to lunch on those. The ones I have ingrained in my head, and maybe I hope I have them right, is the P90 lifespan extension, which in males was 9%, in females was 14%. Do those sound right to you?

I'd have to look it up, but that sounds plausible. Yep.

Directionally that's a big deal because that's total lifespan, isn't it? That's not incremental. Above the point.

You got it. We don't cheat, we don't calculate incremental lifespan.

All right, so let's explain to People what that means. Because these are huge numbers.

Yeah.

So explain to people why that basically translates to 25% more life once you've reached midlife versus 9 or 14% more life. This is a subtle way of doing the math more rigorously.

Yeah. The editors of Nature insisted that we do that calculation. That is the number of additional years or months of life after you give the drug. They thought that would sex up the paper and kicking and screaming, we put it into the paper. But it's not, I think, terribly informative. There's an easy to imagine thought experiment. You've got someone on a ventilator, he or she is on death's doorstep, they're going to die today, and you leave them on the ventilator for one more day and then you give up and they die. So you've extended. You've doubled their lifespan. Right. You've gone from one day to two days. That's 100% increase in lifespan. So when I put it that way, you can see that that's not a very useful statistic to calculate. If you're an insurance agent, it's nice to know at the time you sell the insurance policy what the life extension will be from the time the person buys the insurance policy. But in terms of biology and pathobiology and being able to compare lab A to lab B, drug A to drug B, I think keeping it on the level of what is the change in the overall median is important. There's one exception to that, and that's the fairly obvious one. If you start a drug really, really late, you know, if you start a drug at an age when 35 or 40% of the animals have already died, you're asking a lot of that drug to extend the median lifespan. So if you start rapamycin at 19 or 20 months of age in males at 20 months of age, 15 or 20% of the mice of Oregon, depending on what site you're talking about. So if you get any extension of the median lifespan, it sort of had to start working immediately. Nobody could die for the next few months. That's the situation where the statistics on how many of them make IT to the 90th percentile become more informative than a percentage change in median.

And to my eye, rapamycin is the only drug you've started at 20 months. I know you repeated it once at 20, but that's a big ask of a drug, right?

Well, I'm pleased that it worked. It's not the only drug that works.

Well, I'm saying starting at 20 months of life.

I know, and that's what I'm talking about. Yeah. We've used acarbose starting at 20 months of age.

You did? Okay. I thought acarbose started earlier.

Okay, yes, you're right. The initial paper was a carbos at four or five months. Yeah. But then we repeated it, and that's published too, where we started a carbos at, I believe 20 months of age. And it worked half as well as a carbo started in youth. That is, we got a statistically significant change in the maximum lifespan by our statistical procedure in both sexes and a statistically significant change in the males, even in the median lifespan in the females. The P value for median was 0.07, if memory serves. So a carbos started in late middle age is effective in both sexes, not as effective as if you started in youth. And by the time this podcast airs, our paper on 17 alpha estradiol, late start will have appeared. And 17 alpha estradiol, which only works in males, the new paper just now at the last stages of revision, show that if you started at 16 months of age in males, it's just as good as if you start it at an earlier age. Even if you started as late as 20 months of age, it works great just as much. It's not statistically distinguishable from the 16 month start. So that means of the drugs that we've tested for late start, rapamycin works perfectly. Acarbose works about half half as well as an early start, and 17 alpha estradiol seems to work just as well in late middle age. A former student of mine, now an independent researcher named Mike Garrett, together with his colleagues Charlene Day and John Herrera, have already published a paper using these same mice treated with 17 alpha estradiol, where they started at 16 months or 20 months of age. And they found that the muscles got stronger, their glucose tolerance got better, muscle structure changed for the better. So they published that before we could publish our lifespan data. But it's quite consistent. Even late start for 17 alpha estradiol seems to have highly beneficial effects. Just as good and in some cases even a little better than if you start in earlier ages.

Now, I definitely want to spend some time on acarbose and 17 alpha estradiol, along with a few others. But I want to go back to something in rapamycin. So we've talked about the first experiment. Tell me about the dosing. Was the dosing done daily for rapamycin? Yes, in that first experiment, sure.

It's in the food. They Eat all they want every single day.

So this is interesting, right, because what we've learned about rapamycin since that time is it inhibits two complexes of TOR complex one and complex two. Now, today, we believe most of the longevity benefits of rapamycin come from the inhibition of Complex I1, not Complex 2. And furthermore, we believe that some of the negative consequences of constitutive use of rapamycin come from its inhibition of Complex 2. And so when you look at some of the more recent human data using rapalogs such as Everolimus, they seem to favor an intermittent strategy. So dosing rapamycin, say weekly, which is enough to inhibit, if you give a big enough dose, it's enough to inhibit complex one, but not enough to inhibit complex two. And then by the time you come back to redose, you're in this situation of you're just knocking down one, but never two. So of course, I've always found it just interesting that the study even worked with daily dosing of rapamycin. What are your thoughts on that?

You raised five or six or seven or eight really interesting points there. So let's go back and do the one at a time. Rapamycin does not directly inhibit MTOR complex 2. What it does is it leads to feedback circuits which destabilize MTOR Complex 2 and eventually cause it to be degraded. So the subtleties of the kinetics of when to start it, when to start it, when to get one. So that's. And it may well differ from cell type to cell type. So those are things that pharmacologists need to work out carefully. The second thing to point out is that a guy in my lab, Gonzalo Garcia, has very carefully looked at Tor complex 1 and Tor complex 2 activity for the last three or four years in not in these drug treated mice, but in mutant mice that have one of two mutations, the Snell dwarf mutation and the growth hormone receptor knockout mutation that extend longevity by 30%. And Gonzalo found something very interesting. Both of those mutations move Tor Complex 1 down, that's good. But they move Tor Complex 2 up in the opposite direction. So this raises the possibility that is actually the elevation of TOR complex 2 in these mutant mice. It's good for them. And Gonzalo has found several examples of mice where there's a mutation that blocks the MTOR complex 2 elevation. They do not live a long time. So most recently, and this is a paper that's just submitted, it's not published yet, Gonzalo has taken these drug treated mice rapamycin or acarbose or 17alpha estradiol, and look to see what's happening there. With MTOR Complex 1 and MTOR Complex 2, there are changes with rapamycin, both males and females. There are changes with acarbose, both males and females, and their changes with 17 alpha estradiol. But it's the males only, and that's really cool because it's the males only that get a lifespan benefit from 17 alpha estradiol. So to summarize all of that, it's going to be really complicated. The basic idea that knocking down TOR complex one might be a good thing and that knocking down Tor Complex 2 might be a bad thing, that was the premise of your question. I think that's a good initial, sound foundation for further work. But the interactions between them, the ways in which different cell types may have different responses, it's going to be much more complicated than that. The last complication I'll point out is that Gonzalo has also found that EMTOR complex 2 has four different substrates. That was widely known before he began this work. What he found is that the good stuff turns on three of these targets for MTOR Complex 2 and turns the other one off. So it's really, if he's right, it's really a change, not so much in the amount of Tor Complex 2, but in its target specificity, which particular substrates it modifies and which ones it stops modifying. That's a subtlety. That may be the whole story there as to why changes in mtor complex 2 should be a part of any sophisticated study of this drug and other drugs.

When you repeated the rapamycin study one year later in cohort three, obviously, having worked out the delivery system, the average age at treatment was now nine months instead of 12 months. And obviously this was interesting because nine.

Instead of 20, you mean.

Sorry, sorry, nine instead of 20. So now this is interesting in that A, you're going to repeat a very important finding, and two, you're going to see if the age effect was significant. And it's kind of remarkable both groups got a little bit better, which I guess you would expect. But it's worth noting you still see these remarkable effects, right? You saw a lower median extension because of the point that you mentioned earlier. But the maximum extension in males went from 9 to 11%, and in females it went from 14 to 16%. So again, females had an advantage with rapamycin. But both of these numbers are really quite impressive. I mean, these are pretty significant increases in. In lifespan. And you're now functionally starting this in people in their early 30s. Basically, yeah.

So there are two things to say. One is that the curves for the early start 9 months or late start 20 months were compared very carefully by Scott Pletcher, who's a mathematical biodemographer, and they're indistinguishable. That is, there's no statistical difference between the early start mice and the late start mice. Small differences like 9% versus 11% or 9% versus 12%. There's a lot of variance from mouse to mouse and group to group. You don't want to take minor changes like that with put too much weight on them. They're more or less the same thing from the point of view of replicability. The other point you made that we referred to a moment ago had to do with the male versus female differences for rapamycin response. Randy Strong and Marty Javers and their colleagues at the University of Texas did a really useful study. They gave rapamycin for a short period of time to male and female mice and then took blood samples every, I don't know, every hour, every two hours or something. And the rapamycin blood content in the female mice was two or three times higher than it was in the male mice. So it was higher and it stayed up longer. We don't know why rapamycin in the blood of female mice is greatly elevated compared to rapamycin in the blood of male mice. But when you're comparing two curves at the same dose in food and the females do a bit longer than the males, get a bigger benefit than the males, it's hard to know how that comparison would come out if you're actually adjusting them to the same blood level.

Have you seen that with any other drug rich where going into the food, same quantity going in, you see a 3x difference in plasma level?

No, we haven't looked very carefully. I mean, we did it for rapamycin because we were so surprised that the females were doing better in the males. And it's a difficult and expensive study to do. I think we should do it for other drugs as well. We, Marty and Randy do check the drug concentration as a part of a pilot project of mice given any one of our drugs, they're given the drug for eight weeks of age. Then Marty and Randy get blood samples, but that gives you just a sort of a steady state dose. What if you're eating it every single day for eight weeks? What is the blood level? The study that was most informative for rapamycin was to take Fresh virgin mice and give them the drug at a defined date, at a defined amount, and then quickly measure the blood levels after that. That has to be set up with a separate cohort of mice. If you want to be obsessive, as you should, you have to do it at young mice, middle aged mice, old mice. It becomes kind of involved and it's. We don't do it on a routine basis.

So the next drug I want to talk about. Well, actually let's go to acarbose. You brought it up. So let's go to that before we get to resveratrol. I mean, it's amazing. Like Rich, we could sit here and talk for the next month and go through every single drug in the ITP history because there's not one of these that isn't interesting both in its success and failures. Obviously we don't have the time for that. We're going to, in the show notes to this put out a spreadsheet that covers every single single result. But nevertheless, it's just an amazing body of work. So let's talk about acarbose now. Was David Allison one of the first people involved in that? I know David very well and I know that David's been involved in a number of suggestions around the ITPs. He sponsored some. Was he originally involved in the acarbose work?

Yeah. The acarbose application came in from David Allison and a colleague of his, Daniel Smith.

Okay. This is an off the shelf drug typically used in people with diabetes. And it basically blocks the absorption of glucose in the gut. More or less well tolerated, unless you take too much of it and in which case you're going to get some GI distress. But it works. Well, I've certainly taken it a number of times. I've got bottles of it laying around the house. I probably haven't taken it in a couple of years, but I used to use it as kind of my cheat meal drug. If I wanted to have a pizza, I'd take 100 milligrams of a carbose and it would manage to to prevent my glucose from spiking.

They should give it away with every pizza. I agree. That's a good marketing strategy.

There you go. Take your a carbose with your carby junk food.

Sprinkle it all the pepperoni.

Now, of course, what you guys were saying is, hey, we're not just giving the mice the a carbose with their pizzas. They're going to eat this stuff every day. Right. It's part of the chow Yep. Okay, so what was your thinking? What was going to happen here?

Well, Daniel and David said caloric restriction is good for you. A carbos is sort of like caloric restriction. It doesn't block glucose absorption, but it does block the digestion of starches to sugars and so glucose wouldn't go up so much. That would be good for you, just like caloric restriction in their view. And that was their rationale.

Sorry, just to be clear, I understand that because there's two different hypotheses that you could be testing here. One hypothesis is you will functionally consume or absorb fewer calories and therefore this will be a CR play. The other is no, the animal will continue to eat. You know, they'll make up for it with more calories, but they will have lower glucose. Which of those two was the thinking here?

I don't think Daniel and David did a particularly compelling job of discriminating those two ideas. The sort of framework for their application, which, you know, I haven't read for what, 12 or 13 years now. I'd have to go back to see exactly what they argued, but they were saying it's sort of like caloric restriction a little. Why don't we try it now? It's not like, as it happens, it's not like caloric restriction. There are a lot of reasons for saying it's not reproducing caloric restriction. But anyway, that was the original rationale. My current interpretation is that it probably is operating by blocking very highest levels of glucose in the mice did not lead to a change in the integrated glucose level. There's a clinically useful measure which is used in human diabetics too. Hemoglobin A1C, which gives you a measure of, over the last few weeks, how much average glucose has been in the serum. If a person with diabetes takes a carbose that hemoglobin A1C goes down. That's one of the ways in which, you know, it's working in a person. Now in mice, it turns out it doesn't go down. We've done this twice now, so it's probably not an overall change in the amount of glucose that gets in, but a change in the kinetics. If I eat five slices of pizza, my blood glucose is going to shoot way up and then come down. But if I've taken a lot of acarbose first, blood glucose will go up, it'll just go up slow, it won't reach that big peak. It'll stay up longer overall, but it won't hit the big peak. And so currently our guess was mine and the other colleagues was that it was working by blocking the peak glucose. And now we think the evidence for that is quite good. Because of a drug you haven't mentioned yet, canagliflozin, which was just published a few weeks ago. Canagliflozin, which is also used for diabetes in people, also blocks peak glucose in mice and in people, and it also extends lifespan and it also works preferentially in males. So it's a reasonable guess that both acarbose and canagophosin are working by eliminating the huge peak of glucose you get after you eat a meal with a lot of starch in it.

And of course, David also was part of the suggestion team for Canna. So I know that David and I have spoken at length about SGLT2 inhibitors for quite some time and we are absolutely going to get to that. So did the mice experience any GI distress in any way that you could assess that?

None that they complained about.

They didn't fill out the questionnaire.

So we don't have any way of detecting how the mice are feeling. They didn't stop eating or show obvious signs of distress. You know, their GI tract and our GI tract are very different. We're omnivores, grew up as scavengers 200,000 years ago. We'll eat almost anything, animal or vegetable. And mice are evolved to be interested in mostly a grain diet. So direct one to one comparisons for what our GI tract can handle and theirs may be kind of tricky.

Did those animals, both the treatment and the control groups, weigh the same?

No. Acarbos led to weight loss or a lack of weight gain. We don't understand it. We see that in some populations, but not all. In the original ITP paper, the acarbose treated mice were lower in weight than the control mice. But then my own laboratory made a whole batch of acarbose treated mice for another purpose and we didn't see a dramatic change in weight and we really don't know why. So it's a little bit frustrating and a little bit mysterious and embarrassing. We don't always see the same weight change.

I assume the rapamycin treated in controls did not have a difference in weight.

It's complicated and it depends on the age at which you start. If you start in youth, then when the mice are still gaining weight. The mice on rapamycin do not gain weight as much as the control mice, which does not surprise me because rapamycin stops or slows cell Growth and cell division, etc. If you start in late age, like 20 months of age, you have a whole series of factors going on. Mice, when they get old, they often get sick. When they get sick, they often lose weight. And if rapamycin is extending their lifespan, then they're getting sick later. So they're not losing weight because they're healthier. And there's several different things going on in old age at 20 months of age where rapamycin in principle would increase some of them, decrease some of them, it's harder to interpret. Whatever you see, before we leave acarbose. I mean one really important point that I'm sure you're aware of, but I think deserves some emphasis. Since Acarbos is FDA approved and has long safety history, hundreds of thousands of people take it, particularly in Asia and in some parts of Europe. It's an obvious candidate for human clinical trial. The safety profile is quite good. It's very well documented with many cases having been evaluated for safety. So if one was thinking about what drug that works pretty well in mice would be safe for a human clinical trial, a carbos has to be on that list. I've been arguing with your friends. My friend too, Matt Kaeberlein, he's giving rapamycin to dogs. As you know, that's a very important, very interesting, excellent study and I've been trying desperately to get him to give a carbos to dogs too. And we're still not yet having reached a consensus on that point.

Let's come back to this point specifically, Rich, when we get to metformin because to me the metformin a carbo story is a very interesting story. One as we suss out the difference between very compelling human data and failures in the ITP versus successes in the ITP absent the out of the ballpark human data. And I think that the acarbose metformin story is a great example of that. So let's park that, but let's visit now. Resveratrol, right? So for reasons that aren't entirely clear to me, I mean, I suppose it's because of the relationship between resveratrol and grapes and wine. This just was one of those things that caught the world by storm and it's never really gone away, right? So if I had a dollar for every time I got some stupid Google alert telling me to drink more wine because of its anti aging benefits, I'd have a lot of dollars. So David Sinclair, I don't remember the original paper would have been 0304 ish.

You probably know this better than I do, but that sounds about right.

Yeah. So less than 20 years ago, basically, David's lab at Harvard published work showing that when resveratrol was given to metabolically ill mice that were being basically overfed, it produced a longevity benefit, sort of.

We can come back and examine that claim in more detail if you wish.

Yeah, well, let's do that now and then let's talk about how you. How it informed your design.

Yeah. So the mice in question were being given a diet consisting of 60% coconut oil. They were being poisoned. The original paper did not mention what they died of. They did mention this in a subsequent paper published in Cell about two or three years later. The mice that were dying because they were on 60% coconut oil were dying because their livers got so big, so filled with fat, that it compresses the chest cavity and crushes the lungs and the mice cannot breathe. That's the cause of death in mice in this particular study that are given a 60% high fat diet, so to speak, that was used in the original resveratrol paper. So understand, they're not really studying aging. They're studying a bizarre pathological process where the liver gets so fat that it crushes the lungs, preventing breathing. And then in addition, the paper that they published reported median lifespan. There was a statistically significant increase in median lifespan, but only in the animals that were on this highly toxic diet. Now, it turns out that when you have the whole curve and you look at other indices of maximum lifespan, the resveratrol was not a benefit. And when you look at the lifespan effects in the mice that were not on this toxic diet, there weren't any. So paper in nature which was widely misinterpreted over interpreted as a demonstration of a drug that slowed aging, was a drug that did not extend maximum lifespan, except in mice that were dying of this extremely unusual lipid specific poisoning. The second part of your question was, how did this influence our decision to test resveratrol? The answer, this is behind the scenes gossip, but it's completely true, is that we were ordered to test it. Richard Hodes, the director of the National Institute on Aging, was very impressed with resveratrol. And like you, he was getting, I'm sure, hundreds or thousands of questions a year requests like, why don't you guys test resveratrol? So resveratrol did not go through our usual screening process.

This was a directive from the top.

This was the only time this has happened. We were instructed, you will be testing resveratrol. And I called David Sinclair and said, what dose should we use? The same dose that you used in your paper. And he said, no, no, no, no, no, no, that's much too low. Use at least five times, 10 times, 20 times higher. So we followed David's advice and also checked with his close friend and colleague Rafa de Cabo to get advice on dose. So the two doses that we used for resveratrol were. I'd have to look it up either three times and ten times higher than in the original Sinclair paper. Maybe it was 10 and 30 times. I have to check my notes to get that exactly right. But we use both doses, both high. And then we started at two different ages because we were told to do start some of the mice in youth and others of the mice in middle age. And we did that and had no effect on longevity. And we were, I think, the first of three or four groups, including Sinclair and Dicavo later on to show that resveratrol given to mice on a normal diet does not extend their lifespan.

Yeah. So you, in Cohort 3, you started at 12 months of age. So functionally these are mice in their 30s to 40s, and then you had a cohort that started at four months of age. You're starting these, you know, in teenagers, no change in median or maximum lifespan. Although I believe you had a small, statistically significant increase in maximum extension and maximum lifespan expansion in the late onset resveratrol by 3%. Does that sound about right?

I don't think any of them were statistically significant. Again, I'd have. I don't want to put my hand in a bible until I've reread the paper. But I think there was no change in any of the statistics.

It clearly speaks to the point earlier, which is in the spirit of looking at this through the lens of studying longevity as opposed to studying protection from a sort of artifactual diet. This appeared to be pretty clear. What was the response like to that?

There are people who are, who were making a living by selling stuff that was related to resveratrol. There was a company, actually Surtris, that was sold to Glaxo for 600 million bucks for pushing sirtuin inhibitors. So the reaction from people who had just sold the company was probably, I don't care, I've got the money. The reaction from people who sell things in health food stores which are guaranteed to raise your resveratrol levels or that have purpley grapes on their cover, they were raking in money Hand over fist. They don't really care who reads papers. Failing to show unaffected mice, it's not going to be a problem for their bottom line. You know, I was always very irritated, over and over again. Pretty good scientists whom I really respected. The first slide was a bottle of wine. The second slide with some nice grapes. The third slide was a picture of resveratrol. Now, by that time, they knew that resveratrol was not an important component of wine. Sinclair had written a beautiful review article showing that if you really want to eat resveratrol, it's easy. You eat rhubarb. Rhubarb is where all the resveratrol is at. And red wine has got so little that to get the mouse dose of resveratrol, you'd have to drink 600 bottles of wine a day.

Even Andre the Giant couldn't have done that.

The beautiful satirical article in the New Yorker about what it's like to be a mouse drinking 600 bottles of really good French champagne a day that I said to friends when they ask about resveratrol. So I think eventually there's been a lot of controversy as to whether the sirtuins are important or not in the aging process. There's a big fight between two worm groups, one of which said mutants that modified the activity were highly effective. And others said, well, maybe just a little bit. Little bit. And leaving all that stuff aside, I don't think there's any evidence yet that resveratrol is good for you or that sirtuin activators slow the aging process. I was invited to give a talk at Searchers.

Interesting.

At the height of all of this, David Stipp, who's a wonderful science journalist, has written a wonderful book about the whole surtris resveratrol story.

The youth pill or the youth fountain or the youth pill or something like that.

That's about right. Yeah. It's a great book. David's a fine writer. So I went to Search Risk. I gave my talk on rapamycin, which is what they wanted to hear. And at the end, I went to talk to their director, and I said, so what are you guys doing for resveratrol? Is this going to work? He said, oh, well, no. I'm leaving the company. I'm going to work for a company that works on. And he named another chemical, highly different. He was going to work on TOR inhibitors. He wasn't going to work on resveratrol. They'd already sold the thing or about to sell the thing to glaxo and he knew that that was tied to move on to inhibitors or compounds that had a greater likelihood of actually working.

Well, we're gonna come to one of those in a little while, which is basically things that provide substrates for sirtuins as opposed to activators. But nevertheless, one of the other ones that sort of I thought was interesting that you studied, not that you shouldn't have, in the spirit of this being kind of a community effort, was green tea extract tract. Does that. Do you remember that one at all?

I sure do.

Tell me a little bit about that. What was your expectation going into that itp?

Well, anybody in the world can suggest a compound or a group of compounds or a mixture of compounds.

You realize I'm going to be doing this now, you're going to be getting these suggestions from me, so just.

Oh good, yeah, that's good news. I'm sorry. In advance, each of these suggestions is evaluated by a group of five scientists. Initially our access committee and then secondly, following their advice, the steering committee, which I'm on, makes the final cut. So when people are recommending green tea extract, they say there's a central compound that's probably the main player in green tea extract. Many people have said it's good for you, traditional Chinese medicine, etc. Please test it. And so when we evaluate this, there are all sorts of pros and cons. We strongly prefer single compounds, purified synthetic compounds to complex mixtures. But we can see sometimes a ground for testing a complex mixture. It may be that five or six or seven things in balanced amounts might together be good for you, even if no one of them used singly might be beneficial. We also like to test things that have acquired a popular reputation for being healthful. If the popular reputation is even if it's based on a thousand year old tradition of traditional medicine, it still might be right? And so some of the time if we stumble on something that will work, that would be great news. And if it doesn't work, then we feel we've done a public service. We certainly have not proven that green tea extract might not be good for people, but it'll, we hope, slightly take the wind out of the sails of that discussion. If we demonstrate when we give it to animals, it didn't do any good to them.

The other one that caught my eye, by the way, on the much longer list of failed compounds, and you made a good point, right? Which is the failures are just as important as the successes, right? I mean some of the most important insights are these things that don't work. That we've long held beliefs around Methylene blue caught my eye. I was sort of interested to see methylene blue curcumin, many of these other things that have been tested. You go back and revisit rapamycin again through a high, medium, low dose. I mean, you really put the screws to rapamycin to continue to demonstrate its efficacy. Really. Rapamycin is in some ways probably one of the three poster children of the.

ITP4 now counting Canagoflows.

Yeah, yeah, yeah, exactly. Let's talk about 17 alpha estradiol. I know more about the estrogens than the average person. I gotta be honest with you. I don't have a clue what 17 alpha estradiol is. I'm very familiar with 2, 4 and 16 hydroxy estradiol and estrone and all these other things. I can't for the life of me figure out what 17 alpha. Can you help me understand where this shows up? Is there an actual molecule, is there a drug that we have? What was the genesis behind this plan?

Yeah, the genesis was a guy named Jim Simpkins. He's a pharmacologist. He's always interested in estrogen receptors. And he tried to make a compound that was like estrogen, I.e. 17 beta estradiol, generically crudely referred to as estrogen 17 beta estradiol, but which had much lower affinity for the classical receptors for the estrogens, er, alpha and beta. So he synthesized 17 alpha estradiol, which is very much the same compound except for one of the bonds tilts up instead of tilting down. And in Jim's work, mostly in tissue culture cells, he found that, yes, it was tenfold less active at least than 17 beta estradiol. And in fact, whatever activities it was having on his cells and culture, it worked just as well if you deleted the classical estrogen receptors. So when he gave it to mice, he found it was non feminizing. When you give it to male mice, it did not induce the secondary sexual characteristics that the stronger estrogens, like 17 beta estradiol, estrogen would do. So his idea, and I'm not endorsing this, I'm just reporting it, is that giving an estrogen to male mice might make them live as long as female mice. The female mice live 5 or 10% longer than males. He said maybe it will turn them into females, but it would be a much better compound for people because no man wants to live look feminine. So if it works, people will give it to men and they'll live as long as women, but they won't be embarrassed by looking like a girl. That was enough to get it through the steering committee. And when we gave it to mice, it indeed had a terrific effect on male lifespan. It had no effect whatever. Now, in three separate studies in female lifespan, in fact, the males given the drug live not only longer than regular old males, they live longer than regular females. It's not merely moving the male curve to be coincident with the female curve, it's moving it way past where the normal females live. So why it works only in males, we have a clue, but we don't know for sure. And we certainly want to know why the effect is so dramatically sex specific so that we can try to figure out a way to get around that obstacle and to get it to work in females as well.

Now, if I'm looking at my table correctly, the 17 alpha estradiol in Cohort 5 Extended Median Survival in males, but not maximum.

And of course, no, it does. Well, that was used at a low dose. You're right. At a low dose. When we went back and we did it at a three times higher dose, it extends our measure of maximum longevity as well as median longevity and at all three sites.

Yeah, I see. Okay. So that's the difference is at the 14.4 parts per million versus the lower dose.

Yeah. When we got the first result, somebody said quite plausibly, hey, I'll bet it would work better at a higher dose. So we tried that.

Now, again, I just don't know enough about mice, but what's the menopausal situation of mice? How long are you the female mice exposed to estrogen to 17 beta estradiol.

I don't know about when the estro cycle ceases in mice, it depends a lot on the strain. As you know, dozens to hundreds of inbred strains, and they each have slightly different characteristics. In general, production of new healthy pups tends to slow down when the females are 8 or 9 or 10 or 11 months of age and the cycles start to get irregular and then to stretch out and then decease altogether. But this is something I don't know very much about, and to quote details, I'd have to talk to somebody who actually knows a lot more than I do about mouse reproductive endocrinology.

Is there an actual molecule of 17 alpha estradiol? Is there an IND for this?

Very good question, and the answer is not really well known. There is one obscure paper by a Swedish or Finnish group maybe 15 years ago that said they had detected 17 alpha estradiol, but only in the brain. And in fact they said that they had isolated in the brain a receptor which they called the estrogen X receptor that was relatively specific for 17 alpha estradiol, the one that's working in our tests. But I have never seen that repeated. I actually, following up sort of your line of thinking, corresponded a little bit with two or three of the American scientists who are authentically steroid chemists and who spend their whole life synthesizing various derivatives. And I kept, I asked them what is 17 alpha estradiol? Is it anywhere? Does it do anything? And they all said, good question, I don't know. So. So I know nothing about it. But they should know something about it. And apparently it's distinctly under investigated.

So even after your findings, which are quite impressive, it hasn't been pursued as an investigational new drug.

I'm over my head here. I don't know the answer to that question. Certainly nothing has come to my attention about its being investigated. But I could well have missed something. There is an awful lot of, of attention to developing novel estrogenic compounds. Really good academic labs and I'm sure 10 times that much work in pharmaceutical firms. There are major papers in important journals about here are 30 or 40 different steroid compounds that do not have that side effect or do not have this, do not bind to this receptor, etc. I don't know that literature very well and I'll bet a great chunk of it is kept quiet by pharmaceutical firms. So I'm not confident that anyone outside those firms really knows the answer to your question.

Interesting. Do you remember the rationale for ursolic acid?

I will have to double check. I believe this feels almost exactly like my qualifying exam.

Sorry.

Ursolic acid I think is the one that changed retinoid receptor activation. There was a claim from Gretchen Darlington and some of her colleagues saying that the long lived mice, the Ames dwarf mice, had turned on a whole bunch of genes that related to bile acids. And the bile acids work through some molecules related to the retinoid acid receptor or something like that. Ursolic acid was proposed as a way of getting at the bile acid issue. But the chances that I'm right on this point are only about 50. 50. I'd have to check my notes to be sure.

The reason I ask is there was a human based supplement for this that was all predicated around improving body composition. But my one recollection of it was it was not orally bioavailable. It was only you had to use it in sort of a topical fashion, which sort of rendered it not particularly helpful. Rich, can you tell me a little bit about the work being done on hydrogen sulfide?

There's a colleague of mine who died recently, a guy named Jay Mitchell, a wonderful scientist. Jay was really interested in the idea that hydrogen sulfide might be an important controlling element in the aging process and had published a long and really impressive series of papers on that. So he suggested that we give to mice a drug that would break down to produce hydrogen sulfide, a drug called SG1002. So we tried it, and we messed up. We had control mice that were on the wrong control food, and the mice given Jay's drug did better than the mice or the wrong control food. But we didn't publish that because we were not sure what would have happened if we'd use the right control food. So we're repeating that now. We've fixed our mistake, and we don't have data yet. In about a year, we'll know whether Jay's hunch that giving mice a. That would increase their internal generation of hydrogen sulfide will be good for them. I'm certainly hoping it's correct. When we did it wrong, the drug seemed to work. So I'm hoping that when we do it right, the drug will continue to work, but we don't know yet.

Okay, so let's go to one of your more recent findings. This SGLT2 inhibitor.

Canagliflozin. Yeah.

Yeah. So this is a drug which is widely available. This class of drug has been around for, what, probably a decade, right?

I think that's right, yeah.

Used pretty readily in patients with type 2 diabetes. Blocks the reuptake of glucose. So that more glucose is excreted in the urine.

Yes, but only when you're hyperglycemic. One of the nice things about it in clinical practice is you can't make somebody hypoglycemic and put them at risk that way. So it affects the process in the kidney that deals with very high glucose levels without having much scope for causing toxicity.

Now, this increased median lifespan in the male mice by 14%, maximum extension of 9%. It had no effect on the females.

That's right.

Is this another one of those drugs where you have any insight into differential plasma levels between the males and females?

We actually did measure that, and it's in the paper. The females, if I recall, actually had higher levels, slightly higher levels than the males. The absence of an effect in females is not explicable on the idea that they don't get the drug into their blood.

What do you think could explain the difference in the effect here?

Gosh, I wish I knew. When you take this together with the acarbose results, you're led to the inference, and this is only a tiny baby step forward. Something about aging in the male mice depends a lot on staying away from really high glucose levels. Now, whether that means that high glucose in the males triggers a circuit in the hypothalamus, which is bad for you or something. I mean, from here on in, it's all hand waving. The mice are mostly dying of cancer, both the males and females. About 80% of the deaths are due to some form of cancer. A wide range of different kinds of cancers. But cancer in these mice and most mice is the predominant cause of death. So you could probably start to think about why is it, if you want to avoid all sorts of cancers, high glucose is a really bad thing for you if you're a male? That's the sort of line of argumentation you'd want to consider. But that's certainly unsatisfying, and we don't know. We would love to know that.

Well, to me, why high glucose is bad for cancer is a relatively straightforward question, at least compared to. To why disproportionately for males and females. So let me start by asking another question. If you do nothing and just observe these animals as though you were an actuary, what is the distribution of death for females versus males?

You mean the age of death?

No, the cause of death?

Yeah, about 80% cancers in both. Oh, it's 80% in both different kinds of cancer, Both. The leading cause for both males and females is hematopoietic cancers of one sort or another. Lymphomas are related histiocytic sarcomas and other sorts of lymphoid tumors. Then in the females, the next leading cause is mammary cancer, and the third leading cause is either liver cancer or fibrosarcoma. In the males, the second is lung cancer cancer, and then liver cancer or fibrosarcoma. So all of them. It's cancers, cancers, cancers.

And how much is hematopoietic in each?

About 30% of the deaths.

Okay, so that's hard to make sense of, I guess. I mean, are there enough data on the glucose and insulin sensitivity of each of those relative types of cancer that you could make the case that. I mean, I gotta be honest with you, that's a little counterintuitive based on human data, because obviously, in humans, the mammary cancers Would be quite glucose and insulin sensitive, Whereas the lung would not be. So you would actually expect the opposite if that were true in mice, that you would expect the females to disproportionately have the benefit. But I mean that's a lot of hand waving, right? So I'm just curious if we have more molecular data about PI3K activity or other activity IGF activity in these animals as it pertains to their cancer.

These are really good questions that I've been trying to urge my cancer biology friends to start digging in and answering those questions As a sort of the framework you can think as you were thinking. Maybe the glucose thingy works on the cancers, but it's also possible that the glucose excursions are working on an anti cancer defense, Some aspect of immune defenses. Or it could be, and this is my hope, but it's only a hope. It could be that the glucose specific excursions are modifying some fundamental, still undiscovered element of the aging process in the hypothalamus. Maybe it has to do with the susceptibility of the hypothalamus to inflammatory change or something like that, differentially in males and in females. And that this nebulous sort of change has an impact on the cancer or an impact on anti cancer deficit. Any of that is in my view about equally likely to be true. But at present none of it gives us much of a hint as to why these drugs have a much more striking effect in the male mice.

Did you see the same effect here that you did with acarbose in that the hemoglobin A1C was unaffected?

I'm not sure. I'd have to go back and check the paper. My memory, which may be incorrect is that there was a figure showing no difference, no effect on hemoglobin A1c in the Canangal flows in paper, which is surprising because it is affected in human diabetics. You can really move hemoglobin A1C into the down into the healthy direction by canagliflozin and the other different SGLT2 inhibitors. But if I recall it did not produce that effect in the mice. We don't know whether the canagula flows in effect is on glucose. We don't know whether it's on SGLT 2. It also has an effect on SGLT 1. And there are now reports appearing in the literature, Quite striking ones, Some of them in clinical settings showing effect on tumors, specifically on other aspects of other kinds of disease. Human effects on heart attacks, independent of whatever's happening in terms of glycemic control. So the shortest, simplest explanation, its effect in glucose, may not be right. There are other ideas that deserve a lot of exploration.

Yeah, we are very excited about this class of drugs clinically in our practice, the cardio protection, renal protection. Obviously the glycemic benefits are all pretty exciting. And of course, this brings us to an interesting contrast. Right. I think if. If Rapamycin is one of the most remarkable success stories of the ITP in terms of its consistency. Personally, for me, the biggest surprise of the ITP is the failure of metformin. Can you say a little bit about that? How was the study done? What are the possible blind spots? I mean, obviously metformin succeeded when paired with rapamycin, but you could argue that's just rapamycin. But metformin alone did not succeed. And has it only been run by itself once or twice?

So I'm not surprised. There are all sorts of reasons in which metformin failure is unsurprising. One is it might be really good for people and not good for mice. Mice and people are organized very differently. It's also possible that we use the wrong dose. If we use a dose that was twofold lower or twofold higher, it might have been great. Or if we had given it for a few months and then taken it away for a month and then given it for a few months, that might have worked as well. You know, we tried it at one dose with one continuous dosing schedule, it did not produce a significant effect. But that's not to say that it could not work in mice. And even if it could not work in mice, that's not to say that it might not be good for people. So there are all sorts of ways in which you might get a disparity. You asked how many times metformin has been tried. It's been published twice, once by Rafa de Cabo, once by us. In our study, it led to no significant lifespan extension in either males or in females. In Rafa's study, which I'm about to be critical of, it produced no effect in females, and Rafa alleged that it produced a significant effect in the males. Now, the p value was 0.046, so it was really close. And it was. It was not the test that he ordinarily uses our lab, and almost everybody, including Rafa, always uses the log rank test. And he didn't use the log rank test. Instead, he picked another test which gives special weight to early deaths. And if you use that special specific test, then it is statistically significant at a P value of. Of 0.046.

My hunch is it didn't work with the log rank.

Yeah, it tried, it failed. So they went statistics shopping, found one where it got to.046 and published that. I have asked Rapha for the data privately two or three times. He's always agreed to send it to me over the last three or four years, but it hasn't actually arrived yet.

Well, maybe this podcast will help speed that up.

I hope so. In any case, whether or not his was just over or just under the border of statistical significance, it was a small effect at best. And in itp, it didn't work at all. But it might work, might be good for people.

Yeah, well, I guess that's the challenge here, right, is you've got these agents that work under so many different circumstances. And again, you can probably tell Rich, my bias is I think rapamycin is an amazing agent here. And I. And I think it's one that warrants lots of investigation in humans because I think that the non human literature are so compelling, both in terms of the actual drug use and also the mechanism of action and the genetic manipulation all the way from worms, fruit flies, obviously mice, yeast, Matt Kaeberlin's dogs. It just over and over and over again demonstrates efficacy. And even when you look at the softer endpoints in humans, softer is maybe the wrong word, but less longevity specific, but things like immune function, things like that, it always seems to be pointing in the right direction. Now, in metformin, we have this undeniable data of diabetics that take it versus diabetics that don't. And you can slice that 10 ways to Sunday, it always seems to favor metformin in that diabetic group. But as Nir Barzilai and many others are interested in, what does that tell us about non diabetics? Is there a way we can better get at that answer while we await the results of a TAME study to get funded and executed? Is there another ITP we should be doing?

Well, the principal stimulus for, at least for me, and I assume for those who've been spending their time developing this test of metformin in aging people. The TAME study was a paper doubtless familiar to you, an epidemiological paper. The point of the paper was not so much that metformin was good for diabetics, which everybody knew it was, it was that the mortality risk of diabetics on metformin was actually better than non diabetics of the same age and sex. So that doesn't prove that metformin is good for non diabetics, but it's certainly a hint in that direction. Now, it was an observational study, not a controlled randomized clinical trial, of course, so it might not be right. But that was a really big hint that if you're a person, non diabetic, and you're not taking metformin, your chances of death over a wide range of ages are actually worse than a diabetic person as long as the diabetic person is being protected by metformin. So my view, even though it's a single study and others now need to be done, and all sorts of caveats need to be kept in mind during the interpretation, that's a hint that metformin may well be pretty good for non diabetic people as well. And studying that in terms of a randomized clinical trial strikes me as justified. It's very expensive, it's very ambitious to meet the FDA criteria. And in order to try to get results before they all become emeritus or retire, they need to have endpoints that can be measured five to, oh gosh, as many as seven years in the future. So for human lifespan, it's kind of a short term, but it may work. And it's not my money and I can't wait to see what they come up with.

So there's one more compound I want to talk about which by the time this podcast will come out will be at least in a preprint, and that's nicotinamide riboside. So I guess it's safe to say that that if resveratrol was the in molecule of the early 2000s, the next big in molecule has been anything that points to nad. Again, I don't know how many times I have to get an email or a text message saying, hey, what do you think about these NAD IV clinics? I really have heard great things about this. And what do you think about NR and NMN and these sorts of things? So. So I guess I'll give a really, really, really fast. Or actually I'm going to let you give a really, really, really fast explanation of what NAD is, why it might matter, and why we give NR orally as a precursor to it.

I'll be glad to. I feel a little embarrassed because you know 10 times more than I do about this, particularly in a clinical context. If you had to make a list of the 10 or 15 or 20 molecules that are part of intermediary metabolism, that may well be really important elements in the control of Aging rate. NAD and its chemical derivatives would surely be on that list. It's at the top of your list. For some people, it's sort of in the middle, for others at the bottom. But it deserves a lot of attention. And there's a great deal of pretty strong data suggesting that aspects of aging and age sensitive diseases, age associated diseases, can be altered by making NAD more or less available. One of the reasons it's gotten so much attention is because of clever marketing. There are scientists who have made a great deal of money by making more or less unsupported. You can't sue them. They're not lying exactly, but they're sort of winking and nodding, saying, hey, take this formulation. It'll sort of do some goodish stuff to your nad. We're not allowed to say we'll treat any disease because we would run afoul of the fda. But wink, wink, try it. It's a nutritional supplement and they've made a lot of money and gotten a lot of attention. Some of it positive and some of it negative. The reason that NR nicotinamide riboside was recommended to us by a company that wants to sell it is that it's orally bioavailable and more stable than some of the other ways that have been proposed by mao.

Can you disclose which company asked? There's probably only two or three public.

I think it's Chemadex. I think it's public knowledge is who suggests things to the itp. So they said, well, we'll provide it to you. You guys test it. You can publish results, positive or negative, just in the interest of science. Let's find out. That's a highly honorable and sensible thing for them to have done. So we tested it. It's a bioavailable form. We used a dose that they suggested and the. The paper that will have come out by the time this podcast becomes available suggested that NR did not extend lifespan in our mice. There are all the usual caveats. Maybe it would have if we'd given it at a higher dose or at a lower dose. In addition, we tried a little bit to detect it in various tissues. There's a collaborator who looked at the brains and the muscles and the hearts, and we were not able to demonstrate a consistent major change in NAD levels in these tissues. NAD is a highly active molecule, as of course, you know, that turns over constantly and with a time scale of a few seconds to a few minutes, can go way up or way down in given cells, depending on how stressed one is or how high the glucose is, or whether you're exercising, all that stuff. So it's possible that we fail to measure major changes in NAD in our treated mice because we should have fasted them, or we should have let them rest for a few hours, or we should have gotten the blood out in a different way or something like that. But in any case, the observations that we're confident of at the dose we used, it did not produce a lifespan effect.

Did it produce any benefit?

Well, we would have to go and check for other things. Right. The first time we do a study with a new drug, lifespan is the only thing we measure. We would love to be able to measure immune function and vision and hearing and cardiovascular function, et cetera, but those are expensive. And if we did those, we would have to spend money on them and cut back by a factor of whatever two or three or five, the number of drugs we test. So we generally reserve those health outcomes only for those drugs that give a lifespan benefit. And NR did not, in our hands, give a lifespan benefit.

You started the NR at about 8 months of age. So again, this is a reasonably young animal in its twenties, the equivalent of a human. Do you have a sense of what the dose was on the scale of where humans are taking this commercially? I believe humans are typically. Oh, God, I don't even know. Taking probably a couple of grams of this stuff a day.

I do not know. I'd have to look it up.

But the company basically had you utilizing a dose that they felt was consistent with what humans are being encouraged to take.

We followed their recommendations. We picked a dose that they said would be an appropriate, good starting dose for our tests, and we followed their guidance on this. It's a little tricky when you're going to give something for lifespan. If they're doses that are often of drug X, I'm not talking about nr, but doses of drug X that are typically given to people for a month or two, and it's really good and doesn't produce side effects. You still worry that might not work for mice because if you're going to give it to them, the whole lifespan, side effects may accrue. So if there is an ambiguity, we want to start usually with a pretty low dose, because if they're going to be on it from 4 months of age until death, keeping the dose nice and low is probably an important precaution against unanticipated side effects that might not be seen in an acute.

How often do you guys see unanticipated consequences of exactly what you described? Where Side effects don't show up in the first year of the mouse receiving the drug, but show up later.

Well, we have not yet, and this surprised me somewhat, we have not yet found any drug suggested to us that caused a shortening of the lifespan. None of the drugs that we've used has caused a significant decline in lifespan. There were one or two where the lifespan went down 5% or 8%, but it did didn't achieve statistical significance. Now, there could be some side effects that we don't look for. Maybe the mice have problems with hearing or sight or memory or muscle function or something like that, which we would never see unless we look for them.

It really seems that there are three really important takeaways from what are you at now, 18, 20 cohorts of ITPs. I mean, almost 100 studies here, basically, or 100 experiments, which is mtor matters. Less glucose is better than more and sex specific steroid hormones probably do something relevant, right? I mean, is that if you were at a dinner party and someone asked you to give them the most salient findings of the last two decades, would that be a fair assessment?

I think the points you're making are really good ones, but it's not quite the way I would put it if I were allowed to pick only three.

Okay, give me your three. Yeah.

The first point I'd make is the one we began at the beginning of this conversation, which is that by gosh, you actually can put something in the food that extends healthy lifespan and it's an enormous effect, ten times better than a cure for cancer. So in terms of the fundamental imagining of where medical research ought to go and what people who want to make, make their friends stay alive and healthy longer should think about, our program shows that what they should think about is finding drugs that slow aging. It seems obvious, you know, you and I are talking about it because we sort of believe that we sort of think it's cool and interesting and important. But that's most of my friends, scientific and non scientific. They haven't decided that that's the case. So I would take that as the first, most important take home message. The second take home message would be, to my mind, the really surprising finding that some of the time, maybe even most of the time, these drugs work even when you start them in middle age. I would have guessed against that when we started. I would have said you have to start in youth, period. But that was a bad guess. And two drugs that we've tested so far work just as well in middle age, and one works half as well. So that was a bad guess. And that's really important, too, in terms of both the understanding of aging and also thinking about transition to human usage. The third point I would make of broad general applicability. I would have guessed any drug that works is going to work in both sexes. And that's mostly wrong. We have two drugs, rapamycin and one we didn't mention, the amino acid glycine, which have equally strong effects in males and females. We didn't talk about glycine because it works in males and females. Significant, but it's a really tiny effect. And so it's sort of fallen by the wise side as we've moved on to drugs with greater effectiveness. But the sex specificity and the ability to start in late life were both surprises and open up new ideas about how can you discover why it's working in old age, and how can you discover why it's working in one sex, but not both. Those are important research questions. And we could easily take three or four or five times as many really good new labs in aging and devote them to trying to tease out those puzzles. And then the fourth set of questions would be the ones you brought up. You're pointing to specific molecular clues. Is it tor? Is it glucose? Are there sex specific hormone receptors in the brain that are important? I think those are really interesting as well. And in fact, any postdoc grad student who comes into my lab, they get one of those to work on. That's the cutting edge of turning this into a molecular and cellular understanding of what slows the aging process and where we ought to pay attention when we want to find new drugs that will work in just the same way. So the points you've made strike me as highly pertinent, very important for anyone who wants to work on something in this area. But the three more general points, that it works at all, that it can work late in life, and that it's often sex specific, I don't want those to get lost in the shuffle. Because they fundamentally rearrange our ideas about aging, about late life, disease, and about what one ought to do if one was, say, in charge of America or the universe's medical research apparatus.

On that point, Rich, if money were absolutely no object, and the reality of it is it doesn't need to be as extreme as one thinks. But let's just say that the NIH came along and said, look, we are going to give you the budget you need to study these questions in any animal. What would you choose as the best model organism that would give you the trade off of duration and. And closeness to humans. Would it be staying where you are, where you have an advantage, that you get answers quickly, but the disadvantage is you're quite far from humans. Or would you move more towards what we saw in the very famous caloric restriction experiments at Wisconsin and nia in the 80s and 90s, sort of 90s through 2000s, where you were very close to humans, but obviously it took a long time to get the answer.

So I have two answers for you. The first is the mouse, and the other is. It's a bad question. If you ask a carpenter what is the most important tool, you'll get the same sort of answer, because the answer is the tool that I select depends upon the problem I have to solve. I think that each of the major experimental animals that is currently receiving a lot of attention is doing something good. They're molecular and genetic questions that are best addressed by flies and worms, even though they don't resemble me and my mouse in more than cellular aspects. But they're things that they can only solve that we can't solve. And that's an important thing to understand in primates. I think there are a few people who are starting to understand that rhesus is not where it's at, marmoset is where it's at, because marmosets are primates. They are organized much like humans, but they have a lifespan, an average lifespan of maybe eight or ten years, something like that. So, like dogs, they have the advantage of a lifespan which is longer than we'd like, but short enough that you can think about doing an experiment and completing it within about a decade. Some of the work has to be done in humans, too, for reasons I don't think I need to convince you of. The mouse is the correct answer to the. You know, if you have to pick one, it's got to be the mouse. You said they're not much like people, but of course, they're extremely like people. They have the hypothalamus, they have the pancreas, they have the beta cells. There are very few things, organs or cells or circuits or hormones or whatever, that mice have, but people don't have, or vice versa. There's important subtleties.

Yeah, I guess I'd push back a little bit in terms of. Some of the differences are profound. Right. So one profound difference is herbivores versus omnivores. Now, that might pose a lot less of a difference for the ITPs than it does for one of my pet peeves, which is the never ending nutritional studies of mice, which I find generally unhelpful for that reason, and we've even discussed it today in the context of looking at acarbose. Right. That's a nutritional intervention in some ways. And it might be a little bit difficult to extrapolate. But I think the other area where we do need to be a little bit circumspect with the use of mice is that 80% of them die of cancers, a third of which are hematologic. Whereas when you look at us, all humans die with atherosclerosis, about a third die from it. And I think that is an important distinction that basically says this is just my own personal philosophy and any effort to thwart human aging must be able to punch atherosclerosis squarely between the eyes.

So let me give you a different way of thinking about this. Let me try to change your personal philosophy here. We gave a lot of mice and rapamycin and then instead of letting them die, we euthanized them at 22 months of age, when most of them were still alive and fairly healthy. And then with the aid of a pathologist, J. Irby Wilkinson, Irby looked at dozens of different organs and Sue Brooks also looked at their tendons. So their tendons were youthful, their kidneys were youthful. They did not have changes in the heart, they did not have changes in the endometrium, they did not have changes in the liver, they did not have changes in the adrenal that were characteristic of 22 month old control mice. So the implication here is that even though the leading cause of death was cancer, the rapamycin was actually slowing a very wide range of age associated changes. Now, some changes, and we don't know which ones they are, produce cancer in mice and in people, some changes, and we don't know which they are, produce atherosclerosis or strokes or diabetes in people and in a small fraction of mice. So from my perspective, the details of what you're going to die of, whether it's mostly a stroke, mostly atherosclerotic in nature, or mostly neoplastic, they're certainly important if you're, if you've got a patient, you want to help that patient, or if you're looking for disease specific drugs, but my interest is not related to that, it's to what can we do to slow the aging process. And I think it's a very plausible, at least plausible idea that when you slow the aging, whatever diseases, affliction, the members of that species will be slowed down as well. Alzheimer's disease, for instance, does not ordinarily occur in mice. It's obviously a major hassle for people. Yet if you make a mouse that has some aspects of Alzheimer's disease and you give them these drugs, the disease is postponed. That's consistent with the notion that whatever diseases aging leads to in either species, anti aging medications will help retard. Cause of death is an important wrinkle, but only an elaboration of the aging process in a species specific way.

And I think this also speaks to the importance of whenever possible, starting younger, although again, rapamycin almost flies in the face of that. But the earlier you would start, the more opportunity you would have to beneficially impact those tendons, those nephrons, those cardiac myocytes and those neurons.

That's a really good general point. I used to agree with that completely and now I think one has to answer that question directly point by point by point. The fact that giving 17 alpha estradiol or acarbose or rapamycin to middle aged mice and you still get a nice postponement of whatever the lethal process is, that argues against the notion that many, many things have to start early. In addition, this is again published now. This is Mike Garrett's paper with John Herrera and Charlene Day. When they started treating some old mice with 17 alpha estradiol, they postponed or reversed. It's hard to tell. Muscle changes, muscle grip strength changes, rotor rod, the ability to stand on a rotating rod, and changes in old age and glucose tolerance. So it may be that despite my personal intuition, which is similar to the intuition that you just expressed, the evidence starts to suggest some aspects of aging can be reversed or prevented. Even if you're starting in late middle age. That would be cool. If it was true, I would have bet against it. But I may be wrong.

Has your conviction about any of the molecules you've tested in the ITP led to you taking any?

I prefer not to answer that because I'm not a doctor and I would never ever wish to recommend drugs to someone else. Take them too seriously.

I wasn't going to ask you which, but we'll leave it at that. We'll leave it at that. Rich, this has been an awesome discussion. I mean, I was really looking forward to this and I honestly think you have one of the coolest jobs in the world because you get to basically hear from anybody and everybody who has an idea about something that might extend life and a very straightforward intervention, which is if a person, in this case, if a mouse, takes this drug, will it extend their life? And you get to test this in a highly reproducible manner with two great colleagues. And you've basically got this well oiled machine that produces these fantastic studies which both positive and negative, continue to add to the body of knowledge we have. I want to close with something really fun and you probably don't want me to do this, but I'm going to do it anyway. So 15 years ago you wrote something really hilarious at the MIT Technology Review. Do you know where I'm going with this?

I remember the piece pretty well. Yes. I wrote it right here at this desk.

Okay.

Early morning, it was six in the morning, I couldn't sleep. I had an hour. I just felt in the mood to be Mead, so I wrote the piece.

So Aubrey De Grey must have have done something on television that you must have watched. Had he basically gone on TV and.

Said, no, I made that part up. I'd been hearing Aubrey speak over and over and over and over again for many years. I fantasized that he had made a television appearance and I wanted to enlist his help.

And basically what you said is, I'm going to read just the first part of this to you, Aubrey. So the MIT Tech Review published this. I saw you on TV the other day and I was hoping that now that the aging problem has been solved, you might have time to help me in my publicity campaign to solve a similar engineering challenge. One that has been too long ignored by the ultra conservative fraidy cat mainstream scientific community. The problem of producing flying pigs. A theoretical analysis of the problem using the fastest available modern computers shows that there are a mere seven reasons why pigs cannot at present fly. One, they do not have wings. Two, they are too heavy to get off the ground. Three, the so called law of gravity. Four, they cannot climb trees. Five, hair instead of feathers. Six, they do not wish to fly. And seven, they do not tweet. Now this is really funny because you wrote this prior to Twitter. When I read this, I was like, what does tweeting have to do with anything? And of course I had to remind myself Twitter didn't even exist in 2005. I'm not going to read the rest of this, but we're going to link to it in the show notes because it is a brilliant. You go into great detail, elaborating on each of your points, but this is to me really a great example of who you are, right? Which is you're grounded in reality and you don't buy into this idea that immortality is in our future. And you view it as an enormous win if there's a drug out there that can extend human life by 25%. I'm in your camp, Rich, so I'm highly biased and I've seen nothing to suggest that immortality is in our future, and frankly, I'm not sure it should be, even if we were ever given the choice. But boy, 10 to 15% improvement in our lifespan and health span would be a remarkable achievement. And if there are molecules that can do that, I'm glad you're working on them.

Well, I'm delighted to have been invited. A friend of mine said that he'd been on your podcast a number of times and had a wonderful, wonderful time and told me if I ever got a chance, I would enjoy it too. And this is a very good, good prediction. It's a real treat to be interviewed by someone who really knows the stuff and also asks the very best questions. So I I'm honored to have been added to your roster of guests. This has been a great pleasure for me, Rich.

Thank you for taking the leap of faith and sharing your work with us. This was fantastic.

Good deal. Well, thank you for inviting me.

Thank you for listening to this week's episode of the Drive. If you're interested in diving deeper into any topics we discuss us, we've created a membership program that allows us to bring you more in depth exclusive content without relying on paid ads. It's our goal to ensure members get back much more than the price of the subscription.

Now.

To that end, membership benefits include a bunch of things. One totally kickass comprehensive podcast show notes that detail every topic, paper, person, thing we discuss on each episode. The word on the street is nobody's shownotes rival these monthly AMA episodes or Ask Me Anything episodes. Hearing these episodes completely access to our private podcast feed that allows you to hear everything without having to listen to spiels like this the Qualys, which are a super short podcast that we release every Tuesday through Friday highlighting the best questions, topics and tactics discussed on previous episodes of the Drive. This is a great way to catch up on previous episodes without having to go back and necessarily listen to everyone. Steep discounts on products that I believe in but for which I'm not getting paid to endorse and a whole bunch of other benefits that we continue to trickle in as time goes on. If you want to learn more and access these member only benefits, you can head over to Peteratti md.com subscribe. You can find me on Twitter, Instagram and Facebook, all with the ID PeterAttiaMD. You can also leave us a review on Apple Podcasts or whatever podcast player you listen on. This podcast is for general informational purposes only and does not constitute the practice of medicine, nursing, or other professional health care services, including the giving of medical advice. No doctor patient relationship is formed. The use of this information and the materials linked to this podcast is at the user's own risk. The content on this podcast is not intended to be a substitute for professional medical advice, diagnosis or treatment. Users should not disregard or delay in obtaining medical advice from any medical condition they have, and they should seek the assistance of their healthcare professionals for any such conditions. Finally, I take conflicts of interest very seriously. For all of my disclosures and the companies I invest in or advise, please Visit Peter attention tamd.com about where I keep an up to date and active list of such companies. It.