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

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

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

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

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

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

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

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

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

00:00:47.740 here's today's episode. I guess this week is Richard Miller. Rich is a professor of pathology

00:00:55.980 at the University of Michigan and the director of Michigan's Paul F. Glenn Center for Aging

00:01:01.480 Research. He served on a variety of editorial and advisory positions on behalf of the American

00:01:06.360 Federation for Aging Research, AFAR, and the National Institutes of Aging, NIA, which we

00:01:11.840 talk about a little bit. He's also served as the editor-in-chief of Aging Cell. He's the recipient

00:01:18.120 of the Nathan Schock Award and the Allied Signal Award, the Irving Wright Award, and an award from the

00:01:24.640 Glenn Foundation, along with a number of other awards for aging research. Dr. Miller's research

00:01:29.740 focuses on the problems of the basic biologies of aging, mostly in mice, but sometimes using other

00:01:36.020 cell lines. Though we speak in this podcast almost exclusively about one of the amazing products of

00:01:43.060 his life's work called the NIA-funded interventions testing programs, the ITPs. Now, over time, you'll

00:01:52.080 probably hear me or you have heard me talk about the ITPs on various podcasts, certainly with respect

00:01:58.300 to the podcasts around rapamycin and metformin. The ITPs are a set of studies that are done

00:02:06.640 concurrently in three labs, the Jackson Labs, Rich Miller's lab in Michigan, and Randy Strong's lab

00:02:15.320 in UT San Antonio. We go into great detail about what the criteria are and how these are done,

00:02:21.480 but suffice it to say that based on the types of mice that are used and the rigor with which these

00:02:27.040 are done statistically and otherwise, the ITPs represent effectively the gold standard of testing

00:02:34.180 molecules for longevity in arguably the most important subset of mice we could study. And basically what we

00:02:43.100 do in this podcast is go through the long list of molecules that have been tested, what the results

00:02:50.440 were a number of these are quite surprising. They're all very interesting, both in positive

00:02:55.260 and negative findings. I've been looking forward to speaking with Rich for about a year, but the reason

00:03:02.480 we delayed this until now is I wanted to make sure that a couple of the findings that were in the

00:03:08.640 pipeline were close enough to publication that we could speak about them confidently. And this podcast

00:03:14.440 can now be released as soon as those manuscripts have been accepted, in particular, one around an

00:03:19.960 SGLT2 inhibitor and another one around nicotinamide riboside, which is a very popular supplement that

00:03:26.460 many people ask about as a precursor to NAD. So without further delay, please enjoy my conversation

00:03:32.980 with Dr. Rich Miller. Rich, thank you so much for making time to chat today. I've been looking forward

00:03:43.480 to this one for almost a year now, but we wanted to wait until some of these exciting results we're

00:03:49.600 going to get to were far enough outside the pipeline that we could speak about them.

00:03:54.500 Good to chat with you.

00:03:55.680 But I think before we get to the most recent findings of the ITP and even explaining what an

00:04:01.500 ITP is, I want to give people a bit of a sense of who you are and how you've played kind of an

00:04:07.740 amazing role in the field of longevity, along with a couple of your colleagues. So where'd you grow up,

00:04:13.900 by the way? I was born in Philadelphia. And when we were five, my mom and dad moved me and my two

00:04:19.340 brothers to the northern suburbs of Philadelphia, Cheltenham Township. I went to high school with Ben

00:04:25.420 Netanyahu of all people who joined our soccer team as a sophomore. I was just born slightly too late.

00:04:35.760 If I'd been born three years earlier, I would have had Reggie Jackson as a teammate.

00:04:40.700 Wow. Interesting. Split your loyalties there between what city you're going to choose for sports.

00:04:46.440 So when did you take an interest in science growing up?

00:04:49.320 When I was about five.

00:04:50.400 That's pretty unusual for a five-year-old, right?

00:04:53.660 Well, I really liked science. I was decent at school and I always liked getting good grades

00:04:59.960 and stuff like that. And I decided early on that aging was bad for you. It made people sick and

00:05:04.740 then die. I was kind of against that. And the best way to fight against that was to learn something

00:05:10.160 about aging and to develop ways of slowing the aging process down. So you can't do that unless

00:05:15.280 you're a scientist. So you end up doing your MD and PhD degrees at Yale in the 1970s. At that time,

00:05:25.800 I think you went off and did a postdoc at MSK. Is that correct? At Memorial Sloan-Kettering?

00:05:30.840 Yes, that's exactly right. I actually did a postdoc for three years first at Harvard,

00:05:36.080 which did not work out at all. And then I was rescued by a mentor at Sloan-Kettering.

00:05:41.960 And tell me more about what you mean that it didn't work out at all, the first part.

00:05:46.620 I don't want to be sued for slander, but let's just say that my postdoctoral mentor

00:05:50.520 had some defects as a mentor and I wasn't very productive and he wasn't very happy with me and

00:05:57.240 I wasn't very happy with him. So as a beginning postdoc, you're always scared of offending the

00:06:02.400 big boss and you tend to put off the decision to leave as long as you possibly can. And eventually,

00:06:09.280 we both got fed up with one another. And so I moved to Sloan-Kettering where things went better.

00:06:14.940 Did your field of study change when you made that switch?

00:06:17.460 It changed not at that point, but at the beginning of my postdoc, I knew I'd always wanted to work on

00:06:22.520 aging. And the PhD project that I worked on was involved somatic cell hybridization. I wanted to

00:06:28.820 mix together two kinds of cells, one that could divide and one that couldn't divide

00:06:33.160 and immortalize the T lymphocyte that couldn't divide. And that project went well, but the more

00:06:39.500 I learned about the Hayflick hypothesis and all of that cells and essence stuff, the more convinced

00:06:44.740 I was that it wasn't going to teach me anything I wanted to learn about aging. And so I decided as a

00:06:51.120 postdoc to become an immunologist. The lab I worked at at Harvard was an immunology lab and the one at

00:06:56.920 Sloan-Kettering was an immunology lab too. Then when I set up my own lab at Boston University,

00:07:01.400 it was principally to study immunology and how aging modified the immune system.

00:07:07.940 When I came to Michigan in 1990, I was still mostly an immunologist focusing on aging and writing book

00:07:13.260 chapters on how aging modified immunity and the like. Over the next decade or so, I switched

00:07:19.400 gradually into focusing on aging more broadly with immunology becoming only a minor interest. And now we do

00:07:25.920 very little immunology. It's nearly all the kind of work that is focused on what is controlling

00:07:31.300 the aging process and how you can use that to develop interventions.

00:07:35.440 Give folks a little bit of the history around Hayflick and the division of cells and some of

00:07:41.940 these observations. These are kind of the, some of the foundational things in aging, right?

00:07:46.920 Unfortunately, yes. I don't know how, whether you want the two minute or the 20 minute or the three

00:07:52.260 hour rant about this, but the phenomenon that Hayflick more or less discovered was that if you take

00:07:59.000 normal human cells, which are not cancer cells, and you let them grow in tissue culture, they're

00:08:03.900 divide, but only a limited number of times, about 50 times, and then they stop. And if you listen to

00:08:09.880 Hayflick, he then decided that was sort of like aging. He had sort of found a way to study aging in

00:08:16.640 culture. Now that's nuts. It's nothing like aging in the slightest, but he thought it might be like

00:08:24.220 aging. So at that time, people were really excited by tissue culture. It was the brand new thing.

00:08:30.460 Everybody wanted to work on tissue culture and it seemed like a wonderful, exciting opportunity to

00:08:36.080 be able to use tissue culture to study aging. So they bought this idea and a whole generation of

00:08:43.120 superb cell biologists spent their life investigating the Hayflick system under the illusion that they were

00:08:49.420 studying aging. Parenthetically, about 15 or 20 years later, when I was just beginning to become

00:08:58.300 well-known, I was invited to give a talk at New York University, whose chair of pathology was Vittorio

00:09:03.620 Defendi. And Defendi had been a colleague of Hayflick when Hayflick was making these discoveries.

00:09:10.780 And Defendi told me that he and Hayflick had had lunch over and over again. And Hayflick had been a kind

00:09:15.840 of a grump and had said over and over again, oh, my cells, they just won't grow. Oh, I tried them,

00:09:20.980 I thawed them out only two weeks ago and they already stopped growing. And Defendi, if you believe

00:09:25.780 his story, which I do, told Hayflick, well, ha-ha, Len, maybe they're just getting old. And Hayflick,

00:09:33.160 whose sense of humor is notoriously absent, did not understand he was being joshed. He thought it was a

00:09:40.480 hypothesis, a scientific hypothesis. And after a while, he persuaded himself that it was his hypothesis

00:09:46.860 and then that it was the correct hypothesis. So he's a persuasive fellow. And many people bought

00:09:53.640 into the notion that the idea that when cells stopped growing in culture, it's sort of like or

00:09:59.820 caused by or something related to aging. Too many people, in my view, accepted that as the truth of the

00:10:07.160 matter. And then when it later became established that the limitation of growth, the Hayflick limit

00:10:12.360 was actually due to the shortening of telomeres, a very important finding and certainly true,

00:10:18.000 people convinced themselves because they thought it was like aging, that telomeres were something to do

00:10:22.860 with aging too. And it was that way in which a whole new generation of people sort of bought into the

00:10:29.960 Hayflick and the telomeres thing as a central cause of the aging process, despite all the evidence

00:10:37.140 to the contrary. Well, it's interesting. You bring this up, of course, because at least at the time

00:10:41.880 that we sit here recording this, it's only been about a week since a story came out of Israel,

00:10:47.260 which I'm sure you've been emailed about as many times as I have. You haven't. Okay. Well,

00:10:52.160 consider yourself lucky because I've only been emailed at about 57 times, which is apparently

00:10:57.960 the most rigorous study in the history of humans demonstrating the anti-aging benefits of

00:11:05.860 hyperbaric oxygen. So you see, Rich, in this amazing study done in Israel, a group of volunteers were

00:11:13.140 exposed to hyperbaric oxygen. And wouldn't you know it, their telomeres lengthened a bit.

00:11:19.480 And so as a result of that, I've had, I don't know, I'd probably need scientific notation to

00:11:24.780 count the number of people that have emailed me that story to tell me, why are we not doing hyperbaric

00:11:30.000 oxygen? You see, this is the fountain of youth. And I think I forwarded it to Joan Manick and Matt

00:11:36.600 Kaberlein because the three of us have a grumpy old men, women club where we groan on about the

00:11:43.180 idea of until you have really good biomarkers for aging, it's really hard to study aging. And we can't

00:11:49.180 use telomere length as a biomarker of aging, but nobody really wants to listen to this.

00:11:53.720 All three of you. Yeah. And you know, it's funny. It's entirely possible that hyperbaric oxygen could

00:12:01.060 well have health benefits under certain circumstances. Certainly I'm not endorsing it.

00:12:05.720 There's no evidence that that statement is true, but it's not silly. There are certainly oxygen sensing

00:12:11.540 circuits in multiple cells, which at least in worms can trigger anti-aging programs, pro-longevity

00:12:17.980 programs. So it's not a dumb idea. And there's some really good labs pursuing it. The weakness,

00:12:23.220 I haven't read the paper, but from what you've told me, the weakness is just the one you've cited.

00:12:28.160 Aging and telomere length are not the same thing. And if you want to prove that hyperbaric

00:12:33.600 conditions at suitable doses and at suitable time intervals, et cetera, might be good for you,

00:12:39.060 that's a very plausible idea, well worth testing. It's just that you don't test it by measuring

00:12:43.000 telomeres. Yeah. I mean, it's interesting. Obviously, I think that the telomere, I don't want to say

00:12:48.880 story story, I would say the telomere concept or notion as a biomarker, even though it's not really

00:12:54.220 a biomarker, I think gets a lot of credibility based on the fact that a Nobel prize was awarded

00:12:59.580 for its elucidation. So it's interesting. I've heard people say, look, and I think this is the

00:13:06.200 most accurate thing I've heard stated, which is look, the work that Blackburn did is really amazing

00:13:12.580 biology and it's worthy of a Nobel prize, but it doesn't really explain aging. And those two

00:13:18.380 statements can coexist. Oh, sure. But you've left out a third important statement. Telomere biology

00:13:24.000 is a critical element in cancer biology. And the amazing work that several people, including Blackburn,

00:13:30.920 did to work out the telomere story is, I think, really a fundamental advance in our understanding of

00:13:36.980 cancer in people. And it may also have an interesting sidelight in evolutionary biology.

00:13:42.980 There are some species, mice, for instance, with extremely long telomeres that if they need an,

00:13:48.860 they don't need much anti-cancer defense at all. They're going to get eaten in six months.

00:13:53.040 Vera Gorbunova at Rochester has done some lovely work on the ways in which different species with

00:13:59.320 different body size and different lifespans differ in their ways of stopping the cancer process.

00:14:05.840 Some relying, like people do, on the telomere alarm clock, where the telomeres get too short.

00:14:12.200 This puts a whole lot of anti-cancer defenses into play. And others, like mice, where this does not

00:14:18.340 happen. The mice have very long telomeres. They don't need telomeres to tell them when their cells

00:14:22.100 are getting cancerous. As a fundamental step in our understanding of cancer biology in people and the

00:14:28.980 way in which anti-cancer defenses evolve, it's great stuff. It's just the notion that it's

00:14:35.220 sort of a shortcut for actually working on aging, in my view, does not hold up very well.

00:14:42.260 Yeah. Well, I think I couldn't agree with you more. And it's certainly on the short list of

00:14:45.960 frustrations I have in this space where people very quickly revert to, well, but look, but intervention

00:14:52.220 X shortened telomeres, you know, lengthened telomeres. Therefore, it's somehow a cure for

00:14:58.280 longevity and, or a cure for aging. And despite all evidence of the contrary. So what, what prompted

00:15:03.980 the move from BU to Michigan in 1990? My wife got a job as a professor of English at the University

00:15:10.320 of Michigan. And she said, would you like to move with me to Michigan? Why don't you look for a job?

00:15:14.920 So that, that worked out pretty well. I mean, I'm being a little coy. I had always wanted to move

00:15:20.460 to Michigan. Michigan was very strong university and very strong in aging research. And I had said

00:15:25.860 to her, you know, she was coming up for tenure at Harvard. Harvard had never tenured a woman in 372

00:15:31.840 years in English. They had never found a competent woman who could be an English professor. So we know

00:15:37.220 they were not going to start with Patsy. And I said, Hey, why don't you apply to Michigan? I'd love to

00:15:41.980 move to Michigan. And it worked out for both of us. Excellent. So what is it that brought to your

00:15:49.060 mind? This idea of the ITP? This is partly your brainchild, right? I was the midwife or something.

00:15:57.420 So the national aging Institute in its division of aging biology at that time was headed by a

00:16:03.480 visionary guy named Huber Warner. And Huber thought it would be a great idea for the aging research

00:16:10.140 community generally to have a program in which drugs were tried on mice or mice and worms or

00:16:17.260 mice, worms and flies and dogs or something. So he commissioned a committee of eight or 10 or 12 of

00:16:23.660 us to sit down with him and Nancy Nadon, who was his deputy at the time. And we spent a day or two

00:16:29.940 talking about what the NIA might do to develop a program in which interventions, potential interventions

00:16:36.120 were tested directly. And after the first night, I went into a room with Arlen Richardson, who at the

00:16:42.200 time was at Texas and is now at Oklahoma. And Arlen and I put together what we viewed would be

00:16:48.840 the best way to do this, focusing on mice. And we sold it to Huber. So Huber then asked me to

00:16:58.220 write down a more formal plan, which he and I wrote up as an article for mechanisms in aging and

00:17:03.640 development. And when it was time for the NIA to formulate the rules for the program and to start

00:17:09.140 saying, would you please everybody apply for this program? We have some money. The program that Arlen

00:17:14.980 and I had sketched out and which Huber had helped to refine and had by then been through three or four

00:17:19.820 versions became the sort of foundation on which the program was developed. And then three grants were

00:17:26.580 awarded, my own, David Harrison's at the Jackson Labs and Randy Strong at Texas. So the three of us,

00:17:32.840 it took us a while to sort of get to the point where we all agreed as to what should be done and

00:17:38.420 how it should be done. But the foundation of the program is one that Huber endorsed and which Arlen

00:17:43.520 and I had helped to develop by now. It's about 18 years ago. This is around the time you wrote a

00:17:50.300 paper, Extending Life, Scientific Prospects and Political Obstacles, right? That paper coincides

00:17:57.780 almost with when you and Strong and Harrison kind of came up with what the principles of the ITP would

00:18:04.480 be, right? Yes. I mean, the two events were not really related in any cause and effect fashion,

00:18:09.340 but it's about the same time. I had been asked to give a talk at the law school at the University of

00:18:13.980 Michigan. They were putting a seminar together, an all day symposium in the end of life and the other

00:18:19.540 people chatting. One was how to take care of old people. One was how to make decisions, triage decisions

00:18:25.360 at the end of life. And I was the only biologist on the program. So I gave them the talk, which then

00:18:30.060 became that article in the Millbank Quarterly that you've been mentioning. Now you make three big

00:18:35.260 points in this article that I think are worth spending a minute on before we jump into the ITP

00:18:40.260 because I think it really sets an adequate stage, right? So again, keeping in mind, this is an article

00:18:45.240 written about 18 years ago. And it's important, I think, for people to realize that the world we live

00:18:49.660 in today is much more open to the discussion we're going to have than it was, I think, 20 years ago.

00:18:54.160 That's for sure.

00:18:55.100 So you said, look, discussions of anti-aging medicine aren't silly. The development of an

00:18:59.600 anti-aging strategy is starting to make headway and further work here would be worthwhile. That

00:19:05.540 was basically the thesis of your paper, yes? I agree. And it's, I think, a really important point

00:19:10.660 that deserves stress. It's still an evolving process. We're not there yet, but 18 years ago,

00:19:16.340 we certainly weren't even close to there. Around that time, the Gordon Research Conference on Aging,

00:19:21.660 which is an invitation-only conference, about 120 more or less distinguished researchers,

00:19:27.500 I got into a fight with my friend George Martin as to whether aging could be considered

00:19:31.480 a single process, a drug might slow that process down, or whether it was a very complex

00:19:37.760 spaghetti bowl filled with 30 or 40 or 50 different processes, each one of which had its own independent

00:19:43.440 control. And George and I did not agree about that. So as a game, he asked the assembled conferees to

00:19:51.340 vote. And you could vote zero to 10, zero if you agree that there were too many processes ever to be

00:19:58.260 controlled together, and 10 if you happen to take my position. The average score was about two or three.

00:20:03.660 Very few people at that time, even among the sophisticated, best-educated, professional aging

00:20:09.540 researchers would check the box that said, yeah, there's a single aging process. And until you

00:20:15.820 check that box, the notion that you could interrupt the process makes no sense, because if there is no

00:20:22.240 process, then interrupting it is a hopeless goal. I think you really sort of need to see it as a unitary

00:20:28.500 process in order to devote much time to trying to slow that process down. It's confusing because there

00:20:35.820 are dozens and hundreds of things that go wrong in old age, and learning about any one of them can be

00:20:40.340 very productive. It's just that in addition to looking at any one of them individually, the one

00:20:46.140 that leads to cataracts, the one that leads to muscle failure, etc., the concept that there's some

00:20:51.700 underlying mechanism that speeds them up a lot in mice, or a little bit in horses, or slows them down

00:20:59.540 a lot in people or an awful lot in whales. Once you get that idea in your head, then suddenly it

00:21:05.600 becomes permissible to think about ways to slow it down. Now, there were a couple of things that

00:21:10.180 we already, quote unquote, knew at this time that had to give you some hope. Obviously, there was the

00:21:17.340 vast literature on caloric restriction. So almost without exception, and there are some very notable

00:21:23.760 exceptions, but almost without exception, some form of caloric restriction, and we could get into all

00:21:29.640 the details of different species, and when it's applied during life, and to what extent, and wild

00:21:34.920 types versus not. But there was clearly a path towards extending life with caloric restriction.

00:21:41.060 And then, of course, we had the DAF mutants in C. elegans, Cynthia Kenyon's work, which I think

00:21:49.040 was probably early 90s, right? That was 93. 1993. Yeah. 1993. And if I recall, her first—God,

00:21:56.220 it's been a while since I've looked at this—her first observation was with—was it DAF-16 was the

00:22:01.980 first one, or was it DAF-16? The first paper, the critical paper, made two big discoveries. One was

00:22:06.240 that mutants of DAF-2 extended lifespan, and that mutants of DAF-16 blocked. When combined—yeah. And

00:22:13.200 now, you could combine with CR and then get a better outcome, but yes. And DAF-2 ended up being

00:22:18.840 the analog of the IGF receptor, and I think DAF-16 was basically an analog of FOXO-3A or 3B. I can't

00:22:26.840 recall. Yeah, it wasn't quite the receptor. It was the thing that the receptor turned on, but it's one

00:22:31.580 step downstream from the receptor. One step downstream inside, yeah. So explain to folks why that was

00:22:37.260 relevant, as a proof of principle, at least. Well, first of all, it's just the point you were

00:22:42.380 making. Those of us who knew the caloric restriction literature were screaming our heads off, saying,

00:22:46.860 see, look, you can slow the whole thing down. And no one quite bought into that. Cynthia's

00:22:53.260 demonstration, which was also, you know, pioneered by other people like Tom Johnson and Gary Rufkin

00:22:59.040 and some others at the same time, pointed out that a single gene mutation could extend longevity

00:23:05.120 dramatically in a worm. The funny thing is, you know, Nature, early in 1993, published a very

00:23:12.760 erudate, detailed six-page paper by Linda Partridge and her colleagues proving that there could never

00:23:18.960 be a single gene mutation that extended lifespan. Aging was too complicated, too many feedback circuits.

00:23:24.100 It could never, ever be done. And Linda's paper was a statement of any species.

00:23:28.840 It wasn't explicitly stated, but it was the theoretical proof that you could never have a single gene

00:23:33.420 mutated that extended lifespan. Absolute mathematical proof. Lots of equations. I couldn't follow that.

00:23:39.740 It can't follow now. So about six months later, Cynthia's paper came out saying, hey, look, here's a gene

00:23:45.520 that does that. So that same issue, Linda was asked to comment on, Linda Partridge. And she wrote a paper

00:23:52.180 saying, well, of course, C. elegans is different. C. elegans has dour mutants. They're very, very special.

00:23:57.340 But there could never, ever be, in any other species, a gene that does this. Trust me on that.

00:24:02.860 So it took three more years for the second shoe to drop, so to speak, which is Andrzej Bartke's paper

00:24:08.640 with Holly Brownborg and two other colleagues showing that the Ames dwarf mutation could extend

00:24:14.540 lifespan by about 40% in a mouse. Once it was true for mice and true for worms, the idea that, you know,

00:24:22.600 there might be something to this became defensible. One of my favorite conference stories, about a year

00:24:29.600 or two after the Ames dwarf paper came out, also in Nature, I was at a conference and Robin Holiday,

00:24:36.320 an Australian scientist who was very well respected, gave a keynote lecture explaining why you could

00:24:41.760 never have a gene that it slowed the aging process and extended longevity. And I raised my hand to say,

00:24:48.680 but you know, there are two papers here, one in worms, one in mice, saying that there are single

00:24:54.740 gene mutations that extend longevity, and that's not compatible with your theory. And he said, yeah,

00:25:00.580 but there could never be a single gene mutation that could extend longevity. And so we went through

00:25:05.840 that loop about twice, and then I went out to get a cup of coffee. It's a notion, a zombie notion,

00:25:11.100 that sort of was hard to kill. And many scientists still are distinctly uncomfortable with the notion,

00:25:17.920 not all of them anymore. I think our team is winning. But many people still are reluctant,

00:25:22.860 even if they're sophisticated and know the literature to accept the general principle,

00:25:27.160 that there are aging processes which can be slowed down if you consider them as a unitary

00:25:33.660 group of changes with a common controlling factor.

00:25:37.760 And I always think that part of the debate is fixating on the wrong point, right? I mean,

00:25:42.500 whether or not it's a single gene or multiple genes is probably less the point. The bigger point,

00:25:48.240 at least for me as an outsider, is aging malleable, yes or no? To me, that's the question

00:25:55.400 that matters. And if prior to the 1990s, if the answer was believed, no, aging is not malleable,

00:26:02.020 it is what it is, then it's a very uninteresting field to be in.

00:26:06.580 Yeah. No, I agree with that. I think you're right on point here. I would add one more,

00:26:11.220 Philip, to that. The question everybody was asking, and which I used to think was pertinent,

00:26:15.780 is what causes aging? What is the cause of aging? And now I don't think that anymore. I think the

00:26:21.740 real question that we're all trying to answer, or ought to be trying to answer, is what is it that

00:26:25.840 can slow aging? My own current sort of philosophical framework is that aging is caused by an awful lot of

00:26:34.540 different things. Some cells die. Some cells become mutant. Some tissue structures get cross-linked,

00:26:41.180 or heavy metals accumulate in a key cell. All of those are really bad for you. And as they accumulate

00:26:47.300 and accumulate and accumulate, you start to feel older and then more diseased. But the key point,

00:26:53.540 I think, is that there are biological processes that can postpone all of that stuff together,

00:26:58.380 that can postpone it for five decades in people, or almost a year in a mouse, or 25 years in a chimp.

00:27:07.220 So I don't really care what causes aging. What I care about is what is the process that can postpone

00:27:14.140 all the different aspects of aging? I think once you rephrase the question in that way,

00:27:20.300 you're well on the way towards designing experimental paradigms to address the serious question,

00:27:26.500 which is the coordination and eventually the postponement of these multiple aspects of the

00:27:31.960 aging process. I think it's worth restating that, Rich, because that might be one of the most

00:27:38.720 important teaching points, I think, for anybody who thinks about science. And certainly my mentor in

00:27:45.420 the lab, I also studied immunology, would constantly make this point, which is if you don't ask the right

00:27:51.400 question, you are guaranteed to flail. You might get lucky, but you're going to flail. Now, if you

00:27:58.760 ask the right question, you may still fail, but it's the difference between starting on your own

00:28:04.440 20-yard line and the other guy's 20-yard line. Yeah, I agree. And I agree. It's a really critical

00:28:09.780 point. A lot of the time at an aging meeting, usually someone fairly new to the field will say,

00:28:16.900 oh, well, you don't understand what aging is. Let me explain to you what aging is. And everybody

00:28:21.940 who's heard this line of argument 40 or 50 or 60 times will sigh. Sometimes we'll go out and get

00:28:27.820 another cup of coffee and say, well, let's talk about that at the break or at the bar after the

00:28:33.180 meeting. But they're wrong. The answer is you have to think about it just as you were saying. You have

00:28:39.260 to think about these fundamental issues. What is the best way to frame this question? And it can be

00:28:45.500 tricky if the world is filled with people, including myself, who are convinced we're right that you're

00:28:51.840 surrounded by people who aren't right. They're wrong, but they're convinced they're right. And

00:28:55.280 sort of talking them into viewing the matter from another framework takes a while.

00:29:01.620 Let's dive into the ITP, the Interventions Testing Program, this amazing, amazing scientific

00:29:11.560 framework. Think of it as a program. I don't know. How do you describe it to people? I sort of think of

00:29:17.980 it as a gold standard by which we test single drug often, but sometimes not just single drug

00:29:26.180 interventions in a robust, rigorous manner. And when things pass the ITP test, we take them really

00:29:35.240 seriously. When things fail in an ITP, even if they've succeeded elsewhere, we look at them very

00:29:42.620 closely. So what's the word you would use to describe the ITP to somebody on the outside?

00:29:48.900 Well, I agree with most of what you've said. And I'm certainly very flattered to hear you describe

00:29:54.000 it in that way. Although I think you're being a little too optimistic and a little kind. So let me

00:30:00.040 sort of back, go back over what you said. The goal of the ITP, and we've tried our best to meet this

00:30:05.660 goal, is to develop a fundamentally sound way of testing one question. Does this drug extend mouse

00:30:11.680 lifespan? And we do our best by designing the program in ways so that our answers will be reliable and to

00:30:19.640 the best we can, reproducible and believable. For example, we use an awful lot of mice because that

00:30:26.460 gives us a lot of statistical power. We want to be sure that we will have enough mice present that

00:30:32.280 if a particular drug extends lifespan by 8% or 10%, we'll pick it up almost all the time, 80% and 90%

00:30:40.000 of the time. That's sort of an ambitious thing to do. To give you an example, Jay Olshansky and Bruce

00:30:45.820 Carnes and their colleagues demonstrated years ago that if you abolish cancer in people, nobody ever gets

00:30:51.740 cancer again. Human lifespan goes up by only 3%. So the drugs that we are interested in by designing

00:30:59.440 our program are drugs that are designed to win at our program. You've got to be at least three times

00:31:04.020 better than a complete conquest of cancer. I'd like to come back to Jay's methodology on that because

00:31:10.520 I've never fully understood that calculation because I know he's also made the claim that if you

00:31:15.400 completely abolish atherosclerosis, it's about a 3% bump. And if it's cancer and atherosclerosis,

00:31:21.280 it's like a 7% bump. Yeah, that's right. And if you got rid of cancer, heart disease,

00:31:26.360 stroke and diabetes, it's about an 18% bump. So we'll revisit that. But let's go back to these

00:31:31.960 principles because there really are four remarkable principles of the ITP. And that's what it is that

00:31:37.540 for me personally gets me excited. So you alluded to one already, which is from a statistical standpoint,

00:31:45.120 you power to 80% for an 8% to 10% detection rate. And to put that in English, it means you need an

00:31:54.020 awful lot of animals. Yes. Each year, each of the three sites will have 50 males and 50 females on each

00:32:00.720 drug and also 100 male controls and 100 female controls. We double the number of controls because

00:32:08.540 it gives us a lot better statistical power at relatively little cost. Then we pull all the

00:32:14.840 data from all three sites together for our analysis. And then you've alluded to now the second feature

00:32:19.980 of the ITP, which is they are run in parallel at three sites. So the Miller lab is doing this at

00:32:25.940 Michigan. The Strong lab is doing this in San Antonio at UT and Harrison is doing this at Jackson

00:32:31.680 labs, correct? Yeah, that's right.

00:32:33.540 Okay. The next point that makes this pretty special is you are using genetically heterogeneous mice. These are

00:32:42.040 not homogeneous. So explain to folks why using these genetically heterogeneous mice is an important feature

00:32:50.440 of the ITPs. Yeah, I will. Let me just say one more point about that second point you raised. Doing it at all

00:32:55.380 three sites not only gives us power, but it also tells us that we're getting an effect that is reproducible at

00:33:01.520 three sites. If, for instance, and we've certainly seen this, we get a drug that does just terrific at one of the

00:33:06.940 sites, but not at all at the other two, something may have gone wrong.

00:33:10.960 That's right. That's a methodologic issue right there, potentially.

00:33:13.860 Yeah. If you get a drug that works pretty darn well at all three sites, people have very good reason for believing

00:33:19.480 it could work at their site too. That sort of instantaneous reproducibility is a major reason for doing it at

00:33:25.780 three sites. The literature is filled with reports of a drug that did something good to a mouse at one

00:33:32.180 site, and then no one wants to wait four years to test it out. Doing three tests at the same time

00:33:38.860 gets us around that. Then the question you asked was, why use the genetically heterogeneous mice? About 90%

00:33:45.440 of the work in aging with mice, and actually of medical research with mice, uses a single inbred

00:33:52.940 genotype, where every mouse is the same. There's no variation from mouse to mouse in their genetics.

00:33:59.380 In addition, it's inbred, and everybody knows if you have a choice, you'd rather not be inbred. That's

00:34:03.780 why we don't allow people to get married to their brothers and sisters, because it tends to produce

00:34:09.640 weak offspring. Every one of the inbred strains has its own set of bizarre peculiarities. Many of them

00:34:16.840 are blind by the time they're grownups. Most of them are deaf, et cetera, et cetera.

00:34:20.620 So what we wanted to do in developing a genetically heterogeneous stock was avoid that. We did not

00:34:26.660 want to trick ourselves into picking a drug that only worked on black six mice. And conversely,

00:34:33.380 we did not want to miss a really good drug that just happened not to work on black six mice.

00:34:38.960 So what we do is we obtain from the Jackson Laboratories two different kinds of mice,

00:34:44.740 mice, which have represent four different grandparents, inbred grandparents. When we cross

00:34:49.640 them together, just like when your mother and father produce offspring, each one of those children is

00:34:55.820 genetically unique, but each one will share half of its genes with all the others at random, a random

00:35:01.940 half. So all of the mice we've produced for the IDP, of which there are now more than 20,000,

00:35:07.420 have that in common. No two mice are ever the same, but any two mice share half their genes.

00:35:15.240 The other advantage of doing it this way is that the Jackson Labs, sort of the gold standard in

00:35:20.260 inbred stocks, their black six this year, and their black six 10 years from now, and their black six 10

00:35:26.500 years ago are as close to the identical as science can guarantee. And what that means is that the four-way

00:35:34.040 cross mice we use, the HET3 mice, are the same year after year after year. We can make 100 or 1,000 or

00:35:41.280 10,000 of them with the same characteristics. Anyone else anywhere in the world that wants to make them

00:35:46.800 can do so very inexpensively. And be sure that their population characteristics, the average numbers of

00:35:53.340 mice that have this gene or this pair of genes or this triplet of genes, will be the same as the ones

00:35:59.600 that we are using. It gives a sort of reproducible heterogeneity, which I wish almost everyone would

00:36:07.520 adopt. I think that's the way to do science. And people who were at the origins of the IDP agreed

00:36:15.080 with me about that. So that's why the IDP does it that way. Do you have a sense of what fraction

00:36:20.520 of research is done in mice utilizing genetically heterogeneous mice in the fashion you describe?

00:36:28.860 Are we talking less than 10%? Oh yeah, much less than 10%. And the reason is that when people want

00:36:34.280 to work in aging, they call the NIA, the National Aging Institute, and they say, can I please have some

00:36:40.120 old mice? And the NIA does not have genetically heterogeneous mice. They had them for, I begged and

00:36:47.080 pleaded and kicked and screamed, as did some colleagues. And they set up a colony of genetically

00:36:52.140 heterogeneous mice. And after three or four years, they were not receiving many requests. Everybody was

00:36:57.300 saying, I want black six mice. My scientific parents used black six mice, and everybody always uses black

00:37:04.140 six mice. I need black six mice. So the NIA, for lack of demand, folded up the colony. And now, I think

00:37:12.120 largely because the IDP people are beginning to realize the folly of using black six mice, and the

00:37:18.700 good reasons for using genetically heterogeneous mice. And the NIA keeps getting a sprinkling of phone

00:37:25.400 calls every now and then saying, please, please, can you carry these genetically heterogeneous mice?

00:37:30.820 But it's momentum. They have contracts with their mouse producers, and it would take at least Richard

00:37:36.700 Hodes, the director of the NIA, at two or three or four of his trusted associates to say, oh, yeah,

00:37:43.540 we should do that. And so far, they haven't. All right. So there's one more feature of the ITP,

00:37:49.540 which is that basically anybody can make a suggestion for a molecule, right? Yeah.

00:37:55.660 So what was the first candidate molecule in terms of the first pool that was suggested?

00:38:02.560 I don't know. I have to go back and check my notes. And the reason I'm a little foggy on this

00:38:06.600 is that actually, we had a sort of a proto-embryonic ITP at the University of Michigan

00:38:11.700 only, a single site, for about two years. At that time, we had a Nathan Shock Center

00:38:16.160 funded by NIA. And I devoted my chunk of the Nathan Shock Center to testing four or five or six

00:38:22.960 drugs over a period of two years. None of them worked. And then the ITP got formally funded. It

00:38:28.780 became a multi-institutional program. That first year, there were four suggestions that we accepted.

00:38:34.120 One of them was aspirin. One of them was a molecule called NDGA, which is nor-dihydro-guiuretic

00:38:41.880 acid, which actually does work. It's worked three times in a row, although it works only in males.

00:38:46.800 One of them was nitrofluorobuprofen. And I forgot the fourth. I'd have to check my notes.

00:38:52.420 But those were the first four that were accepted for the initial round of the ITP.

00:38:57.360 Now, aspirin is interesting because it initially showed that it worked in males. But this is one of the

00:39:03.060 examples of something that didn't replicate after the ITP. Is that correct?

00:39:07.700 Yeah. It was not actually a formal attempt at replication. The initial dose of aspirin was very

00:39:13.860 low. That is one one-hundredth of the dose a person would take, even if adjusted for mouse body weight.

00:39:20.060 And it gave only an 8% or 10% increase, and it was in males only.

00:39:24.540 Why the decision to use such a low dose?

00:39:26.580 The sponsor, that is the person who says, please use this drug, we take his or her suggestion very

00:39:33.560 seriously. And the sponsor of this was a guy named Christian von Levenberg, who's now at the

00:39:39.760 University of Florida. And he gave a strong rationale for why he thought this dose was

00:39:45.420 appropriate. And we accepted his suggestion. And it seemed to be a fairly good suggestion in the sense

00:39:50.960 that it worked. But then several years later, we said, maybe it would have worked even better if

00:39:56.540 we'd used a higher dose. Maybe it would have worked even better, maybe even in females, if we used a

00:40:01.960 dose that approximates the sort of 83 milligrams a day thing that I used to take to prevent heart

00:40:07.960 attacks, that a lot of people take to prevent heart attacks. So we tried it at two doses higher than

00:40:13.880 the one we had originally evaluated. And it did not extend longevity. So it was not an exact

00:40:21.160 replication. It was a replication at higher dose. It succeeded the first time, sort of. It did not

00:40:28.540 succeed the second time. And by that point, we had enough things that really worked great that we

00:40:33.580 weren't about to devote additional time and energy to trying intermediate doses or something like that.

00:40:40.200 To give you a sense of the budgetary commitment, the total program budget each year in direct costs

00:40:48.120 is about a million bucks. So if we test six drugs a year, something like that, $3 million divided by

00:40:55.920 six, about a half a million buck investment for each drug accepted for an initial test into the program,

00:41:03.560 crudely speaking. So we have enough money now because the NIA has been extremely generous to us

00:41:09.000 that we can test five or six or seven drugs each year for the first time and also go back and recheck

00:41:15.740 one or two of these a year. But we can't go much above that without running out of money. So if we were

00:41:22.680 to decide to try aspirin at a few more doses, we'd be eliminating from our program some other drugs that

00:41:28.760 have never been tested but look really promising. So we make that decision each year as a group with the

00:41:35.380 help of five other scientists who are not part of the ITP but give us advice on what drugs to pick.

00:41:41.620 Is part of your decision to include a drug or not include a drug at all based on clinical data in

00:41:48.740 humans? Oh yeah, sure. I mean, we know that mice are not humans and the things that work in mice might

00:41:54.320 not work in humans and vice versa. But if someone says, see this drug shows potential for helping to

00:42:01.180 prevent cancer in humans or prevent some aspects of neurological decline in humans, that's a very

00:42:09.480 strong argument as to why it might well be worth testing in mice. Testing a drug in humans with

00:42:15.500 lifespan as an endpoint is almost hopelessly slow and almost hopelessly expensive. The ideal animal for

00:42:23.660 making a test of that is the mouse. So whenever we have a suggestion, we consider a wide range of

00:42:30.040 different kinds of evidence and clinical evidence about a health effect in people can be given a

00:42:36.320 very high weight. In addition, even if a drug is used in humans for something that's not arguably

00:42:43.440 unrelated to the aging process, the fact that it's FDA approved also is a big selling point because

00:42:48.580 if one of our drugs works in mice and it's already FDA approved, then the barriers towards evaluating it

00:42:55.920 in human clinical situations become much easier to get over because of the safety tests that have

00:43:03.560 already gone into obtaining FDA approval for some other indication, like metformin, for instance.

00:43:08.320 Yeah. We'll come to metformin because there's a few drugs on the ITP hit list. I want to really spend

00:43:13.660 some time on metformin as one of them, of course, rapamycin another. But I'm curious, like using aspirin as an

00:43:19.600 example, if you were today going to study aspirin for the first time, would you be swayed by the fact

00:43:26.740 that the evidence for use of aspirin to prevent cardiovascular disease is quite weak? Or would you

00:43:32.980 chalk that up to, hey, maybe for that indication it is, but nobody's really evaluated all-cause mortality

00:43:39.460 over a long enough period of time in humans, we think it's still worth studying.

00:43:43.540 You're thinking about this in a very good, sophisticated way. You could join our compound review committee.

00:43:48.720 So all of these points can and do come up in the discussion of which drugs to give highest priority

00:43:55.820 to. The fact that something might or might not modulate longevity or all-cause mortality in people

00:44:01.920 is worth knowing, but there are lots of things that test interesting ideas in mice that may not be of

00:44:10.600 primary relevance in the human clinical setting. Aspirin, for instance, is a famous anti-inflammatory,

00:44:16.540 and there are lots of good reasons, which I'm sure you know, suggesting inflammation may go up as you

00:44:21.700 get older and it's probably bad for you and it may contribute to lots of diseases. So there's a very

00:44:25.660 strong argument for saying, let's test some anti-inflammatory drugs in the mice. We're not

00:44:31.540 testing aspirin in mice to prevent heart attacks. Mice don't get heart attacks. But the notion that it's a

00:44:37.760 strong anti-inflammatory and that the processes of inflammation contribute to aging in both species,

00:44:44.360 that makes some sense. And that's a good reason to test it.

00:44:47.300 Yeah. Let's move to your second cohort here, because this is where you sort of hit one out

00:44:51.880 of the park, right? So in cohort two in 2005, started in 2005, rapamycin is one of the candidate

00:45:00.640 drugs. Who brought that to the committee? Dave Sharp. Dave Sharp was a colleague of Randy

00:45:05.660 Strong's at Texas. He's an expert on TOR, the target of rapamycin. He was aware of invertebrate

00:45:13.720 data in worms and in flies, saying that if you had a mutant that inhibited the function of the enzyme,

00:45:21.600 that led to longevity increases in these two invertebrate species. And so he said, look,

00:45:28.140 I've got a drug that can inhibit the enzyme. Why don't you give the drug to mice?

00:45:31.460 Yes. Rapamycin is that drug, and it's actually safe enough that you can use it in certain high

00:45:37.900 risk situations in people. So Dave Sharp suggested it, and we tried it. Randy Strong noticed, first of

00:45:45.480 all, that when you gave it to mice in the usual form, 95% of it was eaten up in the stomach. So

00:45:53.260 Randy and Dave and their collaborators then spent a year successfully trying to reformulate the

00:45:59.840 rapamycin by coating it in a kind of a shell that would get it through the stomach into the small

00:46:05.040 intestine where it would be absorbed. That's why for the first time around, instead of starting at

00:46:09.960 four months, which was our goal, we could only start at 19 months of age or 20 months of age. It took us

00:46:15.320 that long, took Randy that long to get the drug into a form which was effective when given by mouth.

00:46:22.320 But yes, then that was the first drug that gave a very strong signal in both males and females.

00:46:27.660 And it's still the only drug that we've tested so far, which gives a very strong signal in females.

00:46:33.220 We have two others that work repeatedly in females, but they're not as strong in terms of the size of

00:46:40.120 the effect as rapamycin in females. The other thing that's really noteworthy about this is what you

00:46:46.380 just alluded to, which is based on the need to reformulate rapamycin to improve its bioavailability.

00:46:53.000 You, without planning it this way, effectively, we're giving rapamycin to 60-year-old mice.

00:47:01.260 Yep, 20-month-old mice, which are sort of roughly the equivalent of a 60-year-old person.

00:47:05.720 So this is actually very interesting because the gold standard intervention for treating mice,

00:47:13.800 which is caloric restriction, typically only works when initiated earlier in life.

00:47:19.280 And so typically, when you wait until a mouse is two years old to begin calorically restricting it,

00:47:26.120 it's too little too late.

00:47:27.820 That's right. Richard Weindrich found that it works great, and caloric restriction works great

00:47:32.960 at sort of four or five months of age. 12 to 14 months of age, it still works, but not so great.

00:47:38.280 And then Weindrich found that if you start as late as 19 or 20 months in his hand, it failed,

00:47:43.100 just as you were saying.

00:47:43.780 So how do you put the rapamycin finding in? So one way that you can think about this, I guess,

00:47:51.240 is to think of it as here's a 50-year-old person. How many years do they have left in life, right?

00:47:58.840 And if you do nothing, we would say the average 50-year-old person has 30 years left. Yes? Is

00:48:05.460 that about the right number?

00:48:06.740 If she's a white woman, yes.

00:48:09.020 And a man wouldn't be too far off from that, right?

00:48:11.640 Five years less, but yes, that's right. That's generally correct.

00:48:15.360 So what did the rapamycin data suggest if they were to translate to humans, which is a big if?

00:48:20.980 The point you're making is exactly on point. I bet that it was a waste of time and money

00:48:26.300 to give rapamycin to 20-month-old mice, to middle-aged mice. I said, this can't possibly work.

00:48:32.040 Look, caloric restriction doesn't work at this age, just as you were saying. And so it's a waste of time.

00:48:36.820 And somebody said, yeah, but we've already got the mice. We've already paid for them.

00:48:40.960 Let's try it. And they were right, of course, and I was completely wrong. And that's a big surprise.

00:48:46.520 That, I think, is the reason why the paper was attractive enough to be published in Nature.

00:48:50.540 The notion is, and here this becomes hand-waving because we don't have a molecular explanation,

00:48:56.040 but the notion is that there are some processes really bad for you when you're a 20-month-old mouse

00:49:02.640 or a 60-, 65-year-old person that are still not irreversible, that can be reversed,

00:49:08.360 and that that reversing can be dependent upon inhibition of TOR signals.

00:49:14.260 So giving rapamycin to a 20-month-old mouse extends longevity dramatically,

00:49:20.500 just as dramatically as if you had started the rapamycin at a much younger age.

00:49:24.380 That's a fundamental, I think, reformulation of how we thought aging worked.

00:49:30.680 And independent of the notion that, say, you can take a drug and it can make you live a long time,

00:49:35.840 the observation that some, though not all of the drugs that work for us,

00:49:39.720 are fully effective when started late in life suggests two things.

00:49:44.880 One is the theoretical thing. Some things are going on late in life,

00:49:47.920 which are still reversible and have a major effect on your health.

00:49:51.060 And secondly, of course, if you have any friends who are 50 and 60 and 70-year-olds

00:49:55.600 and they would like to take a drug that makes them live longer, that's fairly good news.

00:50:00.720 So let's explain a slight semantic point that's going to be interesting or, I guess,

00:50:05.480 relevant as we kind of go through these things.

00:50:07.660 Can you explain to folks the difference between increasing median survival and maximum survival?

00:50:13.420 Like, how do you think about it?

00:50:14.820 What moves the needle for you when you're thinking about something?

00:50:17.900 Yeah, I mean, just technically, median survival is the age at which half of them have died

00:50:22.060 and half of them haven't died yet.

00:50:24.700 Maximum survival is a folk notion.

00:50:27.260 It's basically, how long is the oldest mouse or person who's ever lived?

00:50:32.280 And that's a value with very limited statistical appeal,

00:50:35.560 because the oldest person in a group of 100 people is not going to live as long

00:50:40.080 as the oldest person in a group of a million people or 100 million people or 10 billion people.

00:50:45.700 So the statistic that is probably intuitive and a best substitute for that is,

00:50:52.200 what is the age at which 90% of the people have lived?

00:50:55.600 We crudely refer to that as maximum lifespan, though it's really not maximum lifespan.

00:51:02.640 The reason it's an important concept in terms of understanding how to interpret survival statistics

00:51:09.240 and survival curves is that if you happen to have, let's say, a group of animals,

00:51:14.300 many of which are dying, let's talk about in terms of human years,

00:51:17.720 most of them are dying in their 30s and 40s and 50s or something,

00:51:21.240 and you have something that extends their median lifespan up to 60,

00:51:27.440 you can see as a public health benefit.

00:51:29.840 That's why a lot of people get immunized and don't want to have their kids smoke, etc.

00:51:33.720 But it hasn't worked on the aging process.

00:51:36.700 It hasn't, in particular, had any real effect on how much longer you're going to live when you're 70.

00:51:43.280 But if you had a drug or a diet or some intervention that authentically slowed the aging process,

00:51:52.300 then it would extend expectancy of additional healthy years of life,

00:51:57.920 even for those who are already quite old in their 50s, 60s, 70s, and 80s.

00:52:01.600 And that will modify the age of death at the 90th percentile.

00:52:07.060 So drugs that only extend the median and don't affect the age of death of the longest live, 5% or 10%,

00:52:16.760 they might be interesting in some ways,

00:52:21.060 but they are considered less plausible as candidates for anti-aging drugs.

00:52:26.420 If you have a drug that's authentically slowing aging,

00:52:30.160 one of the things you most want to see is that the very oldest animals in this drug-treated group

00:52:36.480 are living longer than the very oldest animals in the untreated control group.

00:52:41.140 And so when you look at the rapamycin data, they were very interesting.

00:52:45.080 The median extension, which we've already said, not as interesting.

00:52:49.320 That's just saying what's the age at which half the animals have died.

00:52:54.280 In the males, it went up 20%.

00:52:56.600 In the females, it went up 13%.

00:52:58.880 These are huge numbers.

00:53:00.340 I think you may have the numbers backwards.

00:53:02.180 In the females, the median goes up more than in the males.

00:53:05.220 Oh, I thought that was for the 90th percentile.

00:53:08.780 It went up more in the females.

00:53:10.560 For both of them.

00:53:11.420 It's a technical artifact because when you give the same dose of rapamycin in chow,

00:53:16.540 the blood levels in the females are three times higher than the males.

00:53:19.780 So we don't know why.

00:53:21.180 I'd like to come back to that in a second.

00:53:22.760 The numbers I've always remembered are actually not the median ones.

00:53:26.200 I could be out to lunch on those.

00:53:27.620 The ones I have ingrained in my head, and I hope I have them right,

00:53:31.980 is the P90 lifespan extension, which in males was 9% and females was 14%.

00:53:39.820 Do those sound right to you?

00:53:41.240 I'd have to look it up, but that sounds plausible.

00:53:43.420 Yep.

00:53:43.540 That's a big deal because that's total lifespan, isn't it?

00:53:48.560 That's not incremental above the point in life where they are.

00:53:51.280 You got it.

00:53:51.460 We don't cheat.

00:53:52.580 We don't calculate incremental lifespan.

00:53:54.400 All right.

00:53:54.920 So let's explain to people what that means because these are huge numbers.

00:54:00.440 Yeah.

00:54:00.700 So explain to people why that basically translates to 25% more life once you've reached midlife

00:54:09.360 versus 9% or 14% more life.

00:54:12.460 This is a subtle way of doing the math more rigorously.

00:54:16.700 Yeah.

00:54:16.820 The editors of Nature insisted that we do that calculation.

00:54:20.080 That is the number of additional years or months of life after you give the drug.

00:54:25.160 They thought that would sex up the paper and kicking and screaming.

00:54:28.980 We put it into the paper, but it's not, I think, terribly informative.

00:54:33.380 There's an easy to imagine thought experiment.

00:54:36.600 You've got someone on a ventilator.

00:54:38.460 He or she is on death's doorstep.

00:54:40.820 They're going to die today.

00:54:42.860 And you leave them on the ventilator for one more day, and then you give up and they die.

00:54:47.900 So you've extended, you've doubled their lifespan, right?

00:54:50.440 You've gone from one day to two days.

00:54:52.720 That's 100% increase in lifespan.

00:54:56.520 So when I put it that way, you can see that that's not a very useful statistic to calculate.

00:55:02.260 If you're an insurance agent, it's nice to know at the time you sell the insurance policy,

00:55:06.380 what the life extension will be from the time the person buys the insurance policy.

00:55:10.700 But in terms of biology and pathobiology and being able to compare lab A to lab B, drug

00:55:17.260 A to drug B, I think keeping it on the level of what is the change in the overall median

00:55:23.540 is important.

00:55:25.280 There's one exception to that, and that's the fairly obvious one.

00:55:28.620 If you start a drug really, really late, if you start a drug at an age when 35 or 40%

00:55:34.720 of the animals have already died, you're asking a lot of that drug to extend the median lifespan.

00:55:42.760 So if you start rabamycin at 19 or 20 months of age, in males at 20 months of age, 15 or

00:55:47.620 20% of the mice have already died, depending on what site you're talking about.

00:55:51.100 So if you get any extension of the median lifespan, it sort of had to start working immediately.

00:55:57.180 Nobody could die for the next few months.

00:55:59.320 That's the situation where the statistics on how many of them make it to the 90th percentile

00:56:07.960 become more informative than a percentage change in median.

00:56:12.180 And to my eye, rapamycin is the only drug you've started at 20 months.

00:56:18.160 I know you repeated it once at 20, but that's a big ask of a drug, right?

00:56:24.000 Well, I'm pleased that it worked, but it's not the only drug that works.

00:56:27.160 Well, I'm saying starting at 20 months of life.

00:56:29.620 I know, and that's what I'm talking about.

00:56:31.020 Yeah, we've used a carbose starting at 20 months of age.

00:56:33.660 Well, you did, okay.

00:56:34.560 I thought a carbose started earlier, okay.

00:56:36.860 Yes, you're right.

00:56:37.600 The initial paper was a carbose at four or five months, yeah.

00:56:41.020 But then we repeated it, and that's published too, where we started a carbose at, I believe,

00:56:45.940 20 months of age.

00:56:47.200 And it worked half as well as a carbose started in youth.

00:56:51.800 That is, we got a statistically significant change in the maximum lifespan by our statistical

00:56:57.000 procedure in both sexes, and a statistically significant change in the males, even in the

00:57:02.940 median lifespan.

00:57:04.340 In the females, the p-value for median was 0.07, if memory serves.

00:57:08.360 So a carbose started in late middle age is effective in both sexes, not as effective as

00:57:17.500 if you started in youth.

00:57:19.300 And by the time this podcast airs, our paper on 17-alf estradiol late start will have appeared.

00:57:27.040 And 17-alpha estradiol, which only works in males, the new paper just now at the last

00:57:33.760 stages of revision, show that if you started at 16 months of age in males, it's just as

00:57:40.440 good as if you started at an earlier age.

00:57:42.820 Even if you started as late as 20 months of age, it works great, just as much.

00:57:48.040 It's not statistically distinguishable from the 16-month start.

00:57:51.700 So that means of the drugs that we've tested for late start, rapamycin works perfectly.

00:58:00.740 A carbose works about half, half as well as an early start.

00:58:05.540 And 17-alf estradiol seems to work just as well in late middle age.

00:58:10.720 A former student of mine, now an independent researcher named Mike Garrett, together with

00:58:16.140 his colleagues, Charlene Day and John Herrera, have already published a paper using these

00:58:21.240 same mice treated with 17-alf estradiol, where they started at 16 months or 20 months of age.

00:58:27.760 And they found that the muscles got stronger, their glucose tolerance got better, muscle structure

00:58:33.620 changed for the better.

00:58:35.500 So they published that before we could publish our lifespan data, but it's quite consistent.

00:58:40.040 Even late start for 17-alf estradiol seems to have highly beneficial effects, just as good,

00:58:47.000 and in some cases, even a little better than if you start in earlier ages.

00:58:50.160 Now, I definitely want to spend some time on acarbose and 17-alf estradiol along with a

00:58:54.420 few others, but I want to go back to something in rapamycin.

00:58:56.880 So we've talked about the first experiment.

00:59:00.300 Tell me about the dosing.

00:59:01.640 Was the dosing done daily?

00:59:03.780 For rapamycin?

00:59:04.820 Yes, in that first experiment?

00:59:06.180 Sure.

00:59:06.480 It's in the food.

00:59:07.180 They eat all they want every single day.

00:59:08.620 So this is interesting, right?

00:59:10.620 Because what we've learned about rapamycin since that time is it inhibits two complexes

00:59:18.660 of TOR, complex one and complex two.

00:59:21.540 Now, today we believe most of the longevity benefits of rapamycin come from the inhibition

00:59:27.640 of complex one, not complex two.

00:59:30.460 And furthermore, we believe that some of the negative consequences of constitutive use of

00:59:36.400 rapamycin come from its inhibition of complex two.

00:59:41.400 And so when you look at some of the more recent human data using rapalogs such as Everolimus,

00:59:48.780 they seem to favor an intermittent strategy.

00:59:51.400 So dosing rapamycin, say, weekly, which is enough to inhibit, if you give a big enough

00:59:58.140 dose, it's enough to inhibit complex one, but not enough to inhibit complex two.

01:00:02.820 And then by the time you come back to re-dose, you're in this situation of you're just knocking

01:00:07.440 down one, but never two.

01:00:10.440 So of course, I've always found it just interesting that the study even worked with daily dosing

01:00:17.860 of rapamycin.

01:00:19.020 What are your thoughts on that?

01:00:20.920 You raised five or six or seven or eight really interesting points there.

01:00:25.040 So let's go back and do the one at a time.

01:00:27.900 Now, rapamycin does not directly inhibit mTOR complex two.

01:00:31.880 What it does is it leads to feedback circuits, which destabilize mTOR complex two and eventually

01:00:36.640 cause it to be degraded.

01:00:39.140 So the subtleties of the kinetics of when to start it, when to start it, when to get it

01:00:43.600 once, and it may well differ from cell type to cell type.

01:00:48.180 So those are things that pharmacologists need to work out carefully.

01:00:51.600 The second thing to point out is that a guy in my lab, Gonzalo Garcia, has very carefully

01:00:57.060 looked at TOR complex one and TOR complex two activity for the last three or four years

01:01:01.500 in, not in these drug treated mice, but in mutant mice that have one of two mutations,

01:01:07.880 the Snell-Dwarf mutation and the growth hormone receptor knockout mutation that extend longevity

01:01:11.780 by 30%.

01:01:12.820 And Gonzalo found something very interesting.

01:01:15.680 Both of those mutations move TOR complex one down.

01:01:18.640 That's good.

01:01:20.460 But they move TOR complex two up in the opposite direction.

01:01:24.580 So this raises the possibility that is actually the elevation of TOR complex two in these mutant

01:01:30.540 mice, it's good for them.

01:01:32.640 And Gonzalo has found several examples of mice where there's a mutation that blocks the mTOR

01:01:38.320 complex two elevation.

01:01:40.260 They do not live a long time.

01:01:41.660 So most recently, and this is a paper that's just submitted, it's not published yet, Gonzalo

01:01:47.120 has taken these drug treated mice, rapamycin or acarbose or 17-alf estradiol and look to

01:01:53.680 see what's happening there with mTOR complex one and mTOR complex two.

01:01:58.400 There are changes with rapamycin, both males and females.

01:02:02.520 There are changes with acarbose, both males and females.

01:02:06.060 And there are changes with 17-alf estradiol, but it's the males only.

01:02:10.880 And that's really cool because it's the males only that get a lifespan benefit from 17-alf

01:02:16.020 estradiol.

01:02:17.180 So to summarize all of that, it's going to be really complicated.

01:02:21.420 The basic idea that knocking down TOR complex one might be a good thing and that knocking

01:02:27.360 down TOR complex two might be a bad thing.

01:02:30.200 That was the premise of your question.

01:02:31.480 I think that's a good initial sound foundation for further work.

01:02:36.220 But the interactions between them, the ways in which different cell types may have different

01:02:41.640 responses, it's going to be much more complicated than that.

01:02:44.400 The last complication I'll point out is that Gonzalo has also found that mTOR complex two

01:02:50.700 has four different substrates.

01:02:52.540 That was widely known before he began this work.

01:02:55.100 What he found is that the good stuff turns on three of these targets for mTOR complex two

01:03:00.840 and turns the other one off.

01:03:02.520 So it's really, if he's right, it's really a change, not so much in the amount of TOR

01:03:08.520 complex two, but in its target specificity, which particular substrates it modifies and

01:03:14.500 which ones it stops modifying.

01:03:16.120 That's a subtlety that may be the whole story there as to why changes in mTOR complex two should

01:03:23.880 be a part of any sophisticated study of this drug and other drugs.

01:03:27.320 When you repeated the rapamycin study one year later in cohort three, obviously having worked

01:03:34.320 out the delivery system, the average age at treatment was now nine months instead of 12

01:03:40.200 months.

01:03:41.480 And obviously this was interesting because-

01:03:44.060 Nine instead of 20, you mean.

01:03:45.260 Sorry, sorry.

01:03:45.820 Nine instead of 20.

01:03:46.740 So now this is interesting in that, A, you're going to repeat a very important finding, and

01:03:51.040 two, you're going to see if the age effect was significant.

01:03:53.680 And it's kind of remarkable, both groups got a little bit better, which I guess you would

01:03:58.660 expect, but it's worth noting, you still see these remarkable effects, right?

01:04:03.260 You saw a lower median extension because of the point that you mentioned earlier, but the

01:04:09.920 maximum extension in males went from nine to 11%.

01:04:13.360 And in females, it went from 14 to 16%.

01:04:16.840 So again, females had an advantage with rapamycin, but both of these numbers are really quite

01:04:22.720 impressive.

01:04:23.360 I mean, these are pretty significant increases in lifespan.

01:04:26.600 And you're now, you know, functionally starting this in people in their, you know, early thirties,

01:04:31.580 basically.

01:04:32.420 Yeah.

01:04:32.960 So there are two things to say.

01:04:34.340 One is that the curves for the early start, nine months, or late start, 20 months, were compared

01:04:39.800 very carefully by Scott Pletcher, who's a mathematical demographer, biodemographer.

01:04:43.780 And they're indistinguishable.

01:04:46.480 That is, there's no statistical difference between the early start mice and the late start

01:04:51.340 mice.

01:04:52.000 Small differences like 9% versus 11% or 9% versus 12%.

01:04:56.740 There's a lot of variance from mouse to mouse and group to group.

01:05:02.040 You don't want to take minor changes like that with, put too much weight on them.

01:05:07.000 They're more or less the same thing from the point of view of replicability.

01:05:10.720 The other point you made that we referred to a moment ago had to do with the male versus

01:05:15.660 female differences for rapamycin response.

01:05:18.760 Randy Strong and Marty Javers and their colleagues at the University of Texas did a really useful

01:05:24.460 study.

01:05:24.900 They gave rapamycin for a short period of time to male and female mice and then took blood

01:05:29.020 samples every, I don't know, every hour, every two hours or something.

01:05:31.820 And the rapamycin blood content in the female mice was two or three times higher than it was

01:05:38.860 in the male mice.

01:05:40.240 So it was higher and it stayed up longer.

01:05:42.400 We don't know why rapamycin in the blood of female mice is greatly elevated compared to

01:05:48.600 rapamycin in the blood of male mice.

01:05:50.560 But when you're comparing two curves at the same dose in food and the females do a bit

01:05:55.640 longer than the males, get a bigger benefit than the males, it's hard to know how that

01:06:00.700 comparison would come out if you're actually adjusting them to the same blood level.

01:06:04.040 Have you seen that with any other drug, Rich, where going into the food, same quantity going

01:06:09.740 in, you see a 3x difference in plasma level?

01:06:14.840 No, we haven't looked very carefully.

01:06:17.600 I mean, we did it for rapamycin because we were so surprised that the females were doing

01:06:22.300 better than the males.

01:06:23.540 And it's a difficult and expensive study to do.

01:06:26.780 I think we should do it for other drugs as well.

01:06:30.760 We, Marty and Randy, do check the drug concentration as a part of a pilot project of mice given any

01:06:38.720 one of our drugs.

01:06:39.740 They're given the drug for eight weeks of age, then Marty and Randy get blood samples.

01:06:43.980 But that gives you just a sort of a steady state dose.

01:06:47.200 What if you're eating it every single day for eight weeks?

01:06:49.560 What is the blood level?

01:06:51.000 The study that was most informative for rapamycin was to take fresh virgin mice and give them

01:06:56.660 the drug at a defined date, at a defined amount, and then quickly measure the blood levels after

01:07:01.760 that.

01:07:02.600 That has to be set up with a separate cohort of mice.

01:07:05.320 If you want to be obsessive, as you should, you have to do it at young mice, middle-aged mice,

01:07:10.340 old mice.

01:07:11.040 It becomes kind of involved.

01:07:12.500 And we don't do it on a routine basis.

01:07:15.820 So the next drug I want to talk about, well, actually, let's go to acarbose.

01:07:19.240 You brought it up.

01:07:19.920 So let's go to that before we get to resveratrol.

01:07:22.640 I mean, it's amazing.

01:07:23.840 Like, Rich, we could sit here and talk for the next month and go through every single

01:07:28.880 drug in the ITP history because there's not one of these that isn't interesting, both in

01:07:33.940 its success and failures.

01:07:35.980 Obviously, we don't have the time for that.

01:07:37.320 We're going to, in the show notes to this, put out a spreadsheet that covers every single

01:07:42.200 result.

01:07:43.420 But nevertheless, it's just an amazing body of work.

01:07:47.320 So let's talk about acarbose.

01:07:48.700 Now, was David Allison one of the first people involved in that?

01:07:51.280 I know David very well, and I know that David's been involved in a number of suggestions around

01:07:55.480 the ITPs.

01:07:56.140 He's sponsored some.

01:07:57.500 Was he originally involved in the acarbose work?

01:08:00.020 Yeah, the acarbose application came in from David Allison and a colleague of his, Daniel

01:08:03.960 Smith.

01:08:04.540 Okay.

01:08:04.800 This is an off-the-shelf drug typically used in people with diabetes, and it basically

01:08:10.060 blocks the absorption of glucose in the gut.

01:08:13.140 More or less well-tolerated unless you take too much of it, and in which case, you know,

01:08:17.620 you're going to get some GI distress.

01:08:19.720 But it works well.

01:08:21.340 I've certainly taken it a number of times.

01:08:23.340 I've got bottles of it laying around the house.

01:08:25.280 I probably haven't taken it in a couple of years.

01:08:27.740 But it, you know, I used to use it as kind of my cheat meal drug.

01:08:32.400 If I wanted to have like a pizza, I'd take, you know, 100 milligrams of acarbose, and it

01:08:37.440 would manage to, you know, prevent my glucose from spiking.

01:08:41.100 They should give it away with every pizza.

01:08:42.880 I agree.

01:08:43.260 That's a good marketing strategy.

01:08:44.680 There you go.

01:08:45.380 Take your acarbose with your carby junk food.

01:08:47.920 Sprinkle it on the pepperoni.

01:08:49.380 Now, of course, what you guys were saying is, hey, we're not just giving the mice the

01:08:53.260 acarbose with their pizzas.

01:08:54.440 They're going to eat this stuff every day, right?

01:08:56.560 It's part of the chow.

01:08:57.540 Yep.

01:08:57.700 Okay.

01:08:58.340 So what was your thinking?

01:09:00.480 What was going to happen here?

01:09:02.120 Well, Daniel and David said caloric restriction is good for you.

01:09:07.100 Acarbose is sort of like caloric restriction.

01:09:09.520 It doesn't block glucose absorption, but it does block the digestion of starches to sugars.

01:09:14.460 And so glucose wouldn't go up so much.

01:09:16.440 That would be good for you, just like caloric restriction in their view.

01:09:19.680 And that was their rationale.

01:09:21.620 Sorry, just to be clear, I understand that because there's two different hypotheses that

01:09:25.340 you could be testing here.

01:09:26.120 One hypothesis is you will functionally consume or absorb fewer calories, and therefore this

01:09:34.040 will be a CR play.

01:09:35.880 The other is, no, the animal will continue to eat.

01:09:38.300 You know, they'll make up for it with more calories, but they will have lower glucose.

01:09:41.860 Which of those two is the thinking here?

01:09:45.020 I don't think Daniel and David did a particularly compelling job of discriminating those two ideas,

01:09:50.700 the sort of framework for their application, which, you know, I haven't read for what,

01:09:54.140 12 or 13 years now.

01:09:55.380 I'd have to go back to see exactly what they argued.

01:09:57.800 But they were saying it's sort of like caloric restriction a little.

01:10:00.140 Why don't we try it?

01:10:01.560 Now, it's not like, as it happens, it's not like caloric restriction.

01:10:04.440 There are a lot of reasons for saying it's not reproducing caloric restriction.

01:10:08.080 But anyway, that was the original rationale.

01:10:10.800 My current interpretation is that it probably is operating by blocking very highest levels of

01:10:17.080 glucose.

01:10:18.460 It, in the mice, did not lead to a change in the integrated glucose level.

01:10:24.680 There's a clinically useful measure, which is used in human diabetics too,

01:10:28.580 hemoglobin A1c, which gives you a measure of over the last few weeks how much average

01:10:34.520 glucose has been in the serum.

01:10:36.820 If a person with diabetes takes a carbose, that hemoglobin A1c goes down.

01:10:41.820 That's one of the ways in which you know it's working in a person.

01:10:44.480 Now, in mice, it turns out it doesn't go down.

01:10:46.980 We've done this twice now.

01:10:48.780 So it's probably not an overall change in the amount of glucose that gets in, but a

01:10:54.440 change in the kinetics.

01:10:55.620 If I eat five slices of pizza, my blood glucose is going to shoot way up and then come down.

01:11:02.420 But if I've taken a lot of acarbose first, blood glucose will go up.

01:11:06.020 It'll just go up slow.

01:11:07.660 It won't reach that big peak.

01:11:09.940 It'll stay up longer overall, but it won't hit the big peak.

01:11:14.120 And so currently, our guess was, mine and the other colleagues, was that it was working

01:11:20.840 by blocking the peak glucose.

01:11:23.080 And now we think the evidence for that is quite good because of a drug you haven't mentioned

01:11:27.300 yet, canagliflozin, which was just published a few weeks ago.

01:11:31.020 Canagliflozin, which is also used for diabetes in people, also blocks peak glucose in mice

01:11:36.680 and in people.

01:11:37.900 And it also extends lifespan.

01:11:39.500 And it also works preferentially in males.

01:11:42.160 So it's a reasonable guess that both acarbose and canagliflozin are working by eliminating

01:11:49.520 the huge peak of glucose you get after you eat a meal with a lot of starch in it.

01:11:55.600 And of course, David also was part of the suggestion team for Canna.

01:11:59.720 So I know that David and I have spoken at length about SGLT2 inhibitors for quite some

01:12:03.580 time, and we are absolutely going to get to that.

01:12:06.120 So did the mice experience any GI distress in any way that you could assess that?

01:12:11.340 None that they complained about.

01:12:14.800 They didn't fill out the questionnaire?

01:12:16.840 So we don't have any way of detecting how the mice are feeling.

01:12:21.040 They didn't stop eating or show obvious signs of distress.

01:12:25.820 You know, their GI tract and our GI tract are very different.

01:12:29.420 We're omnivores, grew up as scavengers 200,000 years ago.

01:12:33.800 We'll eat almost anything, animal or vegetable.

01:12:37.180 And mice are evolved to be interested in mostly a grain diet.

01:12:41.120 So direct one-to-one comparisons for what our GI tract can handle and theirs may be kind

01:12:47.920 of tricky.

01:12:49.200 Did those animals, both the treatment and the control groups weigh the same?

01:12:54.080 No, acarbose led to weight loss or a lack of weight gain.

01:12:57.940 We don't understand it.

01:12:59.300 We see that in some populations, but not all.

01:13:02.220 In the original ITP paper, the acarbose-treated mice were lower in weight than the control mice.

01:13:09.220 But then my own lab made a whole batch of acarbose-treated mice for another purpose, and we didn't see a

01:13:14.500 dramatic change in weight.

01:13:15.740 And we really don't know why.

01:13:17.560 So it's a little bit frustrating and a little bit mysterious and embarrassing.

01:13:22.560 We don't always see the same weight change.

01:13:25.420 I assume the rapamycin treated and controls did not have a difference in weight?

01:13:30.200 It's complicated, and it depends on the age at which you start.

01:13:34.080 If you start in youth, then when the mice are still gaining weight, the mice on rapamycin do

01:13:40.520 not gain weight as much as the control mice, which has not surprised me because rapamycin stops or

01:13:47.880 slows cell growth and cell division, etc.

01:13:50.820 If you start in late age, like 20 months of age, you have a whole series of factors going on.

01:13:58.460 Mice, when they get old, they often get sick.

01:14:00.800 When they get sick, they often lose weight.

01:14:02.560 And if rapamycin is extending their lifespan, then they're getting sick later, so they're not losing weight because they're healthier.

01:14:11.740 And there's several different things going on in old age, 20 months of age, where rapamycin in principles would increase some of them,

01:14:20.180 decrease some of them.

01:14:21.380 It's harder to interpret whatever you see.

01:14:23.420 Before we leave ACARBOS, I mean, one really important point that I'm sure you're aware of, but I think deserves some emphasis.

01:14:32.340 Since ACARBOS is FDA-approved and has a long safety history, hundreds of thousands of people take it,

01:14:38.960 particularly in Asia and in some parts of Europe, it's an obvious candidate for a human clinical trial.

01:14:45.580 The safety profile is quite good.

01:14:48.180 It's very well documented, with many cases having been evaluated for safety.

01:14:54.600 So if one was thinking about what drug that works pretty well in mice would be safe for a human clinical trial,

01:15:01.200 ACARBOS has to be on that list.

01:15:03.160 I've been arguing with your friend, Matt K...

01:15:05.720 My friend, too, Matt Kaberlein.

01:15:07.420 He's giving rapamycin to dogs, as you know.

01:15:10.080 That's a very important, very interesting, excellent study.

01:15:13.060 And I've been trying desperately to get him to give ACARBOS to dogs, too.

01:15:16.960 And we're still not yet having reached a consensus on that point.

01:15:22.160 Let's come back to this point specifically, Rich, when we get to metformin.

01:15:27.220 Because to me, the metformin ACARBOS story is a very interesting one,

01:15:32.080 as we suss out the difference between very compelling human data and failures in the ITP

01:15:40.840 versus successes in the ITP absent the out-of-the-ballpark human data.

01:15:48.600 And I think that the ACARBOS metformin story is a great example of that.

01:15:52.200 So let's park that.

01:15:54.600 But let's visit now resveratrol, right?

01:15:56.520 So for reasons that aren't entirely clear to me,

01:15:59.280 I mean, I suppose it's because of the relationship between resveratrol and grapes and wine.

01:16:04.720 This just was one of those things that caught the world by storm,

01:16:09.980 and it's never really gone away, right?

01:16:11.960 So if I had a dollar for every time I got some stupid Google alert telling me to drink more wine

01:16:17.920 because of its anti-aging benefits, I'd have a lot of dollars.

01:16:21.880 So David Sinclair, I don't remember the original paper, would have been 03, 04-ish?

01:16:27.940 You probably know this better than I do, but that sounds about right.

01:16:31.080 Yeah, so less than 20 years ago, basically, David's lab at Harvard published work showing

01:16:37.820 that when resveratrol was given to metabolically ill mice that were being basically overfed,

01:16:47.440 it produced a longevity benefit.

01:16:51.580 Sort of.

01:16:52.160 We can come back and examine that claim in more detail if you wish, yeah.

01:16:56.820 Well, let's do that now, and then let's talk about how it informed your design.

01:17:03.760 Yeah.

01:17:04.460 So the mice in question were being given a diet consisting of 60% coconut oil.

01:17:08.520 They were being poisoned.

01:17:09.900 The original paper did not mention what they died of.

01:17:13.080 They did mention this in a subsequent paper published in Cell about two or three years later.

01:17:17.880 The mice that were dying because they were on 60% coconut oil were dying because their livers

01:17:23.680 got so big, so filled with fat, that it compresses the chest cavity and crushes the lungs and the

01:17:29.840 mice cannot breathe.

01:17:31.500 That's the cause of death in mice in this particular study that are given a 60% high-fat diet,

01:17:38.380 so to speak, that was used in the original resveratrol paper.

01:17:42.420 So understand, they're not really studying aging.

01:17:44.840 They're studying a bizarre pathological process where the liver gets so fat that it crushes

01:17:50.640 the lungs, preventing breathing.

01:17:53.460 And then, in addition, the paper that they published reported median lifespan.

01:17:58.280 There was a statistically significant increase in median lifespan, but only in the animals that

01:18:05.320 were on this highly toxic diet.

01:18:07.840 Now, it turns out that when you have the whole curve and you look at other indices of maximum

01:18:14.940 lifespan, the resveratrol was not a benefit.

01:18:17.900 And when you look at the lifespan effects in the mice that were not on this toxic diet,

01:18:22.880 there weren't any.

01:18:24.580 So paper in Nature, which was widely misinterpreted, over-interpreted, as a demonstration of a drug

01:18:31.740 that slowed aging, was a drug that did not extend maximum lifespan, except in mice that were dying

01:18:38.420 of this extremely unusual lipid-specific poisoning.

01:18:43.180 So the second part of your question was, how did this influence our decision to test resveratrol?

01:18:49.480 And the answer, this is behind-the-scenes gossip, but it's completely true,

01:18:54.000 is that we were ordered to test it.

01:18:56.100 Richard Hodes, the director of the National Institute on Aging, was very impressed with

01:19:00.820 resveratrol.

01:19:01.740 And like you, he was getting, I'm sure, hundreds or thousands of questions a year, requests

01:19:07.600 say, why don't you guys test resveratrol?

01:19:10.040 So resveratrol did not go through our usual screening process.

01:19:14.140 This was a directive from the top.

01:19:15.720 This was the only time this has happened.

01:19:18.120 We were instructed, you will be testing resveratrol.

01:19:21.260 And I called David Sinclair and said, what dose should we use?

01:19:26.060 The same dose that you used in your paper?

01:19:28.140 And he said, no, no, no, no, no, that's much too low.

01:19:30.400 Use at least five times, 10 times, 20 times higher.

01:19:33.580 So we followed David's advice and also checked with his close friend and colleague,

01:19:39.700 Rafa DiCavo, to get advice on dose.

01:19:42.380 So the two doses that we used for resveratrol were, I'd have to look it up,

01:19:47.300 be it the three times and 10 times higher than in the original Sinclair paper,

01:19:51.540 or maybe it was 10 and 30 times.

01:19:53.000 I'd have to check my notes to get that exactly right.

01:19:55.760 But we used both doses, both high.

01:19:57.700 And then we started at two different ages because we were told to do,

01:20:02.160 start some of the mice in youth and others of the mice in middle age.

01:20:05.560 And we did that and it had no effect on longevity.

01:20:08.580 And we were, I think, the first of three or four groups, including Sinclair and DiCavo later on,

01:20:14.920 to show that resveratrol given to mice on a normal diet does not extend their lifespan.

01:20:20.660 Yeah.

01:20:20.880 So you, in cohort three, you started at 12 months of age.

01:20:24.760 So functionally, these are mice in their 30s to 40s.

01:20:28.360 And then you had a cohort that started at four months of age.

01:20:31.360 You're starting these, you know, in teenagers.

01:20:33.720 There's no change in median or maximum lifespan.

01:20:38.760 Although I believe you had a small statistically significant increase in maximum extension and

01:20:45.820 maximum lifespan expansion in the late onset resveratrol by 3%.

01:20:49.460 Does that sound about right?

01:20:50.920 I don't think any of them were statistically significant.

01:20:52.940 Again, I'd have, I don't want to put my hand in a Bible until I've reread the paper,

01:20:57.100 but I think there was no change in any of the statistics.

01:20:59.980 It clearly speaks to the point earlier, which is in the spirit of looking at this through

01:21:07.060 the lens of studying longevity, as opposed to studying protection from a sort of artifactual

01:21:14.620 diet, this appeared to be pretty clear.

01:21:17.840 What was the response like to that?

01:21:20.720 There are people who are, who were making a living by selling stuff that was related to

01:21:25.480 resveratrol.

01:21:26.260 There was a company actually, Sirtris, that was sold to Glaxo for 600 million bucks for

01:21:32.520 pushing Sirtuin inhibitors.

01:21:35.320 So the reaction from people who had just sold the company was probably, I don't care, I've

01:21:39.620 got the money.

01:21:40.520 The reaction from people who sell things in health food stores, which are guaranteed to

01:21:45.560 raise your resveratrol levels or that have purple-y grapes on their cover, they were raking

01:21:50.940 in money hand over fist.

01:21:52.220 They don't really care who reads papers failing to show an effect in mice.

01:21:57.140 It's not going to be a problem for their bottom line.

01:22:00.820 You know, I was always very irritated over and over again.

01:22:03.940 Pretty good scientists whom I really respected.

01:22:06.540 The first slide was a bottle of wine.

01:22:08.380 The second slide was some nice grapes.

01:22:10.040 The third slide was a picture of resveratrol.

01:22:13.160 Now, by that time, they knew that resveratrol was not an important component of wine.

01:22:18.580 Sinclair had written a beautiful review article showing that if you really want to eat resveratrol,

01:22:24.060 it's easy.

01:22:24.640 You eat rhubarb.

01:22:25.600 Rhubarb is where all the resveratrol is at.

01:22:28.260 And red wine has got so little that to get the mouse dose of resveratrol, you'd have to

01:22:33.580 drink 600 bottles of wine a day.

01:22:36.420 Even Andre the Giant couldn't have done that.

01:22:38.980 There's a beautiful satirical article in The New Yorker about what it's like to be a mouse

01:22:43.340 drinking 600 bottles of really good French champagne a day that I said to Fred's when

01:22:48.480 they asked about resveratrol.

01:22:51.040 So I think eventually there's been a lot of controversy as to whether the sirtuins are

01:22:55.780 important or not in the aging process.

01:22:58.460 There's a big fight between two worm groups, one of which said mutants that modified the

01:23:03.920 activity were highly effective, and others said, well, maybe just a little bit.

01:23:07.880 And leaving all that stuff aside, I don't think there's any evidence yet that resveratrol

01:23:13.160 is good for you, or that sirtuin activators slow the aging process.

01:23:18.120 I was invited to give a talk at Sirtris.

01:23:20.480 Interesting.

01:23:21.160 At the height of all of this, David Stipp, who's a wonderful science journalist, has written

01:23:24.740 a wonderful book about the whole Sirtris-Resveratrol story.

01:23:29.080 The Youth Pill or The Youth Fountain or The Youth Pill or something like that?

01:23:31.920 That's about right.

01:23:32.420 Yeah.

01:23:32.580 It's a great book.

01:23:33.340 David's a fine writer.

01:23:34.680 So I went to Sirtris.

01:23:35.640 I gave my talk on rapamycin, which is what they wanted to hear.

01:23:39.480 And at the end, I went to talk to their director.

01:23:43.000 And I said, so what are you guys doing for resveratrol?

01:23:46.460 Is this going to work?

01:23:47.800 And he said, oh, no, I'm leaving the company.

01:23:49.840 I'm going to work for a company that works on it.

01:23:51.720 He named another chemical, entirely different.

01:23:53.880 He was going to work on tor inhibitors.

01:23:55.940 He wasn't going to work on resveratrol.

01:23:57.620 They'd already sold the thing or about to sell the thing to Glaxo.

01:24:00.400 And he knew that it was tied to move all up to inhibitors or compounds that had a greater

01:24:06.480 likelihood of actually working.

01:24:09.480 Well, we're going to come to one of those in a little while, which is basically things

01:24:13.760 that provide substrates for sirtuins as opposed to activators.

01:24:18.000 But nevertheless, one of the other ones that sort of I thought was interesting that you

01:24:22.180 studied, not that you shouldn't have in the spirit of this being kind of a community

01:24:25.640 effort was green tea extract.

01:24:28.100 Does that, do you remember that one at all?

01:24:29.840 I sure do.

01:24:30.640 Tell me a little bit about that.

01:24:31.720 What was your expectation going into that ITP?

01:24:34.460 Well, anybody in the world can suggest a compound or a group of compounds or a mixture of compounds.

01:24:39.580 You realize I'm going to be doing this now.

01:24:41.000 You're going to be getting these suggestions from me.

01:24:43.240 So just.

01:24:44.000 Oh, good.

01:24:44.440 Yeah, that's good news.

01:24:45.660 FYI, I'm sorry, in advance.

01:24:46.900 Each of these suggestions is evaluated by a group of five scientists initially, our access

01:24:52.600 committee, and then secondly, following their advice, the steering committee, which I'm

01:24:57.500 on, makes the final cut.

01:25:00.000 So when people are recommending green tea extract, they say there's a central compound that's

01:25:05.100 probably the main player in green tea extract.

01:25:08.060 Many people have said it's good for you, traditional Chinese medicine, et cetera.

01:25:13.120 Please test it.

01:25:13.880 And so when we evaluate this, there are all sorts of pros and cons.

01:25:17.900 We strongly prefer single compounds, purified synthetic compounds to complex mixtures.

01:25:24.380 But we can see sometimes a ground for testing a complex mixture.

01:25:28.460 It may be that five or six or seven things in balanced amounts might together be good for

01:25:33.440 you, even if no one of them used singly might be beneficial.

01:25:37.320 We also like to test things that have acquired a popular reputation for being healthful.

01:25:44.680 Well, if the popular reputation is, even if it's based on a thousand-year-old tradition

01:25:50.620 of traditional medicine, it still might be right.

01:25:53.640 And so some of the time, if we stumble on something that will work, that would be great

01:25:58.380 news.

01:25:58.880 And if it doesn't work, then we feel we've done a public service.

01:26:02.260 We certainly have not proven that green tea extract might not be good for people, but

01:26:08.280 it will, we hope, slightly take the wind out of the sails of that discussion if we demonstrate

01:26:14.180 when we give it to animals.

01:26:15.960 It didn't do any good to them.

01:26:17.840 The other one that caught my eye, by the way, on the list of, on the much longer list of

01:26:21.540 failed compounds, and you made a good point, right?

01:26:24.360 Which is the failures are just as important as the successes, right?

01:26:27.800 I mean, some of the most important insights are these things that don't work that we've

01:26:31.800 long held beliefs around.

01:26:34.200 Methylene blue caught my eye.

01:26:35.560 I was sort of interested to see methylene blue, curcumin, many of these other things that

01:26:40.140 have been tested.

01:26:41.760 You go back and revisit rapamycin again through a high, medium, low dose.

01:26:46.120 I mean, you really put the screws to rapamycin to continue to demonstrate its efficacy.

01:26:49.960 Really, rapamycin is in some ways probably one of the three poster children of the ITP.

01:26:55.940 Four now, counting kinagoflozin.

01:26:58.160 Yeah, yeah, yeah, exactly.

01:26:59.880 Let's talk about 17-alpha estradiol.

01:27:03.180 Yeah.

01:27:03.340 I know more about the estrogens than the average person.

01:27:07.240 I got to be honest with you.

01:27:08.460 I don't have a clue what 17-alpha estradiol is.

01:27:12.700 I'm very familiar with 2, 4, and 16-hydroxyestradiol and estron and all these other things.

01:27:19.220 I can't for the life of me figure out what 17-alpha.

01:27:21.920 Can you help me understand where this shows up?

01:27:25.040 Is there an actual molecule?

01:27:26.740 Is there a drug that we have?

01:27:28.540 What was the genesis behind this plan?

01:27:30.540 Yeah, the genesis was a guy named Jim Simpkins.

01:27:33.400 He's a pharmacologist.

01:27:34.960 He's always interested in estrogen receptors, and he tried to make a compound that was like

01:27:40.060 estrogen that is 17-beta estradiol, generically, crudely referred to as estrogen, 17-beta estradiol,

01:27:47.200 but which had much lower affinity for the classical receptors for the estrogens, ER-alpha and beta.

01:27:52.960 So he synthesized 17-alpha estradiol, which is very much the same compound, except for one

01:27:58.820 of the bonds tilts up instead of tilting down.

01:28:01.940 And in Jim's work, mostly in tissue culture cells, he found that, yes, it was tenfold

01:28:06.660 less active, at least, than 17-beta estradiol.

01:28:11.160 And in fact, whatever activities it was having on his cells and culture, it worked just as

01:28:15.880 well if you deleted the classical estrogen receptors.

01:28:18.560 So when he gave it to mice, he found it was non-feminizing.

01:28:22.640 When you give it to male mice, it did not induce the secondary sexual characteristics that the

01:28:28.060 stronger estrogens, like 17-beta estradiol, estrogen, would do.

01:28:33.800 So his idea, and I'm not endorsing this, I'm just reporting it, is that giving an estrogen

01:28:39.820 to male mice might make them live as long as female mice.

01:28:44.240 The female mice live 5% or 10% longer than males.

01:28:47.740 He said, maybe it will turn them into females, but it would be a much better compound for

01:28:51.820 people because no man wants to look feminine.

01:28:54.860 So if it works, people will give it to men and they'll live as long as women, but they

01:28:59.860 won't be embarrassed by looking like a girl.

01:29:03.180 That was enough to get it through the steering committee.

01:29:05.800 And when we gave it to mice, it indeed had a terrific effect on male lifespan.

01:29:10.520 It had no effect, whatever, now in three separate studies in female lifespan.

01:29:17.580 In fact, the males, given the drug, live not only longer than regular old males, they live

01:29:23.940 longer than regular females.

01:29:26.080 It's not merely moving the male curve to be coincident with the female curve.

01:29:32.180 It's moving it way past where the normal females live.

01:29:35.900 So why it works only in males, we have a clue, but we don't know for sure.

01:29:42.240 And we certainly want to know why the effect is so dramatically sex specific so that we

01:29:48.380 can try to figure out a way to get around that obstacle and to get it to work in females as

01:29:53.160 well.

01:29:54.140 Now, if I'm looking at my table correctly, the 17 alpha estradiol in cohort five extended

01:30:01.800 median survival in males, but not maximum.

01:30:05.380 And of course-

01:30:05.700 No, it does.

01:30:06.260 Well, that was used at a low dose.

01:30:08.880 You're right.

01:30:09.120 I see.

01:30:09.580 At a low dose.

01:30:10.840 When we went back and we did it at a three times higher dose, it extends our measure of

01:30:15.600 maximum longevity as well as median longevity and at all three sites.

01:30:19.680 Yeah, I see.

01:30:20.380 Okay.

01:30:20.580 So that's the difference is at the 14.4 parts per million versus the lower dose.

01:30:25.680 Yeah.

01:30:26.040 When we got the first result, somebody said quite plausibly, hey, I'll bet it would work

01:30:30.860 better at a higher dose.

01:30:32.080 So we tried that.

01:30:33.380 Now, again, I just don't know enough about mice, but what's the menopausal situation of

01:30:38.740 mice?

01:30:39.100 How long are the female mice exposed to estrogen to 17 beta estradiol?

01:30:45.020 I don't know about when the estrous cycle ceases in mice.

01:30:49.940 It depends a lot on the strain.

01:30:51.400 As you know, dozens to hundreds of inbred strains, and they each have slightly different

01:30:56.400 characteristics.

01:30:57.620 In general, production of new, healthy pups tends to slow down when the females are eight

01:31:04.600 or nine or 10 or 11 months of age, and the cycles start to get irregular and then to stretch

01:31:10.540 out and then to cease altogether.

01:31:12.600 But this is something I don't know very much about.

01:31:15.580 And to quote details, I'd have to talk to somebody who actually knows a lot more than

01:31:20.660 I do about mouse reproductive endocrinology.

01:31:23.560 Is there an actual molecule of 17 alpha estradiol that, is there an IND for this?

01:31:28.540 Very good question.

01:31:29.320 And the answer is not really well known.

01:31:31.860 There is one obscure paper by a Swedish or Finnish group maybe 15 years ago that said they

01:31:39.940 had detected 17 alpha estradiol, but only in the brain.

01:31:43.100 And in fact, they said that they had isolated in the brain a receptor, which they called

01:31:49.000 the estrogen X receptor, that was relatively specific for 17 alpha estradiol, the one that's

01:31:58.240 working in our tests.

01:31:59.940 But I have never seen that repeated.

01:32:02.880 I actually, following up sort of your line of thinking, corresponded a little bit with two

01:32:08.260 or three of the American scientists who are authentically steroid chemists and who spend

01:32:13.280 their whole life synthesizing various derivatives.

01:32:16.840 And I kept, I asked them, what is 17 alpha estradiol?

01:32:19.980 Is it anywhere?

01:32:20.820 Does it do anything?

01:32:22.100 And they all said, good question.

01:32:23.520 I don't know.

01:32:24.680 So I know nothing about it, but they should know something about it.

01:32:28.980 And apparently it's distinctly under-investigated.

01:32:31.600 So even after your findings, which are quite impressive, it hasn't been pursued as an investigational

01:32:39.780 new drug.

01:32:40.960 I'm over my head here.

01:32:42.080 I don't know the answer to that question.

01:32:43.800 Certainly nothing has come to my attention about it's being investigated, but I could well

01:32:49.840 have missed something.

01:32:50.760 There is an awful lot of attention to developing novel estrogenic compounds.

01:32:56.800 Really good academic labs, and I'm sure 10 times that much work in pharmaceutical firms.

01:33:03.120 There are major papers and important journals about here are 30 or 40 different steroid compounds

01:33:09.220 that do not have that side effect or do not have this, do not bind to this receptor, et cetera.

01:33:14.420 I don't know that literature very well.

01:33:16.280 And I'll bet a great chunk of it is kept quiet by pharmaceutical firms.

01:33:21.320 So I'm not confident that anyone outside those firms really knows the answer to your question.

01:33:26.380 Interesting.

01:33:27.320 Do you remember the rationale for your solic acid?

01:33:31.100 I will have to double check.

01:33:33.040 I believe this feels almost exactly like my qualifying exam.

01:33:37.340 Sorry.

01:33:38.660 Our solic acid, I think, is the one that changed retinoid receptor activation.

01:33:42.940 There was a claim from Gretchen Darlington and some of her colleagues saying that the long-lived

01:33:49.520 mice, the Ames-Dwarf mice, had turned on a whole batch of genes that related to bile acids,

01:33:54.420 and the bile acids work through some molecules related to the retinoid acid receptor or something

01:34:01.400 like that.

01:34:01.800 Or solic acid was proposed as a way of getting at the bile acid issue.

01:34:06.420 But the chances that I'm right on this point are only about 50-50.

01:34:09.940 I'd have to check my notes to be sure.

01:34:12.580 The reason I ask is there was a human-based supplement for this that was all predicated

01:34:16.980 around improving body composition.

01:34:18.460 But my one recollection of it was it was not orally bioavailable.

01:34:22.460 It was only you had to use it in sort of a topical fashion, which sort of rendered it

01:34:27.720 not particularly helpful.

01:34:29.360 Rich, can you tell me a little bit about the work being done on hydrogen sulfide?

01:34:32.600 There's a colleague of mine who died recently, a guy named Jay Mitchell, a wonderful scientist.

01:34:37.940 Jay was really interested in the idea that hydrogen sulfide might be an important controlling

01:34:43.220 element in the aging process and had published a long and really impressive series of papers

01:34:48.800 on that.

01:34:50.260 So he suggested that we give to mice a drug that would break down to produce hydrogen sulfide,

01:34:57.860 a drug called SG1002.

01:35:00.140 So we tried it and we messed up.

01:35:03.600 We had control mice that were on the wrong control food.

01:35:07.900 And the mice given Jay's drug did better than the mice or the wrong control food.

01:35:15.000 But we didn't publish that because we were not sure what would have happened if we'd use

01:35:19.440 the right control food.

01:35:20.860 So we're repeating that now.

01:35:23.160 We've fixed our mistake.

01:35:25.080 And we don't have data yet.

01:35:26.960 In about a year, we'll know whether Jay's hunch that giving mice a drug that would increase

01:35:33.720 their internal generation of hydrogen sulfide will be good for them.

01:35:38.020 I'm certainly hoping it's correct.

01:35:40.040 When we did it wrong, the drug seemed to work.

01:35:42.860 So I'm hoping that when we do it right, the drug will continue to work.

01:35:46.120 But we don't know yet.

01:35:47.360 Okay.

01:35:48.380 So let's go to one of your more recent findings, this SGLT2 inhibitor, CANT.

01:35:55.200 Canagliflozin, yeah.

01:35:56.380 Yeah.

01:35:56.820 So this is a drug which is widely available.

01:36:00.500 This class of drug has been around for, what, probably a decade, right?

01:36:04.240 I think that's right, yeah.

01:36:05.480 Used pretty readily in patients with type 2 diabetes, blocks the reuptake of glucose

01:36:11.320 so that more glucose is excreted in the urine.

01:36:14.920 Yes.

01:36:15.560 But only when you're hyperglycemic.

01:36:17.840 One of the nice things about it in clinical practice is you can't make somebody hypoglycemic

01:36:22.800 and put them at risk that way.

01:36:24.900 So it affects the process in the kidney that deals with very high glucose levels without

01:36:30.480 having much scope for causing toxicity.

01:36:34.600 Now, this increased median lifespan in the male mice by 14%, maximum extension of 9%.

01:36:40.840 It had no effect on the females.

01:36:42.960 That's right.

01:36:43.660 Is this another one of those drugs where you have any insight into differential plasma levels

01:36:48.120 between the males and females?

01:36:49.580 We actually did measure that, and it's in the paper.

01:36:52.380 However, the females, if I recall, actually had higher levels, slightly higher levels than

01:36:57.460 the males.

01:36:58.700 The absence of an effect in females is not explicable on the idea that they don't get

01:37:03.660 the drug into their blood.

01:37:05.480 What do you think could explain the difference in the effect here?

01:37:09.340 Gosh, I wish I knew.

01:37:11.660 When you take this together with the acarbose results, you're led to the inference, and this

01:37:16.660 is only a tiny baby step forward.

01:37:18.800 That something about aging in the male mice depends a lot on staying away from really high

01:37:24.440 glucose levels.

01:37:26.140 Now, whether that means that high glucose in the males triggers a circuit in the hypothalamus,

01:37:32.860 which is bad for you, or something.

01:37:34.960 I mean, from here on in, it's all hand-waving.

01:37:37.380 The mice are mostly dying of cancer.

01:37:39.500 Both the males and females, about 80% of the deaths are due to some form of cancer, a wide

01:37:45.020 range of different kinds of cancers, but cancer in these mice, and most mice, is the predominant

01:37:50.480 cause of death.

01:37:51.520 So you could probably start to think about why is it, if you want to avoid all sorts of

01:37:57.820 cancers, high glucose is a really bad thing for you if you're a male.

01:38:03.420 That's the sort of line of argumentation you'd want to consider.

01:38:07.280 But that's certainly unsatisfying, and we don't know.

01:38:09.480 We would love to know that.

01:38:10.700 Well, to me, why high glucose is bad for cancer is a relatively straightforward question, at

01:38:16.700 least, compared to why disproportionately for males and females.

01:38:20.560 So let me start by asking another question.

01:38:23.340 If you do nothing and just observe these animals as though you were an actuary, what is the distribution

01:38:33.340 of death for females versus males?

01:38:36.220 You mean the age of death?

01:38:37.700 No, the cause of death.

01:38:39.440 Yeah, about 80% cancers in both.

01:38:42.300 Oh, it's 80% in both.

01:38:44.180 Yeah, it's different kinds of cancer.

01:38:46.640 Both the leading cause for both males and females is hematopoietic cancers of one sort or another.

01:38:52.140 Lymphomas are related histiocytic sarcomas and other sorts of lymphoid tumors.

01:38:57.560 Then in the females, the next leading cause is mammary cancer.

01:39:02.940 And the third leading cause is either liver cancer or fibrosarcoma.

01:39:07.820 In the males, the second is lung cancer and then liver cancer or fibrosarcoma.

01:39:14.580 So all of them, it's cancers, cancers, cancers.

01:39:17.840 And how much is hematopoietic in each?

01:39:20.060 About 30% of the deaths.

01:39:22.380 Okay.

01:39:22.660 So that's hard to make sense of, I guess.

01:39:26.280 I mean, are there enough data on the glucose and insulin sensitivity of each of those relative

01:39:31.520 types of cancer that you could make the case that, I mean, I got to be honest with you,

01:39:35.920 that's a little counterintuitive based on human data.

01:39:38.200 Because obviously in humans, the mammary cancers would be quite glucose and insulin

01:39:42.660 sensitive, whereas the lung would not be.

01:39:45.320 So you would actually expect the opposite if that were true in mice, that you would expect

01:39:49.160 the females to disproportionately have the benefit.

01:39:52.320 But I mean, that's a lot of hand-waving, right?

01:39:54.700 So I'm just curious if we have, you know, more molecular data about PI3K activity or other

01:40:01.620 activity, IGF activity in these animals as it pertains to their cancer risk.

01:40:06.080 These are really good questions that I've been trying to urge my cancer biology friends

01:40:10.580 to start digging in and answering those questions.

01:40:13.680 As a sort of the framework, you can think, as you were thinking, maybe the glucose thingy

01:40:19.900 works on the cancers.

01:40:20.940 But it's also possible that the glucose excursions are working on an anti-cancer defense, some

01:40:26.700 aspect of immune defenses.

01:40:28.460 Or it could be, and this is my hope, but it's only a hope, it could be that the glucose

01:40:34.660 specific excursions are modifying some fundamental, still undiscovered element of the aging process

01:40:42.520 in the hypothalamus.

01:40:43.860 Maybe it has to do with the susceptibility of the hypothalamus to inflammatory change or

01:40:49.020 something like that, differentially in males and in females.

01:40:52.660 And that this nebulous sort of change has an impact on the cancer or an impact on anti-cancer

01:40:59.400 defenses or something.

01:41:01.120 Any of that is, in my view, about equally likely to be true.

01:41:04.220 But at present, none of it gives us much of a hint as to why these drugs have a much more

01:41:09.520 striking effect in the male mice.

01:41:11.840 Did you see the same effect here that you did with acarbose in that the hemoglobin A1c

01:41:16.320 was unaffected?

01:41:18.500 I'm not sure.

01:41:19.740 I'd have to go back and check the paper.

01:41:22.300 My memory, which may be incorrect, is that there was a figure showing no difference, no effect

01:41:30.040 on hemoglobin A1c in the canagliflozin paper, which is surprising because it is affected in

01:41:35.900 human diabetics.

01:41:37.680 You can really move hemoglobin A1c down into the healthy direction by canagliflozin and the

01:41:44.100 other different SGLT2 inhibitors.

01:41:46.940 But if I recall, it did not produce that effect in the mice.

01:41:50.360 We don't know whether the canagliflozin effect is on glucose.

01:41:54.620 We don't know whether it's on SGLT2.

01:41:56.340 It also has an effect on SGLT1.

01:42:00.360 And there are now reports appearing in the literature, quite striking ones, some of them

01:42:06.260 in clinical settings, showing effect on tumors specifically, on other aspects of other kinds

01:42:12.220 of disease, human effects on heart attacks, independent of whatever's happening in terms

01:42:16.760 of glycemic control.

01:42:18.400 So the shortest, simplest explanation, it's affecting glucose, may not be right.

01:42:23.760 There are other ideas that deserve a lot of exploration.

01:42:27.200 Yeah, we are very excited about this class of drugs clinically in our practice.

01:42:32.720 The cardio protection, renal protection, obviously the glycemic benefits are all pretty exciting.

01:42:38.100 And of course, this brings us to an interesting contrast, right?

01:42:41.160 I think if rapamycin is one of the most remarkable success stories of the ITP in terms of its

01:42:50.240 consistency, personally for me, the biggest surprise of the ITP is the failure of metformin.

01:42:58.260 Can you say a little bit about that?

01:42:59.940 How was the study done?

01:43:01.260 What are the possible blind spots?

01:43:03.520 I mean, obviously metformin succeeded when paired with rapamycin, but you could argue that's

01:43:08.440 just rapamycin, but metformin alone did not succeed.

01:43:13.080 And has it only been run by itself once or twice?

01:43:17.900 So I'm not surprised.

01:43:20.320 There are all sorts of reasons in which metformin failure is unsurprising.

01:43:26.420 One is it might be really good for people and not good for mice.

01:43:30.520 Mice and people are organized very differently.

01:43:32.600 It's also possible that we use the wrong dose.

01:43:35.100 If we'd use a dose that was twofold lower or twofold higher, it might've been great.

01:43:39.720 Or if we had given it for a few months and then taken it away for a month and then given

01:43:43.380 it for a few months, that might've worked as well.

01:43:46.180 You know, we tried it at one dose with one continuous dosing schedule.

01:43:51.140 It did not produce a significant effect, but that's not to say that it could not work in

01:43:56.340 mice.

01:43:57.080 And even if it could not work in mice, that's not to say that it might not be good for people.

01:44:00.760 So there are all sorts of ways in which you might get a disparity.

01:44:04.520 You asked how many times metformin has been tried.

01:44:06.500 It's been published twice, once by Rafa de Cabo, once by us.

01:44:10.700 In our study, it led to no significant lifespan extension in either males or in females.

01:44:17.280 In Rafa's study, which I'm about to be critical of, it produced no effect in females.

01:44:24.340 And Rafa alleged that it produced a significant effect in the males.

01:44:28.660 Now, the p-value was 0.046, so it was really close.

01:44:34.280 And it was not the test that he ordinarily uses.

01:44:37.560 Our lab and almost everybody, including Rafa, always uses the log rank test.

01:44:43.080 And he didn't use the log rank test.

01:44:44.840 Instead, he picked another test which gives special weight to early deaths.

01:44:49.840 And if you use that special specific test, then it is statistically significant at a p-value of 0.046.

01:45:00.140 My hunch is...

01:45:01.620 It didn't work with the log rank.

01:45:03.540 Yeah, it tried.

01:45:04.500 It failed.

01:45:05.040 So they went statistics shopping, found one where it got to 0.046 and published that.

01:45:11.020 I have asked Rafa for the data privately two or three times.

01:45:15.580 He's always agreed to send it to me over the last three or four years, but it hasn't actually arrived yet.

01:45:20.960 Well, maybe this podcast will help speed that up.

01:45:23.340 I hope so.

01:45:23.800 In any case, whether or not his was just over or just under the border of statistical significance,

01:45:29.360 it was a small effect at best.

01:45:32.240 And in ITP, it didn't work at all.

01:45:34.720 But it might work.

01:45:35.400 It might be good for people.

01:45:36.180 Yeah, well, I mean, I guess that's the challenge here, right, is you've got these agents that work under so many different circumstances.

01:45:45.620 And again, you can probably tell, Rich, my bias is, you know, I think rapamycin is an amazing agent here.

01:45:51.260 And I think it's one that warrants lots of investigation in humans because I think that the non-human literature are so compelling, right,

01:45:59.300 both in terms of the actual drug use and also the mechanism of action and the genetic manipulation all the way from, you know, worms, fruit flies, obviously, you know, mice, yeast, Matt Caberlin's dogs.

01:46:12.320 It just over and over and over again demonstrates efficacy.

01:46:16.140 And even when you look at the, you know, softer endpoints in humans, softer is maybe the wrong word, but less longevity specific.

01:46:23.940 But, you know, things like immune function and things like that, it always seems to be pointing in the right direction.

01:46:27.700 Now, in metformin, we have this undeniable data of diabetics that take it versus diabetics that don't.

01:46:34.280 And, you know, you can slice that 10 ways to Sunday.

01:46:36.500 It always seems to favor metformin in that diabetic group.

01:46:40.200 But as Nir Barzali and many others are interested in, what does that tell us about non-diabetics?

01:46:46.260 Is there a way we can better get at that answer while we await the results of a TAME study to get, you know, funded and executed?

01:46:55.120 Is there another ITP we should be doing?

01:46:57.000 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.

01:47:13.100 The point of the paper was not so much that metformin was good for diabetics, which everybody knew it was.

01:47:17.640 It was that the mortality risk of diabetics on metformin was actually better than non-diabetics of the same age and sex.

01:47:26.360 So that doesn't prove that metformin is good for non-diabetics.

01:47:31.100 But it's certainly a hint in that direction.

01:47:34.720 Now, it was an observational study, not a controlled, randomized clinical trial, of course.

01:47:39.060 So it might not be right.

01:47:40.500 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 a diabetic person is being protected by metformin.

01:47:58.500 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.

01:48:14.700 And studying that in terms of a randomized clinical trial strikes me as justified.

01:48:21.220 It's very expensive.

01:48:22.820 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.

01:48:39.440 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.

01:48:48.260 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.

01:48:58.960 So I guess it's safe to say that if resveratrol was the in molecule of the early 2000s, the next big in molecule has been anything that points to NAD.

01:49:14.060 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?

01:49:22.120 I really have heard great things about this.

01:49:24.420 And what do you think about NR and NMN and these sorts of things?

01:49:29.060 So I guess I'll give a really, really, really fast, or I'm actually 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.

01:49:41.420 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.

01:49:49.100 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.

01:50:06.960 It's at the top of your list. For some people, it's sort of in the middle. For others, it's at the bottom.

01:50:11.740 But it deserves a lot of attention.

01:50:14.240 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.

01:50:25.800 One of the reasons it's gotten so much attention is because of clever marketing.

01:50:31.020 There are scientists who have made a great deal of money by making more or less unsupported.

01:50:37.800 You can't sue them. They're not lying exactly, but they're sort of winking and nodding, saying, hey, take this formulation.

01:50:45.100 It'll sort of do some goodish stuff to your NAD.

01:50:48.260 We're not allowed to say, well, treat any disease because we would run afoul of the FDA, but wink, wink, try it.

01:50:54.260 It's a nutritional supplement.

01:50:56.220 And they've made a lot of money and gotten a lot of attention, some of it positive and some of it negative.

01:51:01.720 The reason that NR, nicotinamide riboside, was recommended to us by a company that wants to sell it

01:51:08.960 is that it's orally bioavailable and more stable than some of the other ways that have been proposed by MALC.

01:51:16.040 Can you disclose which company asked? There's probably only two or three.

01:51:19.940 I think it's public. I think it's Chemodex.

01:51:20.980 I think it's public knowledge who suggests things to the ITP.

01:51:25.020 So they said, well, we'll provide it to you. You guys test it.

01:51:27.940 You can publish the results, positive or negative, just in the interest of science.

01:51:31.020 Let's find out. That's a highly honorable and sensible thing for them to have done.

01:51:36.100 So we tested it. It's a bioavailable form.

01:51:40.320 We used a dose that they suggested.

01:51:42.080 And the paper that will have come out by the time this podcast becomes available

01:51:47.880 suggested that NR did not extend lifespan in our mice.

01:51:52.360 There are all the usual caveats.

01:51:54.820 Maybe it would have if we'd given it at a higher dose or at a lower dose.

01:51:58.780 In addition, we tried a little bit to detect it in various tissues.

01:52:04.560 There's a collaborator who looked at the brains and the muscles and the hearts.

01:52:07.380 And we were not able to demonstrate a consistent major change in NAD levels in these tissues.

01:52:16.240 NAD is a highly active molecule, as of course you know, that turns over constantly.

01:52:21.540 And with a timescale of a few seconds to a few minutes can go way up or way down in given cells,

01:52:28.000 depending on how stressed one is or how high the glucose is or whether you're exercising,

01:52:32.220 all that stuff.

01:52:32.780 So it's possible that we fail to measure major changes in NAD in our treated mice because

01:52:39.920 we should have fasted them or we should have let them rest for a few hours or we should

01:52:44.680 have gotten the blood out in a different way or something like that.

01:52:47.680 But in any case, the observations that we're confident of at the dose we used, it did not

01:52:52.820 produce a lifespan effect.

01:52:55.000 Did it produce any benefit?

01:52:56.540 Well, we would have to go and check for other things, right?

01:52:58.860 The first time we do a study with a new drug, lifespan is the only thing we measure.

01:53:04.520 We would love to be able to measure immune function and vision and hearing and cardiovascular

01:53:09.760 function, et cetera, but those are expensive.

01:53:12.160 And if we did those, we would have to spend money on them and cut back by a factor of whatever,

01:53:17.660 two or three or five, the number of drugs we test.

01:53:19.960 So we generally reserve those health outcomes only for those drugs that give a lifespan benefit

01:53:25.840 and NR did not in our hands give a lifespan benefit.

01:53:28.860 You started the NR at about eight months of age.

01:53:31.800 So again, this is a reasonably young animal in its 20s, the equivalent of a human.

01:53:37.220 Do you have a sense of what the dose was on the scale of where humans are taking this

01:53:43.580 commercially?

01:53:44.180 I believe humans are typically, oh God, I don't even know, taking probably a couple of

01:53:49.000 grams of this stuff a day.

01:53:50.800 I do not know.

01:53:51.540 I'd have to look it up.

01:53:52.380 But the company basically had you utilizing a dose that they felt was consistent with

01:53:58.880 what humans are being encouraged to take.

01:54:01.020 We followed their recommendations.

01:54:02.640 We picked a dose that they said would be an appropriate, good starting dose for our tests.

01:54:07.660 And we followed their guidance on this.

01:54:09.940 It's a little tricky when you're going to give something for a lifespan.

01:54:12.740 If there are doses that are often of drug X, I'm not talking about NR, but doses of drug

01:54:18.160 X that are typically given to people for a month or two and it's really good and doesn't

01:54:22.640 produce side effects, you're still worried that might not work for mice because if you're

01:54:26.280 going to give it to them the whole lifespan, side effects may accrue.

01:54:29.720 So if there is an ambiguity, we want to start usually with a pretty low dose because if they're

01:54:35.420 going to be on it from four months of age until death, keeping the dose nice and low is

01:54:42.040 probably an important precaution against unanticipated side effects that might not be seen in an acute

01:54:47.880 treatment.

01:54:48.900 How often do you guys see unanticipated consequences of exactly what you described where side effects

01:54:55.220 don't show up in the first year of the mouse receiving the drug, but show up later?

01:54:59.720 Well, we have not yet, and this surprised me somewhat, we have not yet found any drug

01:55:05.120 suggested to us that caused a shortening of the lifespan.

01:55:08.920 None of the drugs that we've used has caused a significant decline in lifespan.

01:55:14.260 There were one or two where the lifespan went down 5% or 8%, but it didn't achieve statistical

01:55:19.080 significance.

01:55:20.200 Now, there could be some side effects that we don't look for.

01:55:22.320 Maybe the mice have problems with hearing or sight or memory or muscle function or something

01:55:28.600 like that, which we would never see unless we look for them.

01:55:32.040 It really seems that there are three really important takeaways from, what are you at now,

01:55:39.620 18, 20 cohorts of ITPs.

01:55:42.320 I mean, almost 100 studies here, basically, or 100 experiments, which is mTOR matters, less

01:55:50.380 glucose is better than more, and sex-specific steroid hormones probably do something relevant,

01:55:57.280 right?

01:55:57.600 I mean, is that, if you were at a dinner party and someone asked you to give them the most

01:56:03.720 salient findings of the last two decades, would that be a fair assessment?

01:56:09.660 I think the points you're making are really good ones, but it's not quite the way I would

01:56:13.000 put it if I were allowed to pick only three.

01:56:15.940 Okay.

01:56:16.320 Give me your three.

01:56:17.080 The first point I'd make is the one we began at the beginning of this conversation, which

01:56:22.000 is that, by gosh, you actually can put something in the food that extends healthy lifespan,

01:56:26.720 and it's an enormous effect, 10 times better than a cure for cancer.

01:56:31.380 So, in terms of the fundamental imagining of where medical research ought to go and what

01:56:38.300 people who want to make their friends stay alive and healthy longer should think about,

01:56:43.180 our program shows that what they should think about is finding drugs that slow aging.

01:56:48.720 It seems obvious, you know, you and I are talking about it because we sort of believe

01:56:52.240 that, we sort of think it's cool and interesting and important, but that's most of my friends,

01:56:56.760 scientific and non-scientific, they haven't decided that that's the case.

01:57:00.520 So, I would take that as the first most important take-home message.

01:57:04.680 The second take-home message would be, to my mind, the really surprising finding that

01:57:11.220 some of the time, maybe even most of the time, these drugs work even when you start

01:57:16.800 them in middle age.

01:57:18.520 I would have guessed against that when we started.

01:57:21.080 I would have said, you have to start in youth, period, but that was a bad guess, and two drugs

01:57:26.240 that we've tested so far work just as well in middle age, and one works half as well.

01:57:30.320 So, that was a bad guess, and that's really important, too, in terms of both the understanding

01:57:33.920 of aging and also thinking about transition to human usage.

01:57:39.820 The third point I would make of broad general applicability, I would have guessed any drug

01:57:45.620 that works is going to work in both sexes, and that's mostly wrong.

01:57:49.720 We have two drugs, rapamycin and one we didn't mention, the amino acid glycine, which have equally

01:57:54.940 strong effects in males and females.

01:57:57.600 We didn't talk about glycine because it works in males and females, significant, but it's a

01:58:01.900 really tiny effect, and so it's sort of fallen by the why side as we've moved on to drugs with

01:58:06.860 greater effectiveness.

01:58:08.220 But the sex specificity and the ability to start in late life were both surprises and open

01:58:14.680 up new ideas about how can you discover why it's working in old age, and how can you discover

01:58:21.800 why it's working in one sex but not both.

01:58:24.200 Those are important research questions, and we could easily take three or four or five times as

01:58:29.620 many really good new labs in aging and devote them to trying to tease out those puzzles.

01:58:36.240 And then the fourth set of questions would be the ones you brought up.

01:58:39.420 You're pointing to specific molecular clues.

01:58:43.020 Is it TOR?

01:58:43.940 Is it glucose?

01:58:45.360 Are there sex specific hormone receptors in the brain that are important?

01:58:50.000 I think those are really interesting as well.

01:58:53.000 And in fact, any postdoc or grad student who comes into my lab, they get one of those to

01:58:58.060 work on.

01:58:58.540 That's the cutting edge of turning this into a molecular and cellular understanding of what

01:59:04.280 slows the aging process and where we ought to pay attention when we want to find new drugs

01:59:09.620 that will work in just the same way.

01:59:11.600 So the points you've made strike me as highly pertinent, very important for anyone who wants

01:59:16.480 to work on something in this area, but the three more general points, that it works at

01:59:21.960 all, that it can work late in life, and that it's often sex specific, I don't want those

01:59:26.480 to get lost in the shuffle because they fundamentally rearrange our ideas about aging, about late life

01:59:32.460 disease, and about what one ought to do if one was, say, in charge of America or the universe's

01:59:38.380 medical research apparatus.

01:59:40.120 On that point, Rich, if money were absolutely no object, and the reality of it is it doesn't

01:59:47.700 need to be as extreme as one thinks, I mean, but let's just say that the NIH came along and

01:59:54.000 said, look, we are going to give you the budget you need to study these questions in any animal.

02:00:02.160 What would you choose as the best model organism that would give you the trade-off of duration

02:00:09.880 and closeness to humans?

02:00:14.220 Would it be staying where you are, where you have an advantage that you get answers quickly,

02:00:18.140 but the disadvantage is you're quite far from humans?

02:00:20.540 Or would you move more towards, you know, what we saw in the very famous caloric restriction

02:00:25.820 experiments at Wisconsin and NIA in the 80s and 90s, or sort of 90s through

02:00:32.140 2000s where you were very close to humans, but obviously it took a long time to get the

02:00:37.520 answer.

02:00:38.560 So I have two answers for you.

02:00:40.580 The first is the mouse, and the other is it's a bad question.

02:00:45.000 If you ask a carpenter, what is the most important tool, you'll get the same sort of answer,

02:00:49.420 because the answer is the tool that I select depends upon the problem I have to solve.

02:00:55.060 I think that each of the major experimental animals that is currently receiving a lot of attention

02:01:00.720 is doing something good.

02:01:02.680 They're molecular and genetic questions that are best addressed by flies and worms, even

02:01:06.860 though they don't resemble me and my mouse in more than cellular aspects, but they're

02:01:13.740 things that they can only solve that we can't solve, and that's an important thing to understand.

02:01:20.100 In primates, I think there are a few people who are starting to understand that rhesus is not

02:01:25.140 where it's at. Marmoset is where it's at, because marmosets are primates.

02:01:29.700 They are organized much like humans, but they have a lifespan, an average lifespan of maybe

02:01:35.100 eight or ten years, something like that.

02:01:36.820 So like dogs, they have the advantage of a lifespan which is longer than we'd like, but short

02:01:43.360 enough that you can think about doing an experiment and completing it within about a decade.

02:01:48.540 Some of the work has to be done in humans, too, for reasons I don't think I need to convince

02:01:53.100 you of. The mouse is the correct answer. If you have to pick what, it's got to be the

02:01:57.160 mouse. You said they're not much like people, but of course they're extremely like people.

02:02:01.920 They have the hypothalamus, they have the pancreas, they have the beta cells. There are very few

02:02:06.700 things, organs or cells or circuits or hormones or whatever, that mice have but people don't

02:02:12.980 have or vice versa. They're important subtleties.

02:02:16.480 Yeah. I guess I'd push back a little bit in terms of some of the differences are profound,

02:02:21.880 right? So one profound difference is herbivores versus omnivores. Now that might pose a lot less

02:02:29.940 of a difference for the ITPs than it does for one of my pet peeves, which is the never-ending

02:02:34.820 nutritional studies of mice, which I find generally unhelpful for that reason.

02:02:39.880 That's right.

02:02:40.360 And we've even discussed it today in the context of looking at a carbose, right? That's a nutritional

02:02:46.860 intervention in some ways, and it might be a little bit difficult to extrapolate. But I think

02:02:51.200 the other area where we do need to be a little bit circumspect with the use of mice is that 80%

02:02:57.920 of them die of cancers, a third of which are hematologic. Whereas when you look at us,

02:03:04.740 all humans die with atherosclerosis, about a third die from it. And I think that is an important

02:03:12.480 distinction that basically says, this is just my own personal philosophy, any effort to thwart human

02:03:20.300 aging must be able to punch atherosclerosis squarely between the eyes.

02:03:25.520 So let me give you a different way of thinking about this. Let me try to change your personal

02:03:29.620 philosophy here. We gave them a lot of mice and rapamycin. And then instead of letting them die,

02:03:34.840 we euthanized them at 22 months of age when most of them were still alive and fairly healthy.

02:03:39.060 And then with the aid of a pathologist, J. Irby Wilkinson, Irby looked at dozens of different

02:03:44.360 organs. And Sue Brooks also looked at their tendons. So their tendons were youthful. Their

02:03:50.940 kidneys were youthful. They did not have changes in the heart. They did not have changes in the

02:03:56.220 endometrium. They did not have changes in the liver. They did not have changes in the adrenal that

02:04:01.740 were characteristic of 22-month-old control mice. So the implication here is that even though the

02:04:08.560 leading cause of death was cancer, the rapamycin was actually slowing a very wide range of age-associated

02:04:16.800 changes. Now, some changes, and we don't know which ones they are, produce cancer in mice and in people.

02:04:24.800 Some changes, and we don't know which they are, produce atherosclerosis or strokes or diabetes in

02:04:30.720 people and in a small fraction of mice. So from my perspective, the details of what you're going

02:04:38.420 to die of, whether it's mostly a stroke, mostly atherosclerotic in nature, or mostly neoplastic,

02:04:44.860 they're certainly important if you've got a patient, you want to help that patient,

02:04:48.880 or if you're looking for disease-specific drugs. But my interest is not related to that. It's to

02:04:53.600 what can we do to slow the aging process? And I think it's a very plausible, at least

02:05:00.420 plausible idea that when you slow the aging, whatever diseases afflict the members of that

02:05:05.320 species will be slowed down as well. Alzheimer's disease, for instance, does not ordinarily occur

02:05:10.660 in mice. It's obviously a major hassle for people. Yet if you make a mouse that has some aspects of

02:05:16.800 Alzheimer's disease and you give them these drugs, the disease is postponed. And so that's

02:05:22.440 consistent with the notion that whatever diseases aging leads to in either species, anti-aging

02:05:30.780 medications will help retard. Cause of death is an important wrinkle, but only an elaboration of the

02:05:40.460 aging process in a species-specific way. And I think this also speaks to the importance of whenever

02:05:46.800 possible starting younger. Although again, rapamycin almost flies in the face of that, but the earlier

02:05:52.140 you would start, the more opportunity you would have to beneficially impact those tendons, those

02:05:58.060 nephrons, those cardiac myocytes, and those neurons. That's a really good general point. I used to agree

02:06:03.760 with that completely. And now I think one has to answer that question directly point by point by point.

02:06:10.740 The fact that giving 17-alf-estradiol or acarvos or rapamycin to middle-aged mice, and you still get a

02:06:17.300 nice postponement of whatever the lethal process is, that argues against the notion that many, many

02:06:24.480 things have to start early. In addition, this is again published now, this is Mike Garrett's paper with

02:06:30.600 John Herrera and Charlene Day. When they started treating some old mice with 17-alf-estradiol,

02:06:37.740 they postponed or reversed. It's hard to tell. Muscle changes, muscle grip strength changes,

02:06:45.100 rotor rod ability, the ability to stand on a rotating rod, and changes in old age and glucose

02:06:49.720 tolerance. So it may be that despite my personal intuition, which is similar to the intuition that

02:06:57.020 you just expressed, the evidence starts to suggest some aspects of aging can be reversed or prevented,

02:07:04.660 even if you're starting in late middle age. That would be cool if it was true. I would have bet

02:07:09.140 against it, but I may be wrong. Has your conviction about any of the molecules you've tested in the

02:07:14.980 ITP led to you taking any? I prefer not to answer that because I'm not a doctor and I would never,

02:07:21.240 ever wish to recommend drugs to someone else, take them too seriously. I wasn't going to ask you which,

02:07:26.920 but we'll leave it at that. We'll leave it at that. Rich, this has been an awesome discussion. I mean,

02:07:32.060 I was really looking forward to this and I honestly think you have one of the coolest jobs in the

02:07:35.840 world because you get to basically hear from anybody and everybody who has an idea about

02:07:43.640 something that might extend life in a very straightforward intervention, which is if a

02:07:50.320 person, if a person in this case, if a mouse takes this drug, will it extend their life? And you get

02:07:55.720 to test this in a highly reproducible manner with two great colleagues and you've basically got this

02:08:02.380 well-oiled machine that produces these fantastic studies, which both positive and negative continue

02:08:09.300 to add to the body of knowledge we have. I want to close with something really funny and you probably

02:08:14.800 don't want me to do this, but I'm going to do it anyway. So in 15 years ago, you wrote something

02:08:19.680 really hilarious at the MIT technology review. Do you know where I'm going with this?

02:08:23.720 I remember the piece pretty well. Yes. I wrote it right here at this desk.

02:08:28.680 Okay. So early morning, it was six in the morning. I couldn't sleep. I had an hour.

02:08:33.560 I just felt in the mood to be meed. So I wrote the piece. Yeah.

02:08:36.680 So Aubrey de Grey must've done something on television that you must've watched.

02:08:43.220 Had he basically gone on TV and said,

02:08:45.920 No, I made that part up. I'd been hearing Aubrey speak over and over and over and over again

02:08:52.660 for many years. I fantasized that he had made a television appearance and I wanted to enlist his

02:09:00.020 help. And basically what you said is, I'm going to read just the first part of this to you, Aubrey.

02:09:04.280 So the MIT tech review published this. I saw you on TV the other day and I was hoping

02:09:09.980 that now that the aging problem has been solved, you might have time to help me

02:09:14.620 in my publicity campaign to solve a similar engineering challenge. One that has been

02:09:20.160 too long ignored by the ultra conservative fraidy cat mainstream scientific community,

02:09:25.600 the problem of producing flying pigs. A theoretical analysis of the problem using the fastest available

02:09:33.660 modern computers shows that there are a mere seven reasons why pigs cannot at present fly.

02:09:41.260 One, they do not have wings. Two, they are too heavy to get off the ground. Three, the so-called law

02:09:50.480 of gravity. Four, they cannot climb trees. Five, hair instead of feathers. Six, they do not wish to fly.

02:10:01.360 And seven, they do not tweet. Now, this is really funny because you wrote this prior to Twitter. When

02:10:08.380 I read this, I was like, what does tweeting have to do with anything? And of course, I had to remind

02:10:13.160 myself Twitter didn't even exist in 2005. I'm not going to read the rest of this, but we're going to link

02:10:18.100 to it in the show notes because it is a brilliant, you go, you go into great detail elaborating on

02:10:23.580 each of your points. But this is to me really a great example of who you are, right? Which is

02:10:30.140 you're grounded in reality and you, you don't buy into this idea that, that immortality is in our

02:10:35.560 future. And you view it as an enormous win. If there's a drug out there that can extend human life

02:10:40.900 by 25%. I'm in your camp, Rich. So I'm highly biased and I've seen nothing to suggest that

02:10:48.120 immortality is in our future. And frankly, I'm not sure it should be, even if we were ever given the

02:10:52.720 choice. But boy, 10 to 15% improvement in our lifespan and healthspan would be a remarkable

02:11:00.300 achievement. And if there are molecules that can do that, I'm glad you're working on them.

02:11:04.540 Well, I'm delighted to have been invited. A friend of mine said that he'd been on your podcast a number

02:11:10.760 of times and had a wonderful, wonderful time and told me if I ever got a chance, I would enjoy it

02:11:15.400 too. And this is a very good prediction. It's a real treat to be interviewed by someone who really

02:11:20.600 knows the stuff and also asks the very best questions. So I I'm honored to have been added to

02:11:26.520 your roster of guests. This has been a great pleasure for me. Rich, thank you for taking the leap of faith

02:11:32.180 and sharing your work with us. This was fantastic. Good deal. Well, thank you for inviting me.

02:11:36.640 Thank you for listening to this week's episode of The Drive. If you're interested in diving deeper

02:11:40.580 into any topics we discuss, we've created a membership program that allows us to bring you

02:11:44.980 more in-depth exclusive content without relying on paid ads. It's our goal to ensure members get back

02:11:50.760 much more than the price of the subscription. Now to that end, membership benefits include a bunch of

02:11:56.080 things. One, totally kick-ass comprehensive podcast show notes that detail every topic, paper,

02:12:01.660 person, thing we discuss on each episode. The word on the street is nobody's show notes rival these

02:12:07.100 monthly AMA episodes or ask me anything episodes, hearing these episodes completely access to our

02:12:13.760 private podcast feed that allows you to hear everything without having to listen to spiels like

02:12:18.660 this. The Qualies, which are a super short podcast that we release every Tuesday through Friday,

02:12:24.540 highlighting the best questions, topics, and tactics discussed on previous episodes of The Drive.

02:12:28.780 This is a great way to catch up on previous episodes without having to go back and necessarily

02:12:33.660 listen to everyone. Steep discounts on products that I believe in, but for which I'm not getting

02:12:39.220 paid to endorse, and a whole bunch of other benefits that we continue to trickle in as time goes on.

02:12:44.720 If you want to learn more and access these member-only benefits, you can head over to

02:12:48.120 peteratiamd.com forward slash subscribe. You can find me on Twitter, Instagram, and Facebook,

02:12:55.160 all with the ID, peteratiamd. You can also leave us a review on Apple Podcasts or whatever podcast

02:13:01.680 player you listen on. This podcast is for general informational purposes only and does not constitute

02:13:07.240 the practice of medicine, nursing, or other professional healthcare services, including the

02:13:12.140 giving of medical advice. No doctor-patient relationship is formed. The use of this information

02:13:17.680 and the materials linked to this podcast is at the user's own risk. The content on this podcast is not

02:13:24.220 intended to be a substitute for professional medical advice, diagnosis, or treatment. Users should not

02:13:30.660 disregard or delay in obtaining medical advice from any medical condition they have, and they should

02:13:36.820 seek the assistance of their healthcare professionals for any such conditions. Finally, I take conflicts

02:13:42.780 of interest very seriously. For all of my disclosures and the companies I invest in or advise, please

02:13:48.840 visit peteratiamd.com forward slash about where I keep an up-to-date and active list of such companies.

02:14:06.820 for all of my disclosures and the companies I invest in, please visit peteratiamd.com forward slash about where I keep going.

The Peter Attia Drive - February 08, 2021

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

Episode Stats

Length

Words per Minute

Word Count

Sentence Count

Misogynist Sentences

Hate Speech Sentences

Summary

Transcript