The Peter Attia Drive - #66 - Vamsi Mootha, M.D.： Aging, type 2 diabetes, cancer, Alzheimer's disease, and Parkinson's disease – do all roads lead to mitochondria？

00:00:00.000 Hey everyone, welcome to the Peter Atiyah drive. I'm your host, Peter Atiyah. The drive

00:00:10.880 is a result of my hunger for optimizing performance, health, longevity, critical thinking, along

00:00:15.940 with a few other obsessions along the way. I've spent the last several years working

00:00:19.660 with some of the most successful top performing individuals in the world. And this podcast

00:00:23.620 is my attempt to synthesize what I've learned along the way to help you live a higher quality,

00:00:28.360 more fulfilling life. If you enjoy this podcast, you can find more information on today's episode

00:00:33.020 and other topics at peteratiyahmd.com.

00:00:41.440 Hey everybody, welcome to this week's episode of the drive. I'd like to take a couple of minutes

00:00:46.000 to talk about why we don't run ads on this podcast and why instead we've chosen to rely entirely on

00:00:51.660 listener support. If you're listening to this, you probably already know, but the two things I care

00:00:56.480 most about professionally are how to live longer and how to live better. I have a complete fascination

00:01:02.760 and obsession with this topic. I practice it professionally and I've seen firsthand how

00:01:07.480 access to information is basically all people need to make better decisions and improve the

00:01:12.120 quality of their lives. Curating and sharing this knowledge is not easy. And even before starting the

00:01:17.600 podcast, that became clear to me. The sheer volume of material published in this space is overwhelming.

00:01:22.820 I'm fortunate to have a great team that helps me continue learning and sharing this information

00:01:27.880 with you. To take one example, our show notes are in a league of their own. In fact, we now have a

00:01:33.840 full-time person that is dedicated to producing those and the feedback has mirrored this. So all of this

00:01:39.720 raises a natural question. How will we continue to fund the work necessary to support this? As you

00:01:45.980 probably know, the tried and true way to do this is to sell ads. But after a lot of contemplation,

00:01:51.100 that model just doesn't feel right to me for a few reasons. Now the first and most important of

00:01:56.300 these is trust. I'm not sure how you could trust me if I'm telling you about something when you know

00:02:02.100 I'm being paid by the company that makes it to tell you about it. Another reason selling ads doesn't

00:02:07.580 feel right to me is because I just know myself. I have a really hard time advocating for something

00:02:12.840 that I'm not absolutely nuts for. So if I don't feel that way about something, I don't know how I can

00:02:18.500 talk about it enthusiastically. So instead of selling ads, I've chosen to do what a handful

00:02:23.280 of others have proved can work over time. And that is to create a subscriber support model for my

00:02:29.320 audience. This keeps my relationship with you both simple and honest. If you value what I'm doing,

00:02:36.260 you can become a member and support us at whatever level works for you. In exchange, you'll get the

00:02:41.840 benefits above and beyond what's available for free. It's that simple. It's my goal to ensure that no

00:02:47.860 matter what level you choose to support us at, you will get back more than you give.

00:02:52.960 So for example, members will receive full access to the exclusive show notes, including other things

00:03:00.380 that we plan to build upon, such as the downloadable transcripts for each episode. These are useful

00:03:06.760 beyond just the podcast, especially given the technical nature of many of our shows. Members also get

00:03:12.900 exclusive access to listen to and participate in the regular ask me anything episodes. That means

00:03:19.840 asking questions directly into the AMA portal and also getting to hear these podcasts when they come

00:03:25.800 out. Lastly, and this is something I'm really excited about. I want my supporters to get the best

00:03:30.780 deals possible on the products that I love. And as I said, we're not taking ad dollars from anyone,

00:03:35.620 but instead what I'd like to do is work with companies who make the products that I already love and would

00:03:40.600 already talk about for free and have them pass savings on to you. Again, the podcast will remain

00:03:47.740 free to all, but my hope is that many of you will find enough value in one, the podcast itself, and two,

00:03:56.260 the additional content exclusive for members to support us at a level that makes sense for you.

00:04:01.680 I want to thank you for taking a moment to listen to this. If you learn from and find value in the

00:04:06.180 content I produce, please consider supporting us directly by signing up for a monthly subscription.

00:04:11.960 My guest this week is Dr. Vamsi Muta. Vamsi is a professor of systems biology at the Harvard

00:04:17.220 Medical School. He has an appointment at the Broad Institute, which is actually where we met to conduct

00:04:21.480 this interview. And we talk a little bit about what makes the Broad so special. He specializes in

00:04:26.320 rare mitochondrial diseases as opposed to longevity per se, something that I love to talk about. But I think

00:04:32.780 you'll see in this interview why it's so interesting to talk to someone who specializes in rare orphan

00:04:37.580 mitochondrial diseases about longevity. His laboratory uses a blend of genomics, computational

00:04:42.440 biology, biochemical physiology, and systems biology to study mitochondrial function and dysfunction.

00:04:49.460 He received his bachelor's in mathematics and computer science at Stanford before going on to

00:04:54.360 Harvard as part of the joint MIT Harvard program in medical school. He stayed in Boston to do his

00:05:01.600 training in internal medicine, though, as we discussed, he now focuses exclusively on research.

00:05:06.680 He's received more honors and awards than I could name here, but it's always worth mentioning it when

00:05:12.080 someone is a genius award recipient. So he won the MacArthur Foundation Award in 2004, which obviously

00:05:18.700 puts him in pretty rarefied earth. In this episode, we talk about a lot of things. We start with,

00:05:24.820 at least for me, one of the best discussions I've ever had on the mitochondria. And you might think at

00:05:30.760 first that some of this recapitulates things that were discussed on previous podcasts. You may recall,

00:05:35.840 we had a great discussion on the mitochondria with Navdeep Chandel. But we go a little bit deeper and

00:05:40.080 we talk more about some of the evolutionary pressure around the mitochondria. And even though at first

00:05:45.840 you might think, well, gosh, this seems awfully scientific. Where's the application? If you stick

00:05:49.800 with it, you're going to see where it comes and how understanding these rare orphan diseases that

00:05:57.020 most of us have never heard of can give us an insight into aging. We do eventually start to talk

00:06:02.460 about metformin, which I know many people ask about. And he provides a great insight into, or several great

00:06:07.960 insights into what metformin may or may not be doing. And again, for at least for me, this was like

00:06:13.140 just the master's class in mitochondrial biogenesis, electron transport chain, et cetera. Talk about ways

00:06:20.400 that we can target mitochondrial proteins and complexes to treat disease. But perhaps the single

00:06:26.600 most insightful and interesting thing that we talked about that completely blew my mind was the role of

00:06:33.160 hypoxia as a treatment. I want to repeat that again. The role of hypoxia, oxygen deprivation as treatment.

00:06:41.820 Now, we're going to cover this in a way that I think is very interesting. So I don't want to say

00:06:46.420 any more about it, but I do want to add a disclaimer to this that Vemsi and I spoke about

00:06:51.640 after the podcast. And I want to make sure it's here, right? So Vemsi emphasizes that their published

00:06:57.540 research to date on the beneficial effects of hypoxia in animal models of mitochondrial disease

00:07:02.460 is still in its early stages and it's restricted to animal studies. It is not yet ready to be applied to

00:07:07.940 humans. To extend these ideas to humans would be premature and irresponsible since hypoxia can have

00:07:14.100 life-threatening implications. If and when the concept is extended into humans, it will need to

00:07:19.480 be done so in a clinical trial setting with the appropriate ethical, regulatory, and safety measures

00:07:24.460 in place. So with that caveat and without further delay, please enjoy my conversation with Dr. Vemsi Muka.

00:07:30.980 Vemsi, thank you so much for making time today.

00:07:36.980 Oh, absolutely. Welcome.

00:07:38.340 Yeah, yeah. Sight unseen too. I'm always impressed when people I don't know in advance agree to sit

00:07:44.460 down with me. I'm kind of honored, humbled. So thank you very much.

00:07:48.180 Thank you for coming and meeting.

00:07:50.400 Yeah, the Broad is quite an impressive place. Tell folks a little bit about where we are right now and

00:07:55.220 what makes this unique, even within the hallowed halls of Boston Biomedical Research Institutes.

00:08:01.400 No, I think it's a really exciting place and I've had the pleasure of actually watching it

00:08:05.660 blossom from nothing. It's really the brainchild of somebody named Eric Lander, who was one of the

00:08:11.680 leaders of the Human Genome Sequencing Project. And as that was being completed in draft form in 2001,

00:08:18.060 and then in a quote, finished form in 2003, he saw the need to create some sort of an entity that would

00:08:25.700 take advantage of the power of genomics for improving biomedicine. So he created this institute and it's

00:08:34.140 unprecedented in a lot of ways. It's joint between Harvard and MIT. It involves all the hospitals here in

00:08:40.880 Boston. And it's basically an incredible forum where people can get together and basically pursue

00:08:47.220 projects that they can't pursue on their own in their own individual laboratories. There's a big

00:08:52.700 computational theme here. There's a big theme on being systematic in one's approach, not focused

00:08:59.200 narrowly on the protein that they've studied in the past, but being systematic and seeing where the

00:09:03.820 data takes you. And there's a very pervasive theme of collaboration as well. So Eric has always said

00:09:09.720 that the Broad Institute is a bit of an experiment, but now it's been about 15, 16 years. And I think

00:09:15.080 without a doubt, it's been a wildly successful experiment.

00:09:18.420 Do most of you here at the Broad have an appointment elsewhere? Like I know you spend

00:09:22.000 much of your time at Mass General. Lots of people here spend, you know, they have an appointment at

00:09:26.560 MIT. Does everybody have another appointment outside of the Broad?

00:09:30.080 Almost. So there's almost two types of people that work at the Broad. There's some people that are

00:09:34.680 actually formally employed by the Broad Institute. And there's others like myself that are employed

00:09:39.380 elsewhere. In my case, I work at Mass General Hospital and Howard Hughes Medical Institute.

00:09:44.360 That's my primary appointment. My paycheck comes from there. But then I spend a day a week over here

00:09:49.420 working on collaborative types of projects that I can't pursue easily in my own lab. So there's

00:09:54.340 almost two types of folks over here. I think one of the neat things about the Broad Institute is,

00:09:59.240 you know, traditionally there's this grad student, postdoc, assistant professor, academic track.

00:10:05.000 But how about for all the other people that want to do research in a nonprofit academic setting,

00:10:09.800 but aren't interested in applying for R01 grants or teaching, the Broad actually has an entire

00:10:15.360 research scientist track as well. And so you can be employed here as a research scientist doing

00:10:21.360 science.

00:10:22.220 And funded here.

00:10:23.020 And funded here. And I think that's a very, very different organizational model compared to any

00:10:27.800 other academic organization.

00:10:30.260 Let's back up for a moment. You studied at Stanford where you, you did, were you a computer science

00:10:35.420 major? What was your major in undergrad?

00:10:37.680 I was math and computer science.

00:10:39.200 Okay. Got it. Did you know you wanted to get into biology and medicine?

00:10:43.000 You know, when I was a kid, I'm an Indian American. My father is a retired surgeon.

00:10:47.320 As it turns out, my three older siblings would end up becoming doctors also. So I like to joke that

00:10:52.360 we had the medicine gene in our family. So growing up, I was convinced I wanted to be a doctor of some

00:10:58.040 sort. But then in high school, I fell in love with math. I did the high school math competitions,

00:11:02.720 was pretty good at those, did the high school science fair competitions in math, did well with

00:11:08.380 those. And so I ended up going to Stanford with the goal of being a math and computer science major.

00:11:13.580 And that's what I was squarely focused on. And then towards the end of college, when I was looking

00:11:19.420 for a research project, my advisor told me about the work of Sam Carlin. He's a statistician that was

00:11:26.700 developing some of the underlying methods for biomolecular sequence analysis. He's a fundamental

00:11:31.620 mathematician and statistician.

00:11:33.080 And this is like early 90s?

00:11:35.040 Early 90s. That's right. That's right. So I ended up working on DNA sequence analysis as a college

00:11:40.560 student, writing code to try to analyze DNA and protein sequences. And I fell in love with that.

00:11:46.240 And then my advisor said, why don't you contemplate a future career in medicine, if for no other reason,

00:11:54.400 going to medical school is a great way of learning physiology. You know, I thought about it from a

00:11:59.180 practical perspective. I applied simultaneously to PhD programs in mathematical biology and also

00:12:05.480 applied to MD-PhD programs. And then I applied to this program joint between Harvard and MIT. I

00:12:10.940 actually thought that it was an MD-PhD program. This is all pre-internet. So I thought it was an

00:12:16.500 MD-PhD program, but it wasn't. It was a straight MD program, but it's a joint program between the two

00:12:21.560 institutions.

00:12:22.440 Yeah. They would set aside like a, I remember like 20 students for this, what was it called? The H- 0.91

00:12:27.080 HST.

00:12:27.820 HST program.

00:12:28.700 That's exactly right. That's right. So it's a part of the HST program and it was kind of catered to

00:12:34.120 students that had slightly quantitative backgrounds, math, computer science types of backgrounds.

00:12:38.780 So it provided a bit of a more gentle introduction to medical school.

00:12:43.300 I could have used that program. My first year of medical school was just unbearable. I mean,

00:12:47.920 it was such a culture shock of, I don't know if I told the story on the podcast before the worst

00:12:53.300 exam I've ever done in my life from the point when I decided to care about school was the first

00:12:59.120 semester histology exam in medical school, because I very arrogantly and naively assumed that if I

00:13:06.960 understood the concepts of histology, I didn't need to memorize anything. So I took this very pure math

00:13:15.120 approach, which was I could just derive everything in my mind if I understood the fundamentals and that

00:13:22.000 got me a 53% on the final exam in histology, which like I just sat there looking at this number

00:13:29.820 thinking, how is this possible? Am I the stupidest human being that has ever walked the face of it? It 0.99

00:13:36.620 was really a wake up call that said, you know, you're going to actually just have to start memorizing

00:13:41.820 things around here. It's a big culture shock. I think when I know you have a background in

00:13:46.760 engineering and applied math, so very similar to my background and it made for a lot of unhappiness

00:13:51.840 for me also that first semester of medical school. At least you were in California. I was facing the

00:13:57.540 Boston winter at the same time I was facing histology. That's a good point. I did have that

00:14:03.000 going for me. So when you finished med school, you did a residency in internal medicine. You stayed in

00:14:09.160 Boston, correct? You stayed at the Brigham? That's right. That's right. I've been here now for,

00:14:13.100 it's hard to believe, but about 25 years. During residency, the Brigham is obviously

00:14:18.140 certainly among the top three most academic medicine programs in the country. I was talking

00:14:23.720 to one of your colleagues yesterday and he made the point that when he looked at sort of his class

00:14:29.100 or the cohort of people that entered medicine at the Brigham, you just fast forward 20 years and

00:14:34.680 they're basically the leaders within each of the different scientific fields. So I assume that was a

00:14:38.960 pretty deliberate decision on your part to really preserve sort of academic optionality.

00:14:44.580 Yeah. You know what? I think at almost every single one of these junctures-

00:14:47.940 The nodes, yeah.

00:14:48.740 Yeah, exactly. From college to medical school, medical school to residency, I think maybe I'm

00:14:54.120 just inherently a bit of an indecisive person, but I was unsure as to whether or not I wanted to do

00:14:58.660 a residency. By then I'd fallen in love with basic research. And the question was, was I going to do a

00:15:04.800 basic science postdoc or was I going to do a residency? And so just as I did at the previous

00:15:10.420 node, I decided to just apply for both. So I simultaneously applied for postdoc positions

00:15:14.900 and for residency programs. That was the era when you would fill out your residency match

00:15:21.000 list with a number two pencil. So I'd fill it out every day.

00:15:24.440 I'd fill it out and then I would rip it up the next day. I'd fill it out and rip it up the next

00:15:32.920 day. And I joke that the only reason I did a residency is because the due date was an even

00:15:37.500 number.

00:15:40.340 So that said, I want to put you on the spot and ask you a question that I get asked all the time.

00:15:45.620 And frankly, I don't know the answer to it. And I feel bad that I always provide

00:15:50.600 sort of nebulous answers. But I do get asked a lot by either college students who are contemplating

00:15:57.400 medicine or not, or medical students who are interested in research. And the question is a

00:16:02.080 variant of one, if you're at the college node, but you're very interested in medical research,

00:16:08.080 is there benefit to doing an MD? And then the second order question is, if you're in medical

00:16:13.520 school and you're going to finish and you know you want to do research, is there a benefit to doing

00:16:18.360 a residency? Now I generally tell people, and I, if you disagree with me, I hope you say so very

00:16:23.780 loudly. I generally say if you're in college, but undecided about medical school or not, I would say

00:16:31.340 don't do it. Pursue the PhD. If you are in medical school and presumably at least somewhat interested in

00:16:38.040 medicine still, do an internal medicine residency because there is no substitute for doing that type

00:16:44.900 of research and at least having the ability to understand clinically why you're doing it. So

00:16:50.320 again, do you disagree with that? And if so, how would you, or how would you modify that?

00:16:54.900 I also get asked this question quite a bit because I think there's a group of people out there that

00:16:59.720 love science, are probably good at clinical medicine if they do it, but they're also good at

00:17:05.200 research if they do it. And maybe they're inherently a little bit risk averse as well. So they're always

00:17:09.940 trying to figure out what's the best path for me. The advice that I had gotten when I was in college

00:17:15.880 was that the reason to go to medical school, if you're interested in research, is that it's one of

00:17:20.540 the best curricula for understanding physiology. How do all of the parts come together and operate?

00:17:27.300 And trying to understand how an entire living system operates, if you don't go through all of

00:17:33.120 medical school, if you don't learn anatomy, if you don't learn histology, if you don't learn

00:17:36.920 cardiovascular physiology, it's a bit tough. And so it's a good point you make, right? Which is

00:17:41.040 in somewhere between 16 and 24 months, which is the preclinical part of medical school,

00:17:47.740 it's almost impossible to get a more well curated view of the human body because it's been optimized

00:17:54.200 for that over a hundred years. That's exactly right. More. And it's a living system, right? I mean,

00:17:59.380 you could study yeast, you could study drosophila, but the human is extraordinarily well

00:18:04.680 investigated. And you're right. In about two years or so, you have a very intense curriculum

00:18:09.080 studying one living system, different aspects of it. So from that perspective, I think it's a great

00:18:14.220 education. It's a broad education for biomedical research. At the other node, again, I spent a lot

00:18:21.540 of time thinking about this. And one of my advisors actually said, at least do an internship because then

00:18:26.280 you'll be licensed. You can prescribe medicines at that point. And there is something special about

00:18:32.220 caring for patients. There's something very special about going through the process of

00:18:37.020 residency with your colleagues. And again, you really get to see the human living system at some

00:18:42.740 of its extremes. So I think it's a great way of continuing to learn physiology and pharmacology

00:18:48.100 as well. So I'm super grateful for the path, of course, and of one experiment, but I'm really happy

00:18:53.860 that I did medical school. And I'm also super happy I did the clinical training as well, because it

00:18:58.600 completely shaped my research focus as well. Okay. No, that makes sense. I generally say the

00:19:05.000 same thing, which is, I don't regret the path I took, though I'm glad I didn't know in advance

00:19:10.940 what it was going to look like, because it would have seemed too indirect, but nevertheless. So

00:19:16.020 right now you spend virtually all of your time doing research. Is that correct?

00:19:21.800 That's right.

00:19:22.260 When did you stop seeing patients?

00:19:24.680 It's been about five to six years or so.

00:19:27.100 Okay. Was that a tough decision?

00:19:29.700 It was tough because we went through this entire medical school and residency process with the

00:19:34.820 goal of actually seeing patients and caring for them. The truth is the number of patients

00:19:39.300 I was seeing was asymptoting as a function of time. Emotionally, it was very, very difficult

00:19:44.260 to go to zero. So even though I was seeing probably one patient a month or so, five years ago or so,

00:19:50.020 going to zero was actually the harder part. But at some point I had to just do the math and say,

00:19:56.000 there's only a certain number of hours per day. I'm running a pretty big research group that's

00:19:59.940 focused on mitochondria and mitochondrial disorders. And the long-term goal is to impact

00:20:05.000 these patients. And so I'm going to let other people be the ones that care for these patients

00:20:09.700 at the front line. And I'm going to try to develop new diagnostics and drugs for these patients.

00:20:13.540 So let's start talking about the mitochondria. When did you come to the realization that that

00:20:19.760 was the area you wanted to focus on? And what is it about the mitochondria that especially drew you in?

00:20:24.200 So as a first year medical student, I was a little bit unhappy here in Boston. It was cold.

00:20:30.060 There's a lot of memorization that first year, that first semester in particular. I wasn't sure if I'd

00:20:35.540 actually made the right decision in coming to medical school from a math background. And then right in

00:20:40.940 the middle of that first semester, when we were taking our histology and pathology class,

00:20:46.200 we had a very, very brief lecture on myopathies, muscle diseases. And there's one slide that

00:20:53.040 basically indicated that there's a rare form of myopathies that are due to mutations in the

00:20:59.660 mitochondrial DNA. Now, truth be told, I don't think I'd appreciate it at that time that we had our own

00:21:06.220 mitochondria with their own genomes. And so that immediately fascinated me. You dig a little bit

00:21:11.920 deeper, you learn that these were once bacteria. They're kind of swimming around in our cells. So

00:21:16.600 that's kind of cool. You see an image of mitochondria in the muscle. It's just beautiful. And there's just

00:21:22.120 something that really captivated me. And I kind of got hooked on that organelle that first semester of

00:21:28.160 medical school. And I've told this story to other people, Peter, but what happened was a few weeks

00:21:33.760 later, my dad's cousin, who is at that time, a postdoctoral fellow in the Harvard research,

00:21:40.160 one of the research hospitals. She heard that I was unhappy. So she invited me over to her house

00:21:44.760 in Somerville for a Friday night dinner. So it started to snow. I took the red line all the way

00:21:50.320 to Somerville. It was a 20 minute walk to her house. By the time I had gotten in, my shoes were soaked. I

00:21:56.060 was freezing. She immediately looked at me and said, Oh my God, you look terrible. And so she took my coat

00:22:01.580 off. She tried to dry me up and she started cooking dinner. And then her boyfriend appeared

00:22:06.920 like an hour later. And now it's about eight o'clock or nine o'clock or so. We had dinner together.

00:22:12.040 And then we found out that the tea had been shut down. And so not only was I an unhappy first year

00:22:17.840 medical student, but now I'm spending Friday night with my dad's cousin and her boyfriend in Somerville.

00:22:25.060 And so they create a makeshift bed for me in their living room. I'm spending the night there at this

00:22:29.540 point. And I look at the bookshelf and there's a textbook of mitochondrial biology on the bookshelf.

00:22:36.380 So just a few weeks earlier, I had this initial encounter with the organelle. I saw this book,

00:22:42.960 so I just took it off the shelf and I just started reading it. I read about a hundred pages that first

00:22:47.140 night. I borrowed it and then I basically devoured that book over the course of the next few weeks.

00:22:51.680 And I just fell in love with the organelle at that point and thought to myself that this is what I

00:22:56.560 want to work on for the rest of my career. That's an amazing story. And I think for people

00:23:01.640 listening, they shouldn't be concerned that they haven't had that eureka moment, right? I mean,

00:23:06.340 I've heard a lot of amazing scientists talk about their passion and not all of them. In fact,

00:23:11.760 most of them don't have such a laser moment like that. Now, of course, the stories of people that have

00:23:18.600 those moments are even more vivid in my mind of that, oh my God, I can't think of anything but this.

00:23:24.080 So let's talk a little bit about the mitochondria. You've already alluded to some of the really

00:23:27.840 interesting features about it. So let's start with the, well, gosh, which one's the most

00:23:31.960 interesting? Let's just start with the lineage, right? So how did these bacteria find their way

00:23:36.740 in to these eukaryotic cells? Yeah, this is, I think, one of the most fascinating aspects of

00:23:43.380 the organelle. So this is a process known as endosympiosis. So the current theory is that there is

00:23:49.420 an organism probably resembling a modern day gram-negative rod, something like a rickettsial

00:23:55.880 species that merged. Folks might not even know what that means. So gram-negative is just a staining

00:24:01.360 that allows us to identify a type of bacteria. So you've got this bacteria that is shaped like a rod,

00:24:07.500 literally. And how does it get its energy prior to this encounter? The thinking is that it had a full

00:24:14.480 electron transport chain and was probably capable of doing aerobic metabolism using things like oxygen

00:24:20.680 as a terminal electron acceptor. And one of the theories proposes that that had an electron

00:24:27.320 transport chain that's not that different than modern day mitochondria. The hypothesis is that that

00:24:33.800 somehow merged with something resembling modern day archaea. So there's three main domains of life.

00:24:41.700 You have archaea, you have bacteria, and then you have eukaryotes. And eukaryotes have nuclei. So the

00:24:49.020 hypothesis is that there's an archaeal species, there's a bacterial species, they form a union of

00:24:54.380 some sort. And this is what gave rise to modern day eukaryotic cells, maybe about one and a half

00:25:00.060 billion years ago. That's kind of remarkable. And is the idea, I mean, what kind of evolutionary pressure

00:25:07.980 must have been placed to create an entirely new species? It's also binary. It seems somewhat

00:25:14.660 binary to me, is it? In other words, when you look at the difference between, I don't know, say

00:25:20.240 a chimpanzee and an ape or an ape and a human, you can see lots of continuous evolution, you know,

00:25:28.000 Neanderthal, et cetera, et cetera. What you're describing sounds much more switched on, switched

00:25:32.960 off. Is that the case? Was this a, or were there lots of iterations that we can't even

00:25:37.800 appreciate today in between the union and the current paradigm?

00:25:44.880 As far as we know, endosymbiosis of the mitochondrion only took place once.

00:25:49.640 Wow.

00:25:50.500 So there are lots of eukaryotes, almost, not all, but almost all eukaryotes have a mitochondrion.

00:25:57.200 And if you look at the DNA of those mitochondria that still have a genome, the phylogenetic

00:26:04.340 analysis tells us that this was a monophyletic event. This took place only once. And that's not

00:26:10.420 true for something like the chloroplast. That also arose through serial endosymbiosis. And that likely

00:26:17.320 took place at least two independent times in evolution. But for the mitochondrion, it probably

00:26:22.940 only took place once as far as we know. And the chloroplast of course is to a plant effectively

00:26:27.980 what the mitochondria is to an animal like us. That's exactly right. Yeah. So how many genes do

00:26:35.040 we think that that gram-negative rod had? Great question. Probably about a thousand to two thousand.

00:26:41.820 And at the time, the species that it merged with would have had how many to the best of our guessing?

00:26:48.320 Probably also a few thousand. So you had two things that had comparable numbers of genes

00:26:53.820 merge. And yet today, you or I would have 30,000, 20,000 genes inside the nucleus. And we'd have,

00:27:04.620 is it 13 genes inside our mitochondria? 13 preserved genes. So in other words, to a first order approximation,

00:27:11.840 the mitochondria lost all of its genes. But a deeper dig says, actually, it somehow hung on to 13.

00:27:20.000 Why do you think that was? This is what we call reductive evolution. Modern day mitochondria actually

00:27:26.020 represent a mosaic. So you need about a thousand proteins total to make our mitochondria. And so some

00:27:34.820 of those are attributable to the original bacterial ancestor. And others are brand new innovations that even

00:27:41.000 that original bacteria did not have. But on the reductive side, approximately a thousand genes from

00:27:47.480 that original bacteria have either been lost altogether or have been transferred to the nuclear

00:27:53.700 genome. So that that genome today is tiny. It's only about 16,000 bases. And it encodes 13 proteins.

00:28:00.800 But that's only if you're looking at animals. There's a lot of different eukaryotes. So there's a lot of

00:28:06.400 mitochondrial diversity. So you and I still retain 13 proteins that are encoded by our mitochondrial

00:28:13.580 genomes. But if you look at malaria, it's also eukaryote. It has a mitochondrion, but its genome

00:28:19.900 only encodes three proteins today. So that's additional reductive evolution. What about things like,

00:28:25.880 you know, flies, yeast, are they also variable? Flies are also animals. And so most animals- You would put

00:28:31.920 them in the category, yeah. That's right. So they would have about 13, sometimes 14 proteins. But if

00:28:36.900 you look at something like Giardia, which causes a terrible diarrheal illness, beaver fever, that's a

00:28:43.780 eukaryote. It's actually lost its mitochondrial DNA altogether. I know we talk about 13 proteins. Is it

00:28:50.320 one-to-one mapping or are there genes that are non-coding or is it- Right. So the mitochondrial genome

00:28:55.440 encodes two ribosomal RNAs, 22 tRNAs, and then 13- 13 proteins. Yep. So obviously to have 13,

00:29:05.080 we pretty much have a good sense of what the function of each of these are. How many diseases

00:29:09.420 that afflict humans result from genetic disorders there, inherited mutations that produce dysfunctional

00:29:17.100 proteins? Right. There's about 250 different syndromes of the mtDNA, of the mitochondrial DNA.

00:29:24.660 And by syndrome, I mean, there's a particular mutation that's associated with a particular

00:29:29.440 clinical phenotype. But that's only if we're talking about the mitochondrial disorders of the

00:29:36.240 mtDNA. That's right. Is that a higher or lower frequency on a probability basis, given the number

00:29:43.080 of base pairs? Because you said something like only, what, 16,000 base pairs? It's relatively tiny.

00:29:47.140 We have genes in our nuclear genome that are 10 times that size. So from a probability basis that

00:29:55.280 you could have 250 mutations in that 16,000 base pair that would each lead to these distinct

00:30:02.120 mitochondrial diseases, is that more or less robust from a DNA perspective than our nuclear DNA?

00:30:08.280 I don't think we have a good answer to that question. There's a little bit of an ascertainment

00:30:14.100 bias in clinical medicine. So, you know, when I was going through residency, probably when you're

00:30:18.460 also going through your clinical training, the answer to almost every single question about

00:30:23.560 mitochondrial disease was maternal inheritance. And that's because the nuclear genome was sequenced

00:30:31.540 in draft format in 2001. But the mitochondrial DNA was sequenced in 1981. So that was almost exactly

00:30:38.780 20 years earlier, and it's small. And so as soon as you saw patients that had a particular clinical

00:30:45.800 phenotype that could be a mitochondrial disease, it was easy to sequence the mitochondrial DNA.

00:30:51.100 So beginning in 1988, when two papers were published reporting the first mutations in the mitochondrial

00:30:56.800 DNA, it's been relatively straightforward to sequence it and associate mutations in the

00:31:02.540 mtDNA with the disease phenotype. And this is why now there's about 250 of these mtDNA mutation

00:31:09.580 clinical phenotype mappings. The nuclear genome has been a little bit slower to follow since 2001.

00:31:18.240 In beginning, even a little bit before the completion of the genome sequence, we've now as a community

00:31:23.520 been able to identify about 300 different genes in the nuclear genome that underlie mitochondrial

00:31:29.800 disorders. And of course, that's 300 different genes, but then there's different allelic variants

00:31:34.880 as well. So I think it's going to be a little while before we can answer your question in a

00:31:40.140 satisfactory way, but it's a really provocative question.

00:31:43.460 And one more number to sort of extract from that is, do we have a sense of how many genes in the

00:31:50.160 nuclear genome are required for mitochondrial function? So in other words, what's the denominator

00:31:54.740 nuclear-wise?

00:31:56.440 Right. That's actually one of the areas that we focused on heavily in our laboratory.

00:32:00.240 So beginning soon after the sequencing of the human genome, we knew that the human genome encodes about

00:32:06.440 22,000 proteins. It's an important question is, which of those find their way to the mitochondria?

00:32:12.700 And it's going to be more than 13. These are elaborate organelles. So we used a lot of methods in the

00:32:19.120 early 2000s, things like proteomics, GFP tagging, microscopy, computation. And we're able to

00:32:26.200 identify about 1,100 proteins that are made by the nuclear genome that find their way into the

00:32:31.840 mitochondrion. So those 1,100 proteins have to work with those 13 proteins in space and time to do all

00:32:39.780 the amazing things that the organelle does.

00:32:41.640 How do they communicate? Do we know how expression of mtDNA is coordinated with nuclear DNA?

00:32:50.220 It's a really, really interesting question. So some of the mechanisms are known, but a lot of

00:32:56.020 mechanisms are not yet known. So when you exercise, for example, if you're deconditioned, and if you do a

00:33:01.560 combination of aerobic training and strength training, you can actually increase the number

00:33:06.480 of mitochondria. And there's an entire transcriptional program that will turn on all those

00:33:12.460 nuclear genes. But that same program will also turn on the replication factors that will go

00:33:18.380 into the organelle and cause the mitochondrial DNA to replicate as well. So it's a really smart

00:33:23.880 transcriptional program that says in response to exercise, make more of the nuclear encoded

00:33:29.680 components and make more of the mitochondrial DNA and make more of the mitochondrial DNA encoded proteins.

00:33:35.040 So that increases mitochondrial density in a given cell. So you have a myocyte, a muscle cell.

00:33:41.740 Do we know what the actual signals are? They basically instruct the cell to make more DNA. So what

00:33:46.700 is the input from the exercise? You mentioned two types of exercise, right? Strength training

00:33:51.040 and aerobic activity. Those have very different physiologic properties. So what is it that at the

00:33:59.200 physiologic level is leading to this nuclear level?

00:34:02.640 One of the important signals is something called AMP kinase, the change in the ATP to ADP ratio.

00:34:08.600 That's known to be one of the activators of this particular transcriptional program that

00:34:14.160 induces mitochondrial biogenesis.

00:34:17.040 So this would be a decrease in AMPK activity?

00:34:22.160 An increase in AMP kinase.

00:34:23.420 An increase. Okay.

00:34:24.540 So that is basically sensing something like the ATP to ADP ratio.

00:34:29.000 Ratio is going down. Yeah.

00:34:30.320 So when we exercise, the ATP levels are usually pretty nicely defended. But then what's happening

00:34:36.000 is there's another reaction that's taking two ADP molecules, making an ATP, liberating AMP.

00:34:42.740 While ATP levels are defended, some of these ratios change and that can be sensed. And that's one of

00:34:48.160 the inputs into this program that says, let's make more mitochondria. 0.99

00:34:51.340 Yeah. Now we'll come back to this far down the line, but just because you mentioned it,

00:34:56.480 one of the effects of a drug called metformin that everybody loves to

00:34:59.980 ask about is it activates AMPK. Does that imply that metformin administration

00:35:05.780 by itself alters mitochondrial density?

00:35:09.160 So I think when we talk about exercise, I think AMP activation is generally regarded as one of the

00:35:16.360 important signals for mitochondrial biogenesis, but I don't think it is sufficient.

00:35:20.880 It's not sufficient. It's necessary though, potentially?

00:35:23.620 I think it is, but I would need to review some of the older literature to really confirm that.

00:35:28.600 So what do you think some of the other signals are?

00:35:30.620 I think some of the other signals are things like calcium. And clearly there's other signals.

00:35:35.280 There's a couple of disease states as well, where we see a massive proliferation of mitochondria.

00:35:41.320 They're not functioning properly, but there's a massive proliferation of

00:35:44.780 malfunctioning mitochondria. So we're trying to work at some of those signals as well.

00:35:50.060 So as of right now, it's a bit of an open question as to exactly how

00:35:53.600 the number of mitochondria is sensed and regulated, but we know some of the inputs.

00:35:58.840 Right. We know crudely what the inputs are, right? Exercise, as you mentioned.

00:36:03.100 What are some of the other things that even hormonally, for example, did nutrients play a

00:36:07.860 role?

00:36:09.040 One of the studies that I was a part of about 10 to 15 years ago or so show that even things

00:36:14.000 like androgens and testosterone can actually influence the amount of mitochondria. Disuse

00:36:19.900 is a great way of rapidly eliminating mitochondria. And then in terms of nutrients, NAD is an important

00:36:28.380 signal as well. This transcriptional regulator that I'm talking about, it's called PGC1-alpha,

00:36:35.220 and upstream of it is something called CIRT1, which of course utilizes NAD as a cofactor.

00:36:42.200 So there's a couple of these inputs.

00:36:44.100 I'm having lunch with David Sinclair later today, so we'll be talking plenty about this.

00:36:48.380 Absolutely. Absolutely. So I think a couple of these signals are beginning to emerge.

00:36:54.020 We're not at that stage right now where in a pill, we can put seven of these things,

00:37:00.360 give it to a patient, and boom, we can replace exercise. We're not there yet, but it'd be

00:37:05.920 remarkable if we understood the process well enough so that we could one day.

00:37:09.860 Yeah. There's been lots of claims of exercise in a pill. I remember the New York Times wrote about

00:37:15.360 something called IRISON. God, they wrote about it in 2011, 12, and then I saw it again recently.

00:37:20.460 And it's funny, of course, you see these stories, which you realize they're being written about in

00:37:25.200 such a crude way that you can infer nothing. So you go back and look at the paper and you realize

00:37:28.840 it's kind of a ridiculous claim at this point. What is the turnover of mitochondria in a cell?

00:37:34.280 So if you have, let's put some scale to things for folks. So I'm going to take, I'm going to biopsy

00:37:39.200 one cell from your quadricep, one muscle cell. How many mitochondria approximately would be in it,

00:37:46.140 assuming you're a relatively well-conditioned individual? Talking about the number of mitochondria

00:37:51.260 is a little bit ambiguous because they're not quantal units. Mitochondria will constantly undergo

00:37:57.760 fusion and fission. So at any given state, you can have a different number of mitochondria.

00:38:04.900 The mitochondrial DNA is a quantal unit. So you can ask how many mitochondrial genomes are there

00:38:10.680 per nuclear genome in a cell type. So a fibroblast will have a few hundred, maybe a thousand copies

00:38:17.320 of the mitochondrial genome for each of its nuclear genome. That's right. But there's a lot of

00:38:22.640 variation. The highest would be what, like a cardiac myocyte or something, or a neuron?

00:38:27.660 The unfertilized egg. Wow.

00:38:30.760 It has half a million copies of mitochondrial DNA.

00:38:35.100 You see, I wouldn't have guessed that. I don't know why. I could have come up with 10 guesses and

00:38:38.980 that would not have been one of them. It's remarkable. And dad's sperm probably has a few

00:38:43.720 hundred at best. And I would have guessed the sperm needed more because it's doing the motion, 0.98

00:38:48.480 right? It has to fight to get to the egg. Part of the reason is that the egg is so big

00:38:54.100 relative to the size of the sperm. So that's to a first order approximation,

00:38:59.580 the explanatory variable. But the unfertilized egg is sort of the Olympic gold champion when it comes to

00:39:05.120 the Michael Phelps of MTDNA. That's right. That's right. And then red blood cells, of course,

00:39:11.000 have no mitochondrial DNA. Then what's the turnover look like, right? So if you have a

00:39:15.920 cardiac myocyte that presumably lasts for a long time, you're in an individual, like these cells

00:39:21.520 are not turning over quickly. Are they turning over new mitochondria constantly? Is that, how often

00:39:28.580 is that mitochondrial DNA churning over? I guess what I'm trying to get at is what's like the half

00:39:35.280 life of these sort of non-discreet mitochondria? I think in most non-dividing tissues, the half life

00:39:42.520 is on the order of a few days. Wow. So we are cranking out mitochondria.

00:39:48.660 That's right. That's right. There's differences between dividing cells versus non-dividing cells.

00:39:54.160 When you have a dividing cell, obviously, as a function of the eukaryotic cell cycle,

00:39:59.640 you need to double the number of mitochondria, and then you partition that into two. Then you

00:40:04.640 double partition. So dividing cells have very different mitochondrial turnover dynamics than

00:40:11.500 non-dividing tissues like muscle or neurons. There's like a hundred other questions I want to

00:40:18.060 ask, but I also don't want to. I want to get us back to the beginning so that we can set the

00:40:22.040 framework to ask some of these other questions. So you alluded to something, right, which was,

00:40:26.680 I believe, something that at least for me until very recently, I didn't think was challenged or

00:40:32.920 questioned, which was the origin of said mitochondria being maternal. So presumably in the early 80s,

00:40:41.120 when we first, we, like I had anything to do with it, when people far smarter than me

00:40:47.240 sequenced mitochondrial DNA, given how few the number of genes were, it was not difficult to

00:40:53.780 realize, hey, this all seems to come from the maternal side, not the paternal side.

00:41:00.160 And as recently as several years ago, that was considered almost an axiom. All the DNA in your

00:41:07.800 mitochondria comes from your mother, not your father. Has that been called into question lately?

00:41:12.120 Have there been exceptions to that?

00:41:13.360 I think the textbook teaching is that the mitochondrial DNA is transmitted exclusively

00:41:19.100 maternally. And the reason for that is related to...

00:41:23.280 This comes back to the egg, I'm sure, right?

00:41:24.620 It gets back to the sexual dimorphism. The egg is huge. It has half a million copies of empty DNA.

00:41:30.060 Dad's sperm is tiny. It only has a few hundred copies of empty DNA. So by sheer dilution,

00:41:36.700 it's very difficult for dad's empty DNA to get transmitted, right? It's outnumbered half a million to a

00:41:41.840 hundred. But on top of that, there's actually active mechanisms that will seek and destroy

00:41:48.580 dad's mitochondria. So the mitochondria coming from the sperm are coated with a protein called

00:41:54.660 ubiquitin. So after fertilization, those mitochondria are actually actively eliminated by a surveillance

00:42:00.180 program. So not only is it very difficult for dad's DNA, empty DNA, to compete with a huge number of

00:42:07.480 empty DNA molecules from mom, but those that do make it inside are actively destroyed as well. So

00:42:12.720 mechanistically, these are the two reasons why empty DNA is passed on almost exclusively from mom 0.91

00:42:20.960 to child. Now with that said, about 15 years ago or so, there's a case report in the New England Journal

00:42:27.360 of Medicine that took a biopsy of a particular individual, and they saw empty DNA molecules with

00:42:35.540 two different haplotypes. And then they looked at the haplotype of the father, and they concluded that

00:42:41.300 this was a rare case of paternal transmission. So that paper from about 15 years ago was sort of the

00:42:48.800 lone exception to this rule that we accept as an axiom. And then about a year ago or so, in 2018,

00:42:56.400 another paper emerged, again, claiming that in a few families, there is paternal transmission of

00:43:02.200 mtDNA. It's a rare event. But this is an active area of research. Rules are made to be broken. And I

00:43:09.740 think two things. Number one, I think other people will have to try to replicate these results to make

00:43:15.060 sure that even if they're rare, they're real and not some sort of a technical artifact. And then number

00:43:21.200 two, if they're not technical artifacts, I think there's an opportunity to learn something very,

00:43:26.540 very deep about the mechanisms of maternal transmission.

00:43:29.540 Right. Because that would suggest, based on the explanation for how this takes place,

00:43:35.600 it's hard to deny that the first one's going to continue to take place, which is just the

00:43:39.520 stochastic sampling. But you could certainly see scenarios under which the find and kill program

00:43:47.320 malfunctions, and therefore you sneak in a little bit of the paternal DNA.

00:43:51.820 That's right. And in fact, in one of these papers that was just published in 2018,

00:43:56.000 they speculated that perhaps there's a mutation on the nuclear genome in that program, so that it's

00:44:04.560 possible for dad's mtDNA to get transmitted.

00:44:08.520 Well, that is interesting. But I think to the first order approximation, we can still assume that

00:44:12.560 mitochondrial DNA is maternal.

00:44:14.160 Yes.

00:44:14.380 So, let's talk a little bit about sort of mitochondria 101. So, you're a first-year

00:44:19.800 medical student learning about the mitochondria. Here's basically what you learn, right? You've

00:44:23.520 got this inner membrane, this outer membrane. I forget what the term is. Do they call it the

00:44:28.900 powerhouse of the cell or something like that? I think there's some sort of glib term that I know

00:44:33.020 that people who study mitochondria like yourself and Navdeep Chandel, who I've interviewed, sort of

00:44:37.460 bristle at the simplicity of such a term. But the idea is, all things being equal, a cell takes in

00:44:44.760 glucose, free fatty acid, substrate for energy. And let's start with the glucose, because we'll

00:44:51.160 probably spend more time talking about glucose today. Through, I can't even remember, 10 steps,

00:44:55.960 we turn glucose into pyruvate, more or less. We then basically have a choice. A cell has a choice

00:45:02.940 based on the availability of oxygen and the rate at which ATP is being demanded. So, if oxygen is

00:45:10.260 scarce relative to ATP demand, you can take an inefficient route, but at least you guarantee to

00:45:17.460 get some ATP, which is you can turn the pyruvate into lactate. And lactate itself is interesting,

00:45:23.020 and maybe we can talk about it, but you generate some ATP. If the demand for ATP is not as great,

00:45:29.680 and there's sufficient cellular oxygen, you can take that pyruvate and make acetyl-CoA. And that

00:45:36.480 acetyl-CoA then becomes one of the substrates leading into this thing called the electron transport

00:45:42.140 chain you alluded to. Walk us through what happens in that latter scenario.

00:45:47.780 I like to think of the mitochondria as being the key place where there's energy transformations.

00:45:53.460 So, in order for a cell to work, it needs energy, but in the same way that in order to live our

00:46:01.100 lives, we need to have one type of battery for our iPhone, a different type of a battery for our

00:46:06.280 laptop, a different type of a battery for our automobile. We need energy packaged in different

00:46:11.860 ways. So, this is sort of the charm of the mitochondria, and what it does is it's going to take

00:46:19.020 fats and carbohydrates and proteins, and it's going to break it up almost like a Cousinart.

00:46:24.320 And as it's breaking it up, it's going to harness the electrons, and that's what's called an

00:46:28.580 electromotive force. That's one type of energy. So, you're probably familiar with things like the

00:46:34.820 NADH to NAD ratio. That's basically an electron carrier. So, that's one form of energy. And certain

00:46:41.200 types of enzymes can be powered directly by NADH and NAD. But then that can also be converted

00:46:48.160 to a gradient, a voltage across the inner membrane, a different type of an energy form.

00:46:55.180 And that energy form can be used to drive transport.

00:46:59.920 So, let's explain what that means to people. So, everybody knows what a battery looks like,

00:47:05.140 like literally a Duracell battery that you stick into whatever, you know, your kid's toy. And that's

00:47:11.140 typically about 1.5 volts, right? A 9-volt battery gets its name because it has 9 volts, but everyone

00:47:17.220 recognizes it as the goofy square one. We take this, I think, for granted because we sort of have

00:47:22.760 these backgrounds in math or engineering and stuff. But I think for the average person, it's helpful to

00:47:27.880 understand what voltage means. And you just alluded to it, right? It's a potential, it's a gradient that

00:47:33.940 is created by disproportionate placement of electrons. And it's only with that that you can

00:47:41.200 generate power. So, why does a battery die? Why is it that, let's pretend I'm still using a Walkman.

00:47:47.580 I put my two 1.5-volt AA batteries in my Walkman. At some point, it stops working. Why?

00:47:55.020 At some point, it's not going to be able to hold charge. And so, I'm probably more familiar with

00:48:00.060 the mitochondrial battery than I am with the Duracell battery. But when you have nicely

00:48:05.140 functioning mitochondria, you can charge them. You can create a nice voltage that can be used for work.

00:48:11.780 You can dissipate it. You can recharge it. But at some point, when it gets older, when the membrane

00:48:17.480 is a little bit leakier, when it can't, it'll stop holding charge.

00:48:21.560 So, it's that ability to keep a difference in charge across the membranes that is the same reason

00:48:27.860 your little 1.5-volt battery in your Walkman. I love that I'm saying Walkman, by the way.

00:48:33.340 There's half the people listening. This won't actually know what that means. Or the 9-volt battery in your

00:48:38.220 smoke detector, or frankly, the fancy, I don't even know how many volts a Tesla battery is. It's like

00:48:43.460 12 volts, I'm guessing, or something like that in an electric vehicle, whatever. But there's a reason

00:48:47.900 you have to keep charging these. But at some point, even when you charge them, they cease to work if

00:48:53.100 it's a rechargeable. And you're basically saying, look, think of a mitochondria as partially being

00:48:58.400 a battery.

00:48:58.840 That's right. That's only one form of energy, right?

00:49:01.580 That's taking chemical to electric.

00:49:03.440 That's exactly right. And so, you can have an electromotor force, you can have a proton

00:49:07.000 motor force, and then you can dissipate that battery to basically catalyze a formation of ATP,

00:49:14.940 which most people know is the energy form that's used to power muscle when you exercise.

00:49:21.180 So, the mitochondria is doing all of these elaborate energy transformations from electrical

00:49:26.300 potentials to proton potentials to phosphorylation potentials. And different enzymes and processes

00:49:32.880 and machines in your cells will use one or the other.

00:49:36.300 This is so exciting. Like, I've tried to have this discussion with my daughter, who's 10, and she's

00:49:41.700 not quite at the point where she sees why I think this is amazing. But she's almost at that point where

00:49:49.100 I guess you can look at a piece of food on the plate and explain why eating that is essential,

00:49:57.060 right? Like, where is the energy in that food? So, maybe we'll use this as sort of the example to

00:50:02.580 go full cycle. So, you are looking at a Cheerio on your plate, right? Now, that Cheerio is mostly

00:50:10.920 carbohydrates. So, we'll simplify this. And it's got glucose in it. It's probably got more complex

00:50:17.920 carbohydrates in it. But at a molecular level, it's a lot of carbons joined to hydrogens,

00:50:24.840 carbons joined to carbons, and carbons joined to oxygens, and oxygens joined to hydrogens. That's

00:50:30.960 probably most of the bonds, correct? So, bonds contain energy. There's chemical energy there.

00:50:37.580 Which of those bonds would be the most energetic? Probably the carbon hydrogen in total number,

00:50:44.540 just given the ubiquity of it, right? A carbon hydrogen has more energy than a carbon carbon,

00:50:48.660 I'm guessing. I'm not too sure.

00:50:51.580 I think that's the case. In the show notes, we'll list what the potential energy is in each of those

00:50:55.680 bonds. But nevertheless, you go through this process of actually eating the thing. You put it in your

00:50:59.420 mouth. You break it down. Mechanically, you've broken it by the time it exits your stomach.

00:51:04.600 But it's really once it gets absorbed out of the bloodstream that you begin this chemical process

00:51:11.420 of breaking those bonds. And then you get something for free, right? When you break those bonds,

00:51:16.700 that's when you're getting the energy to create this electron gradient.

00:51:20.760 That's right.

00:51:21.340 Which you then use at the end to basically do this one thing you alluded to, which is

00:51:27.420 phosphorylate the ATP.

00:51:29.700 So that energy gradient allows you to then put a phosphate back onto an ADP.

00:51:36.860 That's the most obvious one from food. Now you talked about NAD and NADH. Can you say a little

00:51:42.540 bit more about what those are and how they fit into this, in particular in the mitochondria?

00:51:46.860 Again, the classical teaching is that NAD is an electron carrier. When you have two electrons,

00:51:53.440 it's going to be in what's called the NADH form. And if you don't have those two electrons,

00:51:57.500 it's in the oxidized form, which is the NAD form. That's another way of holding the energy

00:52:03.780 that can catalyze reactions. What we're learning over the last few years is that in addition to this

00:52:10.460 role as an electron carrier, the NAD itself can be used as a substrate for other reactions.

00:52:18.440 Outside of the mitochondria or inside?

00:52:20.600 Outside of its role as an electron carrier.

00:52:22.820 Okay. And that biggest role for that is in these, I guess we'll allude to it,

00:52:27.600 but there are these complexes in the mitochondria.

00:52:29.900 But that's still as a redux carrier. So the electrons go from the NADH are transferred to

00:52:35.740 the electron transport chain. And that's basically a wire that's going to conduct the electron

00:52:40.040 until they hit oxygen and make water. That's a downhill process. And during that downhill process,

00:52:46.680 these protons are pumped across the inner membrane, generating about 150 millivolts. And that can then

00:52:53.080 be used to do work, including catalyzing the conversion of ADP to ATP. And that's what we call

00:53:00.220 oxidative phosphorylation.

00:53:02.380 Yep. But then talk about the other example that you were using there.

00:53:05.460 And this is a newer area of biomedicine. And actually, David Sinclair has worked in this area

00:53:11.280 quite a bit. But everything that I just told you relates to electrons coming on and off of NADH.

00:53:17.080 So when we talk about the NADH to NAD ratio, that's sort of the redox potential. But that NADH molecule,

00:53:24.380 it can also participate as a substrate in certain chemical reactions.

00:53:30.140 So sirtuins use NAD.

00:53:31.620 Exactly. Sirtuins, some of the DNA damage response pathways, the PARPs,

00:53:37.040 they'll actually use the NADH as a cofactor. So that's important because it's possible in certain

00:53:44.760 states. Like when you have a damaged cell, for example, if you have a cell with damaged DNA,

00:53:51.440 that NADH can decline very, very rapidly. It's an electron carrier. So you're actually losing the

00:53:57.700 ability to carry some of this charge now.

00:54:00.460 This is interesting, right? Because observations are, as we age, these NADH levels decline.

00:54:05.220 Is that due to greater demand for it? Is it due to a reduction in production? And of course,

00:54:12.480 the clinical question that everybody asks is, is there benefit to replacing it?

00:54:17.120 What's really interesting is there's a couple of different signatures of the aging process. So if

00:54:21.520 you biopsy muscle from individuals of varying ages, you'll see a gradual decline in the NAD content.

00:54:28.740 And if you quantify the amount of mitochondria using any of the different metrics, you'll see

00:54:35.880 a decline. If you look at things like VO2 max in skeletal muscle as a function of age, you'll see

00:54:42.400 a gradual decline. So there's this gradual decline in NAD and in mitochondrial activity as a function

00:54:50.440 of age. And I think the big question in the field is, do you just have an old and sick tissue? So you

00:54:56.800 have sick mitochondria or will targeting the mitochondria actually somehow alleviate age

00:55:03.880 associated decline in tissue function? This is such an interesting question and something that

00:55:10.980 probably until a year ago, I don't think I spent enough time thinking about, which is

00:55:15.280 what does it mean to age at the level of the mitochondria? And what are the implications of it?

00:55:23.300 And perhaps most importantly, what can be done to slow the rate of aging? Now, you study a problem

00:55:31.440 at sort of a different node, which is you are looking very specifically at diseases that people,

00:55:38.700 most people haven't actually heard of. And I don't want to say you're not interested in those diseases

00:55:44.060 per se, because you are, but they're basically a gateway for something else. So later on in this

00:55:51.220 podcast, I suspect we'll come back to more of this mitochondrial fitness, health, inflammation,

00:55:57.240 mitophagy. There are many other topics I want to explore with you, but let's now, having laid the

00:56:01.540 groundwork, go back and talk about your work and what you're learning. So give me an example of some

00:56:06.620 of the diseases that you study in your lab. We've historically placed a lot of emphasis on a very

00:56:12.300 large collection of individually rare inborn errors of mitochondrial metabolism. These are typically

00:56:19.220 single gene disorders. They can be due to recessive mutations in the nuclear genome, or they can be

00:56:27.220 due to mutations in the mitochondrial genome. But at the end of the day, there's a component of the

00:56:33.100 mitochondrion that's defective at birth. And so what we just spoke about is the fact that as all of us age,

00:56:39.900 there's a gradual decline in the activity of mitochondria. The big question in the field is

00:56:44.700 whether that's cause or consequence. These other 300 rare monogenic disorders of mitochondria,

00:56:51.860 there's no doubt. There's no question. The gene did the randomization for you. You know cause and

00:56:56.820 effect. That's exactly right. The mitochondrion is defective at birth. And now we can actually

00:57:01.620 evaluate what the consequences are. You said about 300. What is the phenotypic spectrum? How many of

00:57:07.760 these, for example, are fatal within the first year of life? They tend to follow a bimodal distribution.

00:57:14.420 The recessive mutations in the nuclear genome, they tend to present early in infancy within the

00:57:19.860 first few weeks or months of life. The mutations in the mitochondrial genome, those tend to present a

00:57:25.660 little bit later in life. Wow. So let's focus on the latter group for a moment. I mean, unless you'd

00:57:30.780 prefer to start with the former, which one do you spend more time looking at? We spend a little bit

00:57:34.680 more time on the nuclear. Okay. Because they present earlier and presumably they're more severe.

00:57:39.220 That's right. Okay. So give me an example of what some of those mutations are and what their phenotype

00:57:43.760 is. One of the clinical syndromes that we study is something called Lee syndrome. So there's about

00:57:49.060 80 different genes that can be mutated to give rise to this clinical syndrome, which is basically a

00:57:56.120 subacute degeneration of gray matter. It's a very rapid neurodegeneration. It's a terrible,

00:58:02.620 terrible disease. And what age are people when they start to experience this neurodegeneration?

00:58:07.740 So most of these kids are actually born developmentally okay. And then within the

00:58:12.900 first couple of months of life, there's some sort of a stressor. Sometimes it's some sort of an

00:58:17.420 infection. Sometimes it's dehydration that'll put them into a neurometabolic crisis. And at that point,

00:58:23.840 if you look at their brain MRIs, you'll see lesions in the brainstem, the basal ganglia,

00:58:29.200 sometimes the spinal cord, corresponding to regions of necrosis.

00:58:34.620 Wow. So quickly fatal. And what did you learn? I mean, you have one syndrome, but there are many

00:58:40.740 paths that produce it. Are there common threads to the genetic insults that lead to this awful

00:58:47.720 phenotype?

00:58:48.380 No, great question. That's exactly what we're trying to figure out. So thanks to genetics,

00:58:52.100 we as a community, I mean, have been able to map out genes in the nuclear genome. Some of these

00:58:57.500 are in the empty DNA. But at the end, we get this thing called Lee's syndrome. And we're trying to

00:59:02.260 figure out what exactly is it about the broken mitochondrion that gives rise to this phenotype.

00:59:08.220 And honestly, we don't know what the full answer is right now, but it's a very, very active area of

00:59:12.340 research right now.

00:59:13.760 It might be naive, but just listening to you describe this, I can't help but think,

00:59:18.260 can we learn something about Alzheimer's disease or other forms of neurodegeneration, which I think

00:59:25.940 many people are starting to argue are basically neuronal energy crises. So there are lots of

00:59:33.440 insults, right? You can have an accumulation of toxin. You can have a insulin resistance. Frankly,

00:59:40.180 you can have microvascular disease. All of these things are predisposing people to neurodegeneration.

00:59:46.700 And something that they could all have in common is depletion of energy to the neuron, which would be

00:59:51.580 perhaps the most sensitive cell to an energy reduction. I mean, anybody can think about that

00:59:57.640 for a moment. If you know somebody who's lost their ability to breathe for a period of time,

01:00:03.420 usually the thing we care about the most is their brain, because that's the first thing that

01:00:07.500 you suffer from when you have a hypoxic event. So is that, based on Lee's syndrome, is that the

01:00:16.200 explanation for why you're potentially seeing it disproportionately in the brain versus skeletal

01:00:22.900 muscle? Is it just the sensitivity of the brain to energy withdrawal, or do you think there's

01:00:28.580 something specific about the mitochondria in neurons? Mitochondrial disorders can actually impact almost

01:00:35.380 any organ system. And so Lee's syndrome represents one type of clinical manifestation of mitochondrial

01:00:42.000 disease, but there's another set of disorders that impact the skeletal muscle as well. I think at

01:00:47.960 this point, we don't know why mutations in one subunit of the electron transport chain gives rise to

01:00:55.820 brain disease. A mutation in the neighboring subunit of the same protein complex that's equally

01:01:02.300 evolutionarily conserved will give rise to muscle disease. Wow. There's a massive non-linearity

01:01:09.880 over here that we simply don't understand right now. So if there is one and only one silver lining

01:01:16.640 in these awful diseases, it's that scientists will have no shortage of questions to ask for decades to

01:01:23.080 come. No, I think this is a super active area of research right now, in part fueled by the link

01:01:30.540 between mitochondrial decline and aging. So that's such a complicated problem. You know, trying to

01:01:36.640 understand why a car breaks down after being in service for 25 years, it's hard. Is it some fan

01:01:43.960 belt break? Did the battery stop charging? Did the tire deflate? It's kind of hard to know why an entire

01:01:51.740 car breaks down after 25 years. In these rare Mito disorders, we have 300 different forms of that

01:02:00.080 automobile that was almost broken at birth, if that makes sense. We study them in part because

01:02:05.800 these patients need new treatments and therapies. And so that's enough of a motivation for us. But

01:02:11.500 we also do expect that a subset of them by studying them will inform what's happening in the more common

01:02:16.960 form of aging. Your lab studies oxygen, but not necessarily in the way that most people commonly

01:02:23.360 think about it. Most people, when they think of oxygen and mitochondria, something that comes to mind

01:02:28.720 pretty quickly is ROS. Absolutely. I spoke with a friend of mine, Navdeep Chandel, and we spoke a lot

01:02:33.800 about ROS. And now I've had a great take on it, which was, look, we think of ROS typically only in

01:02:39.780 the negative. And they do lots of negative things. But they may also be a signaling molecule, and

01:02:45.260 therefore an essential thing. Talk to me about the lens through which your group looks at oxygen.

01:02:50.320 So remember a while ago, you asked a very, very astute question. You have these 80 different genes.

01:02:55.240 When they're mutated, they give rise to this thing called Lee syndrome, which is a different type of

01:02:59.720 neurodegeneration. What does that pathology pathogenesis look like? The traditional dogma

01:03:08.060 for mitochondrial pathogenesis is that when the powerhouse of the cell is broken, there's not

01:03:13.500 enough ATP, and there's a power failure. That's the traditional dogma. And without a doubt, there's truth

01:03:19.060 to that in some instances. What we've discovered is that in addition to producing ATP, mitochondria are

01:03:27.600 also consumers of oxygen. Most of the oxygen that you breathe, Peter, is being consumed by your

01:03:37.360 mitochondria. When a patient has a birth defect in the mitochondrion, in addition to not being able to

01:03:44.980 produce sufficient ATP, they also have excess oxygen as well. So oxygen delivery tends to be

01:03:52.280 patent in these patients, but the utilization ends up being poor.

01:03:59.060 And oxygen, just asking for people to understand this, how does oxygen even get to the mitochondria?

01:04:04.400 So we all understand that we're breathing air that has oxygen, and let's even go one step further and

01:04:09.160 just take it for granted, that there's a gradient in the lung that allows oxygen to get into hemoglobin

01:04:15.020 to a red blood cell. Now that you have a fully loaded red blood cell in an arterial that enters a

01:04:20.720 capillary, how many things have to happen for oxygen to get into the mitochondria specifically, not just

01:04:25.940 the cell? Oxygen is not particularly soluble in water or fluids, and so we have an oxygen-carrying

01:04:32.200 protein called hemoglobin that's found in our red blood cells. And so in the lungs, all of the

01:04:37.800 red blood cells in their hemoglobin get loaded with oxygen, and now these red blood cells are

01:04:43.320 delivered to peripheral tissues, and the oxygen gets extracted from the fluid, and as the fluid gets

01:04:50.140 depleted in oxygen, the hemoglobin will basically offload its oxygen. And so there's certain tissues

01:04:56.640 that become extremely, extremely hypoxic. So the oxygen will get extracted by a tissue, largely by

01:05:03.860 diffusion. And the mitochondria is basically consuming most of this oxygen.

01:05:08.540 So a cell that is sitting there doesn't require an active transporter to get oxygen across its

01:05:13.760 outer membrane, so it just diffuses across. That's right.

01:05:17.640 When oxygen enters the cytoplasm, how does it get over to the mitochondria?

01:05:22.300 Depending on the cell type, it either diffuses...

01:05:25.460 How does it know that the mitochondria is the place it needs to go?

01:05:28.440 Well, it's being consumed, so it's a little bit of a sink, basically.

01:05:31.060 Ah, so that's what I want to understand. There are lots of places oxygen could hang out. So what

01:05:36.360 is the force that is drawing it into the mitochondria? Is it simply utilization that

01:05:41.540 basically creates a vacuum?

01:05:42.960 That's right. That's right. Now, in certain tissues like skeletal muscle and heart, we have

01:05:47.060 other oxygen carriers in the tissue. Things like myoglobin are oxygen carriers, so they're

01:05:52.120 almost little buffers of oxygen that are in the tissue.

01:05:54.560 Yeah, they're more like loaders, right?

01:05:56.100 That's right.

01:05:56.280 I think of them as like another storage for oxygen inside, yeah.

01:05:59.940 That's right.

01:06:00.680 But that still has to get off the myoglobin and get sucked into the mitochondria effectively,

01:06:05.860 right?

01:06:06.160 That's right. So it's another little storehouse, if you will. It's another buffer of oxygen so

01:06:10.480 that when you're exercising, for example, you may not want to be oxygen limited. So you have a little

01:06:16.040 extra oxygen in your myoglobin. But basically, your mitochondria is where most of your oxygen is

01:06:21.660 being consumed. So it's a sink. And so that's the reason that we have gradients inside of our cells.

01:06:28.260 It's amazing. I've never really even thought about it this way. But it's sort of interesting

01:06:31.480 to think at how efficiently the body disposes of carbon dioxide. Because as you alluded to earlier,

01:06:37.880 the end of that downhill gradient is a final path of electron acceptance generating H2O and CO2,

01:06:49.680 both of which we managed to largely off-gas. So somehow those things have to exit the mitochondria,

01:06:56.280 weasel out of the cell, cross the gradient, and go back to, well, in the case of CO2,

01:07:01.900 get back to the hemoglobin molecule and get carried back. It's like there's a lot of things going on here.

01:07:06.640 Well, so one of the things that we're discovering by studying these rare diseases, and this happened

01:07:10.660 because of a CRISPR screen that we did a couple of years ago, but what we've now discovered is that

01:07:16.100 one of the consequences of mitochondrial dysfunction is excess unused oxygen.

01:07:23.740 So in other words, if a mitochondrion is failing to do its job, you will be failing to utilize oxygen.

01:07:34.040 Therefore, you would see an excess accumulation of oxygen.

01:07:37.460 That's right. And we believe, this is our hypothesis now, it's that some of that excess

01:07:42.940 unused oxygen is what is contributing to the pathology that we see in some of these rare

01:07:48.260 diseases. It's a very, very different type of an idea. It's not all about the ATP. It's about excess

01:07:54.680 unused oxygen. And we're not necessarily invoking reactive oxygen.

01:07:59.440 Yeah, I was just about to say, is it through free radicals or is it actually oxygen to oxygen,

01:08:07.280 no extra electron oxygen?

01:08:09.740 This is dioxygen toxicity that we're talking about. The way that I like to think about it is that if

01:08:15.320 you have an automobile that's outside and it's rusting, it's rusting in part because the oxygen

01:08:21.080 is directly oxidizing iron, you may produce a radical, but that's not the phenomenon.

01:08:28.820 So the car outside is not rusting because of too much superoxide or hydrogen peroxide.

01:08:33.780 It's direct oxidation of iron centers by oxygen. And so one of our hypotheses is that enzymes

01:08:41.300 are tuned to operate within a particular oxygen range. And when the mitochondrion is not functioning

01:08:49.980 properly, the oxygen levels rise, that excess dioxygen can now oxidize enzymes that will damage

01:08:58.880 them as a consequence. So it's a very, very different way of thinking about mitochondrial

01:09:03.680 pathogenesis.

01:09:04.520 Now, why doesn't the body correct for that and note that, well, the oxygen in the mitochondria is 0.56

01:09:10.660 not being utilized as quickly as my evolutionary prediction would allow, but therefore there's

01:09:17.620 less gradient pressure. So I'm going to off-gas, I'm going to offload less oxygen to the cell

01:09:22.520 with subsequent trips through. In other words, you almost think this would have been corrected

01:09:27.060 in a few milliseconds.

01:09:28.960 Well, I think in a healthy human, that's exactly right. We have pretty solid matching of oxygen

01:09:33.940 delivery and oxygen utilization. But we're talking about patients that have broken mitochondria.

01:09:40.020 So this is one of the things that we're actually trying to investigate. One of the ways that

01:09:43.380 Ron Haller at the University of Texas Southwestern Medical Center has proposed diagnosing patients

01:09:50.300 with mitochondrial diseases to put them on a treadmill, measure their oxygen extraction,

01:09:56.500 and patients with mitochondrial myopathies will often have high venous oxygen.

01:10:01.720 Let's explain that to people because we're going to talk about VO2 max and you alluded to it. So

01:10:05.900 most people associate that test with sort of peak athletic performance, but let's talk about what it

01:10:11.520 looks like. If I came into your lab and you were doing this, you'd hook me up to a device. You'd put

01:10:15.520 me on a treadmill. You'd make me, you know, have to do some work to stress the system. You'd plug my nose

01:10:21.320 and put a little miserable device in my mouth, basically creating a seal that would prevent me from

01:10:27.440 being able to get oxygen or dispose of carbon dioxide in any place other than the gas chamber that is attached

01:10:35.180 to the tube going into my mouth. You'd put me at the, you know, you'd ramp up the speed of the treadmill

01:10:40.680 and you would be measuring essentially two things, the amount of oxygen you're putting in. And if it's room

01:10:47.200 air, we sort of know what that is, but more importantly, the concentration of oxygen coming out. And that

01:10:52.780 difference is what you're talking about. It's that extraction. Now, what happens in a normal person

01:10:59.080 when you do this? So Peter, we actually don't do these types of studies in humans. And so, um,

01:11:04.200 but when one does this, yeah, what would normally happen is a, what would you be measuring as a normal

01:11:09.180 person works harder and harder in the difference between provided oxygen and returned oxygen?

01:11:16.040 So again, we don't do these types of studies, but in these types of physiological studies that

01:11:19.780 people like Ron Haller, other cardiologists will perform, they'll look at cardiac output

01:11:25.260 as well as the oxygen tensions on the arterial and venous sides. And so by looking at all of these

01:11:31.680 numbers, you can actually figure out quantitatively the number of O2 molecules being delivered as well

01:11:36.960 as those that are being extracted versus those that are being returned to the venous system.

01:11:41.080 So you sort of do a, you can do a mass balance on oxygen effectively.

01:11:44.120 That's exactly right. That's exactly right. And as it turns out in these patients with inherited

01:11:48.860 mitochondrial disease, for some reason, the cardiac output is high. The extraction is low

01:11:55.140 in a healthy individual. Usually the homeostasis is such that you're not going to be delivering more

01:12:01.980 oxygen than you really need. So that homeostasis is somehow broken in these patients with, with.

01:12:07.980 So that patient would actually have a very high lactate level as well, because if they're able to

01:12:13.920 produce the cardiac output, but they're not doing it with oxidative phosphorylation, they're using

01:12:20.340 their escape valve. So they're going to disproportionately have high lactate levels relative

01:12:25.860 to a healthy individual. That's right. So Ron Haller would actually argue that a high lactate in

01:12:31.560 combination with a high venous oxygen is suggestive of a mitochondrial myopathy.

01:12:36.760 There's a researcher at the university of Colorado who is looking into this very phenomenon as an

01:12:42.680 early indicator of type two diabetes. And it's really fascinating. I'm going to be going out

01:12:48.060 there to spend some time with them this summer. They've made the observation that in the early

01:12:53.220 stages of insulin resistance, the muscle, the skeletal muscle in particular becomes inefficient

01:12:59.600 at oxfos. So you start to see even at baseline, even a person sitting at rest, their lactate levels

01:13:07.980 could be twice as high as that of a fit individual. And of course he came to this through the thinking

01:13:14.840 that if you want to understand that disease of the mitochondria, look at the exact opposite,

01:13:21.740 look at the fittest people in the world, look at, you know, the endurance athletes and ask the question,

01:13:26.680 what do their mitochondria do well? And then what becomes the polar opposite of that in disease?

01:13:32.580 Now, what you're describing is the most extreme example I've ever heard of this, right?

01:13:37.320 Absolutely. What you're describing is actually very consistent with a series of papers that were

01:13:42.440 published probably in 2003 and 2004, including by myself in collaboration with Leif Group and David

01:13:49.680 Eltshuler, Ron Kahn, who used to be the president of the Jocelyn, as well as Jerry Shulman from Yale.

01:13:56.100 So in Sri Kumar Nair at Mayo Clinic, all of us had sort of papers at the same time.

01:14:03.260 David Eltshuler's here, right?

01:14:04.600 David Eltshuler used to be here. He's now at Vertex, but he was one of the founders of the

01:14:09.320 Broad Institute. But all of us had papers published at about the same time that showed that if you take

01:14:14.680 skeletal muscle from pre-diabetics, healthy individuals that have a family history of diabetes,

01:14:21.780 but are still healthy, they'll all have a reduced number of mitochondria. The expression of those

01:14:29.300 1,000 genes required for mitochondria, it's just a little bit lower. If you look at any one gene,

01:14:36.220 it's not significant. But if you look at the entire pathway, the entire pathway is down. The VO2 max is

01:14:42.980 down. So there's a series of papers back in 2003 and 2004 that led to the mitochondrial hypothesis for

01:14:51.060 type 2 diabetes. The big question still to this date, almost 15, 16 years later, is,

01:14:58.080 is that actually causal for the diabetes? Or is it just an early epiphenomenon? There's something

01:15:05.520 else, X, that lies upstream of the amount of mitochondria and independently the predisposition

01:15:13.500 to diabetes? Or is that a part of the causal path? And that's still unanswered to this date.

01:15:18.680 And there's nothing in a Mendelian randomization that can answer that question?

01:15:23.040 So as you probably know, these types of methods are only now becoming possible. So we're trying

01:15:29.640 to work on those types of projects right now. You need large numbers of heavily genotyped

01:15:34.320 individuals. You also need to have nice biomarkers or proxies for mitochondrial function. So those are

01:15:41.180 some of the types of things we're trying now. Because it comes up from time to time, do you mind

01:15:45.340 just explaining how Mendelian randomization works and why it's so powerful? Again, this is not an area of

01:15:51.180 my expertise. Others are much more expert at it than I am. But the analogy that they'll often use

01:15:56.920 in describing a Mendelian randomization is a little bit like a drug trial. So in a drug trial,

01:16:02.680 what you'll do is you'll take individuals and you'll randomize them either to a drug arm or a placebo arm,

01:16:09.240 and then you'll look for an outcome. There's an intervention, which is the assignment of the drug

01:16:14.080 or the placebo. And then you can actually check to see whether the outcome is correlated to that

01:16:18.820 particular intervention. And if you see an effect, now that effect is attributable to

01:16:24.000 that intervention. Because the decision to give the intervention versus the placebo was done

01:16:28.880 randomly. If you didn't do it randomly, you can't make that assertion. That's right. That's right.

01:16:34.080 Now, in a Mendelian randomization experiment, the idea is that there's a randomization that took place

01:16:39.580 at birth. And so if you look at something like LDL levels, for example, an important question is,

01:16:46.340 are LDL levels causal for heart attack? Or are they simply correlative for heart attack?

01:16:54.120 So thanks to genetics, we have a lot of genetic variants that can help explain the population

01:17:00.640 variation in LDL levels. Once you have a good, solid genetic instrument, now what we can do is we can

01:17:08.060 take a very large number of individuals, and we can actually draw a bell curve for what their

01:17:14.660 genetic LDL levels must look like. We haven't measured LDL in them. But from the genetics,

01:17:20.880 you can actually come up with a bell curve in the population. And now you can say, let's take this

01:17:26.840 tail, they've been randomized to a high genetic level of LDL versus this tail, a low genetic level of LDL.

01:17:34.940 Now let's see what the outcomes are. And based on the statistics, based on the patterns of correlation,

01:17:40.400 you can then make the inference that the LDL is either causal for heart attack, or there's no

01:17:46.440 evidence for causality.

01:17:48.180 I am a, not a, an MR, Mendelian randomization jock, but my recollection in reading some of the

01:17:54.880 papers about this is the one place that either as an investigator or a consumer of this science,

01:18:00.580 where you have to force yourself to look closely, is you can be fooled if the randomization of genes,

01:18:10.320 meaning if the genes that you're looking at can also control something else that is related to

01:18:15.360 the disease that could have a counterbalancing effect. Is that correct? Is that like, that's,

01:18:19.040 that's one place where one has to be quite cautious.

01:18:21.920 I think there's multiple places where, I mean, as soon as you have feedback loops,

01:18:26.420 you can get phenomenon called reverse causation. You can have your genetic instrument also has to

01:18:31.940 be really, really good. So again, this is not my area of expertise. So it has to be pursued with,

01:18:38.180 with care for sure. So there's probably going to be a fair number of false positive results for all

01:18:43.600 true positive results from Mendelian randomization studies.

01:18:47.540 Yeah, but it is interesting. I would love to see the MR when it's done for this particular question,

01:18:53.520 because again, you know, the implications are significant, both from the standpoint of

01:18:59.240 preventing chronic disease or risk reduction in chronic disease. But also as we try to approach

01:19:06.140 the question, the way that you phrased it a moment ago, which I really liked, which is

01:19:10.700 everybody kind of has the gestalt of that car that just sort of breaks down. And sometimes it's

01:19:19.000 attributable to a catastrophic failure. Sometimes you blow the head gasket in the car and it's that

01:19:25.120 it dies at that moment. And it just becomes economically not feasible to put a new head gasket

01:19:30.480 on. Other times it just gets harder and harder to start until it's just not worth driving anymore.

01:19:38.440 So taking this to humans, how important is mitochondrial health to that process? In other words,

01:19:47.640 does it become more often than not one of the drivers of this feeling, this lack of robustness,

01:19:55.560 this lack of stability within the organism? And it seems to me that it's, it's on the short list

01:20:02.940 of candidates, right? I mean, when this process goes awry, you are interrupting one of the more

01:20:08.880 fundamental systems in human biology that affects almost every cell in the body, right?

01:20:13.500 Absolutely. I think it kind of cuts both ways because I think when other processes in the cell

01:20:19.100 fail, I think the mitochondrion is at risk for also failing. So it's a highly reactive

01:20:25.840 organelle as well, if that makes sense. It's going to be a tough question to fully answer. And this is

01:20:33.060 why we like these rare genetic disorders, because they teach us about the pathology pathogenesis of the

01:20:39.480 organelle in a few defined modes. And the big question for our field is, which of those rare

01:20:44.720 diseases, which of those rare forms of pathogenesis bears any relevance to the common form of wear and

01:20:52.900 tear aging, right? So there may be a small subset of rare diseases where the fan belt is broken at birth,

01:20:59.720 right? But it may be the case that the fan belt, as it turns out, is so resilient, you never have to

01:21:05.480 change it. And it's rarely the cause of a car breaking down. But it could be the case that

01:21:11.360 it's actually the spark plug, right? It could be the case that there are certain birth defects where

01:21:17.480 the spark plug is actually not working at birth. And as it turns out, with wear and tear aging,

01:21:23.660 spark plugs have to be replaced. And so we have 300 different forms of monogenic mitochondrial disease.

01:21:30.160 You have 300 single point of failures to study. And that gives you a beautiful picture of what it

01:21:38.700 looks like. Some of these are going to bear no relevance to the common form of aging. That's

01:21:44.800 just going to be a fact. But the hope is that a small number of these will bear some relevance to

01:21:49.880 the common form of aging. And just to be clear, we care about these diseases just because

01:21:54.240 these are terrible diseases. And we need therapies for them. And so that alone is motivation for our

01:22:00.380 work in this area. But it's also our hope that studying some of them will provide insights into

01:22:05.940 the common form of aging as well. You talked earlier about signatures of time. And another one of these

01:22:12.360 signatures of time is inflammation within muscle cells. There was a paper that came out about a year

01:22:17.360 ago that used sting to basically block the ability of a cell to sense breakdown in mitochondrial DNA.

01:22:26.640 Are you familiar with this work? I'm a little bit familiar with the sting pathway.

01:22:30.920 My understanding, it's been about a year since I read this, maybe a bit less, was that serially,

01:22:37.340 if you study muscles as they age, you see more and more inflammation. So the question is,

01:22:41.420 what is drawing inflammatory cells into muscle as we age? And the hypothesis was, going back to

01:22:49.260 something you talked about earlier, the DNA of mitochondria is bacterial in origin. And therefore,

01:22:55.720 if you have a loose fragment of nuclear DNA in the cell, it shouldn't be especially immunogenic.

01:23:01.960 But if you have a loose fragment of bacterial DNA in a cell, that should actually be quite immunogenic.

01:23:08.000 Our immune systems would think of that as foreign. So the hypothesis was, what if the increase in 0.98

01:23:15.160 inflammation we see is due to greater and greater mitochondrial damage that is leading to more and

01:23:21.600 more exposure of mitochondrial DNA? And I believe to test that, it used sting. And remind me, I think

01:23:29.000 sting actually blocks the ability of the cell to sense the mitochondrial DNA. Is that correct?

01:23:33.980 Not just mitochondrial DNA, but I think it's a general nucleic acid sensor. And so there could

01:23:39.260 be different sources. I think- It could block the sensing of DNA, period.

01:23:42.780 I think ordinarily it's designed to sense the DNA of incoming viruses or incoming pathogens. But you

01:23:49.040 could imagine that if a mitochondria ruptures, that same DNA nucleic acids can also be sensed. And

01:23:56.260 you're right, because the mitochondria used to be a bacteria, a lot of the components inside of our

01:24:01.780 mitochondria resemble that of modern day bacteria and can be very immunogenic. I don't know if you've

01:24:08.280 followed, there's a study that came out of the Beth Israel Deaconess a few years ago on sterile crush

01:24:13.280 injury. And as a former surgeon, you'd probably appreciate it. But the observation was that when

01:24:19.140 you had an enclosed crush injury, so the skin has not been violated, there's a massive inflammatory

01:24:25.700 response. So this investigator was actually trying to figure out what the inflammatory nidus was,

01:24:31.440 did a lot of fractionation, and ultimately figured out that there was some mitochondrial-derived

01:24:36.860 molecules. And I believe that they found that the mitochondrial DNA, as well as something called

01:24:42.340 formulated peptides, are highly inflammatory. So remember those 13 proteins that are made by the

01:24:49.200 mtDNA. The translation of those 13 proteins resembles the translation of bacterial proteins.

01:24:56.460 And bacterial protein translation doesn't begin with a methionyl tyranny, it begins with a formal

01:25:02.140 methionyl tyranny. Remember this F-met peptide? Yeah, yeah, yeah. This is, my God, I feel like I'm

01:25:07.740 back in medical school. Can you, just for no other reason than for me to remember that, let's go through

01:25:12.500 that. So your DNA makes your messenger RNA. Your messenger RNA then moves over to be translated and

01:25:18.300 to actually have the amino acids put in place to resemble the code that's being said. So talk about

01:25:24.180 where the meth terminal end shows up there. In UNI, in eukaryotes, protein translation begins with a

01:25:30.660 methionine residue. In bacteria, it's not a simple methionine, but it's a modified molecule. It's called

01:25:37.760 formal methionine. And first order approximation, it's almost a signature of a bacterial derived

01:25:43.320 protein. And we actually have receptors that are designed to detect formulated peptides. It's a sign of

01:25:50.160 some sort of an infection. Now, as it turns out, our mitochondria have bacterial origins, as we have

01:25:56.160 discussed earlier. So these 13 proteins, they still begin with the formal methionine. Presumably, they can

01:26:03.980 get away with it under normal circumstances because they reside within the mitochondria. They're protected

01:26:09.220 and shielded from the immune system. That's right. So in certain injuries, right, this can actually

01:26:14.120 escape and it can basically activate the same inflammatory pathways that we have that are ordinarily designed

01:26:19.920 to detect incoming pathogens. I can't wait to actually find that paper. So that was, who is the author on that?

01:26:25.720 I forget his name, but he's from the Beth Israel Deaconess. He's a surgeon.

01:26:29.880 Oh, two, oh, three ish. No, no, no. This is probably within the last five years ago or so.

01:26:34.880 Well, we'll, we'll find that paper and link to it. But basically that paper and the paper I was

01:26:39.060 referring to from about a year ago, which I think was in science, both point to a similar conclusion,

01:26:45.080 which is you can have a profound inflammatory response by simply damaging the mitochondria.

01:26:51.760 And both of them would point to consistent plausible explanations, which are the body confusing

01:26:59.600 the contents for bacterial contents. I think it's a very reasonable hypothesis,

01:27:05.260 which then begs the question, if we believe that the aging inflammation phenotype is not beneficial,

01:27:13.360 how do we prevent mitochondrial breakdown as we age? I mean, this becomes one of the key aging

01:27:20.960 questions, right? If you believe that, and again, I don't know what the best analogy is, but

01:27:26.160 spark plug failure plays a role more often than not in the overall picture of decline.

01:27:34.080 The longer you can protect those things, preserve those things, the better.

01:27:38.320 Another way of evaluating the causality question is if we had a drug that could somehow rejuvenate

01:27:45.340 mitochondria, then you could ask the question, does directly intervening on this organelle retard the 0.97

01:27:52.900 aging process? And unfortunately, as of right now, we don't have that type of a magic bullet. But

01:27:58.560 exercise is one of the best ways of turning over bad mitochondria and inducing the biogenesis of

01:28:08.240 good mitochondria. But the challenge, of course, is exercise does lots of things.

01:28:12.800 Yeah. Now that said, we always, you know, this is one of the differences between, I think,

01:28:19.060 being a scientist and a doctor. When you're wearing your doctor hat, you just have to know what to do

01:28:23.980 for that patient in that moment. And it's a luxury to know how and why it works, right? So when we think

01:28:31.580 about the importance of exercise, I've always found it ironic that I probably classify or qualify as an

01:28:38.940 exercise addict, like in the true unhealthy sense of the word, you know, I probably meet all the

01:28:44.240 criteria of addiction and all of that stuff. But up until recently, I don't think I really

01:28:50.280 appreciated the value of exercise. I think it was honestly just something I did out of my neurotic

01:28:57.560 pathology. But I actually, I think if asked, would have said that nutrition played a much greater

01:29:04.400 role in health, sleep played a much greater role in health. And then exercise, you know, I mean,

01:29:10.420 it's, it's great, but you know, if you, I'd rather you eat well than exercise a lot or something to

01:29:15.080 that effect. I'm certainly revisiting that. And of course, I also find it silly to do these sort of

01:29:20.400 zero sum games, like it has to be one or the other, you know, presumably doing all of these things well

01:29:24.840 is the optimal strategy. But the deeper I look at exercise, and I'd love to know your framework for this,

01:29:32.500 because I'm still trying to create one. I'm putting exercise very loosely into three buckets,

01:29:37.940 strength training, very specific aerobic training. So this would be maximum mitochondrial output with

01:29:44.980 minimal generation of lactate, and then anaerobic training where you are basically demanding ATP at such

01:29:52.480 a rate that you are really running through that lactate production pathway. Do you think that's a

01:29:57.740 reasonable framework to, of buckets of exercise? Do you, do you divide them even more granularly as

01:30:02.440 you think about it mechanistically? We're really not exercise physiologists. I don't think I can

01:30:08.260 comment in any particular way. Just personally, do you think of it in a way like that? I think that's

01:30:12.360 very reasonable. You know, just in hearing you talk, I mean, again, we're not aging researchers, but

01:30:16.700 so I'll ask you a question if that's okay. Has anyone actually, you know, people have given

01:30:22.000 metformin to mice, they've given rapamice into mice, but has anyone given mice those three flavors

01:30:28.380 of exercise to determine what the impact is on longevity? Yeah, I believe those experiments have

01:30:34.920 been done, and I believe all of them show benefit. I'd have to go back and look at the literature

01:30:40.920 in mice. In fact, I'm in the process of just starting to work on that chapter in a book that I'm writing.

01:30:49.620 The problem is I generally bias against heavy mice literature, but you at least have the advantage

01:30:56.320 of control. So the short answer is definitely each of those as a medication, right? If you think of each

01:31:02.500 of those as a pill, each of those produce a longevity phenotype. Where it gets challenging, I think, in humans

01:31:09.600 is, well, I think there are so many ways to die when you get old that, for example, accidental death

01:31:19.000 would rank in the top five causes of death for people over the age of 60. Now the type of accident can change

01:31:28.880 around, but by the time you're in your eighth or ninth decade, falling becomes such a significant cause of

01:31:34.720 death due to the frailty of the individual that some of that exercise, for example, strength training almost

01:31:44.120 assuredly offers protection from that type of death. So the question is, I think it's a little

01:31:50.820 hard to tease out how much of that benefit is in the cardiometabolic side versus otherwise. The other

01:31:57.220 thing that's really challenging in studying humans is we don't have really good prospective studies in

01:32:03.200 anything that resembles a longevity phenotype. So you are now stuck using something I think I recently

01:32:09.580 described as the most servile trash ever shat into civilization, which is epidemiologic questionnaires 0.78

01:32:17.600 to try to impute based on, you know, you telling me how you've exercised over the past 10 years, 0.98

01:32:25.700 how that's going to predict your longevity phenotype. And again, the problem there is

01:32:30.520 the dose matters, the specificity, the quantity, the quality, these things matter. And they're very

01:32:37.740 difficult to tease out from these retrospective views. So I think the evidence is very compelling

01:32:45.000 that exercise matters. And that's maybe less the question I'm interested in. I think what I'd love

01:32:52.440 to gain insights into, and we may have to rely on non-human models, is just as we now can tailor a drug

01:33:01.680 to do something very specific. Can we tailor our exercise to be as optimal as possible? So if you

01:33:09.260 took an individual who said, Peter, look, I'm willing to exercise five hours a week, or I'm willing to

01:33:14.540 exercise 10 hours a week, but I'm not going to be a professional athlete. How do I take those five

01:33:20.240 hours a week or 10 hours a week or whatever it is and make the best use of it to impact all causes of

01:33:27.500 mortality? Meaning reduction of the risk of atherosclerotic disease, cancer, neoplasm,

01:33:34.400 neurodegenerative disease, and accidental death from strengthening the exoskeleton.

01:33:40.100 So that's clinically the question I'm most interested in as it pertains to exercise.

01:33:44.820 But I'm convinced that at the center of that question is understanding the role of exercise

01:33:51.300 and mitochondrial health. I think this is a very important piece of the puzzle and certainly much

01:33:56.840 more important than I appreciated even two years ago. I think what you describe about these

01:34:02.340 age-appropriate or age-acknowledged declines in VO2 max, mitochondrial density, mitochondrial

01:34:11.340 efficiency, venus O2 concentration, I think there's something really important there. And even if

01:34:18.600 exercise is affecting something upstream that is affecting that, at least we have a great proxy

01:34:23.880 through which to measure. I think there's going to end up being a lot of really interesting

01:34:28.100 nonlinear dynamics of mitochondria as a function of age, as a function of exercise. There's a few

01:34:36.120 vignettes I'll share. As you probably know, if you don't use your skeletal muscle, you lose lean muscle

01:34:41.660 mass, you lose your mitochondria very quickly. How quickly? You have measurable defects in the VO2 max

01:34:48.820 after 10 days of hospitalized bed rest. And to recover the VO2 max that you lose in 10 days,

01:34:54.800 it takes about six weeks or so. Oh my God. Yeah. I knew it was bad. I didn't know it was that bad.

01:34:59.020 So there's quite a bit of hysteresis over here, right? Quite a bit of hysteresis. So it's going to

01:35:03.480 be complex and nonlinear. The programs that turn on mitochondria during exercise, they're really

01:35:09.860 elaborate. And the idea of actually replacing it in a pill may end up being kind of naive. I think that

01:35:15.240 exercise does so many things simultaneously. It's like 17 different inputs into the system.

01:35:23.120 And it may be the case that it's only if those 17 inputs are provided with the right dynamics

01:35:28.260 and the right off rates that you get properly functioning more mitochondria. In certain disease

01:35:34.940 states, some of the muscle disorders that I study, the ragged red fiber that you may remember from

01:35:41.320 your board exams, the ragged red fiber represents an accumulation of poorly functioning mitochondria.

01:35:49.280 So I think that if you try to bottle up just two of these factors or three of these factors,

01:35:53.760 we may be able to produce more malfunctioning mitochondria. But it could be the case that

01:35:58.020 we've evolved to require 17 inputs provided at the right time and place in order to get proper

01:36:04.140 mitobiogenesis. It's a really, really smart program, this PTC1-alpha program, because

01:36:09.500 it simultaneously turns on mitochondrial biogenesis while also turning on some of the

01:36:14.920 autophagy programs. And so you're actually turning over your bad mitochondria while you're

01:36:20.440 turning on your good mitochondria simultaneously. And that's what happens with exercise.

01:36:24.540 Well, you read my mind. And I don't know if you could read my little notes I'm taking over here,

01:36:29.060 because as we're talking, I'm making little notes of things I want to ask you. And that's exactly

01:36:32.380 kind of where I want it to go, which was, let's talk about what autophagy means in the context of

01:36:38.740 mitochondria. So people who listen to this podcast know that I'm a big fan of fasting,

01:36:43.960 periodic fasting, because even though we don't have great ways to measure autophagy clinically,

01:36:49.820 I think we have pretty good evidence that periods of really strict fasting, meaning exclusively consuming

01:36:55.380 water for some period of time, my hypothesis is three to seven days, produces meaningful autophagy.

01:37:02.940 But how does exercise impact that based on what you've seen in the mitochondria?

01:37:08.120 So I don't know too much about fasting, but when you do have proper exercise regimes, what we observe

01:37:13.700 is that there are transcriptional programs with multiple inputs, some of which we discussed earlier,

01:37:19.340 but those are probably not sufficient, that will basically turn on all 1,000 of those proteins to

01:37:24.740 produce more mitochondria. But that same program is also saying, hey, let's turn over some of the

01:37:32.160 previously produced mitochondria. So it's a very, very smart system. It's not going to just produce

01:37:36.920 more good mitochondria in the presence of bad mites. It'll actually cleanse the system as well.

01:37:42.800 And remind me what you said. I know you already answered this. I apologize. When you look at cells

01:37:47.020 that are not turning over quickly, so myocytes, neurons, what did you say was the approximate

01:37:52.180 turnover of mitochondria? Probably a few days. A few days. Unbelievable. So it's just, this is an

01:37:59.280 unbelievable amount of work to create the new and systematically and selectively discard and recycle

01:38:06.520 the old. That's right. That's right. And the signals, given that you like exercise, I'll tell you

01:38:11.060 one study that I thought was really provocative, and maybe you already know it, but it came from,

01:38:15.460 I think, Michael Ristow and Ron Kahn about 10 to 12 years ago in PNAS. Do you know this study?

01:38:23.500 I don't. I don't think I do, but keep going.

01:38:25.420 It was a human study.

01:38:26.260 Okay.

01:38:27.080 It was a two-by-two matrix. They randomized humans either to exercise or no exercise and

01:38:35.300 antioxidants or no antioxidants.

01:38:38.480 Ah, yes. Okay. I know this study, but please keep going. No, this is great.

01:38:41.240 So which of the four quadrants do you think is best?

01:38:45.860 Well, I know the answer, so I want you to, yeah, you keep going, yeah.

01:38:48.200 Yeah, so most people would probably guess.

01:38:49.820 Most people would say exercise with antioxidants must be the optimal health.

01:38:55.260 Absolutely. And what this study showed that was a little bit counterintuitive is that

01:39:00.100 antioxidants on top of exercise almost prevents or erases some of the beneficial

01:39:05.380 effects of exercise. And the authors concluded that things like reactive oxygen species are

01:39:14.080 probably playing an important signaling role as well that helps in the adaptation. You need

01:39:20.080 some of those sparks in order to turn on new programs that are net beneficial. So if you erase

01:39:25.440 those sparks, you actually prevent the full benefit of exercise. So just another reason why the entire

01:39:31.360 system is so complicated. I mean, I think investigating exercise, and again, we don't do that. That's not

01:39:36.280 a core scientific focus of our laboratory, but so many diseases, ultimately, their risk is reduced by

01:39:43.480 exercise. So studying it as it should be a very important objective for all of us.

01:39:48.240 It's so interesting because that's a, that is a great example. I'm glad you brought that up.

01:39:51.340 And Nav and I, though we didn't talk about that study, we talked a lot about this issue of blocking

01:39:57.340 ROS and how if one has cancer, for example, the evidence is becoming pretty clear that the last

01:40:05.300 thing you want to do in a cancer patient is give them an antioxidant as sort of anti sort of dogmatic

01:40:11.800 as that would seem because the ROS actually play an important role in selectively targeting a cancer

01:40:16.080 cell cell versus a non-cancer cell. Listening to these discussions makes you almost wonder how

01:40:22.120 in the world does any drug show up with a benefit in longevity? It's almost a miracle that rapamycin

01:40:30.000 can so ubiquitously across so many species extend life when, as you point out, most of the things that do

01:40:38.900 the heavy lifting in longevity have 17 prongs that can't be replicated by a single molecule. I mean,

01:40:47.360 it just, it, it seems impossible. The one that has me very interested right now, and I, again,

01:40:52.980 I don't know how much you've studied this. My guess is even if you haven't, just your peripheral

01:40:57.900 knowledge will exceed that of anybody's is metformin. So again, I think most people listening to this

01:41:05.180 podcast know a lot about it. I had a interview with Nir Barzilai, who I'm also having lunch with

01:41:09.900 today. And Nir's certainly one of the world's experts on this topic. So we had a great discussion

01:41:14.040 of all of the benefits of metformin. Don't think it's really disputable how big those benefits are

01:41:20.000 in people who have diabetes. I think that is becoming very clear. And then by extension in people who are

01:41:26.480 insulin resistant. What I think is not entirely clear, and I think is the purpose of what Nir is hoping

01:41:33.320 to study with TAME is if you took a non-diabetic, non-insulin resistant individual and gave them

01:41:41.140 metformin, will you enhance their longevity phenotype? And the one area that I'm most interested in this

01:41:48.700 question is what is the impact of metformin on the ability of exercise to improve the phenotype? And

01:41:55.720 that's something that just on a personal level, I've been experimenting with a lot. So doing a lot of

01:42:01.060 lactate testing on myself with and without metformin and using lactate as a proxy for mitochondrial

01:42:07.460 function. So, you know, we were talking about this a little bit before, but just for the listeners of

01:42:12.160 what I do is take a resting lactate level. I shouldn't be using any more ATP that I'm using at

01:42:16.940 this moment. What is my level of lactate? Then on a bicycle that allows me to control the wattage

01:42:23.760 to the nearest watt, basically move in five or 10 watt increments slowly, you know, spend 10 minutes at

01:42:31.820 this wattage, go up by five for 10 minutes, go up by five for 10 minutes and keep measuring lactate

01:42:38.600 levels. And you generate a performance curve, an LPC, a lactate performance curve. And you do this with

01:42:44.440 and without metformin, you see a difference. The question is, does that difference matter clinically?

01:42:50.320 And is it possible that metformin is actually not helping in the context of exercise?

01:42:56.320 Are you seeing that in the presence of metformin, if you are exercising, you're producing less lactate?

01:43:01.540 I'm seeing more lactate in the presence of metformin. Now, again, this is an N of one

01:43:06.540 study on myself, but it makes sense that you could, I mean, that's a plausible.

01:43:13.100 How are you measuring your lactate?

01:43:14.500 It's using blood in the finger. Yeah.

01:43:17.380 We should do metabolomics on you. I mean, with our new instrumentation, we can measure not just

01:43:21.880 lactate, but literally hundreds of metabolites. It gives a little bit more of a comprehensive

01:43:27.120 snapshot of all of your metabolic pathways.

01:43:30.240 Could we get an IRB to do that easily?

01:43:32.100 Let's do it.

01:43:33.120 Let's do it.

01:43:33.840 I'm all in. So we could do it. We could do an on metformin, off metformin snapshot because,

01:43:39.920 so here's my crude thinking on this is if metformin is inhibiting complex one,

01:43:45.280 it wouldn't be beyond the realm of possibility that the body might preferentially not shuttle

01:43:53.640 pyruvate into the mitochondria. I mean, it's still doing so to a great extent,

01:43:57.880 but if it's disproportionately now keeping pyruvate outside and turning it into lactate,

01:44:03.740 that could drive up lactate levels. The thing that surprised me the most

01:44:06.820 is how high my resting lactate levels seem to be. I mean, I remember before I started taking

01:44:13.520 metformin, you would barely check a resting lactate level, but it was usually below one

01:44:18.200 millimolar. Now my resting lactate level on metformin is typically between one and two

01:44:24.000 millimolar. It's about two X. And I'm doing this in as painful and, but hopefully valid a way as one

01:44:32.180 can do it, which is I'm using two separate meters checked in duplicate on the third drop of blood.

01:44:37.640 Like I'm trying to be as systematic as possible. Anyone listening to this who wants to do this,

01:44:42.740 I just want to warn you in advance before you get started. Lactate meters are upsettingly expensive

01:44:47.660 and the strips are the racket. I mean, they'll sell you the device for 300 bucks and each strip costs

01:44:53.880 you about $5. So every time I do one of these dumb tests, which I typically do about once a week,

01:44:59.320 it's like dinner and a movie for five people, but it'd be amazing to see what a broader sequence

01:45:05.360 of metabolomics looks like to understand, is there something that's happening? And by the way,

01:45:09.160 then the next question is do that in somebody who has diabetes and see if you see an improvement

01:45:13.980 or a reduction in performance. You measured your fasting, resting lactate before you started

01:45:21.660 metformin. Do you know if it went up after you started metformin and then it went back down again or?

01:45:28.320 Well, so that's an interesting question. So before I started taking metformin,

01:45:33.180 I would do lactate testing, but I was interested in a different question. So it's not an apples to

01:45:37.660 apples comparison. We always had a resting lactate just because you wanted to basically calibrate

01:45:42.220 the machine and make sure your machine was working. But generally when we did what the,

01:45:46.860 what was called this LPC, the lactate performance curve, it was mostly geared towards identifying a

01:45:52.820 different position on the curve, which is very crudely done, but you do a series of efforts at

01:45:59.220 different power outputs or in a swimming pool at different speeds or on a track, different running

01:46:04.080 velocities. And you'll notice that the curve is nonlinear. So it starts like this, I'm sort of

01:46:09.420 drawing, it goes flat, and then it starts to asymptote, it starts to shoot up very quickly towards a

01:46:14.420 vertical asymptote. That can be approximated by two linear curves. And the intersection of that curve

01:46:21.020 is generally a person's lactate threshold. And that's different. That's usually a higher number

01:46:25.980 than two millimolar. Let's just say to make the math easy, that usually is in about the four millimolar

01:46:31.920 range. And it's that point where that corresponds to on the x-axis, that output is generally about the

01:46:38.800 fastest velocity or output a person can hold for a certain type of race that we're interested in

01:46:45.640 studying. So I have infinite numbers of those data for myself back in the, you know, days long before

01:46:52.480 I took metformin. But it was undeniable how low my lactate level started. So I at least have that data

01:46:59.860 point. I think that's really interesting. We published a paper in PNAS a couple of years ago where we placed

01:47:05.220 either healthy individuals or patients with mitochondrial disease on a treadmill, did a 10-minute

01:47:11.340 exercise test, and then drew their blood at rest, peak exercise, and post-recovery just to look at

01:47:17.460 the metabolome response to exercise. So of course we get lactate. And the mitopatients, some of whom

01:47:23.360 have complex one deficiency, not because of metformin, because of a genetic deficiency, they begin with a

01:47:29.240 high resting lactate. And there's a parallel rise in their lactate that parallels what happens to a

01:47:34.940 healthy human. And it stays high parallel with the healthy individual post-recovery.

01:47:40.100 Do you remember offhand how high their resting lactates were relative to the non-insulted? 0.97

01:47:46.960 It was single-digit millimolar. It wasn't sky high. I want to say something like 2, 3, 4 millimolar.

01:47:53.520 But we should look that up just to confirm. But it'd be interesting to see, this is purely science now

01:47:59.100 talking to you, is whether we could repeat that exact same study, not with genetic complex one deficiency,

01:48:04.680 but with metformin on board. You know, a lot of people ask me about metformin and aging. And again,

01:48:11.520 we don't do aging, real aging research in our lab. We hope to be able to impact that through some of our work.

01:48:18.080 And I'm hoping that my questions are like prompting you to celebrate your, yeah, you've got all these

01:48:23.020 amazing tools to study it, right?

01:48:24.640 Absolutely. It really is a space that really captures anyone's imagination. But if you ask me

01:48:31.500 how I think metformin is working, I think it's probably related to the body's homeostatic response

01:48:39.300 to complex one inhibition. So of course, metformin hits complex one. I think that's undeniable. It may

01:48:47.800 have other targets, but without a doubt, it hits complex one. When complex one has been blocked,

01:48:54.540 the body senses it. And there's a feedback loop. There's a homeostatic response. And that's probably

01:49:00.700 what is net protective or helpful. And it may be the case that throwing a wrench in a complex one,

01:49:07.600 it turns on 15 of those 17 inputs that you need to sort of rejuvenate not just your mitochondria,

01:49:15.240 but other parts of your cell as well. I think there's some really interesting experiments in

01:49:19.800 worms. As you probably know, there's quite a bit of worm longevity work. And there's early studies by

01:49:25.140 my MGH colleague, Gary Ravkin, as well as Cynthia Kenyon, who was at UCSF and is now at Calico.

01:49:32.100 They did RNAi screens to basically look for genes, which when disrupted, would lead to a longevity

01:49:38.180 phenotype. And one of the gene sets that was most associated with a longevity phenotype

01:49:45.260 was the mitochondrial electron transport chain. At the same time, one of the gene sets that was

01:49:52.060 associated with a drastically reduced lifespan was the mitochondrial electron transport chain.

01:49:58.660 You can ask the question, it was a different subset of genes, obviously. There's about 90 genes total

01:50:03.680 required for the electron transport chain and oxidative phosphorylation. So the question is,

01:50:08.860 why do loss of some of them lead to longevity? And why do loss of others lead to a shortened

01:50:15.360 lifespan? One hypothesis is that it's just the strength of the allele. If you, some of those RNAIs

01:50:22.540 really wiped out the electron transport chain, probably led to early death of the worm. But if you just

01:50:28.400 gently block the electron transport chain with the right RNAi alleles, perhaps mimicking what

01:50:35.900 metformin does, you do get a longevity phenotype. So a hypothesis in the field, I'm not alone,

01:50:41.820 but I think there's others in the field that think that maybe one of the ways that metformin works is,

01:50:46.220 sure, it does block the electron transport chain, but then it comes back and causes an entire adaptive

01:50:51.980 or a homeostatic response that is not adaptive at the whole organism level.

01:50:56.480 I'm not going to put you on the spot and ask you if you think people should take metformin.

01:51:01.000 I think the broader question is, do you think, based on what you've seen in the ETC models,

01:51:10.940 that it's quite possible that a drug like metformin can be beneficial to some and harmful to others?

01:51:19.160 Oh, I think without a doubt. I mean, we know that metformin is useful for type 2 diabetes. So I think

01:51:25.000 it's a fact that for a subset of the population, metformin is helpful and beneficial. You know,

01:51:31.360 in a rare subset of cases, you can actually have fatal lactic acidosis from drugs related to metformin,

01:51:39.100 things like fenformin.

01:51:40.340 Right, which is more potent.

01:51:41.580 That's exactly right. That gets back to this idea of the potency or the allelic strength of

01:51:47.000 inhibition of the electron transport chain.

01:51:48.760 But if you took that acute toxicity aside, I mean, I think this is really the question I've

01:51:54.040 now become fixated on. If you take somebody who is already maximizing the benefits of exercise,

01:52:01.720 nutrition, sleep, these things that I think the more we look at them, the more powerful they are.

01:52:10.120 It's one thing to say the addition of metformin offers minimal benefit or incremental benefit.

01:52:15.680 It would be another thing if you're more in the Ross category that you alluded to earlier.

01:52:20.880 Is this actually undoing some of the benefit?

01:52:24.860 That's why I actually think that the experiment that you and I just discussed a few minutes ago,

01:52:28.760 trying to see what the cross product of exercise and metformin look like,

01:52:32.760 I think it could be totally fascinating. Is it going to look like the Ron Kahn study from a decade ago where

01:52:37.620 there is at least one experiment out there that suggests that. But again, I don't know how deep

01:52:43.100 they look at this. So this is a very interesting idea. I'm, uh, the only drawback of this idea means

01:52:48.480 I have to keep coming back to Boston, but now we're entering the right time of year to do it.

01:52:52.940 So that's right. That's fine. Yeah. Yeah. So changing gears, the role of hypoxia is a therapeutic.

01:52:58.020 I mean, based on what you see in the mitochondria, how do you see that as a potential therapeutic option?

01:53:03.660 This is something that we're really excited about on the preclinical level. And I really want to

01:53:09.260 emphasize this, Peter, because oxygen follows the Goldilocks principle, right? I mean, too little

01:53:15.920 is absolutely fatal, deadly. What we're discovering is that too much in certain instances, genetic

01:53:22.800 backgrounds can be damaging as well. And so all of our work to date has been focused in preclinical

01:53:28.240 models. One of the things that we are discovering in these rare mitochondrial disorders is that a lot

01:53:35.600 of the ATP levels are actually nicely defended by glycolysis. And so although the textbook dogma is

01:53:42.740 that a lot of these disorders are disorders of energy deficiency, under resting conditions,

01:53:48.600 ATP levels are okay, but what we're observing is high unused oxygen. And the important question is how

01:53:55.400 can we now try to interdict and somehow try to reduce the delivery of oxygen? So at least in our

01:54:02.100 mouse models, we're using hypoxia chambers. We actually dilute the air that the mice breathe with

01:54:07.880 nitrogen. We use some of the devices that the sports industry has created, nitrogen generators,

01:54:14.460 face masks, tents. We place the mice in those apparati, dilute the air with nitrogen, and then we evaluate

01:54:21.660 the impact. And at least in some, not all, some of our mouse models of mito disease, the benefits are

01:54:27.640 striking.

01:54:28.820 Let's explain again just why this is so profound, right? So going back to the model you described

01:54:33.200 earlier, in a subset of clinical scenarios where mitochondrial function, for lack of a better word,

01:54:40.860 is impaired, you're seeing a much higher level of oxygen return to the lungs, despite a high output

01:54:48.480 of energy imputing that their mitochondria simply aren't working. That high amount of oxygen itself

01:54:56.400 can be problematic. So you're saying, well, rather than putting you in an environment where the

01:55:01.960 ambient oxygen concentration is 21%, we're going to lower that. How much do you lower that to, by the

01:55:07.460 way, in these tents with the mice? At sea level, we're typically breathing about 21% oxygen. We reduce

01:55:13.220 it down to about 11%. How is that? I don't know much about my altitudes, but that strikes me as

01:55:19.580 really low. Is that like the top of Mount Everest low? It's not Everest. It's probably base camp.

01:55:24.780 So it's 18,000 feet. Everest is what, 7% maybe? Exactly. Exactly. So we're talking about certain

01:55:30.660 parts of Bolivia. We're talking about Mont Blanc. Yeah, yeah. But my brother's been to base camp and

01:55:35.940 he did something really funny, which is something only my brother would do. He typed out a series of

01:55:41.620 questions for himself that at sea level, the answers to which are patently obvious, 10 questions.

01:55:49.120 And at 10,000 feet and 15,000 feet, and then at base camp is 18. And then I think when he was there,

01:55:54.520 they ended up not being able to cross the icefalls, but they could still get to like 21,000. And then

01:55:58.420 again, at 21,000 feet, he would video himself answering these questions. And it was actually

01:56:03.360 quite interesting, not just the huffing and puffing that invariably goes into it, but the length of time

01:56:09.640 it took him to think of the answers, which when you consider the fact that he was answering the

01:56:13.340 same questions and over and over again, it should have been the opposite. It should have been easier

01:56:16.240 and easier and easier to come up with what year did such and such happen or whatever. That's no joke,

01:56:21.600 right? I mean, that's still for many people quite a deficit. So how did they improve?

01:56:26.740 Earlier in the conversation, we're talking about this disease called Lee syndrome. So we have a mouse

01:56:31.380 model of Lee syndrome. It's actually due to a loss of one of the subunits of complex one. So it's a

01:56:38.380 recessive loss of one of the nuclear subunits of a complex one. And this mouse is born looking okay.

01:56:45.440 It's developmentally okay. But then right around day 30, 35, it starts looking sick. And by day 55,

01:56:52.600 it'll basically fulfill our hospital's euthanasia criteria. It's lost body weight. It's become

01:56:57.760 hypothermic. It's very, very sick. It has lesions on brain MRI.

01:57:02.980 Is there anything in its periphery that looks lesioned or is all of the insult to the brain? 0.98

01:57:07.320 This initial cause of death is probably brain driven. And if you look carefully,

01:57:12.480 other tissues are affected as well. So that's what happens at 21% oxygen. There's a very stereotyped

01:57:18.980 trajectory. These uniformly fatal at about day 55 or day 60. If these mice are grown at 11% instead

01:57:27.460 of 21%, they now survive to about a median of one year. Oh my God. So you've restored them to half

01:57:34.600 their normal lifespan. That's right. And what's their function level? How do they interact? Are

01:57:38.940 they, do they act like normal mice up until then? When we were doing these experiments, a very talented

01:57:44.380 former graduate student, she's now at UCSF and our lab manager over at MGH, they're doing these

01:57:50.140 experiments. They actually thought that there was a genotyping error because the mice looked so good.

01:57:55.340 We actually thought that we'd misgenotyped them. And so they look, they look great. They put on body

01:57:59.620 weight, they put on body temperature. So the results are striking. But again, I want to emphasize that

01:58:04.500 this is all an animal model still. If you're a parent who one day has a child born with this

01:58:09.280 condition to think that the answer could be your child, instead of living a few months,

01:58:15.780 could live into their forties or fifties by moving to a part of the world, which would be the

01:58:21.000 easiest way to accomplish this. I don't think it would make sense to live at sea level and wear an

01:58:24.420 oxygen deprivation mask for your entire life. But if the answer is, guess what? You're going to go

01:58:29.920 be a Sherpa. I mean, that's unbelievable to think of. I mean, that's, I would not have predicted that 0.95

01:58:36.640 at all based on what we've discussed, that it could be that strong in effect.

01:58:40.700 Well, what's interesting is in the past, people have actually proposed hyperbaric

01:58:44.480 oxygen as a way of rejuvenating one's mitochondria.

01:58:48.240 That's what I would have stupidly suggested also, which is, wait a minute, we got to try harder to

01:58:53.280 get that oxygen in there. Let's go hyperbaric. So have you done that in the mice and do the mice 0.88

01:58:57.420 die even faster under hyperbaric conditions? So we didn't try hyperbaric because that's

01:59:01.800 higher pressure, but we just tried hyperoxic. Okay. Okay. So we went up to 55%, which is what is

01:59:07.200 often given in the operating room, as an example, the mice will die within a few days of exposure to 55%

01:59:14.360 oxygen. So there's something about having- And what about hyperbaric at 21%?

01:59:18.980 We haven't tried that yet, but just hyperoxic at 55%, these mice will die within a few days

01:59:25.220 of exposure. Peter, within a few days of us publishing that paper, we actually got phone

01:59:30.460 calls from across a country of cases where patients that were on the outpatient had been

01:59:37.460 placed in hyperbaric chambers. And then they actually ended up, you know, in some cases dying

01:59:42.340 within 24 hours or going blind in a good eye within 24 hours. And so I think there's some

01:59:47.980 anecdotes that suggest that super high oxygen levels on a broken electron transport chain can

01:59:54.020 be very damaging in humans as well. Well, this is sort of interesting, right? Because on the cancer

01:59:58.560 front, people have talked about hyperbaric oxygen being a very potent tool because the mitochondria of

02:00:06.020 cancer cells are going to be defective on balance relative to the non-cancer cells. That's, again,

02:00:12.460 outside of your wheelhouse. But how do you think about hyperbarics in terms of a tool to selectively

02:00:17.000 target cancer mitochondria? We're super excited about it. And in fact, we're really not a cancer

02:00:21.780 biology lab, but there is a subset, a very, very rare subset of tumors where we're currently exploring

02:00:27.180 that idea. I think others have thought about trying to starve cancers of their oxygen and

02:00:33.420 glucose. Our idea is the exact opposite. Maybe in certain instances, you want to flood them

02:00:38.340 with oxygen. Do you have a sense of which cancers in humans might be more or less susceptible to that

02:00:45.480 pressure? We're looking into that now, but it's going to be a rare subset of cancers where there may

02:00:50.320 be some mitochondrial mutations to begin with. So in other words, it might be less about the given

02:00:54.120 histology. So it's funny. I talked to Keith Flaherty recently, and this is a great example

02:00:58.920 of targeted therapy in cancer, right? Imagine you have your tumor, you get it sequenced, and you

02:01:04.220 realize, oh, look, you have a tumor whose mitochondria are especially weak. You are a great candidate for

02:01:10.500 hyperbaric oxygen. Person B over here, the mitochondria in your cancer cell look perfectly fine. Hyperbaric

02:01:16.020 oxygen, if anything, is not going to do. At best, it's going to do nothing. At worst, it might actually

02:01:20.380 harm your other cells. We've looked a little bit into this literature. We have an oncologist,

02:01:24.120 just in our laboratory that's looking in this direction. And hyperbaric oxygen has definitely

02:01:29.260 been proposed as a cancer therapy in the past, but there have been mixed signals. And exactly as

02:01:34.940 you're seeing, it may be the case that if you know how to precisely target it, maybe you'll see a real

02:01:39.180 signal. Anything else in your work have you thinking about cancer? This is a great example. Is there

02:01:44.600 anything else that you think about with respect to what you've learned and how it pertains to cancer

02:01:48.560 prevention or treatment? We had a series of projects about five, six years ago or so where

02:01:55.400 some of the guys in the laboratory were looking at sort of omic data sets from large numbers of

02:02:01.260 cancers, just asking, what are the most consistently altered metabolic pathways in cancer? So there's about

02:02:07.980 1,500 metabolic enzymes encoded by the human genome. Which one is the most upregulated or downregulated

02:02:14.420 across all cancers? And that pancancer analysis, it wasn't a mitochondria-focused analysis to be

02:02:21.260 clear, but it revealed a few mitochondrial enzymes in the folate pathway that are the most consistently

02:02:28.640 upregulated enzymes across all cancers. So the mitochondrion is the powerhouse of the cell. It does

02:02:35.040 produce ATP, but it's also a biosynthetic machine as well. So there's a few pathways within the organelle

02:02:41.880 that are designed to produce one carbon units for growth, folates, things like that. That pathway

02:02:48.860 was highly, highly upregulated. It gave rise to the seductive idea that maybe mitochondria are not

02:02:55.840 being used for energy in cancer, but rather as biosynthetic machines for cancer. So that was an

02:03:02.780 idea that we and others stumbled upon about five, six years ago or so, but it's not an active area of

02:03:09.100 research in our lab. Well, it's interesting because it's very consistent with other hypotheses that the

02:03:14.220 Warburg effect is less a deficit of the cancer cell due to defective mitochondria or inability to

02:03:23.320 undergo oxfos. And maybe the Warburg effect is the result of a cancer cell wanting to get a higher

02:03:29.240 throughput of substrate to foster growth. Obviously, Matt Vander Heiden was the author on a paper that

02:03:34.680 talked about that several years ago, about 2009. I hadn't heard the folate story. So that's kind of

02:03:40.820 yet another really interesting point. Single carbon biology is pretty interesting stuff.

02:03:48.500 That's right. That's right. And several investigators like Josh Rabinowitz, David Sabatini,

02:03:53.420 our group.

02:03:53.900 I had dinner with David last night.

02:03:55.240 Great.

02:03:55.620 Yeah. We actually were talking about single carbon metabolism and the challenges of it. I went to

02:03:59.340 med school with Josh, by the way.

02:04:00.720 Oh, wow. Oh, wow. Okay. So actually the three of us, about five, six years ago or so, we all had

02:04:06.580 sort of independently stumbled upon this mitochondrial pathway as being dramatically upregulated. Josh with

02:04:13.400 metabolomics, David with RNAi, us with computation. And I think to this date, there's a lot of data that

02:04:19.420 supports the idea that this is upregulated in cancers. Now, whether targeting it is going to be

02:04:25.240 beneficial, that's an open question still. But without a doubt, this is one of the pathways that

02:04:30.160 is powerfully upregulated in cancer states.

02:04:33.500 On the topic of cancer, and we've talked about these other chronic diseases, but it doesn't

02:04:37.540 really appear that there is a chronic disease in which the mitochondria remain normal. If you look

02:04:42.820 at cancer, if you look at Alzheimer's disease, if you look at atherosclerosis, and if you look at

02:04:46.720 type 2 diabetes, all of these diseases have mitochondrial signatures that differ from what we would

02:04:53.160 consider healthy. Well, it gets back to the opening parts of the discussion where we said that if you

02:04:57.240 take any aged tissue, it's going to be associated with dysfunctional mitochondria. And if you take

02:05:02.940 diabetic muscle, if you take Alzheimer's brain, Parkinson's brain, you're going to see dysfunctional

02:05:08.720 mitochondria. And this is why it gets back to, is it cause or effect, or is it going to be some

02:05:14.520 complex nonlinear combination of cause and effect? And this is where using that systems biology

02:05:20.500 approach or trying to gain insights from these rare diseases may inform a subset, but not all of

02:05:26.840 these. I actually think that for Parkinson's, the causal hypothesis is pretty compelling out of all

02:05:33.880 of these disorders. Say more about that, because I don't know anything about Parkinson's that I

02:05:39.920 didn't learn in medical school, which if I recall is more to do with dopamine secretion out of a part

02:05:46.340 of the brain that the substantia nigra, is that it? Where you basically lose that. Those neurons.

02:05:52.140 Yeah. And so the patients that maybe start out with genetically fewer of those, because there's a

02:05:56.140 distribution of how many you get, maybe the ones most susceptible. I'd never thought of that as a

02:06:01.940 problem that shows up in the mitochondria. So expand on that.

02:06:05.180 Again, we have two classes of disorders, right? We have sort of common complex diseases and we have

02:06:09.760 monogenic diseases. And the big question is, does the pathology we see in the rare monogenic forms

02:06:17.420 bear any relevance, right, to the common forms of disease?

02:06:21.640 It's funny. I mean, the way we're talking about it, your hands are showing it in a way that I think

02:06:25.600 is representative, right? You had your one hand out over here and you said, look, these are the very

02:06:29.720 simple monogenic diseases. Most people have never heard of them. They're typically quite brutal,

02:06:35.240 but they kill relatively few people. On the other hand, you have, there's a divide, right?

02:06:39.540 That's right.

02:06:40.020 And it's like, it's quite discontinuous, isn't it?

02:06:42.700 Well, also on this side, you know, the prevalence of these monogenic mitochondrial

02:06:46.860 disorders is about one in 4,000. But then as you cross the continuum-

02:06:51.920 The prevalence is one-

02:06:52.680 One in one.

02:06:53.300 Yeah, exactly.

02:06:54.500 And then you have these other disorders like type 2 diabetes and Parkinson's and Alzheimer's

02:06:59.440 that it's not one in one, but it's also not one in 4,000.

02:07:02.920 So you're saying Parkinson's may be the closest example that you can think of that's

02:07:07.220 spanning these two worlds?

02:07:09.080 I think so.

02:07:09.880 I think so.

02:07:10.180 Well, okay. I want to hear more about this.

02:07:11.840 Well, I think there's a couple of reasons. One is if you take, you know, the common form of

02:07:16.580 Parkinson's disease, and if you take some of the postmortem material, if you biopsy that,

02:07:21.640 or if you take postmortem material and you look, you'll actually see mitochondrial lesions. You'll see

02:07:26.000 an increase in the mutation burden in the mitochondrial genome. You'll see complex one

02:07:31.540 deficiency. Already we're seeing some of the molecular features of mitochondrial dysfunction,

02:07:37.700 but perhaps even more compelling is that there are some toxin forms of Parkinson's. Certain types

02:07:43.780 of herbicides and insecticides are actually toxic to complex one.

02:07:48.900 So you're saying there are people who have Parkinson's, and I apologize for my ignorance

02:07:52.560 on this, where they don't actually have a dopamine deficiency in the brain, but they have a

02:07:57.420 Parkinson's-like phenotype purely from an insult to their mitochondria from, say, a toxin?

02:08:03.020 Right. So in these instances, what we think has happened is that because of an environmental

02:08:08.040 pesticide or insecticide, the mitochondria has been poisoned in some of these dopaminergic

02:08:14.200 neurons, and those neurons actually die. So there is a dopamine loss ultimately.

02:08:17.700 I see. I see. Okay.

02:08:18.440 But the root cause is the mitochondria. That's right. So that's toxin evidence where we know

02:08:26.620 that a direct toxin to the mitochondria in humans can give rise to a Parkinsonian-like

02:08:31.740 disease. In mouse models, if you give a high dose of rotenone, it's a complex one poison.

02:08:39.740 Okay.

02:08:40.320 And this actually gets back to-

02:08:41.800 To metformin.

02:08:42.420 Exactly. So if you infuse into a mouse rotenone, which is a very potent-

02:08:48.040 More potent than phenformin?

02:08:49.300 Yes.

02:08:49.700 Yeah.

02:08:49.940 Right. And there's some potential off-target effects as well.

02:08:53.100 Like micromolar or millimolar potent?

02:08:55.180 Micromolar potency.

02:08:56.120 Wow. Okay.

02:08:56.900 It may hit other things like microtubules as well, but it definitely hits complex one of

02:09:02.060 the electron transport chain. That's been used as a model of Parkinson's disease in rats and in mice.

02:09:10.660 And so I think between the toxin evidence, the fact that sporadic forms of Parkinson's disease

02:09:16.000 can be associated with complex one deficiency or mitochondrial mutations, I think helps to support

02:09:22.280 the idea that mitochondrial dysfunction can play maybe in the causal path for Parkinson's disease.

02:09:29.320 Are you optimistic that we are going to be able to target mitochondrial proteins as therapies?

02:09:38.500 I mean, the more I listen to this, the less optimistic I am, truthfully, just because of

02:09:45.580 the exceptions being exceptions and not rules.

02:09:50.700 I'm actually pretty wildly optimistic.

02:09:52.600 Okay. That's great because I'm coming away like discouraged. Like there are too many moving pieces

02:09:57.340 to be able to use a single molecule. So are you thinking of stacking molecules or tell me where

02:10:04.040 your optimism comes from?

02:10:05.240 Well, it comes from the fact that five years ago in some of these mouse models or cellular models,

02:10:11.260 we had zero ways of alleviating mitochondrial disease in a dish or in a mouse. Now we have in

02:10:17.640 our laboratory, we can use, again, in a preclinical way, we can use hypoxia and it actually helps

02:10:22.980 to restore cellular function and longevity and healthspan in mouse models of mitochondrial disease.

02:10:29.700 We're using other approaches that are evolutionarily inspired. We call these protein prostheses where

02:10:35.240 we take proteins from other organisms, from other kingdoms of life. We transplant them into

02:10:41.100 human cells with mitochondrial disease and we can effectively rescue the cells.

02:10:46.460 Well, but how do you transfect all the cells? Are you just saying that you're doing this ex vivo?

02:10:53.860 This is all ex vivo right now, right?

02:10:55.400 But still, it's a proof of concept that's very powerful.

02:10:57.860 I mean, nowadays we have nucleic acid therapeutics, gene therapeutics, protein therapeutics.

02:11:02.560 So give me an example of one of the protein prosthetics.

02:11:05.640 Getting back to the earlier part of the conversation, we spoke about the fact that

02:11:08.980 the electron transport chain was probably one of the earliest features of the early eukaryotes,

02:11:15.740 probably resembling the electron transport chain of bacteria that can do oxidative phosphorylation.

02:11:22.040 But then during reductive evolution, certain organisms lost parts of their electron transport

02:11:27.540 chain. Now there are certain organisms that have lost their entire electron transport chain. And we

02:11:33.380 think that one of the ways that they're able to survive is that they gained a new protein that

02:11:38.440 basically complements part of the activity of the electron transport chain. So we've identified some

02:11:43.680 of those proteins. If we place those proteins in a human cell, you can poison the mitochondria any

02:11:50.300 of five different ways, and the cell will still proliferate because it has that protein that

02:11:54.720 evolutionarily, we believe, allowed that organism to lose its ETC to begin with. Does that make sense?

02:12:02.380 Yeah, it just seems too good to be true.

02:12:04.340 Well, this is why we call it a prosthesis. It's not 100% fix of the solution. What it does is it

02:12:10.300 probably corrects part of the redox imbalance in the electron transport chain, but not the full

02:12:15.480 proton pumping capabilities.

02:12:18.060 Yeah. So play this forward. You now could argue another treatment for type 2 diabetes

02:12:24.860 is in addition to making the changes we're going to make, right? Because I still believe

02:12:31.120 deep down that you can cure type 2 diabetes with corrective exercise and nutrition and sleep. I

02:12:37.280 think if you get those three things fully optimized, type 2 diabetes goes away with the exception of the

02:12:42.880 late stage cases where the pancreas no longer works. But what if you could add another layer to that,

02:12:47.800 which is, oh, by the way, here's some mitochondrial prostheses.

02:12:51.160 That's exactly right. It's not going to fix all of the functions of the electron transmission.

02:12:55.340 But it can become additive to other things, right?

02:12:57.200 That's right. That's right. That's one of the ideas that we're exploring right now. Can we,

02:13:01.580 actually in the context of things like fatty liver and diabetes, can we use some of these

02:13:05.760 mitochondrial protein prostheses to either augment or bypass some of the broken functions of

02:13:12.680 mitochondria to restore function?

02:13:16.000 I don't know if you've ever thought of this because the idea just popped into my head when you said

02:13:19.380 mitochondrial protein prostheses, have you ever thought just in sort of random sci-fi

02:13:25.260 like thinking of the opposite, which is performance enhancing prostheses for mitochondria? I mean,

02:13:30.060 if you think about the efforts that athletes can go to to enhance aerobic performance, the most

02:13:37.360 obvious of these, of course, is blood doping and use of EPO, which simply deliver more oxygen to the

02:13:43.640 system. But in 30 years, will people be talking about genetic cheating where mitochondrial performance

02:13:51.620 and function has been enhanced?

02:13:53.520 Well, it's funny because the sports industry is so often, I mean, the sort of underground illicit

02:13:58.920 sports world of doping often is so ahead of the medical world. It's actually pretty amazing.

02:14:04.720 Yeah, it really is.

02:14:05.520 I mean, one of the compounds, it's like EPO, it's called FG4592. It's a small molecule that

02:14:10.820 blocks one of the prolyl hydroxylases. This tricks the system into thinking that it's hypoxic and you

02:14:16.280 end up producing more EPO. That drug was in clinical trials. And before the drug was even approved,

02:14:24.020 it was being used by cyclists, which is already amazing. But what's also even more amazing is that

02:14:30.600 the anti-doping agency knew to look for drugs that were not yet even approved by the FDA in some of

02:14:37.280 these athletes. And so it's like this red queen where everyone has to stay ahead of everybody else. 0.65

02:14:43.460 It's an arms race. Do you think like theoretically, there are ways to enhance performance through,

02:14:50.340 again, I'm not a huge advocate on genetic engineering. I still think it's so there's so

02:14:54.500 much sci-fi and the limitations of actually getting a virus that could be taken up ubiquitously. But

02:15:00.120 putting that aside for a moment, could you engineer a better mitochondria? Could one, I mean, I don't

02:15:06.280 want to put you on the spot, but could one engineer an even better mitochondria? How much waste is in

02:15:10.760 the system? It's funny. If I ever did this exercise, it's so long ago that I don't recall. But from an

02:15:17.120 engineering perspective, how much room for improvement is there?

02:15:20.380 I think it's going to take a while before things like that are possible. And I gather from this

02:15:25.740 conversation that you, like me, like automobiles. And so if you have an engine, the way to enhance

02:15:32.620 it is to turbocharge it. But turbocharging it is not taking one spark plug out and replacing it with

02:15:39.460 another spark plug.

02:15:40.420 No, it's completely changing the way air flows through it and is recycled.

02:15:43.080 So you have like 17 different things. You have a turbocharger, you put it on top, but there's like 17

02:15:49.220 other things that you have to alter. So the entire system connects up, you have impedance

02:15:54.520 matching, you have airflow matching, so that in the end, you have a better performing automobile.

02:16:01.120 I think something analogous applies to mitochondria. It's not going to be just one thing that you can

02:16:06.000 turn on. If you want to try to, the system is already pretty optimized, right? And so if you just

02:16:11.520 change one thing, it's probably not going to, you can break it.

02:16:14.640 Yeah. You can break it with one thing, but it's hard to enhance with one thing.

02:16:19.220 That's actually a great way to just think about biology in general. It's pretty easy to break it

02:16:24.860 at a single point. Cyanide, I think the most extreme example, right? Detrodotoxin, so easy to

02:16:32.860 kill and break at a single molecule, which comes back to a point I made earlier. It's mind boggling when

02:16:40.000 something like rapamycin works. Right. Again, I don't, I mean, you had dinner with David last night,

02:16:45.580 so he's clearly one of the world's authorities on rapamycin and its function. But at least if we go

02:16:50.820 to metformin, the only way that I can conceive that it could have this total body impact is if

02:16:57.480 it's doing something that then causes the entire system to respond. Even when we talk about statins,

02:17:04.940 every medical student in the country knows that statins are life-saving. Every single medical

02:17:10.660 student in the country knows that statins hit HMG-CoA reductase. But why is it life-saving?

02:17:16.800 Right. Is there other things that it's doing that are also important beyond the reduction of LDL?

02:17:23.520 The way that statins work, my understanding is that sure, statins directly target HMG-CoA reductase,

02:17:29.640 but then what ends up happening is because you're not producing so much sterols, you turn on an

02:17:35.160 entire SREBP transcriptional program. Right. And you bring more receptors to the surface of the

02:17:41.360 liver. That's right. More of the statin efficacy is due to the LDL clearance than the reduction of

02:17:46.360 cholesterol synthesis. That's exactly right. But there may even be benefits beyond that,

02:17:49.680 is my point. And some argue there are actually a whole camp of LDL denialists who can't deny the

02:17:56.260 efficacy of the drug. And so the sort of hypothesis is, well, all of the statin benefit doesn't come

02:18:02.720 through the LDL reduction through the mechanism you described. The unintended consequence is the

02:18:07.720 wrong way of saying it, but the non-obvious consequence, but it could be some of the

02:18:11.580 anti-inflammatory benefits that come from it or the endothelial protective benefits.

02:18:15.800 But the point is that there's an entire response to hitting HMG-CoA reductase. And a consequence

02:18:22.540 of that program is that you have more sterol production. So in the end, the amount of

02:18:28.920 cholesterol that you're producing is kind of balanced, right? You've inhibited it, but you've

02:18:32.920 turned on more of the enzymes, so it's comparable. But you've turned on these other 17 switches as well,

02:18:39.420 at the same time, one of which is the LDL receptor, which helps to clear LDL levels. There may be other

02:18:43.840 things as well that are turned on that are net beneficial. So I think physiological systems are so

02:18:49.840 complicated. Trying to identify all 17 of those things and turning them on at the same time,

02:18:55.320 in general, I think is going to be hard, even in the next 30 years. But it may be the case that

02:19:00.300 some of the interventions are, you know, what people refer to this. Do you think of this as when

02:19:05.060 you, like, let's think of the three examples you've just used, a statin, metformin, and rapamycin.

02:19:11.460 None of those are purely, I mean, they're synthesized today, but they are all derivatives of

02:19:15.740 naturally occurring compounds. I'm sitting here as you're telling this story, trying to think of

02:19:22.080 examples of purely synthesized de novo molecules that have such benefit. And I'm, there may be

02:19:29.440 examples, but it seems less obvious. Do you think it's a coincidence that some of the most potent agents

02:19:34.520 that we have in medicine, which by definition, it seems, are the ones that have to do multi-pronged

02:19:42.080 inputs tend to come from naturally occurring substances? Does that just speak to our

02:19:46.920 co-evolution with these things? No, absolutely. I think there's been this arms race, right? The

02:19:51.860 bacteria are fighting these fungi, and the fungi are fighting these plants. And so there's all sorts

02:19:57.880 of small molecules that are designed to throw wrenches into other mitochondria or translational

02:20:03.620 programs or nutrient sensing programs. So natural products are remarkable chemistries.

02:20:10.880 And they've evolved over hundreds of millions of years to target physiological systems. And so

02:20:18.880 I'm a huge fan of natural products and the biology that they expose.

02:20:24.020 This is so interesting. We'll close with a question that's admittedly kind of a tough question. So I

02:20:29.820 apologize for putting you on the spot and no qualms if you can't come up with a good or great answer.

02:20:36.160 But when you think about the world you want to explore here with the questions that you want to

02:20:40.880 ask, so much of what happens, even at the level that you're at, is still constrained by resources.

02:20:47.420 Have you ever had the thought experiment or the sort of fantasy of what if you were in a totally

02:20:52.520 resource, unconstrained world? So you never had to apply for another grant. In fact, you were given

02:20:57.480 some lump sum of money that was beyond what you could imagine. And even just from an IRB standpoint,

02:21:02.840 there was nothing that stood in your way of doing kind of something that was still ethical, but

02:21:06.720 maybe today would be logistically too challenging. Do you have a sense of what questions you would

02:21:12.960 want to probe, but specifically what experiment or set of experiments you would do in this dream state?

02:21:19.740 Yeah, it's a tough question. It sounds like you're basically asking me not to be limited by

02:21:23.840 anything except my own imagination. And I think so often in biomedical research, that is what limits

02:21:32.700 us. But it's a great question, Peter. One experiment that would be kind of a fun experiment to try is

02:21:39.800 really motivated by your recent work on oxygen. So what we've observed is that at least in preclinical

02:21:47.540 models, when you have severe mitochondrial decline, breathing thinner air appears to be beneficial.

02:21:56.560 Now, how about in the common form of aging, when there's a subtle decline in mitochondrial activity?

02:22:05.060 Is there excess unused oxygen? And will breathing thinner air be beneficial? And given that I'm not

02:22:14.500 resource limited, I mean, wouldn't it be cool if we could construct a Ritz-Carlton hotel or condominium

02:22:22.380 at 16,000 feet? That's extremely comfortable. And perhaps another Ritz-Carlton that looks identical

02:22:30.720 at the plains. And we could take a very large cohort, not five, not 10, but thousands of people

02:22:38.740 that live at either the plains or at the high altitude for many, many months. And we try to

02:22:46.040 evaluate whether there are any biomarkers of aging, age-associated disorders that actually

02:22:52.220 improve under thin air. Maybe they become worse under thin air, but it would be something like a

02:22:59.380 crossover experiment that would allow us to test the idea that thin air may be beneficial for chronic

02:23:05.520 diseases. I love that idea. That is elegant because one, you could do that. You answer the question,

02:23:12.880 right? Which is in a resource unconstrained world, like that's tens of millions of dollars would allow

02:23:17.480 you to do that. Call it hundreds of millions of dollars could do that longitudinally. And your

02:23:22.380 intervention is elegant in that it's going to touch lots of things. That's interesting. So 16,000 feet.

02:23:28.820 So following up on that, have you done or contemplated doing the less beautiful version

02:23:35.620 of that in mice where, or I'm sort of disdainful of mice, but maybe rats or something that's less

02:23:42.340 inbred where you have models of type 2 diabetes, but they're genetically born with normal mitochondria.

02:23:49.440 Has that experiment been done? We're doing those types of experiments now in, of course,

02:23:54.460 our focus is on some of these rare genetic mitochondrial diseases, but we're going into

02:23:58.900 some of these other conditions, more common conditions that are also associated with mitochondrial

02:24:04.600 dysfunction. So those are currently ongoing. When we get back to that dream experiment,

02:24:09.960 what is interesting, Peter, is experiments like that have actually taken place in humans and they

02:24:15.160 can be- Right. They're natural experiments.

02:24:17.000 And one of the experiments, there's this one paper from the early 1970s that was published by

02:24:23.880 the Indian Army. It was the Indian Army reporting on the health outcomes of a huge number of their

02:24:31.520 troops who India has historically had border disputes with China. I think it was in the 1960s,

02:24:39.260 India deployed more than 100,000 troops on the Indochina border. And don't quote me on the numbers,

02:24:45.140 but about 25,000 of these people were at extremely high altitude and another 100,000 or so were at

02:24:53.280 the plains. They were there for about five to seven years or so. And of course, the food that they ate,

02:24:59.860 the temperature, the activity, all of these were different between low and high altitudes. But

02:25:04.660 one of the variables that was different was oxygen. And after, I think it was seven years,

02:25:10.400 the Indian Army actually wrote this paper. It's only been cited a few, like maybe 20 times or so.

02:25:16.980 They reported the long-term health consequences. And what they showed is that death from things

02:25:23.740 like infections were much higher acutely upon going to high altitude. But if you look at chronic

02:25:29.920 diseases like incidence of diabetes, stroke, cardiovascular disease, they were much reduced

02:25:36.840 in the high altitude arm.

02:25:39.840 Even after they returned to the plains?

02:25:42.760 I think the study-

02:25:43.680 Or only as long as they stayed at the altitude?

02:25:45.480 Only as long as they stayed in the altitude. I don't think they had long-term follow-up actually.

02:25:49.820 But it's this type of data that actually makes me wonder whether what we're observing in some of

02:25:55.820 these rare forms of mito dysfunction and the interaction of mitochondria with oxygen,

02:26:01.740 perhaps some of it could bear relevance to more common conditions as well.

02:26:06.840 Super interesting. That can keep going on, but you've been generous with your time,

02:26:12.180 especially to a total stranger. This is really helpful. I want to thank you for the work you're

02:26:17.080 doing. And I want to thank you for taking the time to explain it to me and to a few people listening.

02:26:22.000 Thank you so much. This has been a lot of fun for me.

02:26:23.940 You can find all of this information and more at peteratiamd.com forward slash podcast. There,

02:26:31.940 you'll find the show notes, readings, and links related to this episode. You can also find my

02:26:36.740 blog at peteratiamd.com. Maybe the simplest thing to do is to sign up for my subjectively

02:26:41.820 non-lame once a week email, where I'll update you on what I've been up to, the most interesting

02:26:46.280 papers I've read and all things related to longevity, science, performance, sleep, et cetera.

02:26:52.020 On social, you can find me on Twitter, Instagram, and Facebook, all with the ID peteratiamd. But

02:26:57.560 usually Twitter is the best way to reach me to share your questions and comments. Now for the

02:27:01.680 obligatory disclaimer, this podcast is for general informational purposes only and does not constitute

02:27:06.320 the practice of medicine, nursing, or other professional healthcare services, including the

02:27:11.220 giving of medical advice. And note, no doctor-patient relationship is formed. The use of this information

02:27:17.280 and the materials linked to the podcast is at the user's own risk. The content of this podcast is not

02:27:22.560 intended to be a substitute for professional medical advice, diagnoses, or treatment. Users should not

02:27:28.020 disregard or delay in obtaining medical advice for any medical condition they have and should seek

02:27:32.720 the assistance of their healthcare professionals for any such conditions. Lastly, and perhaps most

02:27:38.260 importantly, I take conflicts of interest very seriously. For all of my disclosures, the companies

02:27:43.060 I invest in and or advise, please visit peteratiamd.com forward slash about.

The Peter Attia Drive - August 12, 2019

#66 - Vamsi Mootha, M.D.： Aging, type 2 diabetes, cancer, Alzheimer's disease, and Parkinson's disease – do all roads lead to mitochondria？

Episode Stats

Summary

Transcript