The Peter Attia Drive - #216 - Metabolomics, NAD+, and cancer metabolism ｜ Josh Rabinowitz, M.D., Ph.D.

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

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

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

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

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

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

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

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

00:00:47.740 here's today's episode. My guest this week is Josh Rabinowitz. Josh is a professor of chemistry

00:00:54.880 and integrative genomics at Princeton University, where his research focuses on a quantitative,

00:00:59.460 comprehensive understanding of cellular metabolism through the study of metabolites and their fluxes.

00:01:04.240 He's also the director of the Princeton branch of the Ludwig Institute for Cancer Research and a member

00:01:09.360 of the Rutgers Cancer Institute. Josh earned his MD and PhD in biophysics from Stanford, which is how

00:01:15.940 we met. We were in the same graduating class, although he of course started earlier because he did two

00:01:20.720 degrees. In between earning his MD and PhD and joining the faculty at Princeton, Josh worked at

00:01:26.320 Alexa Pharmaceuticals as the co-founder and vice president of research, a topic that we actually touch

00:01:31.340 on in this podcast. Josh is the inventor of over 160 patents, including five drug products that are in

00:01:37.580 the FDA-sanctioned clinical testing pipeline. He has received numerous awards, including an NSF Career

00:01:43.560 Award, an NIH Pioneer Award, and was distinguished as an Allen Distinguished Investigator in 2019.

00:01:50.720 This is a pretty technical episode, I'm not going to lie, and we really focus on three things.

00:01:56.080 Metabolomics, NAD specifically, and of all of its sort of precursors and movements, and cancer

00:02:01.400 metabolism. We open the discussion talking about metabolism, metabolomics, and fluxomics, and this

00:02:07.320 includes a pretty in-depth conversation around glucose, glucose metabolism, lactate as a fuel, movement

00:02:14.140 of lactate, and the regulation of these substrates. From there, we speak in more detail on the electron

00:02:19.080 transport chain and the Krebs cycle and what the implications are, both with respect to drugs and

00:02:24.240 nutrition. This is an important segue then into the second major pillar of our discussion, which is

00:02:30.300 that around NAD. Most of you have heard of NAD. We certainly got a lot of questions about NAD and

00:02:36.080 not so much about NAD as we do probably more about their precursors, NR and NMN. We've also had previous

00:02:41.440 podcasts where we've discussed this, including episodes with David Sinclair and Rich Miller.

00:02:46.100 So in this discussion, we talk about the intravenous use of NAD, the oral use of the precursors,

00:02:53.180 and I'll just give you a little spoiler alert, or I'll try not to spoiler alert it, but I'll point

00:02:57.460 you to something, which is I learned something pretty significant in this episode that I have

00:03:01.820 historically been saying incorrectly for some time. So if this is a topic that's interesting to you,

00:03:06.880 and you've heard me speak on it before, you might want to listen to this because I'm going to come

00:03:10.540 into a big correction. We end our conversation talking about cancer metabolism and particularly

00:03:16.180 one way in which cancer metabolism and immunotherapy might intersect. So without further delay,

00:03:22.060 please enjoy my conversation with Josh Rabinowitz.

00:03:30.260 Hey Josh, great to see you again. It was about three years ago, I think was the last time we saw

00:03:35.580 each other in person at some sort of conference in New York, a cancer metabolism conference, I think,

00:03:40.040 right? The New York Academy of Sciences, I think.

00:03:43.240 And we somehow wound up at some kind of mediocre bar in Brooklyn, and it might have been the dorkiest

00:03:49.800 concentration of people talking about autophagy and metabolomics and all sorts of things.

00:03:56.640 That was actually the AACR Brooklyn conference, that's true.

00:04:01.920 Well, you know, it's really funny. Just recently, I've interviewed a few of our classmates from

00:04:06.600 med school, Max Dean and Carl Deseroth. And I think Carl and I were reminiscing about how you,

00:04:11.920 me, and Carl all started our surgical rotation together in the same day. It's almost 25 years

00:04:16.560 ago, and I remember it like it was yesterday when we were all sitting in the room practicing

00:04:19.560 sewing with our big goofy knots and things like that before we all got divided up into our

00:04:25.800 surgical rotations. Do you remember that?

00:04:27.480 I definitely remember you. I really do not remember Carl from those days, but I certainly

00:04:31.900 remember you practicing a lot while I looked and said, oh, I guess that's what you're supposed to

00:04:36.440 be doing if you want to become a surgeon. The thing I remember most clearly is being in surgery with

00:04:41.020 you. And it was like the very end of surgery rotation, finally being given the bovie, immediately

00:04:47.600 making a mistake. Then I got the nice lecture, you know, we're happy to pass you on this rotation

00:04:53.460 as long as you promise never to use a knife or bovie again.

00:04:58.000 Remind me, whose lab did you do your PhD in?

00:05:01.120 I did my PhD with Hardin McConnell, one of the great physical chemists, together with Mark Davis,

00:05:06.160 the immunologist.

00:05:07.680 Tell me and tell the listeners a little bit about what project you worked on for your thesis.

00:05:12.720 My thesis was about the physical chemistry of T cell activation. At that point in time,

00:05:17.820 people had discovered that different antigens could activate T cells differentially. And so I was really

00:05:25.480 interested in how that happened. And so I studied the process of antigens turning into peptides that

00:05:33.040 could bind to MHC, and then how the kinetics of the interaction between the peptide antigen and the

00:05:39.800 matrix to compatibility complex protein MHC. And then the interaction of that complex with the T cell

00:05:45.840 receptor determined whether people have productive or failed immune responses with the hopes that you

00:05:51.420 could then manipulate those processes to promote better vaccination or disease clearance and also

00:05:57.800 to potentially treat autoimmunity.

00:06:00.160 And did you also have an interest at all in cancer? Because of course, this would be one of the

00:06:03.560 hallmarks of how immunotherapy would be effective in eradicating cancer.

00:06:08.140 It shows how bad a prognosticator I am. At that point, I really felt immunological

00:06:13.440 approach just that cancer didn't hold that much promise. And so I was really more focused on

00:06:18.760 infectious disease and autoimmunity. Shows you what I know. It's wonderful to see that the world

00:06:24.300 has turned out to be a lot better than I dreamed it would be on that dimension.

00:06:28.240 It's such an interesting topic. I had Steve Rosenberg on the podcast last year,

00:06:31.920 and it was a beautiful and fun recap of how the immune system works in general. But we were talking

00:06:37.560 about it obviously through the lens of cancer. And I think the part that will forever humble anyone who

00:06:43.800 tries to think about how amazing the system is, is that these things have to be, you know, nine to

00:06:48.880 11 amino acids long. I mean, the peptides have to be just the right size to be presented and then to be

00:06:55.720 recognized. And that just doesn't seem to leave a lot of margin for error. I mean, it is really a

00:07:00.800 tuned system. Do you have a sense of why evolution ended up with such a narrow fragment of peptides

00:07:08.480 that were recognizable as opposed to a broader range or as opposed to just a range that's

00:07:12.900 different? Like why wasn't it two to three or 150 to 160 amino acids? Like, is there a chemical reason

00:07:20.360 because of your background in chemistry? I feel like you'd be more equipped to offer a teleologic

00:07:24.480 explanation for this. I do have an intuitive sense on this, that our bodies work on the scale of

00:07:31.580 billions of immune cells and billions of immune receptors that are made through recombination.

00:07:38.880 And that naturally pairs with billions of antigens. And so you just think about this number nine,

00:07:44.380 and you think about the number of amino acids, right? There are 20 amino acids. So you're talking about

00:07:48.580 20 to the ninth power of presentable peptide antigens. And so these things are all tuned to

00:07:54.840 be on the same scale, this kind of scale of billions and... Tens of billions in that case.

00:08:00.800 Exactly. If it was like a three peptide, it's really a small number actually.

00:08:06.060 There's not enough information there to selectively respond to a virus or a bacteria.

00:08:11.740 And if it was 25 peptides, it's too much. There's too much information there.

00:08:19.140 There's extra information anyway. And this system I think was built to work on

00:08:24.520 minimal or just the right amount of information.

00:08:28.520 Yeah, that's super interesting. So at what point during either your PhD or the end of medical school

00:08:36.360 where we met up in the clinical portion, did you make the decision that you wanted to be

00:08:41.700 a full-time scientist as opposed to a physician scientist?

00:08:45.320 I applied to do an internship knowing that I really loved research, but I guess I decided as

00:08:53.360 I did more and more medicine that medicine is such a noble profession, but it involves a lot of

00:09:00.080 doing the same thing right over and over again. A lot of following the standard of care, even for the

00:09:05.400 most creative physicians. And that ultimately my passion is to come in and try to do something

00:09:10.940 different every day, to think differently than people ever have before. And so that's what

00:09:15.240 really led me to research. I love the patient interaction part of medicine. It was the doing

00:09:20.660 things right part that was challenging for me sometimes.

00:09:23.820 I think those of us that chose the more medical side of things can also speak to the frustration of

00:09:30.580 how creativity can often be stifled in medicine. And that's in large part for good reason, but I think

00:09:36.180 it comes at a cost as well. I think I'm sure on this podcast, I've told stories about how frustrating

00:09:40.300 that was in surgery, at least. Surgery, probably more so than most other disciplines, tends to

00:09:45.680 sometimes at least frown upon creativity and novel approaches to problem solving.

00:09:50.140 If I remember correctly, before you joined the faculty at Princeton, where you are now,

00:09:55.780 you went into industry straight from medical school. Am I getting my facts right?

00:09:59.600 I was fortunate at the opportunity to work with one of the great early biotech entrepreneurs,

00:10:06.040 Alex Zaffaroni. He and I started a company when I was straight out of medical school that was focused

00:10:11.860 on fast drug delivery. So what could you do by being able to deliver medications non-invasively

00:10:17.820 on the timescale of giving an IV push in the hospital? We did that through inhalation of small

00:10:24.880 molecules, kind of building on the concept, obviously, that if you smoke something like a cigarette,

00:10:30.380 you get incredibly rapid access to the systemic circulation. We built that company. Lexa Pharmaceuticals

00:10:37.480 still exists, has one FDA approved drug. That was my first job.

00:10:42.540 What drugs were you targeting with that?

00:10:44.020 You know, our initial focus was migraine. Unfortunately, we never found the drug that

00:10:48.800 had the perfect combination of safety with rapid delivery and efficacy for migraine. The drug that

00:10:54.780 ultimately got approved was for acute agitation. It's a set of a hypnotic for acute agitation. And

00:11:00.760 I'd say it's really wonderful for the patients who get it. They come in the ER. It's something people

00:11:05.800 don't always know about agitated patients. They're agitated. They're frustrated, but they also want to

00:11:11.140 stop being agitated. They don't want to act like that. And so they're very eager to most cases to take

00:11:16.160 a puff of something and calm down in a couple of minutes. It's wonderful for them.

00:11:19.660 So this is kind of an unusual path. I'm guessing that most people who experience that type of

00:11:26.180 success that you had would want to keep doing it over and over again. I mean, hence the term

00:11:30.680 serial entrepreneur. What made you decide to take this lateral step to go into academics,

00:11:37.300 which would have made a lot more sense if you'd done it coming straight out of your PhD?

00:11:42.800 You know, I think I was lucky to get the chance to go straight from that job into a

00:11:47.940 faculty position at Princeton. That's a rare opportunity to have.

00:11:52.120 Because you didn't do a postdoc in there unless they considered your industry as sort of a

00:11:56.240 grandfathered postdoc.

00:11:58.020 Yeah, it was a very, very weird version of a postdoc. It united me with my wonderful wife

00:12:04.120 at Princeton on the faculty here and turned out well.

00:12:07.880 So you show up at Princeton and they put you on a tenure track position,

00:12:12.280 which means they're giving you the types of resources to now start solving whatever problem

00:12:17.500 you want. What was the first problem that you said, I want to build a lab around?

00:12:22.220 You know, one thing that I learned actually doing Alexa was about drugs because we studied

00:12:28.520 every medication in the pharmacopoeia for whether it could be a candidate for rapid delivery. Could

00:12:33.660 you make a benefit by delivering it rapidly? And one thing I noticed is that so many of the

00:12:39.040 most important medications work via metabolism. Then starting out my lab, I realized that there

00:12:44.560 were relatively few labs looking at metabolism broadly compared to other, you know, really

00:12:49.660 important areas of science like immunology or cancer or neuroscience. And so I started my lab

00:12:56.540 with a really simple question. Could we measure the classic metabolites that you read about in a

00:13:01.540 biochemistry textbook in one shot quantitatively? And so that was the starting point for my lab.

00:13:08.260 And I guess the second question that we always had in our mind is, can we measure

00:13:12.180 the activities of those metabolic pathways? So how fast are those metabolites flowing? Where are

00:13:18.340 they coming from? And where are they going? I had lunch or dinner actually with one of our mutual

00:13:22.800 friends, Navdeep Chandel, who was also a previous guest on the podcast three or four years ago,

00:13:27.460 maybe a week ago. And he was here in Austin for a talk. And we were talking about how in the late

00:13:34.240 90s, he said, and he was obviously studying metabolism. He said, if you were to rank me,

00:13:40.760 meaning Nav was ranking himself, right? If he said, you were to rank me and my work just across the

00:13:46.460 spectrum of the stuff scientists were doing, he said, I'm bottom 10 percentile. Nobody found

00:13:52.860 metabolism interesting. This was, I forget the term he used, but like this was basically the corner where

00:13:59.180 the kids went that nobody wanted to play with. If you weren't doing genomics, if you weren't doing

00:14:04.220 this other sexy stuff, immunology, you were really an uninteresting person. But he just found it

00:14:09.500 interesting. And of course, today, as we're going to discover, this is where the action is. You know,

00:14:15.240 Nav was talking about that in the 90s. We're now in the early 2000s. Was it still a little bit of

00:14:21.900 that stigma that Nav described where he was like a totally underperforming loser in his own words?

00:14:28.960 Or was that transition already starting to happen where people saw, wait a minute, there's something

00:14:33.840 going on here? Yeah. Nav was so funny. And I don't think anyone who's ever met Nav thinks he's a loser.

00:14:39.540 I'll say that right off the bat first. I will say that, you know, metabolism as an area was

00:14:45.180 definitely out of favor. And it was a really strange thing because you had the academic current

00:14:52.560 that metabolism was a solved problem. Krebs was the culmination of metabolism research at the same

00:15:01.260 time that metabolic syndrome was becoming worse and worse and worse in the population. When I started,

00:15:08.440 it was still not a real popular topic, metabolism. But I think two things were beginning to shift.

00:15:15.180 shift. One is the fact that people realized that genomics as a standalone was not going to solve

00:15:21.000 health problems, that it was really going to have to be supplemented by other technologies that looked

00:15:27.060 at biochemistry broadly. And metabolomics proved to be one of those. Been enduring, will be enduring.

00:15:34.020 And the second is that, you know, the metabolic syndrome epidemic just kept becoming more and more

00:15:38.200 obvious. So this is kind of where your training as a physician becomes relevant because perhaps more so

00:15:44.840 than somebody who didn't spend four years also doing medical school, you saw the clinical problem

00:15:50.660 that was sort of kicking you in the face, even if it wasn't top of mind to a scientist.

00:15:57.680 I think it's really true. I just benefited so much from the breadth of biology and medicine that you learn

00:16:05.840 in medical school. And I mean, it allowed me to start my lab working on bacteria, which I had

00:16:10.400 never worked on at all as a PhD student because you learn bacteriology is one of the things in medical

00:16:16.320 school. And that was like the perfect starting point for building these technologies with a view all the

00:16:21.680 way to metabolic syndrome and ischemia reperfusion injury, right? These huge medical problems. But I

00:16:27.960 really wanted to start somewhere attractable where we could get firm proof of concept.

00:16:32.380 So for somebody listening to this, let's assume that they have a greater attention span than somebody

00:16:38.000 you're going to walk into at a cocktail party, but obviously not necessarily the depth of

00:16:42.380 understanding of everything we're about to go into. How would you explain metabolism to that person?

00:16:48.600 Metabolism is the process that converts the food we eat into usable energy and the building blocks

00:16:56.040 our body needs to grow or regenerate itself as well as waste along the way.

00:17:00.480 Yeah. And so now how do you layer on the omic piece of that, which is really what we're starting

00:17:06.440 to talk about when people hear the word genomics, they sort of understand what that means. But I think

00:17:12.720 when people hear the term metabolomics, it becomes a little harder to understand. So how would you now

00:17:18.160 layer in that in the context of everything that you're now beginning in your lab?

00:17:22.700 The bulk of activity and metabolism that makes most of our usable energy involves something like

00:17:29.260 a hundred metabolites. And so the first thing we really wanted to be able to do was measure those

00:17:33.460 hundred metabolites really well.

00:17:35.720 Tell people some examples of those, Josh. We take it for granted, but like glucose would be an example

00:17:40.320 of a metabolite. What are some other metabolites that are important to understand if you're trying

00:17:45.400 to study this system?

00:17:47.220 All the amino acids are other fundamental inputs that we get from the diet, like glutamine is a great

00:17:52.840 example. That's a very important circulating nutrient. Other things that come in from the diet,

00:17:58.120 if you have vinegar, you have acetate, or else your microbiome can make acetate that goes in the

00:18:03.340 body. Fats are obviously important input metabolites. And then there are sets of intermediary

00:18:09.320 metabolites. These are things that are like glycolytic intermediates that people may have heard about in

00:18:15.140 biochemistry. Fructose bisphosphate is a famous one of those. Pyruvate, lactate that a lot of people hear

00:18:22.520 about from exercise, members of the Krebs cycle, like citrate is a famous one that exists in our

00:18:29.740 circulation, obviously in citrus fruits. So these are classic examples of metabolites. And then there

00:18:35.320 are the more effector or energy holding metabolites like ATP, NADH, and ADPH.

00:18:42.740 So I kind of interrupted you there, but you were kind of explaining how first you begin with sort of

00:18:47.100 this survey of all of these metabolites. We want to be able to measure all of these. These are kind

00:18:52.540 of the core components of metabolomics. And part of the beauty of metabolism as a system is that

00:18:58.400 with some modest variation, there's almost a singular solution on earth to how metabolism works.

00:19:06.200 And so when we learn to measure the metabolites in E. coli, at the same time, we were really learning

00:19:13.080 how to measure metabolites all the way up to human. There are these basic components of protein and

00:19:20.680 nucleic acids, basic intermediates like fructose bisphosphate that exist at all of these levels.

00:19:26.620 And if you survey relatively comprehensively, there's on the order of a thousand of them that

00:19:32.040 have clear biological function. So it's a big problem, but it's a problem on the scale of, you know,

00:19:37.700 knowing all the kids who go to your high school, not a problem of, you know, knowing everyone in the

00:19:42.580 phone book in New York City, right? So it's a problem that's right at this interface between

00:19:47.040 the human scale and the computational scale.

00:19:51.140 So I wanted to ask you about that. You already kind of anticipated perhaps almost a question,

00:19:54.840 which is, do we think we have the complete solution set here? I mean, we clearly know all

00:20:00.080 of the amino acids. We clearly know all of the intermediate steps and intermediaries period of

00:20:05.620 the Krebs cycle. Do we think that that means we actually know every single metabolomic element,

00:20:12.680 or is there a chance that there are others out there that we don't know because we haven't looked

00:20:18.900 for them or they're very short-lived, for example, and we haven't stopped and looked at reactions

00:20:25.820 closely enough or studied the kinetics hard enough? This is almost a naive question in a way,

00:20:29.480 but I've never actually thought about it until now.

00:20:31.100 It's a great question. And we keep discovering new metabolites. Groups around the world keep

00:20:37.320 discovering new metabolites. I would say it's an interesting yin and yang because there's this

00:20:42.820 steady accumulation of new metabolites, but I don't really think there's been a completely new

00:20:49.120 and obviously important metabolite yet this century. At some level, in terms of the metabolite

00:20:56.540 finding part of the problem, things were wrapping up around the time of Krebs, the most important part

00:21:03.160 of the work anyway. But in terms of understanding how the system really works and how we can choose

00:21:09.640 the right diet to be healthy, given our genotype, given the disease we're fighting, we haven't

00:21:14.620 scratched the surface yet.

00:21:16.820 How many of these metabolites are really tightly regulated, a la glucose, versus not that regulated

00:21:23.960 at all? They can kind of fall to zero and they can... Meaning, how many of these things can change

00:21:29.000 by log orders all over the place and how many are regulated so tightly that if you just fall a little

00:21:36.400 bit out of that range? One of the things I try to explain to people when I explain regulation and

00:21:41.740 homeostasis is I love using pH as an example because the pH spectrum runs from basically zero to 14,

00:21:47.600 neutral being seven. But anybody who's taking care of a patient in the hospital knows 7.4 is where we

00:21:55.660 live as an organism. Almost unsurvivable to have an acidosis that goes below seven or an alkalosis

00:22:02.420 that exceeds about 7.7. So for a system that runs basically zero to 14, the fact that we can't as a

00:22:10.100 species survive outside of 7 to 7.6 or 7.7 talks about something that is so tightly regulated.

00:22:18.240 So in that field of metabolomics, which ones behave like pH and which ones don't?

00:22:23.940 pH is such a great example, right? You have this giant logarithmic scale from zero to 14, right? And so

00:22:30.480 even when you talk about 7.1 to 7.4, you're talking about something of a two, three-fold change.

00:22:38.280 It's about a two- to three-fold change in acid concentration.

00:22:41.920 A lot of metabolites, the important ones, live typically in that two- to three-fold range as

00:22:48.220 being the preferred range. Some of them, there's a lot more active regulation like glucose. Some of

00:22:56.340 them, there's kind of relatively passive processes that tend to keep them in that range. Then of course,

00:23:03.020 there are all sorts of other metabolites that may be some cool secondary metabolite that's made by a

00:23:07.980 plant. And some of us eat it, some of us don't eat that plant. And so some of us may have a lot,

00:23:12.240 some of us may have none. But for the big ones, the biochemistry textbook ones,

00:23:16.900 this kind of few-fold range in the bloodstream is common, healthy place to be.

00:23:22.660 Are there common and consistent tools that the body uses to regulate? Are there principles that

00:23:28.620 the body just adopts over and over and over again in the form of this regulation?

00:23:31.900 There's one most important principle, and that's when it's there, use it up. In a physics-speak,

00:23:40.380 you could say this is a linear consumption of circulating metabolites. Or in chemistry-speak,

00:23:46.720 people call this mass action. Just whatever mass is there, you tend to take it in. A lot of what you

00:23:52.540 eat after a little bit of processing enters the bloodstream. And then it's the job of tissues that

00:23:58.240 need these ingredients to use them and use them at first blush in proportion to their availability

00:24:04.340 in the blood. Are there examples where that regulatory mechanism is not the preferred way

00:24:11.240 to manage them? Well, there's a lot of regulation layered over top of that in order to make the body 0.97

00:24:17.660 work. The most important regulatory hormone in mammals, I'm pretty convinced, is insulin.

00:24:22.520 You know, there are two ways to look at insulin, I think. And there's probably the way that comes

00:24:28.300 to mind first to you, which is insulin is a hormone that acts to control elevations in blood sugar. And

00:24:36.580 it does that at the highest level by promoting uptake of glucose and preventing production of glucose.

00:24:44.380 glucose. But there's an alternative way to look at insulin. And that's that we've evolved mainly to

00:24:52.060 be able to survive lack of nutrients, okay? That this was the strong selective pressure on animals and

00:24:59.800 mammals. And that storing fat is very precious. And that insulin is a hormone that says you don't have

00:25:07.460 to use fat right now, okay? And so it senses that there's enough carbohydrate around and therefore

00:25:14.360 it's safe to not release free fat from your adipose tissue. Does that mean that you think that

00:25:22.220 an equally important role of insulin is not just the disposal of glucose into muscle and the cessation of

00:25:30.000 glucose production in the liver, but you're saying it's equally important as a signal to stop

00:25:36.880 lipolysis to keep your fat in its fat stores, meaning save this for a rainier day because you

00:25:43.560 actually have the glucose here that I'm going to deal with? That's certainly how I look at insulin

00:25:48.660 right now. There's little doubt biochemically and medically the suppression of lipolysis is a

00:25:56.140 primary, perhaps the primary function of insulin. Going back to the broad strokes of metabolomics,

00:26:03.940 you alluded to it briefly without, I think, using the term, but what did we know about the flux of

00:26:09.840 these things? When I think back to even my biochemistry 25 years ago, Stryer's textbook,

00:26:17.060 which is the classic textbook, at least it was then, I imagine it's one of them still today.

00:26:21.680 We really studied it in a static way. And I'm guessing that that's one of the first questions you

00:26:27.980 went after, right? Which is what's the movement? What are the derivatives with respect to time of all of

00:26:33.180 these things? Maybe expand a little bit on this idea of fluxomics.

00:26:37.540 Metabolism is a system in action. And I think this kind of static view of metabolism, which is

00:26:44.240 probably never a view that Stryer ever had in his mind, but that got codified in the textbooks as one

00:26:49.960 that killed metabolism in a way as a topic of excitement. Metabolites are intermediates in the

00:26:57.200 process of converting what we eat into usable energy and protein biomass and these things.

00:27:04.480 They're really relatively low in abundance and they're flowing very, very fast. So they're

00:27:11.260 completely different than parts of our body like neurons that are going to sit there maybe for our

00:27:15.760 entire lifetime. Here, the metabolites are meant to be made and used somewhere on the timescale,

00:27:21.420 depending on the metabolite, of roughly a second to roughly an hour maybe for metabolites in the

00:27:28.300 bloodstream. And all the action is in the flow. It's really understanding where things are coming

00:27:34.520 from, where they're going, where we can learn about how metabolism works.

00:27:38.660 Let's again just use glucose because one, it's ubiquitous. Everybody gets it. It's essential for life.

00:27:44.640 But it also offers, I think, a beautiful portrait in velocity. I just had my blood drawn yesterday. I draw

00:27:50.540 my blood about every two months and tube of blood comes out and let's say my glucose, because that's

00:27:56.640 a snapshot, right? That's literally in that moment, out comes five tubes of blood and it's going to look

00:28:01.820 for a whole bunch of things. But one of them is glucose. And my glucose, because I did a finger

00:28:04.820 prick at the same time, my glucose was 89 milligrams per deciliter. Can you explain to people what that

00:28:11.480 actually means? What does it mean that my glucose at that moment in time was 89 milligrams per

00:28:17.580 deciliter? Well, I'm sure you were smiling about it. It's a super healthy blood glucose in terms of

00:28:22.960 the level. But when you think of that absolute amount of glucose, right, if you took all the

00:28:27.440 glucose in your bloodstream at that level, that's a few minutes of glucose or energy.

00:28:33.180 Yeah, it's probably what, four or five grams of total glucose, 20 calories worth.

00:28:38.500 Exactly. So that has to be constantly replenished in order to feed your brain and the other tissues,

00:28:44.920 activated immune cells that depend on glucose. Now, this is what to me is remarkable. If I had

00:28:52.940 done that same test, Josh, and I had come back, let's make the math easy and say it was 90,

00:28:58.760 we'd still say, great, you're healthy, your fasting glucose is 90. Now, let's say I had come back and it

00:29:03.520 was 180 milligrams per deciliter, my fasting glucose. There's a disease that I would immediately

00:29:09.400 now know that I have. That disease is called type 2 diabetes. What's the absolute difference in the

00:29:16.140 amount of glucose in my bloodstream? It went from being 5 grams to 10 grams? Seems like a really

00:29:22.780 trivial amount. Why is it that the body in the person without diabetes seems to be able to keep it at,

00:29:30.660 you know, 80 to 100 milligrams per deciliter overnight while you're fasting, but in a disease

00:29:36.520 that's going to more than double your risk of mortality and increase your risk of cancer,

00:29:41.980 Alzheimer's disease, cardiovascular disease, I mean, it's really a problem for your health.

00:29:47.300 It's only doubling the amount of glucose in there and it's still a relatively trivial amount that

00:29:52.020 needs a constant, constant update. How can we explain this delta of that seems so trivial in the

00:29:59.580 absolute amount that could be consumed in just an extra couple of minutes, but yet the steady state

00:30:06.020 is still off by this factor. What's going on and why? When you think about why is that a disease

00:30:12.060 problem with only a two-fold excursion, then you think that, right, the system has been built

00:30:17.960 to have about the most circulating glucose that you can have safely. And I think a lot of the really

00:30:25.760 important metabolites have been kind of pushed to the edge this way. And so we know that there

00:30:32.180 are deleterious protein modification reactions, glycosylation reactions that occur when glucose

00:30:38.960 gets above this point. So we're kind of in evolution pushed right up to the highest and

00:30:44.740 non-problematic glucose. We didn't do that for a lot of other metabolites, and that's why there's not a

00:30:50.360 lot of, part of why there's not a lot of other diseases like diabetes. So that's part of the answer

00:30:56.080 is evolution didn't build a lot of wiggle room for your glucose to safely rise because having that

00:31:02.480 good amount of glucose circulating is really valuable. Every time your heart pumps, it's sending

00:31:08.740 that amount of glucose to tissues, and that's productive. On the flip side, obviously there's

00:31:14.020 a broader set of derangements in the body to produce this two-fold excursion in glucose. And this

00:31:22.160 has to do with things going wrong in fat and fat handling. And so that's part of this whole metabolic

00:31:28.160 syndrome that leads to the full set of downstream health consequences. Again, I think this speaks to

00:31:34.840 why the flux problem is the more interesting problem than the static problem. Because if you just think

00:31:39.700 about this example in the static context, you would say, okay, well, at 7.04 and 3 seconds AM,

00:31:47.680 your blood sugar is 190. But three minutes later, if nothing changes, meaning if your liver doesn't put

00:31:56.040 too much glucose back into circulation, you'll be fine. So the problem is not that your blood sugar is

00:32:03.060 too high in that moment. The problem is the liver assumes in part that that's the right level, and it

00:32:09.680 continues to do it. Because at that moment when you're not eating, that is the only source by which

00:32:13.380 glucose is getting into the bloodstream. So we're maintaining this elevated cycle.

00:32:17.480 It's everything from gluconeogenesis, hepatic glucose output. I mean, all of these things

00:32:21.880 continue to stay unregulated. And I think that's only really appreciated when you think of time and

00:32:26.920 the passage of time. I think one thing that's really interesting is that you can have, depending

00:32:33.600 on the details of how processes are tuned, the same amount of production and the same amount of

00:32:38.760 consumption. And these have to be balanced for your glucose to stay anywhere close to steady.

00:32:43.800 So in a diabetic, production and consumption are balanced, and a healthy person, production and

00:32:48.620 consumption are balanced. And they can even be the same amount of production and consumption.

00:32:53.720 But it can just be that you need a higher amount of glucose to achieve that same balance of production

00:33:01.460 and consumption in the diabetic. And that reflects, in my opinion at least, underlying issues with how

00:33:09.260 fat is handled, that you either need more glucose to induce more insulin in order to suppress lipolysis

00:33:21.580 in the diabetic, or more glucose to out-compete fat to get burnt in tissues. A lot of what's setting

00:33:29.500 blood glucose is competition between glucose and fat. This is a very old idea called the Randall hypothesis.

00:33:36.140 But I think there's a lot of truth to it. We have a lot of new data that's consistent with it.

00:33:42.160 A lot of these issues come back to making room for glucose to be burned by controlling the amount

00:33:49.060 of fat that's being used by tissues. Can you state for folks the Randall hypothesis? I'd love to actually

00:33:55.000 talk a little bit about the more recent data. I mean, this is over 50 years old, isn't it?

00:33:59.940 I'm a terrible historian, so I'm going to trust you on that one. But the essence of the hypothesis is

00:34:06.400 that fat is somewhere between a preferred and the preferred fuel for tissues. And there's competition

00:34:14.220 between carbohydrates, classically glucose, and fat for burning. And so when fat is available,

00:34:22.180 then glucose tends not to be burned effectively. And that's a possible cause of diabetes.

00:34:27.580 When you say fat, you don't mean fat within adipose tissue. You mean fat that's available for use.

00:34:34.800 Fat that's available for use. And so that can come in multiple forms. The simplest way to think about

00:34:39.140 it is free fatty acids that are floating in the bloodstream. And that may be the most important

00:34:43.320 form of it. Also, adipose stores within tissues, not subcutaneous white adipose, which is typically a

00:34:51.040 healthy place to store fat molecules. But you can end up with what people call ectopic fat. For example,

00:34:57.440 droplets of fat building up in muscle. And when those are there, they can compete with carbohydrate

00:35:02.740 for being burned. Or you can have breakdown of lipoproteins from the bloodstream, things like VLDL.

00:35:11.280 We desperately need to have broken down in order to have a good HDL and a low LDL cholesterol.

00:35:17.880 Tell me about some of the more recent evidence around why that hypothesis may be more compelling,

00:35:24.920 even so than when it was proposed.

00:35:27.260 We've been doing experiments that look at what are things that can suppress glucose use in tissues.

00:35:34.880 One thing we see is it's just very clear that fat does this. We're certainly not the only people

00:35:39.880 to do this. I think there's a long history of this, but it maybe hasn't been adequately appreciated,

00:35:45.880 just how fundamental that result is. And if you turn off lipolysis different ways,

00:35:51.660 then you rapidly induce glucose consumption. And if you provide other alternative fuels,

00:35:58.340 and we've learned that lactate is a very important circulating fuel. And so it also will compete with

00:36:04.620 glucose to suppress glucose use. So the fact that you can have multiple different types of fuels,

00:36:12.440 either fat or lactate, and any of them will suppress glucose use really makes me believe in this kind

00:36:18.440 of competitive nutrient environment. And that that plays a central role in determining whether you

00:36:23.580 clear or don't clear glucose and how high your glucose has to go basically in order to be cleared.

00:36:28.720 So let's talk a little bit about lactate, because this is one of those things where

00:36:32.200 now given how much I think about lactate, read about lactate, and had a number of podcasts where we

00:36:37.160 get into some detail. Either I was asleep through part of medical school, or it just really wasn't

00:36:42.400 presented in anything other than the following. When your demand for ATP gets high enough and quick

00:36:49.340 enough, you're going to basically take glucose. And when you turn it into pyruvate, rather than take

00:36:55.640 the efficient path of shuttling pyruvate into acetyl-CoA through the Krebs cycle, where you can

00:37:01.660 generate lots of ATP requiring oxygen, you're going to take a quicker path that's less efficient, but

00:37:08.460 doesn't require the same cellular oxygen, and you'll turn pyruvate into lactate. You won't get

00:37:13.440 nearly as much ATP, and you'll also tend to generate a lot of lactate, which tends to gravitate with

00:37:19.260 hydrogen ions, which tends to kind of poison the muscle a little bit. And that's why it becomes

00:37:24.180 rate limiting in terms of how long you can sustain that level of output. Maybe explain today why that's

00:37:29.780 the tip of the iceberg in a generous sense of the term. I think it's all actually really important

00:37:35.780 stuff. It's just only, as you say, part of the picture. And I think the other part of the picture

00:37:41.160 is that mammals have been wired to use lactate as a major circulating nutrient. It's a super, super

00:37:50.120 fast turnover nutrient. So when you think about that glucose, and you're having, you know,

00:37:55.880 a few minutes supply circulating in your blood, lactate, you have even shorter supply than that.

00:38:02.760 It's constantly being made, released into the bloodstream, and consumed. And it serves as an

00:38:08.900 almost universal nutrient. They're transporters that'll carry it into virtually any cell in your

00:38:16.360 body. These are the MCH, or is it the MCT transporters?

00:38:20.820 These are called MCT transporters. It stands for monocarboxylate, because lactate has one carboxylic

00:38:27.360 acid, if you think of it as a chemistry perspective. And so those transporters are ubiquitously

00:38:32.220 expressed, and they allow lactate basically to go everywhere.

00:38:37.400 Which, by the way, Josh, that already differs from kind of how we learned it in biochemistry class,

00:38:42.480 which was all that lactate goes back to the liver, and the Cori cycle turns lactate back into glucose,

00:38:48.940 and then just exports it down the glucose pathway via hepatic glucose output. And you're saying,

00:38:54.220 I'd like to understand when that happens versus when each other tissue says,

00:38:59.020 oh, great, more fuel, let me take in this lactate.

00:39:02.200 I think the really important thing about lactate is that glucose penetration into tissues is actually

00:39:08.040 heavily regulated. It has to be heavily regulated so that if we go through a period of having low carb

00:39:13.920 intake, there's still glucose preserved for the brain and for other cells that particularly need it.

00:39:20.980 And lactate is the universally available form of carbohydrate. In a healthy heart, at least in

00:39:30.780 the fasted state, it basically will not touch glucose. But it will use lactate as fuel.

00:39:37.160 So it's preferred fuel is free fatty acid? Would that be-

00:39:39.860 It's preferred fuel is free fatty acids. It probably also gets some fatty acids from lipoproteins,

00:39:45.440 and it definitely uses lactate and also things like ketone bodies. This is like a very clear

00:39:51.680 example of a tissue other than liver that net consumes lactate, just using it as a fuel to have

00:39:59.160 access to carbohydrate energy. Now, lactate is another metabolite that I pay a lot of attention to,

00:40:05.040 Josh. So as regularly as I'm checking my glucose, I'm checking my lactate. And unlike glucose,

00:40:11.420 the range is much greater. The lowest glucose I've ever measured in myself is probably 50 milligrams

00:40:18.980 per deciliter. And the highest, not including the time Jerry Riven had me do an insulin suppression

00:40:25.440 test at Stanford. And I almost died, actually. This is actually a ridiculous story because one of the

00:40:31.320 IVs got blown and we didn't know that they were pushing glucose because I was just getting so

00:40:37.340 hypoglycemic. I could feel it. You know, you learn in medical school what hypoglycemia feels like.

00:40:43.100 And when you start sweating really profusely, and this was like nothing I've ever experienced,

00:40:47.900 it felt like a bucket of water got dumped on me. And I was like, they got to push glucose. And I could

00:40:53.760 feel the IV was blown. Anyway, to make a long story short, when they finally corrected it, my glucose got up

00:40:58.900 to 240 milligrams per deciliter. Call it 250. That's a 5X range. But with lactate, I mean,

00:41:06.600 I've measured it as low as 0.3 millimole and as high as 20 millimole. So that's a 60-fold difference.

00:41:14.940 Big range, but it probably depends a lot on your physiological state.

00:41:18.540 Well, of course, the 0.3 would be at rest and fasted. The 20 is kind of an all-out two-minute

00:41:27.680 effort. But the point here is that's a much bigger range. Is this regulated? Is there an upper limit to

00:41:33.800 how high lactate can go? Or is it simply how much pain you can tolerate in terms of what is necessary

00:41:38.900 to generate lactate? Is there truly an upper limit? I'm honestly not sure. You may be right on the pain

00:41:45.760 side of the scale. This has to do with how fast its production and consumption are, right? So you can

00:41:52.100 have that excursion to 20. That can be cleaned up in a few minutes if you're actually, you know,

00:41:57.960 completely resting. But it goes down very quickly. Yeah. This is a very flexible metabolite this way.

00:42:05.480 I remember first reading about this in about 2011 where people were starting to say,

00:42:10.640 hey, neurons might like lactate besides glucose. Because at that point in time,

00:42:14.880 there were really only two fuels that a neuron would ingest, right? So under normal circumstances,

00:42:20.080 it was exclusively glucose. And then George Cahill showed in the 60s, yeah,

00:42:24.920 but if you starve somebody, you can turn up to 60% of that fuel stock into beta-hydroxybutyrate.

00:42:32.100 I think it was BHB. Maybe it was acetoacetate, but it was a ketone. So you'd be maybe 60-40 in favor

00:42:37.180 of a ketone to glucose. But that was really it. And then there were these kind of whisperings in these

00:42:41.960 animal studies that suggested, no, actually neurons will consume lactate. Where are we today on that front?

00:42:47.160 I think it's still very unclear which cell types in the brain are the lactate consumers versus

00:42:53.260 lactate producers. Certainly there's lactate use in the brain.

00:42:57.880 And is it more astrocytes, neurons? Do we know?

00:43:00.040 I think it's really an active area investigation. I bring biases to it, but I don't bring answers.

00:43:06.480 You know, my bias is that we are a neuron-centric form of thinker, right? And we didn't evolve to make

00:43:13.720 glucose a unique brain fuel in order to feed astrocytes. We did it to feed neurons. I do think

00:43:19.980 there's a special neuronal dependence on glucose. But lactate goes everywhere. So it probably goes

00:43:26.240 into both astrocytes and neurons as a fuel in the right circumstances. And it probably can be

00:43:32.520 excreted from both as a waste, depending on exactly what activities are required in the brain at that

00:43:38.060 time. And that's really the beauty of lactate is it allows you a tremendous degree of flexibility

00:43:43.060 that wouldn't exist otherwise. And this was actually thought about a lot by a guy named Brooks

00:43:49.280 at Berkeley, who- George Brooks.

00:43:51.460 Yeah, recognized ubiquitous potential for lactate as a fuel. We were able to contribute to that story

00:44:00.120 by really showing it, using mass spectrometry to make it crystal clear that this usage happens

00:44:06.720 throughout the whole body. What is the evolutionary reason in your mind for why the body would allow

00:44:15.860 most tissues to love lactate as a fuel directly versus just having the liver mop it up at the same

00:44:23.420 kinetic rate, turn it into glucose, and shoot that glucose out? Is there an obvious reason for why the

00:44:28.900 current strategy is a better one? It's not an easy answer, but I think there are strong reasons.

00:44:34.660 And I'll say that we've lately done experiments in yeast, actually. Yeast make ethanol as waste.

00:44:43.380 And people always, I think, assume that yeast face this exact same choice that you talked about when

00:44:50.060 you get to the level of pyruvate. Either do you spit it out as a redox-balanced waste in humans that's

00:44:56.460 done as lactate and yeast that's done as ethanol, or do you take the pyruvate into the TCA cycle?

00:45:01.280 We see that going all the way back to yeast, that's a false choice. The default is to spit out the

00:45:08.440 redox-balanced waste. And then you can always pick the waste up and reuse it if you need energy from

00:45:15.080 the TCA cycle. And so I think this goes back to the very earliest days of eukaryotic life, basically,

00:45:22.660 that you want to be able to run glycolysis whenever you need to run glycolysis. So use glucose whenever

00:45:29.580 you need to use glucose. That takes you to pyruvate. You've created a redox problem because you have

00:45:36.420 electrons from the glucose that are not sitting on the pyruvate. And the first priority is always to

00:45:41.860 solve that redox problem that's achieved in our bodies by spitting out lactate. And you don't really

00:45:47.820 want to hold that problem within cells in your body. You want to get that all the way under the

00:45:52.160 circulation. So every cell in your body can work on this master metabolic challenge of keeping, you

00:45:58.140 know, electrons balanced. Then whoever needs energy, okay, and these electrons are a super valuable source

00:46:04.560 of energy, can pick them up in the form of lactate. There's been kind of this false coupling of oxidative

00:46:13.600 and glycolytic metabolism in the way biochemistry is taught, when really our bodies, eukaryotes all

00:46:21.480 the way back to yeast, are really designed to be much more flexible, to allow these two processes to

00:46:26.240 happen in yeast completely independently, because they really can just spit out ethanol to the

00:46:30.980 environment. In us, quasi-independently. So independently at the level of individual cells

00:46:36.560 in our body. So none of them faces this pressure. And that's really good. So if you have a bout of

00:46:42.360 hypoxia, okay, you can release lactate and elsewhere in the body, the problem can be cleaned up. Now in

00:46:48.060 our bodies, it has to be cleaned up within the body somewhere because we don't have any master release

00:46:53.780 valve for this. So all of our cells together have to solve this problem. Meaning the way that yeast can

00:47:00.920 literally eject ethanol from their cell and get it away, we can't emit lactate from our body. We can

00:47:07.720 emit it from a cell, but it's still part of the broader system. Exactly. And that's when you get into

00:47:12.100 medical problems like lactic acidosis. If you have a very fundamental metabolic deficiency, or if,

00:47:17.440 you know, God forbid someone put a bag over your head and you couldn't breathe, then you end up in

00:47:21.500 this crisis of redox imbalance. But we distribute that problem across the body through this concentrating

00:47:28.800 lactate in and out of cells and letting whatever cells that need carbohydrate energy use the lactate.

00:47:35.620 The system would be way, way less flexible if, you know, only the liver could clean this up. It would 0.83

00:47:41.960 be also way less commensurate with effective burst exercise. You know, the heart is super well perfused.

00:47:48.920 If less well perfused muscle that's far from the heart, okay, so it's hard to get oxygen there,

00:47:54.240 is making a lot of lactate. Of course, it's very advantageous at that moment for the heart,

00:47:59.360 which is sitting on more oxygen than it needs to use lactate rather than fatty acids,

00:48:03.260 which are better long-term things to store for the future anyway. So it's way, way better to have

00:48:08.240 a system designed this way. And is that regulated then locally? Is that regulated at each cell? Like

00:48:13.880 how is that decision made? Because how does that myocyte in the heart know the energy of the entire

00:48:21.640 system so that it can make the decision that in the short run counterintuitive? Seems medically or

00:48:28.100 maybe textbook med school counterintuitive, but it's physical chemistry, pure intuition.

00:48:35.040 Lactate goes up, it gets burnt. But how is the decision made? Are you saying it's just made on

00:48:40.980 mass balance and availability of substrate? It's made on mass action and availability of

00:48:45.120 substrate. So there's no decision. It's basically a gradient problem across the board.

00:48:50.080 If you have too much lactate, it flows out. If you are short on energy, it flows in.

00:48:54.560 One thing that I've become very interested in clinically is the implication of fasting lactate

00:49:02.200 levels in the population. So if you measure a person's lactate level first thing in the morning,

00:49:08.620 you're going to see quite a bit of variability. And it seems to be proportional to their metabolic

00:49:14.320 health. The higher that number, the less metabolically healthy they are. It's not uncommon in

00:49:20.420 someone who's insulin resistant to see fasting lactate levels approaching 2 millimole with no

00:49:26.880 activity. Whereas in a healthy individual, it'll be below 0.5 millimole. What do you think that tells

00:49:33.020 us about fuel partitioning and this problem of metabolomics? I think there's a correlation between

00:49:40.180 fasting glucose and fasting lactate. But lactate is maybe harder to measure, but perhaps even more

00:49:46.500 intimately tied to the essence of metabolic dysfunction molecule. It reads out a few things.

00:49:55.240 When lactate is high, it reflects the fact that during these times of fasting, when glucose is not

00:50:02.200 really supposed to be used much, you're still using too much glucose, converting too much of it to

00:50:08.300 lactate. And at the same time, your lactate clearance system isn't working very well. And typically,

00:50:13.780 that's because you're having competition between lactate and fat to be burned. This all feeds into

00:50:20.300 the syndrome of diabetes. Now, there's another interesting push observation that I've made,

00:50:25.600 which is I wake up in the morning, check a lactate, it's 0.4 millimole. I eat the biggest carbohydrate meal

00:50:33.260 I can ingest. Don't lift a finger other than to feed myself. I recheck my lactate in an hour,

00:50:41.020 it's 1 millimole. Why did that happen? We can understand the biochemistry, which is I have

00:50:46.820 more glucose to metabolize. But this gets back to your point of med school biochemistry would suggest

00:50:53.840 my lactate should not have gone up. I'm taking glucose. I'm making pyruvate. I have endless cellular

00:51:01.140 oxygen. I should be running that pyruvate through the Krebs cycle and I shouldn't see any uptick in

00:51:06.640 lactate. But that's exactly what I don't see. Circulating lactate is an intermediary in glucose

00:51:12.720 catabolism. That's just the way the body works. It's not what we were taught in med school. 1.00

00:51:17.440 You have a sense of how it's being taught today? Do you get the sense that biochemistry students at

00:51:22.680 Princeton and Stanford today are being taught what we were taught with respect to this sort of more

00:51:28.820 rigid model of lactate and as a metabolite? Well, I'm probably chipping away at it at Princeton,

00:51:33.920 but I don't know how much it's shifting at the medical education level yet. Probably thank God

00:51:39.420 I haven't sat through those classes at Stanford again. I think it's something that we should see

00:51:43.920 shift and I hope we see the next generation of biochemistry textbooks talk about circulating

00:51:49.620 lactate as an intermediary in glucose catabolism. I think that's a really fundamental thing for people

00:51:55.460 who want to just think about metabolism accurately to know.

00:51:58.660 And I think it's a very interesting thing to consider from a prognostic standpoint. When you go

00:52:05.560 back and look at Jerry Riven's five criteria for what was then syndrome X and what is now metabolic

00:52:11.280 syndrome, fasting glucose is still one of them. You could make a case that fasting lactate would be

00:52:17.160 more telling. I think the challenge with lactate is that it is a metabolite that can get up and down

00:52:23.660 faster. And one response to stress is to rapidly convert glucose into lactate. That's just part of

00:52:30.640 your body activating. But of course, there are people who have stress at a blood draw. And because

00:52:35.140 lactate is a little bit more fluctuating this way, they're going to be pros and cons medically in terms

00:52:40.160 of using it as a biomarker. I don't think our problem with metabolic syndrome anyway is diagnosing it.

00:52:45.640 Our problem is preventing it.

00:52:47.200 Yeah. Although what I would argue is I think we treat metabolic syndrome too discreetly and I think

00:52:53.740 we come to it too late. I think we should be looking for things far before you actually have

00:53:00.280 hypertension and truncal obesity and dyslipidemia and hyperglycemia. And I do wonder with nothing

00:53:08.560 other than just intuition if lactate dysregulation, for lack of a better word, might be one of the

00:53:14.780 earlier canaries in the coal mine. I totally agree with that.

00:53:18.680 I want to go back to something that we've talked about a couple of times. You've mentioned it in

00:53:23.800 passing. You and I know what it's about. But I think it's such an important part of where we're

00:53:30.460 going to go in a discussion that I almost need you to go into full prof mode and really explain two

00:53:37.900 things, which are obviously highly related in a moment you'll see. The first is how the electron

00:53:44.240 transport chain works. What is the Krebs cycle doing and how is that feeding into this massive

00:53:50.200 generation of energy currency? And specifically, can you talk about it with special attention to

00:53:57.080 the concept of redox? I would encourage you, Josh, to take as much time as you need because the more

00:54:03.140 the listeners understand this, the more they'll be able to understand NAD, NADP, NADPH, NR, NMN,

00:54:12.760 all of these other things that people really care about. But I think unfortunately, they've been

00:54:17.460 conditioned into very glib understandings of these things, which I think are serving no one

00:54:23.980 any benefit without actually going back to understanding the root of this problem.

00:54:28.440 Think of it this way. Fundamentally, you eat three macronutrients, carbs and protein and fat.

00:54:37.700 And in a healthy adult, first approximation, every carbon atom that you eat in any of those three forms

00:54:48.240 needs to exit your body as exhaled carbon dioxide. And all of that exhaled carbon dioxide to first

00:54:57.060 approximation is made in the TCA cycle. The main way that nutrients flow into the TCA cycle to become

00:55:05.360 carbon dioxide is first turning into pieces that are two carbon units in size. And so from

00:55:13.400 carbohydrate, the basic flow is glucose to lactate, and then lactate to pyruvate to a two carbon piece

00:55:22.720 that goes into the TCA cycle. Fat is basically composed of pre-assembled two carbon pieces. So they

00:55:29.200 just get chopped up two carbon pieces at a time. And the protein part is a little more complicated. We can

00:55:34.980 probably skip it. Worth just sort of noting, Josh, that protein really, the primary role of protein

00:55:40.960 is actually the nitrogen side, which we're putting into these amino acids that are building blocks.

00:55:46.480 It's really less of an energy substrate, but it does have that carboxylic acid on it that still

00:55:52.580 has to go through this cycle and be exhaled. In other words, that's why we'll skip it for now,

00:55:57.060 because it's really not a significant energetic component, right? I think it really depends on the

00:56:01.540 kind of diet you eat. True. If you're on a carnivore diet, then it's probably a different

00:56:05.280 situation. It's a very interesting side discussion. But ultimately, unless you're gaining protein mass,

00:56:10.680 which of course, wonderful for us guys anyway, when that happens, typically at least societally

00:56:15.800 smiles on it. But other than that, you know, whatever amino acid carbon you take in in the form of protein

00:56:22.200 has to be balanced with also amino acid catabolism. At that level, it's not that different than carbs and

00:56:28.200 fat. It's just a little different in that it can enter the TCA cycle sometimes also as four carbon

00:56:33.600 pieces. But a lot of amino acids are broken down into these same two carbon pieces. There are just

00:56:38.680 20 of them. So no one wants to hear a discussion of how all 20 of them get chopped up. So ultimately,

00:56:45.160 you end up with these two carbon pieces. They congeal with a four carbon piece, and that makes citrate.

00:56:52.500 One of the problems with the Krebs cycle is that it has three names. The Krebs cycle, in honor of the

00:56:58.940 amazing biochemist who played a key role in figuring it out. The citric acid cycle, and that's named for

00:57:05.980 this condensation molecule of the four carbon and the two carbon piece citrate. Or the TCA cycle, and TCA

00:57:13.980 is tricarboxylic acid, and that's because citrate has three carboxylic acids, okay? So you have this cycle

00:57:21.700 that unfortunately has three names, but it's probably three times as important as anything else in

00:57:26.560 metabolism, so maybe it's fair. Ultimately, as this cycle turns, it's going to spit off the two carbon

00:57:36.400 pieces that came in as carbon dioxide. And in so doing, it's going to take the electrons that were part

00:57:47.340 of those two carbon pieces and pass them to this famous cofactor NAD to make NADH. That H stands for

00:57:57.680 hydrogen, and that hydrogen is really one proton and two electrons. And so this is another confusing

00:58:04.860 nomenclature thing that you just can think of that H, even though it may sound to those who've taken

00:58:11.340 freshman chemistry like H+, like acid. This is an H with two electrons stuck to it, so it's really

00:58:17.340 what we call hydride or electrical form of chemical energy. Then NADH that's made from NAD there is what

00:58:28.100 feeds into the electron transport chain, and those electrons then flow through a series of proteins

00:58:36.400 that sit in the inner mitochondrial membrane. The mitochondria have two membranes. The outer one is

00:58:43.480 kind of leaky and kind of not so important. The inner one is super tight and has a ton of regulation

00:58:50.520 in it, and most importantly, can be used to pump protons to one side or the other. And ultimately,

00:58:57.940 it's the pumping of protons out of the mitochondria that's the function of the electron transport chain.

00:59:03.520 And in this kind of metabolic flux way that we talked about earlier, the protons that get pumped

00:59:07.640 out just flow right back in. But as they flow back in, they turn a turnstile. And as that turnstile

00:59:13.860 turns, it squeezes ADP, an inorganic phosphate, physically squeezes them together to make ATP,

00:59:22.440 the master energy currency that we use to power our neurons for thinking, our muscles for moving,

00:59:28.460 and so on. One of the things about this system that is just so beautiful is the transition from

00:59:34.580 chemical energy to electrical energy back to chemical energy. I've tried to explain this to

00:59:40.800 my daughter. She's 13. She's not fully in love with it yet, but I know at some point it'll be a

00:59:47.200 more fun discussion. But it really is a miracle, right? So much of biology just seems like it's hard

00:59:52.980 to believe it all worked out. But if you were going to rank all the things that I can't believe

00:59:58.040 it worked out, this has got to be in the top five. Let's go back to the basics again. You eat a piece

01:00:04.080 of bread. You're eating glucose. It has these carbon to carbon bonds and carbon to hydrogen bonds

01:00:10.820 and some carbon to oxygen bonds. Now, refresh my memory, but carbon to oxygen is not a very energetic

01:00:17.360 bond, right? CO double bonds are spectacular bonds. They're super high energy, but that's where life,

01:00:25.780 I didn't say life, that's where physics and chemistry want to flow to. They want to make

01:00:31.620 these high energy bonds. And in making high energy bonds, you can release a lot of energy.

01:00:36.760 Those are bonds that are very energetically favorable. So they're the end state. It's the CH bonds,

01:00:44.040 as you're alluding to, that start out energetically loaded, okay? They're less energetically good in

01:00:50.020 and of themselves. So they have the potential to become something better. I'm glad you're adding

01:00:54.680 this level of chemical rigor to this. The point I want to make is these carbon-carbon, carbon-hydrogen

01:00:59.820 bonds have this potential that this entire cycle with three names that's so wonderful basically

01:01:07.260 liberates. It basically says, we're going to take that chemical energy and we're going to liberate

01:01:13.680 it through electron transferring apparati. And then at the last second, we're basically going to

01:01:20.100 quickly shunt it right back into a chemical bond, which is the P binding to the ADP to make the ATP.

01:01:28.460 And now we have this energy currency that is going to go and do its own thing. And it has lots of

01:01:33.020 different ways that it unleashes itself. So explain to people the difference between oxidation and

01:01:38.200 reduction in chemical terms, because I think people have to at least hear once what's an oxidation

01:01:43.620 reaction, what's a reduction reaction, and then why we use the term redox synonymously with these two.

01:01:49.460 Oxidation and reduction are always coupled, okay? And they refer to movement of electrons. And so

01:01:55.940 when electrons go from substance A to substance B, the one that gives up the electrons is oxidized.

01:02:03.700 It's subject to oxidation. The one that receives the electrons is reduced. It's the subject of reduction.

01:02:10.260 Let's talk about redox pairing. So you've already brought up NAD and NADH. So talk about how

01:02:16.680 oxidation reduction pairing works with those two to facilitate the electron transfer down this

01:02:22.120 lovely chain of the inner mitochondrial membrane. This is a pair where NAD is the oxidized form.

01:02:29.320 NADH is the electron holding or reduced form. It normally exists in a quite biased ratio towards a

01:02:38.020 lot of NAD and a small amount of NADH. The way nature works is that whenever any pair of chemicals is

01:02:47.220 skewed in one direction, it's favorable to turn the one that's abundant into the one that's less

01:02:54.100 abundant. And so this makes NAD a decent electron acceptor. And so it's sitting there prepared to pick

01:03:01.580 up electrons from these intermediates, carbon intermediates of the TCA cycle that are coming

01:03:07.740 from carbohydrate and fat, and take the electrons, make NADH, which then can feed into electron transport

01:03:15.320 chain. And that back end has to happen fast in order to keep this ratio skewed so you have mainly

01:03:20.860 NAD and not too much NADH. And that's really important because when that NADH starts creeping

01:03:26.920 up, all sorts of things start going wrong. Such as?

01:03:30.640 NADH going up will drive too many electrons in the electron transport chain. And going back to

01:03:36.700 Nov, you know, he's done a spectacular job showing how that leads to production of free radicals.

01:03:42.160 You need the right amount of this, but this is a clear way to get toxic amounts of free radicals

01:03:48.220 if you have NADH buildup. Secondly, it just gums up metabolism. If you have too much NADH relative to

01:03:56.140 NAD, you can get into problems not having enough ATP. And so it can also make signaling things go awry.

01:04:04.000 Now, are there clinical scenarios in which we see that happen? Or are these more typically

01:04:08.320 there are things that result from toxicities? The classic thing you learn in medical school

01:04:12.940 to explain the significance of this whole system is cyanide. Maybe tell folks how cyanide works. And

01:04:17.920 I don't know if that's too extreme an example of how this system can be hijacked, but let's see.

01:04:23.540 Cyanide is electron transport chain inhibitor. And so that leads to the whole system just backing up

01:04:29.540 a bit by bit. And so you can't then transfer electrons from NADH into the electron transport chain.

01:04:35.000 And so NADH goes way up, NAD falls to the floor, and then you have no way to make ATP. And that

01:04:41.360 unfortunately leads to rapid mortality. That's an interesting point because this again comes back

01:04:46.600 to the kinetics and the flux, which is it's not like cyanide kills you in an hour. I mean,

01:04:51.480 it kills you in seconds. It's really a sobering thought. ATP turnover via this system is on the

01:04:57.840 timescale of a second. So same for NADH. And so these are things you're just whizzing through

01:05:04.120 our bodies all the time and that we're constantly dependent on.

01:05:09.800 Are there less extreme examples, Josh, of things that will put that balance in the wrong direction

01:05:15.460 kind of chronically?

01:05:17.120 I don't want to give the misimpression that, you know, the right thing is to have as much NAD

01:05:21.980 and as little NADH as possible. First of all, it's designed to be a dynamic system. If you undergo

01:05:29.960 intense exercise, you're going to drive NADH up. And this is a very healthy context for doing this

01:05:36.940 transiently. Metformin, of course, is a super interesting medication and probably works

01:05:43.320 mainly by slowing the conversion of NADH back to NAD by impairing the complex one of the electron

01:05:53.360 transport chain, the one that does this initial electron offloading from NADH to make NAD.

01:05:58.640 How well is that understood? I mean, you'll talk to five people who study metformin and they'll tell

01:06:03.500 you five different things, which I think just tells you how much we don't know. But I don't

01:06:07.760 think it's really disputed that metformin inhibits complex one, is it? I think the broader question

01:06:11.980 is how much is that the main attraction versus kind of a sideshow?

01:06:16.700 There are many people who know more about this than me, but one thing that we tend to do in our lab

01:06:22.180 sometimes is take these famous metabolic effects like metformin inhibiting complex one and just do a

01:06:27.740 quick test of it. And I say when we do that, about half the time they look to be true and half the

01:06:33.200 time they look to be dubious. And metformin was a shining star in our hands in inhibiting complex

01:06:39.180 one. It was one of the cases where I really felt like it may do other things, but it certainly does

01:06:44.800 what it's supposed to do there. It does that strongly. And I think it's probably the fact that

01:06:49.460 it does it in a relatively liver-specific way due to the way that metformin enters cells of the body

01:06:54.860 that leads to it, first of all, being safe. There are many things that make it safer than cyanide,

01:06:59.840 but it is really crazy that maybe the world's most widely used medication is at some level

01:07:04.920 inhibits the electron transport chain. It's a mechanistic analog of cyanide.

01:07:10.020 So you have like one of the most acutely lethal substances and most widely used drug working in a

01:07:16.820 remarkably similar way. And so I think the fact that there's a strong liver specificity is probably what

01:07:22.100 makes it net beneficial for at least a subset of people.

01:07:27.080 What's the change that you saw in your lab, Josh? So if you go off metformin, what's your NADH to NAD

01:07:34.280 ratio? And then on metformin, how much did it change that?

01:07:37.180 Depends how much metformin you use.

01:07:39.560 But if you tried to approximate an actual clinical dose of say a couple of grams a day?

01:07:43.760 I'm not sure we did this in a way that I would consider clinically applicables. It's certainly

01:07:49.120 crystal clear that it goes in the right direction. And so it doesn't surprise you that metformin would

01:07:53.840 raise fasting lactate levels, correct? No. I mean, it certainly is aligned to do that.

01:07:59.380 And that's just backing up further. It's just basically creating more of a roadblock into the

01:08:03.940 TCA is going to give you more lactate. Yep. More of a roadblock in electron disposal, basically.

01:08:09.780 Do you think that that's a neutral effect or do you think that that's a potentially deleterious

01:08:15.000 effect of metformin that is probably offset in a patient with diabetes by the benefits that it

01:08:20.640 has on hepatic glucose output? It's a great question. I don't think I know. I think having

01:08:26.440 more circulating lactate can be a bit of a challenge for clearing fat because they have some sort of

01:08:33.840 competition. So from that perspective, I think being in a lower state might have some benefits,

01:08:40.640 but then lactate is a valuable fuel as long as it's not getting too high levels. I'm not sure how

01:08:46.620 that all plays out in terms of long-term health. Anything you want to say about NADP and NADPH,

01:08:52.740 just to round it out so people know the full story? These are super important cofactors.

01:08:59.300 They live just on the edge of what people who take biochemistry, either in undergraduate or med school,

01:09:05.420 learn about or don't learn about. They're fascinating cofactors because in terms of their intrinsic

01:09:12.320 chemistry, all their intrinsic chemistry from the energy point of view is exactly the same as NAD,

01:09:18.620 NADH. But they have a different handle on them chemically that allows biology to use them in a 0.99

01:09:25.440 different way. The ratio is maintained quite different level from NAD, NADH. So NAD, NADH is super biased

01:09:34.960 towards NAD. This is much more of an even pairing, which means there's much more driving force to

01:09:41.820 dump the electrons off rather than to absorb them up. NADPH is really, to me, second only to ATP,

01:09:52.220 a master energetic building material. And it's the building material that's used, for example,

01:09:58.600 to assemble fat. I mean, it's the most important one. So as you take pieces, two carbon pieces from

01:10:04.720 carbohydrate and want to put them together to make fat, you keep dumping in electrical energy in

01:10:11.180 the form of NADPH. And then NADPH is used in all sorts of other really interesting ways to fight

01:10:17.300 reactive oxygen species. It's also used if you're trying to kill bacteria to intentionally make

01:10:23.340 reactive oxygen species. This is where biology is freaking confusing and complicated. And there's

01:10:29.880 definitely the yin and yang that you have this awesome co-factor that's so important for fighting

01:10:35.300 oxidative stress and also can be used to create boatloads of oxidative stress intentionally when

01:10:41.280 it's needed. First of all, that was a fantastic overview of how the Krebs cycle works and specifically

01:10:48.440 with attention to how electrons move through it and move through these redox factors, which then brings

01:10:56.440 us to a part of the discussion where a lot of people have an enormous interest, which is, I don't know,

01:11:03.180 go back seven, eight years, it started to become fashionable and it's only become more fashionable to talk

01:11:09.760 about supplementing with NAD. I say that quote unquote, we're going to talk about why you don't actually

01:11:16.600 supplement with NAD. But is it safe to say that at least part of the impetus for this was the observation

01:11:23.340 that as we age, cellular NAD levels decline. And you've already made a very compelling case for why

01:11:30.340 NAD is important. I almost want to avoid the whole sirtuin side of this because I think that story 1.00

01:11:35.500 keeps changing. So unless you feel strongly or compelled to get into sirtuins, we can put those

01:11:40.960 aside for the moment. Yeah, I love putting sirtuins aside. The first principles in this field are great. 0.99

01:11:49.140 NAD plays this super central role in energy generation that we all want to feel more energetic, whether, you know,

01:11:56.940 you want to be a more extremely successful athlete at age 21, or whether you want to feel at age 50, like I am,

01:12:04.020 or later, like you're 21. So you think we could just turn up the burner capacity, right? This would be

01:12:11.380 absolutely fantastic. And then we have this data that NAD is depleted with aging. Although I'll have

01:12:18.900 to say, when we do those measurements, we agree that NAD is depleted with aging, but it is a lot more subtle

01:12:25.760 than you would think looking at the literature. These are really quite subtle NAD depletions. You know,

01:12:33.520 we were talking about these ranges earlier, the kind of threefold range where a lot of metabolites live

01:12:40.500 on a daily basis. Some of them glucose, that's their worst day, right? And as you pointed out,

01:12:45.300 some of them like lactate, they may do the threefold all the time. And then the 60 fold when you stress

01:12:50.880 them. NAD changes we see with aging are like 10%, 20%. Oh, wow. I didn't realize it was that little,

01:12:58.740 Josh. And I'm not saying that in some tissue of, you know, an aged human, there might not be bigger

01:13:04.260 effects. This is the first caution I would give to people thinking that they're going to fix

01:13:09.460 everything through NAD. On one hand, it's a robust finding that this is something that changes with

01:13:16.580 aging, that with a central metabolic role. On the other hand, it's something that happens with a

01:13:21.520 fair amount of subtlety. Can you explain to folks how this is done? Because we talk about measurement

01:13:26.760 sometimes a little too glibly. Pretty easy to explain how we can measure glucose and hemoglobin

01:13:31.980 and lactate. At the other end of the spectrum, we've talked a lot about ATP. What I think most people

01:13:37.060 don't understand is it's very difficult to measure ATP. It comes back to what you said a moment ago.

01:13:42.280 This ain't sticking around a very long time. You're using MRS and super complicated physics

01:13:47.520 to be able to measure these things. Where does NAD fit on that spectrum? And how do you actually

01:13:53.680 measure it? Good news is that NAD, unlike NADH, is not like one of these super transient metabolites.

01:14:01.840 NADH measurement is wickedly difficult. But most of this NADH, NAD parasites is NAD,

01:14:09.500 and it tends to sit around for an hour-ish time scale.

01:14:14.680 Oh, so you don't necessarily have to flash freeze tissue or things like that.

01:14:17.980 That's right. You have some more flexibility in making those measurements, as long as you're not

01:14:23.080 irritating the tissue in a way that leads to massive NAD degradation, which people may do

01:14:28.340 sometimes by accident. But I think generally, it's not that hard of a measurement, NAD. Obviously,

01:14:35.680 like ATP, it's a tissue metabolite, not a circulating metabolite. So you need biopsy

01:14:40.160 specimens to measure it. I'm not a master of the literature of NAD levels in human tissues, but my

01:14:47.320 not fully informed perspective is that it probably isn't as much as we should have. And that's because

01:14:54.100 it's hard to get biopsies from people. If you take blood, if you take a whole blood and you look at

01:15:00.500 PBMC, can you look at NAD levels in there with relative ease? Or is it too complicated because by

01:15:05.880 the time you separate the PBMC, you've kind of lost your window? I think it's a really good question

01:15:11.160 because there is quite active NAD metabolism in immune cells. I'm not an expert in this. I bet there's

01:15:21.200 a way you could develop a good protocol. I haven't followed, you know, how good the measurements up

01:15:25.700 to now have been. So most of what you've measured has been in tissue. Typically, you know, we work a

01:15:30.940 lot in mouse. Sometimes we measure human, but more typically on the cancer side. There, we just take

01:15:36.800 tissues and freeze them and extract metabolites, do mass spec. And so you're seeing a consistent,

01:15:44.100 clear decline in NAD with the aging animal or human, but it's not a fold reduction. It's a percent

01:15:51.720 reduction, 10%, 20% reduction. That's the most common thing. Yeah.

01:15:56.060 This generates a hypothesis. The hypothesis is if you restore NAD levels in the old organism to the

01:16:03.440 level that they were in the young organism, the old organism will feel and perform like the young

01:16:09.040 organism. That's one hypothesis. Another hypothesis would be if you induce supranormal

01:16:14.080 levels of NAD in any organism, they will feel supranormal. Let's assume that both of those

01:16:19.380 hypotheses are simultaneously testable. What happened five, seven, eight years ago, NAD clinics

01:16:25.960 started popping up all over the place. And they started saying, if you come here, we'll put IVs in

01:16:32.320 you and we'll give you NAD. So let's first explain why did they do this intravenously? Why couldn't

01:16:37.640 they make an NAD pill? NAD and its precursors are broken down in the gastrointestinal tract.

01:16:44.240 And so if you take NR, for example, so nicotinamide riboside orally, it mainly will enter the body in

01:16:52.700 the form of nicotinic acid or niacin, which is a healthy substance. Nothing wrong with it, but except

01:17:00.020 for maybe the epithelium or gastrointestinal tract, the body is not seeing nicotinamide riboside.

01:17:08.380 And NAD, just to be clear, we're going to talk about NR and NMN in a moment, but NAD, there's

01:17:13.240 no way to orally take it. There's no known absorption route for NAD. And I think it'll

01:17:18.720 get broken down probably all the way to nicotinic acid, although I'm not 100% sure anyone has proven

01:17:24.960 that. I certainly don't think it would enter the body any other way than either nicotinamide,

01:17:31.020 which is like a little bit closer to remaking NAD or nicotinic acid.

01:17:36.440 So what happens when a person receives intravenous NAD? What's the fate of that NAD? You know,

01:17:44.480 one of the things in metabolism and biology is anytime you put something in a vein, you bypass

01:17:49.840 the liver with something called the first pass effect, which in your former life was very

01:17:54.400 important because when you had these patients in the ER that you were giving inhaled drugs

01:17:59.560 to, it's not just the speed with which they were getting it, it's that you could actually

01:18:04.020 deliver the exact drug you wanted, not a pro-drug that could be modified by the liver.

01:18:09.180 So this idea of giving intravenous NAD is at least theoretically interesting because you're

01:18:15.220 putting the molecule of interest directly into the venous system. So what's its fate?

01:18:20.080 You may be more up on this than me, but it's going to get broken down partially because there's not

01:18:27.500 clear uptake mechanisms known anyway to get NAD from the bloodstream into cells. Nicotinamide

01:18:35.020 mononucleotide may be able to enter cells directly or nicotinamide ribosides. These are partially

01:18:43.840 broken down forms of NAD, but that are nevertheless meaningfully closer to NAD than the normal things

01:18:51.780 that circulate in good amounts in our bloodstream. That does partially, I would say, short-circuit

01:18:58.320 the route to cells making NAD. So they kind of can break down partially the NAD in the bloodstream,

01:19:06.180 take these partially broken down NAD precursors into cells and rebuild NAD in a shortcut manner

01:19:14.480 that probably has a good chance to bolster NAD levels.

01:19:18.980 So in other words, when you give intravenous NAD, there is no transporter to take NAD into a cell,

01:19:25.000 but that NAD breaks down into things like NR and NMN. And in the vascular system, we know that those

01:19:32.700 things can get taken up at least into some cells. Do we know which cells have the capacity to do that

01:19:37.600 or which cells don't? Certainly at least some important cell types in the body can take those

01:19:42.820 up. Maybe pretty broadly, but I don't know off the top of my head. And so then when the NR or NMN

01:19:49.520 gets into this cell, is it relatively straightforward that it will be reconstituted into NAD? What's the

01:19:56.160 energetic cost of doing that? Or how easy is that? And is that the favored reaction at that point?

01:20:02.840 Yeah, I think it's the favored reaction, which is the important thing. And this is not a big demand,

01:20:08.240 relatively speaking. The big energy flow is through this NAD and NADH exchange, but the making of NAD

01:20:16.080 itself is not an expensive process per se. NAD stands for nicotinamide adenine dinucleotide.

01:20:24.820 It's two kind of nucleotide pieces put together. And when you take an NR or NMN,

01:20:31.100 it's one of those two pieces, but the more interesting side. And the other side

01:20:36.620 comes from ATP and it's there all the time because all your cells have ATP or you got much,

01:20:41.680 much deeper problems. And so you just snap it together. And I think you end up with probably

01:20:47.100 effective NAD supplementation when you go the IV route. In other words, taking IV NAD will probably

01:20:55.080 increase intracellular NAD levels, though not directly because there's not a transporter,

01:21:00.840 but it goes through this sort of circuitous route to get there. So it might not be the most efficient

01:21:04.960 way to do it, but this certainly corrects a statement I've made in the past, which is

01:21:09.860 intravenous NAD is not a good way to get NAD because we don't have a transporter. That's correct,

01:21:16.660 but incomplete.

01:21:17.620 That's a fair summary from my perspective.

01:21:21.400 In other words, we don't really know how much, if you take a hundred units of NAD

01:21:25.000 intravenously infused, we don't really know how many units ultimately make their way into a cell,

01:21:29.580 but it's probably not a hundred.

01:21:31.720 I'm sure there's a fair amount of loss in the process. There's a very interesting protein called

01:21:36.700 CD38 that's, I think, designed to control these kinds of pathways. It's a suppressor of NAD levels.

01:21:45.560 It works, I think, by breaking down NAD that's outside of cells. And mainly, there's not normally

01:21:53.540 in physiology NAD in meaningful amounts, probably outside of cells, but there is NMN in meaningful

01:22:00.080 amounts. This is a protein that's super good at breaking down NMN. It still leaves you with NR.

01:22:05.940 So it's one step further away from being NAD, but it's still meaningfully closer than your typical

01:22:12.060 physiological precursor. And so I think it's positioned to, as you say, at least some places

01:22:17.800 in the body boost NAD.

01:22:21.540 Today, I think the majority of efforts to increase intracellular NAD are done through oral precursors,

01:22:28.660 and the two are NR and NMN, which, as you said, are pretty similar. And are you aware of a more

01:22:37.320 convincing argument for why one might be a more preferred substrate? I haven't particularly seen

01:22:42.740 arguments that one is superior than the other. I've seen some unpublished data that suggests

01:22:48.360 one can be made more temperature and moisture stable than the other. But let's put that aside

01:22:54.300 for the moment. Would you consider these, to a first approximation, equivalent approaches?

01:22:59.200 Yeah.

01:22:59.420 Okay. So now talk about something else that you said that it also kind of news to me, which is

01:23:04.640 what is the effect of the gastrointestinal tract on these agents?

01:23:09.480 I mean, they get broken down and they get broken down all the way to the level of nicotinic acid or

01:23:14.800 niacin, basically. This is the main way they enter the body. It doesn't mean that there can't be a

01:23:20.740 trickle of them entering some other way that has a physiological effect, whether there's some local

01:23:26.160 effect or some effect on the microbiome of taking them. Biology is super complicated. There are ways

01:23:31.780 that these could be doing interesting health-supporting things, but I don't really think

01:23:37.560 they're fundamentally different than taking a physician-prescribed niacin pill from the perspective

01:23:42.680 of providing NAD precursors.

01:23:47.600 Now, a physician-prescribed niacin pill, when people used to take niacin for hyperbeta-lipoproteinemia,

01:23:54.100 it wasn't uncommon to get a real flush from the medication. Now, I don't remember what doses

01:23:59.140 people were taking, but I feel like it was on the order of grams, not milligrams. Do you recall how

01:24:05.540 much niacin you would need to give somebody for them to experience an actual flush?

01:24:09.720 I think it was a few grams. That's where my recollection is too.

01:24:13.640 Is that the reason people don't experience a flush with NR and NMN? Because they're typically taking

01:24:19.780 500 milligrams to one gram, and that's simply not going to produce enough niacin to reach that

01:24:24.800 threshold? Yeah, they're also kind of niacin prodrugs, so they probably are delayed absorption

01:24:30.160 forms of niacin, so that may smooth things out enough. So they may be better tolerated,

01:24:36.840 but I think this is how I'd fundamentally think about them, is that they're niacin prodrugs.

01:24:41.800 Your lab has done some of the flux work on this. What are some of the most interesting things you've

01:24:47.760 learned about how NR and how NMN, when given orally, end up in different tissues, and what the effect is

01:24:54.940 in the liver versus the muscle versus the plasma? I think the main thing you see is that these are

01:25:01.740 converted to niacin. They will raise niacin, particularly niacin, heading to the liver out of

01:25:06.800 the gastrointestinal tract, so in the, we call the portal circulation, that connects your intestine

01:25:11.960 to the liver very effectively. And that other than that, their effect on like boosting their own

01:25:18.500 circulating levels is somewhere between subtle and vanishing. I'm still not sure which of those two it

01:25:25.080 is. They certainly remain in the bloodstream much less abundant than nicotinamide, which is the thing

01:25:30.740 the liver is normally producing to feed NAD precursor to the tissues of the body. From what

01:25:36.740 we have seen, no clear route for oral NR or oral NMN to produce circulating levels of NR or NMN that are

01:25:48.480 high enough to compete, at least at a standard concentration level, with nicotinamide, the

01:25:55.400 physiological precursor, as a way of feeding NAD precursors to tissues. So basically, at some level,

01:26:01.380 they don't change what's happening, what most of your tissues are seeing that much if our

01:26:06.900 measurements are correct. Where do you think you could be fooled on this? I mean, I know that's a

01:26:12.440 question every scientist or every good scientist asks themselves that question, right? How can we

01:26:17.220 be fooled by our measurements? I'm sure you've thought about this. Where do you see the opportunities

01:26:21.820 in this particular case to be misled? It could be that there are local effects of NR or NMN on like

01:26:29.340 the intestine that are really important. It could be that their availability impacts the microbiome

01:26:36.160 in important ways. The microbiome can have big effects on health. It could be that even though

01:26:41.480 the amounts of like NR that may reach the liver or even lower amounts that may reach, you know,

01:26:47.300 the heart or something are really small, that there's a subset of cells there that are really

01:26:53.320 NR preferring because maybe they're really deficient in using nicotinamide and maybe getting even small

01:27:01.060 amounts of NR to those cells is meaningful. I think these are all possibilities that we're very much

01:27:06.180 open to. My base assumption is that often the obvious is true and here the obvious would be

01:27:13.440 the physiological system just isn't that impacted by this particular type of oral supplement.

01:27:19.280 Do you think there's any chance that with chronic administration you'd see something different?

01:27:25.540 Because I'm assuming in these experiments you're not seeing the effect of these chemicals being

01:27:32.160 ingested chronically or are you? Human can be different than mouse, okay? First of all,

01:27:36.880 it's another important thing that I say. We haven't done these experiments in human. Someone,

01:27:42.380 if they aren't already, should do these experiments in human. Yeah, chronic versus acute. So there's a bunch

01:27:48.340 of variables that could alter things. Based on what you know now, if the hypothesis is true,

01:27:55.420 if restoring intracellular NAD levels at 50 to the level they were at when you were 20 would improve

01:28:02.340 some measure of performance, based on what you know today, what do you hypothesize would be the

01:28:09.440 most efficient way to restore NAD levels? IV is the promising way to do the restoration.

01:28:16.900 I'm not very convinced about the first hypothesis. I think the big history of medicine, you and I can

01:28:24.520 debate it, is that things are way more complicated than people can envision. Hormone replacement therapy

01:28:31.880 is like one of the great examples, right? It didn't turn out to... Although it's being overturned.

01:28:37.460 I think if you go back and look at the Women's Health Initiative, I think it got it wrong. It was the

01:28:41.940 randomized experiment, but it was really misinterpreted. Say a bit more about that.

01:28:46.660 Maybe I picked up on the wrong thread of where you were going, but I assumed what you were going

01:28:50.620 to say was, look, the epidemiology in the 80s and 90s was that giving women hormones post-menopause

01:28:57.960 was a good thing. And then the Women's Health Initiative came along and said, no, it's a bad thing. I 0.98

01:29:02.920 assume that's what you were going to say. Yeah. And what I was going to say was, actually, no,

01:29:06.460 I think that's actually misleading. I think if you actually go back and look at the Women's Health

01:29:09.760 Initiative, it was just an awful example of how to misinterpret a study. I think there was no

01:29:15.040 increase in the risk of breast cancer. And if there was any increase in the risk of breast cancer,

01:29:19.280 it probably had nothing to do with the estrogen that the women were given. When you actually look,

01:29:23.460 for example, at the relative risk and absolute risk different in those cohorts. So remember the

01:29:28.720 Women's Health Initiative had three, well, technically it was two parallels, right? So you had the placebo 0.97

01:29:33.440 versus estrogen only in women who did not have a uterus. And then you had placebo versus

01:29:39.460 estrogen plus MPA, the synthetic progesterone. So in the estrogen only versus the placebo,

01:29:46.980 there was a non-statistical significant reduction in the risk of breast cancer. So there was a hazard

01:29:52.640 ratio of about 0.8 or 0.79 or 0.81, something like that, but it didn't quite reach statistical

01:29:57.480 significance. But trending towards estrogen actually reduced the risk of breast cancer. In the estrogen

01:30:04.180 plus MPA group, there was a barely statistical significant increase in the risk of breast

01:30:09.660 cancer. I think the hazard ratio was, I want to say it was about 1.24, 1.25, and the p-value was

01:30:16.960 exactly 0.05 or 0.049 or something like that. So at the surface, you'd say, gosh, this is increasing

01:30:22.260 the risk of breast cancer. And what was talked about was a 25% increase in the risk of breast cancer.

01:30:28.000 To talk about the relative risk increase without talking about the absolute risk is obviously

01:30:32.740 irresponsible. If you look at the absolute risk change, it was 0.1%. It was one in a thousand.

01:30:39.540 And that says nothing about a lot of other methodologic issues with the study, including

01:30:43.280 the fact that, in my opinion, a more plausible hypothesis was that the MPA was more the issue

01:30:48.400 than the estrogen. But the estrogen gets all the attention, right? So estrogen causes breast cancer, 0.99

01:30:53.420 gets the attention. If you look at subsequent studies, I don't think we see that to be the case.

01:30:58.940 So I'm going to hypothesize or predict that in 10 years, we'll look back at what happened to a

01:31:05.480 generation of women, which I think is really unfair. Basically, an entire generation of women 1.00

01:31:09.680 got deprived of hormones because of, I think, a really poorly interpreted study. But your point

01:31:15.600 notwithstanding, sometimes the obvious is not obvious. Sorry for the digression.

01:31:20.240 No, I mean, that was super interesting. I thought there was some cardiovascular risk data in that

01:31:24.400 study that was surprising, but you know it much better than me. Yeah, I think on the cardiovascular

01:31:29.080 front, there probably is a slight increase in risk with oral estrogen because of the hypercoagulability.

01:31:35.820 I also think it speaks to understanding the use case. Today, very few women on hormone replacement 0.93

01:31:42.980 therapy are given oral estrogen. The preferred route of administration is a patch, you know,

01:31:46.960 something like a Vavelle dot where you're given topical estradiol and you get all of the benefits of

01:31:52.040 the reduction of vasomotor symptoms, the incredible benefits that you see on bone health without any

01:31:57.520 of the hypercoagulability and cardiovascular. So now we actually see the reverse. Now there's a very

01:32:01.540 clear trend, not just trend, it's statistically significant. There's a very clear reduction in

01:32:06.260 cardiovascular mortality. So that's a great example where you had to give by the right route of

01:32:12.000 administration in order to get the net positive health benefit. NSAIDs, Advil, another one, right?

01:32:19.100 People knew they had a lot of side effects, but everybody assumed that they were kind of

01:32:22.460 counterbalanced by the fact they were reducing coagulability and that this was going to be

01:32:26.220 cardioprotective. I don't know if you're going to tell me that you still believe that, but most people

01:32:31.040 don't. So really where you were going, I think, is you were saying, look, you might even just reject

01:32:36.960 the outright hypothesis. Like this idea that, yeah, we do observe a 20%, 10 to 20% reduction in NAD

01:32:43.300 levels as we age. You haven't even bought the first hypothesis, which is even if I could magically

01:32:48.040 deliver 20% more NAD to a 50-year-old, you're not sounding very convinced that that's going to

01:32:53.820 improve quality or length of life. Not at this point in time. These things involve such complicated

01:32:59.400 interplay of different organ systems. And it may turn out that NAD supplementation is super

01:33:05.120 valuable medically. I am completely open to that and I would love that to be the case. But I think

01:33:11.680 if so, it's going to be because there are select cell types that are genuinely severely NAD depleted

01:33:19.540 and that we will need to figure out how to restore NAD in those cell types. And then we may see big

01:33:25.960 health benefits. So I think that would be fantastic and it's completely possible. And it's possible

01:33:31.940 that the general intravenous supplementation is hitting those cells and doing that. But I think

01:33:37.380 it's equally possible that it's having some adverse effect that's going to be net negative for people.

01:33:41.860 And we don't know the science well enough and we certainly haven't done the clinical experiment well

01:33:46.180 enough to give good health guidance yet. So two things. First, a statement. This is really

01:33:51.860 interesting for me because I really stand corrected and I just want to apologize to all the people over

01:33:56.220 the years that I've said intravenous NAD is not getting in your cell. I stand corrected. It indirectly,

01:34:02.660 based on everything you're saying, may actually be getting into at least some cells.

01:34:05.920 The second is a question, which is, how would you even begin to tackle that question? Which is,

01:34:15.360 are there certain cell subtypes that may indeed benefit from NAD boosting? I haven't really seen

01:34:22.480 a single convincing clinical study in humans using either NR or NMN that has made me excited about this.

01:34:30.660 And I'm not a stranger to putting things in my body without absolute perfect information. I mean,

01:34:37.440 I take rapamycin. I've been taking rapamycin for four or five years. I will be the first to admit,

01:34:42.980 I think we have very good evidence for that. It's not perfect. It's far from perfect. We're never going

01:34:48.780 to have a definitive human clinical trial. But if I'm willing to take rapamycin, why am I not taking NR

01:34:54.920 and why am I not taking NMN? And the reason is, I just can't find a shred of compelling evidence to

01:35:01.080 tell me to do so. And I'm in the same boat you're in. I'd love it if there was, because it's a pretty

01:35:06.540 easy, safe thing to take. So what study needs to be done to help someone like me, a reluctant NADer,

01:35:15.200 get on the NAD train?

01:35:16.380 I think we need to, first of all, map better the basic pharmacology of NAD in animals and human.

01:35:24.920 That's quite doable. And this is something that we as a field are doing. And there are a lot of

01:35:30.720 great people doing this, but we can do more and better. We need to have the technologies to look

01:35:36.980 at this at cellular resolution rather than bulk resolution. This is something we're pushing

01:35:42.840 very hard to develop, the ability to take a slice of tissue and say, what's the heterogeneity across

01:35:49.260 cells in NAD levels. And I think that'll be very helpful. Because if we see that that's really

01:35:56.460 scattershot in aging and homogeneous in young, then you have your answer that all that we need is for

01:36:03.100 one in 10 cells at any time to be really NAD depleted. And we view that 10% reduction,

01:36:08.760 not as some tiny wiggle down, but as one in 10 cells being on the road to a catastrophic outcome.

01:36:16.280 I think this is going to be a really important measurement. I think the field will get there

01:36:20.980 over the next few years, not instantly. And then ultimately, we need successful clinical experiments.

01:36:27.960 There, there have been some really persuasive experiments in animals. For example, I think

01:36:33.620 there are experiments on reversing, you know, bad outcomes after renal ischemia. And so it'd be good

01:36:39.120 if we could find niche experiments where there's a very strong effect in animals, a very quick clinical

01:36:45.580 readout, ischemic renal event, and you do the supplementation and you get a benefit or you don't.

01:36:52.260 How was it administered in that experiment? I don't think I remember that one.

01:36:55.640 I may not get the details of that right. I think I would just say conceptually,

01:37:00.020 we need to find the strongest animal proof of concept that can be translated into a small but

01:37:06.080 definitive clinical trial and prove that this really can do something beneficial in the right

01:37:12.080 context. And then from there, you can think about kind of expanding the indication to general health

01:37:17.240 betterment. Who's the natural owner funding wise of this? Is this a question NIH is interested in? I mean,

01:37:23.680 indirectly through the ITP, Rich Miller, Randy Strong et al. have already done an NR test in their very

01:37:31.520 rigorous tried and true model. As you know, that failed. So NR did not extend life. Is NIH still interested

01:37:38.640 in this question enough to continue funding it? Where is your funding for this level of investigation

01:37:44.280 coming from? We mainly try to help the real NAD expert labs by doing the flux studies, facilitating

01:37:54.660 measurements, but it's not the bread and butter of my life. I'm probably about an observer at the same

01:38:01.180 level as you of this field, broadly speaking. There's money in NIH for interesting science. And so

01:38:07.920 this is too central to metabolism and too much public health interest for the waters to run dry. 0.89

01:38:15.520 You're not worried this is not going to run up against that funding?

01:38:18.380 No. And I think biotech is interested in this. There's very interesting ways to do this

01:38:22.820 pharmacologically. And so we may see those mature faster. CD38 inhibitors. Obviously, there's a whole

01:38:29.380 different financial structure there. If you can make a patent approved medicine, all that economic

01:38:35.440 incentive is great for driving first science answers and then clinical answers.

01:38:41.180 What are the top labs right now in your mind in studying NAD and its precursors or ways to increase

01:38:46.580 it? I'm too much of an outsider to get into naming names on that. I'm going to only get myself in

01:38:52.360 trouble other than to credit, you know, Joe Bauer for being a fantastic collaborator on it.

01:38:57.680 Okay. Let's pivot to the final thing I want to really get into, Josh, which is cancer metabolism.

01:39:02.160 It kind of ties in so much of what we've talked about. And that's how you and I reconnected five

01:39:06.980 or six years ago at a conference. And then obviously a number of times since then. So the

01:39:11.520 irony of it is, right, you do your PhD in the inner workings of how the immune system works,

01:39:17.520 but you're not particularly interested in cancer at the time. You come back to academia as a

01:39:23.020 metabolomics expert, and now you've kind of wound your way back to oncology in a way. Tell me a little

01:39:27.800 bit about that journey, right? How did you go from this profound interest in metabolism,

01:39:32.420 metabolomics, and fluxomics to realizing a beautiful application for this is in the field of oncology?

01:39:39.560 Part of it is really the human connection. I was so fortunate to be at Princeton, which is this kind

01:39:47.060 of academic bubble where I could do my experiments on E. coli and yeast and really set up these good

01:39:53.940 metabolic measurements unmolested. Also, we're close to Penn. At some point, I got a call from the head

01:40:02.040 of the Penn Cancer Center at that point in time, Craig Thompson, saying he wanted to visit. Of course,

01:40:07.200 I just say yes. That was kind of a life-changing call for me because it brought me into the world of

01:40:13.680 biomedicine again, basically, in the context of working on cancer metabolism. It was very natural

01:40:21.040 because if you look at the history of cancer therapy, first great rational triumph in treating

01:40:27.080 cancer was antifolates and Sidney Farber. Name memorialized on the Dana-Farber Cancer Center,

01:40:34.560 right? This is really the origins of how cancer was rationally treated by targeting metabolism,

01:40:41.140 and it just got understudied for so many years, and it was a very natural re-entry point for me

01:40:47.640 because cancer you can study as isolated cells in a culture dish, much as we were studying E. coli

01:40:53.280 and yeast, and so that was much more comfortable to me in the 2008 or whatever time frame trying to

01:41:00.260 work on mice. Delighted that we got back to mice a few years after that. What year did Craig leave Penn

01:41:06.760 to go to Memorial Sloan Kettering? It must have been shortly after he invited you over, right?

01:41:11.280 A few years after.

01:41:12.400 So people like Craig, who I have not had on the podcast, but I'd love to, but Lou, who I have,

01:41:17.540 when you think about cancer metabolism today, I mean, it's just a booming field. I would argue,

01:41:22.080 and no disrespect to people in different fields, but cancer metabolism and immunotherapy are really

01:41:26.940 two of the most promising and exciting areas in the field today, which were two things we didn't

01:41:32.300 have a single word about in medical school, right?

01:41:35.140 Yeah, that's absolutely true. Maybe some anti-metabolites for cancer hidden somewhere in

01:41:40.400 the pharmacology book. You know, one of the most exciting things is going to be the interface

01:41:44.300 of those two fields, and we see this with, you know, microbiome composition being predictive of

01:41:49.960 whether immunotherapy works. Amazing work from Jennifer Wargo showing that, you know, fiber can promote

01:41:56.560 the effectiveness of immunotherapy, and so we're seeing that connection also being made.

01:42:01.960 Soluble or insoluble?

01:42:03.060 TBD. There are different flavors of soluble fiber, and I have my pet dreams for how this

01:42:09.600 may work mechanistically, but I think it's going to be an incredibly important interface.

01:42:14.500 So tell folks a little bit about what it is about cancer cells that makes their metabolism distinct

01:42:19.540 from their non-cancer counterparts, but within the same tissue even. Like if you want to compare

01:42:23.800 adenocarcinoma of the colon or breast or prostate and look at the perfectly normal non-cancer cell

01:42:30.120 sitting right next to it. We typically talk about two hallmarks of cancer, right? We talk about

01:42:33.740 the inability to respond to cell cycle signaling. So this is why these silly things just keep growing 0.91

01:42:38.820 even when they're told stop growing. And then the capacity to metastasize, to basically pick up,

01:42:43.320 leave, go grow in a new site. But what is it about them metabolically that also is a piece of their

01:42:48.540 signature?

01:42:50.040 Cancers tend to be glucose users. Once you step back, and in the fasted state in particular,

01:42:56.600 using glucose is a weird thing rather than a default thing. The default thing is to use fat

01:43:04.440 and lactate. Then the fact that cancer uses glucose is very distinctive. And they do this in large part

01:43:12.500 because they're programmed internally to basically feel like they're always seeing insulin. And this

01:43:19.380 is through mutations and something called the PI3 kinase pathway that Lou Cantley, who you mentioned,

01:43:23.840 you know, pioneered. And this leads to the fact they're positive on this FDG PET scan. So they'll

01:43:29.740 constantly take up and phosphorylate and trap glucose or glucose analogs. And this is actually

01:43:34.560 the most sensitive way to detect most types of cancer. Downstream from this, though, there's a ton of

01:43:42.020 metabolic changes in the cancer cells. The most fundamental of these is the fact that in order to do the

01:43:49.780 uncontrolled growth, they have to do uncontrolled nucleic acid synthesis. And this is the vulnerability

01:43:56.740 that was targeted initially by Farber, but has been targeted by a lot of very important medications

01:44:02.320 that are widely used still today. You know, pemetrexate is first-line treatment for lung cancer. And if you get

01:44:09.440 these medications right, you can induce mutations through the metabolic stress on the nucleotide system,

01:44:15.760 and this can make immunotherapy work better. That's a very exciting part. And I think that part

01:44:22.080 has gotten understudied as cancer metabolism has returned to the fore. There's been a lot of focus

01:44:27.020 on fuel usage cutoff, which is tough because like most cells in the body, cancer cells can use a lot

01:44:33.220 of different types of fuels depending on what's available. Yeah, this is an important point, Josh,

01:44:37.260 because a lot of people, I think, would hear the first part of what you're saying.

01:44:40.900 And their natural conclusion would be, well, wait a minute. If you do a PET scan on somebody and it

01:44:46.320 lights up with glucose, that tells me cancer loves glucose. Ergo, the way to treat cancer,

01:44:53.480 don't eat glucose. Problem with that logic is no matter how little glucose you eat, you still have

01:44:59.740 plenty of glucose in your circulation. I mean, even if you're in a complete state of starvation, again,

01:45:05.160 going back to George Cahill, 40 days of starvation, they still had three millimole of glucose in

01:45:10.740 their circulation 40 days out. So there is no way to eliminate glucose. Now, an argument could be,

01:45:16.140 but you're going to minimize insulin. So I guess the question becomes, is minimizing insulin actually

01:45:20.780 more important than minimizing glucose? But the idea of starving cancer seems potentially overly

01:45:27.060 simplistic, right? Based on everything we've already talked about. I think starving cancer is

01:45:31.880 very, very hard. And as you say, getting circulating glucose to go meaningfully down below, you know,

01:45:38.180 the healthy 89 where you last measured yourself is very, very difficult. Even if you could do that,

01:45:45.680 it's not going to prevent the cancers from having access to internally stored fuel for a while in the

01:45:50.940 form of glycogen, and then ultimately to amino acid fuel and fat fuel and lactate fuel, ketone body fuel.

01:45:58.580 And we've shown very clearly that cancer can use all of those things. They're all valid inputs.

01:46:04.920 They can't replace glucose in the test tube, but it's not easy to cut off the cancer fuel supply,

01:46:12.300 especially not without cutting off some other critical fuel supply, the immune cell fuel supply,

01:46:16.780 right, which would be a disaster, or the brain fuel supply, which would be an even more acute disaster.

01:46:22.380 This idea that there's a way to exploit the metabolic, I don't want to say limitation,

01:46:29.200 but I would just say quirk of cancer in a way that also augments the immune system. Say a bit more

01:46:34.960 about that because that's both fascinating intellectually, but also elegant in that it's

01:46:40.460 mechanistically in line with a bias I have, which is cancer is really going to be hard to get under

01:46:48.500 control. So hoping for a stalemate where you use multiple modes of action is probably a better

01:46:55.660 strategy than hitting really hard on one lever. Again, it's a bias of mine, but at least I can

01:47:00.980 acknowledge it. I think we're seeing a lot of moves to try to make cancer into a chronic disease where

01:47:06.760 the therapies are not so terrible. Hitting the nucleic acid side of cancer is something people

01:47:13.480 are trying for maintenance therapy. And we need to think about that whole side of cancer metabolism

01:47:19.720 fresh because I think there are targets waiting to be developed. If they create nucleotide imbalances,

01:47:27.860 which is a natural thing to do when you hit that system, then nucleotide imbalances are drivers of

01:47:33.020 mutations and mutations in cancer cells are drivers of immune response to cancer. So that's one very

01:47:40.600 appealing avenue. Another... So in other words, interrupt their ability to synthesize DNA,

01:47:45.740 they will create more mutations. More mutations is more shots on goal for the immune system. 0.98

01:47:52.040 That's one really exciting avenue. Another exciting avenue is to apply a very strong stress to the

01:47:58.660 cancer while putting pressure on their fuel supply. I think it's very hard to think that you're going to

01:48:05.260 put so much pressure on the fuel supply that that alone is going to make the tumor slow or even more

01:48:12.480 optimistically regress or something. But if you come in with chemotherapy, for example, that's already

01:48:18.640 targeted preferentially, not perfectly, to the tumor. And then you pair that with something like

01:48:25.020 ketogenic diet, which is lowering insulin, lowering glucose. Then we at least see in animal models that this

01:48:32.900 can be a very powerful combination. We see that the tumors start to deplete glucose in response to

01:48:38.820 the chemotherapy, whether that's because their vasculature is breaking down or whether that's

01:48:43.300 because they have heightened glucose demand because they're having mitochondrial damage from the

01:48:47.760 chemotherapy. I'm not sure. But we see that chemotherapy lowers glucose in the tumor intrinsically.

01:48:54.340 And then if you come in with a diet that lowers glucose availability, this becomes stronger.

01:48:59.740 All right. And then you can get to really low tumor glucose. And we see pretty big improvements

01:49:06.300 in outcome in mouse experiments. Hopefully they'll translate to the clinic. We have a clinical trial

01:49:12.280 open on this now. And what's the best tool we have besides the conventional, and maybe it is simply

01:49:18.120 the conventional, in terms of ways to interfere with their nucleic acid synthesis? Is it literally just

01:49:22.480 going back to old school chemotherapeutics that do that? Well, for the moment, yes. I mean,

01:49:27.140 for the moment, pemetrexate is probably the most successful clinical agent. Gemcitabine,

01:49:32.780 other things of this sort are all well used as part of the armamentarium. But I think we need to

01:49:38.740 think fresh. It's really interesting to me that when we were in medical school,

01:49:45.300 I thought we would not see a cure in our lifetimes for hepatitis. And look at that now, huh? And look at

01:49:50.900 that. And that was mainly nucleoside analogs. I mean, we were told that because I remember talking about

01:49:56.460 this as, why can't there be a vaccine for hep C? And it's like, you'll never be able to vaccinate a

01:50:01.440 flavivirus. Okay, well, that still turns out to be true, but you'll never cure hep C. And yeah,

01:50:07.260 lo and behold. And this is just nucleoside analogs as the centerpiece of this. And so the fact that

01:50:13.700 there's clearly untapped potential there. Now, maybe that potential was maxed out 40 years ago

01:50:20.020 when people were doing this hardcore for cancer from the cancer, but not hepatitis perspective.

01:50:25.040 But my guess is that that's all chemistry that's evolved a lot. And that this is a ripe area for

01:50:30.080 rediscovery. Well, it's interesting you mentioned gemcitabine because of course, that's one of the

01:50:35.500 first line agents for pancreatic cancer. And if I'm not mistaken, you have a particularly keen

01:50:39.720 clinical interest in pancreatic adenocarcinoma. Is that correct? Yeah, it's the cancer that I've

01:50:43.820 worked on the most. It's obviously just a horrible disease. It's definitely one of the cancers that gives

01:50:49.540 cancer a bad name. It's the fourth leading cause of cancer death in both men and women.

01:50:54.360 Yet by incidence, it's a fraction of that. It just speaks to how lethal it is. You know,

01:50:59.680 the last time I looked, Josh, I would say that adenocarcinoma of the pancreas is 95% lethal.

01:51:04.900 And I've heard people argue that the 5% who don't die are misdiagnosed,

01:51:09.680 almost suggesting that it's pretty much impossible to survive pancreatic adenocarcinoma,

01:51:14.140 which is the worst thought in the world. So you certainly picked a tough one to study.

01:51:20.220 I feel a very strong commitment to it because of a bunch of reasons. But just the fact that it's so

01:51:26.420 terrible is a motivation. And I think it was a disease that for a long time, it's just so terrible,

01:51:33.460 we just give up. I don't think that's the right attitude for terrible diseases. And one of the hardest

01:51:41.160 things in making biomedical progress is getting a clinical readout. And the hidden positive and

01:51:48.040 how terrible this disease is, is the clinical readout is just sitting there itching to be

01:51:52.480 improved. There's the capacity to do really compelling clinical tests of any idea. And it's

01:51:59.180 a terrible disease. Most clinical efforts are going to fail, but they can be done relatively fast,

01:52:05.120 relatively cost effectively. And we're seeing progress. Fulfurinox is progress. A lot of

01:52:11.100 patients' tumors respond. Even more will respond if you combine two agents that were approved,

01:52:17.560 gemcitabine and abraxane, albumin-balanforma paclitaxel, with a platinum agent. That triple

01:52:25.740 combination produces regressions in most patients' tumors.

01:52:30.920 But they're not durable. I mean, that's the thing that's killing us.

01:52:33.980 They're not durable. Okay. But the duration of response is terrible right now. But the

01:52:40.440 fact there's response is promise. Normally, from my perspective, once you can start seeing

01:52:46.080 response, you're on the road. We have to figure out how to make the response durable. I hope that's

01:52:51.880 where the metabolic part becomes important. It's going to be some interface of the metabolic part

01:52:58.320 part or the immune part, or a yet harder hit with chemororadiotherapy, earlier diagnosis.

01:53:04.720 These are the hopes for fixing this. Is there something about pancreatic adenocarcinoma

01:53:10.440 that you've observed metabolically that is distinct from other gastrointestinal adenocarcinomas?

01:53:18.560 You know, colon is also a terrible disease. Liver is also a terrible disease. So all the

01:53:22.940 gastrointestinal adenocarcinomas are unfortunately really bad diseases.

01:53:28.640 But at least with those, stage one, certainly stage one colorectal cancer is survivable.

01:53:34.960 Hepatocellular is a bit tougher, but you're better than a coin toss. But again, coming back to pancreatic,

01:53:40.300 stage one is 80 to 85% not survivable. I've always wondered, what is it about pancreatic

01:53:46.480 adenocarcinoma that is so difficult? And is it simply that its rate of early metastasis is so

01:53:54.160 early that stage one is just sort of a misnomer term? There's such a thing as stage one?

01:53:59.220 I think that's a big part of it. It's kind of a soft organ, the pancreas. It's a very invasive

01:54:05.340 cancer, and you can have local invasion so many places from that site in the body.

01:54:12.360 Plus immediate access to the portal system that's just seeding the liver constantly.

01:54:16.480 It's just seeding the liver, and so it's anatomically a really problematic location

01:54:22.260 for keeping the cancer self-contained. Metabolically, it's a very tricky cancer.

01:54:30.000 It's almost solely driven by mutations of the RAS oncogene. Not saying there aren't other drivers,

01:54:35.600 but almost every patient has this RAS driver. And this is an instruction manual for the cancer

01:54:42.080 cells, not just to divide, but to do a bunch of metabolic things that involve scavenging nutrients

01:54:49.560 from the environment and taking in nutrients in non-standard ways. And so it actually instructs

01:54:54.880 the cancer cells to reach out arms, pull in nutrients, internalize them, degrade macromolecule

01:55:02.020 nutrients from the environment, and use this as a garbage recycling form of nutrient access that makes

01:55:10.000 them very metabolically pernicious. The other thing that we see that's really interesting in new work

01:55:16.280 in mouse models of pancreatic cancer is that they don't have to be very metabolically active

01:55:23.320 in order to be horribly lethal. So the pancreas is a master protein producing organ. You may think

01:55:33.440 insulin is most famous protein to come out of the pancreas, right? But the bulk of the pancreas is not

01:55:39.000 beta cells that make insulin. It's exocrine pancreas. Yeah, that's only 5% of it. The

01:55:43.940 exocrine is the real gland. And the exocrine pancreas is just making digestive enzymes like 1.00

01:55:48.800 crazy. It does by far the fastest protein synthesis in the body. The cancer turns that protein synthesis

01:55:55.120 way down. So it's not hypermetabolic. It's just that it has this huge capacity to make stuff that even

01:56:04.460 when it turns it down still has enough biosynthetic capacity to grow and divide and grow and divide

01:56:11.280 because it's turned down its main energy consuming normal function of protein synthesis. It can function

01:56:17.820 with much reduced TCA activity, reduced ATP synthesis rates. So it's very efficient then?

01:56:24.340 Very, very efficient. So it can turn down all these normal functions and it still has the capacity to

01:56:33.560 reproduce the cells and make these horribly invasive and metastatic cells, which are ultimately lethal.

01:56:39.160 Of all the epithelial cancers today for which we don't have a cure, which is almost every one of

01:56:46.460 them, right? Shy of like a gist or something like that, or certain testicular cancers. Is there one that

01:56:51.880 you're more optimistic about in terms of metabolic approaches to therapy? You see pemetrexid being used

01:56:59.580 effectively in lung cancer. I think you see the cancers with mutational burdens being the ones where

01:57:05.560 you're getting the good immunotherapy responses. Whether they're ones that are particularly susceptible

01:57:11.580 intrinsically to metabolic effects, I don't know. I don't think that's going to be the standalone

01:57:17.700 heart of treating any of these cancers. I think it's more going to be a key piece of the puzzle in

01:57:26.020 getting enough either drug killing by preventing their metabolic escape mechanisms or enough immune

01:57:33.500 activity. And those may be opposites. So you may need also kind of cyclic therapies where you go through

01:57:40.380 rounds of metabolic suppression in order to keep things calm while you can, and then periods of

01:57:48.800 metabolic augmentation that are really directed at augmenting the immune system. I'm a big believer

01:57:54.280 that, you know, there's metabolic limitations on immune response to cancer and that if we can overcome

01:57:59.720 them, we will have major therapeutic benefits. You know, you mentioned RAS in the context of

01:58:05.860 pancreatic carcinoma. RAS is rarely immunogenic in the pancreas. It's a driver mutation, but doesn't

01:58:12.020 give us a beautiful little 9 to 11 amino acid peptide that gets presented on an MHC class molecule,

01:58:18.340 right? It's the great irony of this whole thing. You need more shots on gall. You need more antigens.

01:58:23.700 Do you have any sense of how many tumor infiltrating lymphocytes are typically identified

01:58:27.920 at all in resected pancreatic specimens? It's typically quite the lymphocyte desert. There's a ton

01:58:35.840 of macrophage activity in pancreatic cancer, and so I think macrophage rewiring is going to be a big

01:58:42.820 part of allowing lymphocytes to enter, and these are areas where I think metabolism can be quite

01:58:49.020 impactful. Well, Josh, this is super interesting. I hate that we're ending on somewhat a depressing

01:58:53.940 note. Is there anything more optimistic that we want to talk about than pancreatic adenocarcinoma,

01:58:58.760 beyond the statement that, hey, look, this is why we have the smartest people working on the hardest

01:59:02.800 problems, but anything else within cancer metabolism specifically that you think, boy, 10 years from

01:59:08.840 now, like, I'm really optimistic that we're going to have a new way to hack into their DNA synthesis

01:59:17.960 pathway in a corrupt manner that just spits out mutations using kind of novel systems. I don't know,

01:59:24.440 antisense oligonucleotides, like just something that totally disrupts them in a selective manner.

01:59:29.380 My big hope on this front is that we're going to be able to have some combination of directed

01:59:37.660 metabolic immune supplements and diet that really work with therapy to treat cancer. I mean,

01:59:44.300 cancer is such a discrete disease. The clinical trials are so manageable, and the fact that things

01:59:51.260 are metabolically messed up in the tumors so incredibly clear, and they're so clearly messed up

01:59:56.360 in a way that's favoring the wrong kind of immune cells. And so, ultimately, through either some

02:00:02.360 sort of supplement or diet, we're going to be able to reverse that, and we're going to make immunotherapy

02:00:08.080 work instead of for 10% of patients, for a majority of patients. So, that's my uplifting thought for you

02:00:15.060 is that metabolism will be part, along with, you know, better pure immunological therapies

02:00:20.840 of getting immune control of cancer over the coming decade.

02:00:25.640 And besides reducing insulin, which is such an obvious strategy, are there other metabolic

02:00:33.080 levers to pull with the diet? Because really, between ketogenic diets and cyclic fasting,

02:00:39.560 those are kind of the two ways that you do that. Do you see evidence of amino acid restriction or any

02:00:45.960 other nutrient restriction that could potentially play a role? I think amino acids are complicated,

02:00:51.700 but they hold a lot of potential. I think the type of fat can be important. Saturated and unsaturated

02:01:00.520 fat are really different, and in cancer, they're going to play, you know, different roles. So,

02:01:05.420 it's very nice work from Matt van der Heiden's lab showing that a higher saturated fat ketogenic diet

02:01:11.160 could be more tumor suppressive in some contexts because the tumors have trouble making unsaturated

02:01:17.580 fat in the context of hypoxia. Say more about that. I wasn't aware of that. I know Matt's worked

02:01:22.480 very well, of course. Is he still at Dana-Farber? He's at the Koch at MIT. So, he's actually head of

02:01:27.580 the Koch now, the MIT Cancer Institute. And was this in prostate cancer? This was, oh, I forget what

02:01:34.880 cancer background he did it in. I think it was in pancreatic, at least in part, but I'm not 100% sure.

02:01:40.640 So, that's super interesting. So, a ketogenic diet that was higher in saturated fat posed a greater

02:01:46.480 problem for the cancer cells because they couldn't make, presumably, the essential unsaturated fats.

02:01:55.000 Exactly. I think this is an interesting strategy. I mean, those effects were relatively subtle up to

02:02:02.180 now, but, you know, it could be part of the picture. I think the really exciting part of the diet is

02:02:09.760 also the parts that connect to the microbiome. So, I think the fiber part really working that out.

02:02:16.820 And maybe total protein matters in ways that we don't understand and needs to be. Maybe we need to

02:02:22.880 not just think about cycling, fasting, and feeding, but cycling, you know, for example, there's a time

02:02:30.260 maybe when you want to come in with a lot of carbs in the absence of protein, and that may achieve

02:02:34.320 something that creates a particular immune milieu. And then, you know, you need protein in the right

02:02:40.260 timing after that. So, there's a lot of things we can do with timing of macronutrients that can be

02:02:44.820 interesting. It seems like an eternity before we'd be ready to study this in a human clinical trial

02:02:49.760 because the permutations are so many. So, do you feel like we have high-throughput animal models

02:02:55.720 where we can test these hypotheses and say, we've looked at 10 ways to do this in animals, but these

02:03:02.200 are the three most promising, so we're going to kind of go ahead and do these now? You know, the good

02:03:06.600 animal models of cancer are still not that high-throughput, and there's a lot of challenges converting

02:03:12.940 animal diet and human diet. We'll come out with some work showing that, you know, some of the most

02:03:19.260 exciting dietary combinations are absolutely effective in animals, but they're not effective

02:03:25.600 through the mechanisms that people thought before. Because even in animals, trying to get the diets

02:03:30.660 aligned so that you really isolate variables is tough. I think the fact that we're asking these questions

02:03:35.900 that haven't been asked before is going to build momentum, and we're going to build this interface out

02:03:43.600 over the next five-year period in animals, and we'll do clinical work as a field in parallel with that

02:03:50.840 that has an impact, and it has an impact on patients' lives within the five- to ten-year timeline, I hope.

02:03:57.980 Do you worry that the challenges of, even if you came up with the right diet, so let's just assume

02:04:04.180 10 years from now, the answer is a cyclic ketogenic diet that has this much saturated fat,

02:04:10.900 this much monounsaturated fat, this much polyunsaturated fat, this much glucose on this

02:04:16.300 day, this much protein on that day, like the formula exists. So this is almost an impossible

02:04:20.880 thing, but that it's impossible to adhere to in the way that a pill or a drug or an infusion

02:04:27.340 is much easier? Or do you think that in cancer, because the stakes are so high, adherence will be

02:04:33.940 unlike it is in any other field of medicine? No, we have to make it simple. This has to be

02:04:40.360 clinically actionable. I would go much more back to the first question that you asked maybe when I

02:04:46.100 first mentioned the potential of immunotherapy and fiber, like, is it soluble fiber? Is it

02:04:50.680 insoluble fiber? And then there really are different flavors of soluble fiber. Now, maybe it's the

02:04:56.360 gamish of them, or maybe it's one in particular. Maybe it's one isolatable molecule that relates to

02:05:02.040 that. And then we have one isolated molecule, tiny molecule, that will more than double the number of

02:05:08.600 complete responses you get to immunotherapy in mice. One tiny molecule smaller than glucose.

02:05:14.340 You already have that?

02:05:15.680 We already have that as a metabolite.

02:05:17.840 So it could almost be nutritional supplements as opposed to wholesale dietary changes.

02:05:22.400 These can be supplements. The dietary changes may be in a very acute way, just the way the patient

02:05:28.540 comes in the hospital for a tough bout of chemotherapy or a tough surgery. Maybe we're going to go to a

02:05:33.960 place where we take people's glucose in the hospital almost down to zero for 12 hours with a deep ketosis

02:05:42.960 and some pharmacotherapy at the time that we hit them really hard with chemo. And 24 hours of that,

02:05:49.320 it's like night and day in terms of the overall effect. But it can't be asking patients to give up

02:05:54.280 eating and giving up the joy of food. Or another trial that we're starting now is a trial

02:05:58.980 with SGLT2 inhibitor plus a low carbohydrate, but not fully ketogenic diet to see if it can put people

02:06:06.120 in ketosis. And then just looking forward to, would this be a convenient way to get the benefits of

02:06:11.700 ketosis in cancer patients while still allowing them to have a little bit of breaking bread with

02:06:17.060 the family? Last question for you is totally random. I don't know what made me just think of

02:06:20.820 this. Princeton is the only Ivy League school that doesn't have a medical school, correct?

02:06:24.720 I think that's right.

02:06:25.780 There's got to be a deliberate reason for that. Princeton is fantastic in everything. Do you know

02:06:30.060 why it doesn't have a medical school? And is it just the proximity to Penn or it was assumed that

02:06:36.280 that's where the collaborations would be? Princeton doesn't have a medical school because at its heart,

02:06:42.180 Princeton is a hybrid of a college and a university. And it is an institution that has the ultimate

02:06:49.640 priority on undergraduate education. It's committed to that. It's best in world in that.

02:06:55.780 In assessing how to be best in world at undergraduate education, the Princeton administration many

02:07:02.580 times has asked the question, is having medicine on campus part of that? And the answer has always

02:07:09.040 been, no, let's have a somewhat more pure intellectual environment and keep our focus

02:07:15.920 on doing the very best training for undergraduates.

02:07:18.700 It doesn't have a business school either. And I guess it's the same argument.

02:07:21.220 There's no business school. There's no law school. That's Princeton. It makes it very special. And

02:07:26.440 it's good to have special places that are distinct. I think that's a wonderful thing.

02:07:30.660 Are you at all a fan of Richard Feynman's work?

02:07:32.940 At a very light level, I guess I would say. That's the best way to put it.

02:07:36.760 There aren't places around Princeton that you go to see where,

02:07:39.880 what eating club he was in or things like that. You don't look for the old places.

02:07:43.360 I read a giant biography of Oppenheimer and a little film of that on campus. So Matt Damon was

02:07:49.600 apparently on campus. I'm guessing he's playing Oppenheimer, but I haven't checked anyway. So

02:07:54.220 I guess I've become a small Oppenheimer fan after reading that intensive book.

02:07:58.940 I have three of Feynman's books, meaning like books that were actually his. So I have his table of

02:08:04.900 integrals from when he was in high school, and then two more advanced calculus books,

02:08:09.220 one from when he was at Princeton, and then one from when he was his first professorship at Cornell.

02:08:15.220 It's just, they're sacred to me, right? It's like his scribblings, like his notes all over these

02:08:19.800 things, signed Richard P. Feynman and what his address was and things like that. I've never made

02:08:24.760 the pilgrimage to look for his eating club and things like that. It probably doesn't even exist

02:08:28.080 anymore, but just wondered if you'd been on the tour. No, but you should come down. There

02:08:32.280 definitely have been some awesome Princetonians. I guess my kids all went to nursery school in the

02:08:39.920 building where von Neumann built the first computer. So it is amazing the amount of stuff

02:08:45.160 that happened around here. Well, Josh, so great to see you again. I hope it's not too long before I

02:08:49.440 see you in person again, but really appreciate, first of all, the amazing work that you've done

02:08:52.960 over the past 20 plus years and sitting down to share it with us today. It's been fun. I appreciate

02:08:57.720 the opportunity. Thank you for listening to this week's episode of The Drive. If you're interested

02:09:02.440 in diving deeper into any topics we discuss, we've created a membership program that allows

02:09:07.120 us to bring you more in-depth exclusive content without relying on paid ads. It's our goal to

02:09:12.320 ensure members get back much more than the price of the subscription. Now to that end, membership

02:09:17.420 benefits include a bunch of things. One, totally kick-ass comprehensive podcast show notes that detail

02:09:23.180 every topic, paper, person, thing we discuss on each episode. The word on the street is nobody's

02:09:28.740 show notes rival these. Monthly AMA episodes or ask me anything episodes, hearing these episodes

02:09:34.440 completely. Access to our private podcast feed that allows you to hear everything without having to

02:09:40.500 listen to spiels like this. The Qualies, which are a super short podcast that we release every Tuesday

02:09:46.200 through Friday, highlighting the best questions, topics, and tactics discussed on previous episodes of

02:09:51.120 The Drive. This is a great way to catch up on previous episodes without having to go back and

02:09:56.000 necessarily listen to everyone. Steep discounts on products that I believe in, but for which I'm not

02:10:01.740 getting paid to endorse, and a whole bunch of other benefits that we continue to trickle in as time

02:10:06.840 goes on. If you want to learn more and access these member-only benefits, you can head over to

02:10:10.840 peteratiamd.com forward slash subscribe. You can find me on Twitter, Instagram, and Facebook,

02:10:17.600 all with the ID peteratiamd. You can also leave us a review on Apple Podcasts or whatever podcast

02:10:24.400 player you listen on. This podcast is for general informational purposes only and does not constitute

02:10:29.960 the practice of medicine, nursing, or other professional healthcare services, including the

02:10:34.880 giving of medical advice. No doctor-patient relationship is formed. The use of this information

02:10:40.420 and the materials linked to this podcast is at the user's own risk. The content on this podcast is not

02:10:46.960 intended to be a substitute for professional medical advice, diagnosis, or treatment. Users should

02:10:53.140 not disregard or delay in obtaining medical advice from any medical condition they have, and they should

02:10:59.560 seek the assistance of their healthcare professionals for any such conditions. Finally, I take conflicts of

02:11:05.720 interest very seriously. For all of my disclosures and the companies I invest in or advise, please visit

02:11:12.020 peteratiamd.com forward slash about where I keep an up-to-date and active list of such companies.

02:11:42.020 Thank you.

The Peter Attia Drive - August 01, 2022

#216 - Metabolomics, NAD+, and cancer metabolism ｜ Josh Rabinowitz, M.D., Ph.D.

Episode Stats

Summary

Transcript