"Energy Constraints and Tradeoffs" by Martin Picard

Watch Episode Here

Listen to Episode Here

Show Notes

This is a ~57 minute talk titled "Energy Constraints and Tradeoffs: the mitochondrial daily energy expenditure MDEE study" by Martin Picard (https://www.neurology.columbia.edu/profile/martin-picard-phd) and a brief set of ideas by me at the end, presaging a full discussion of this for our next meeting.

I mentioned these 2 papers:

https://www.sciencedirect.com/science/article/pii/S0303264722001435

https://onlinelibrary.wiley.com/doi/epdf/10.1002/bies.201900245

CHAPTERS:

(00:00) Mitochondria energy tradeoffs

(52:14) Scarcity coordinates biology

(54:44) Anthrobot aging evidence

(56:31) Follow-up ideas

PRODUCED BY:

https://aipodcast.ing

SOCIAL LINKS:

Podcast Website: https://thoughtforms-life.aipodcast.ing

YouTube: https://www.youtube.com/channel/UC3pVafx6EZqXVI2V_Efu2uw

Apple Podcasts: https://podcasts.apple.com/us/podcast/thoughtforms-life/id1805908099

Spotify: https://open.spotify.com/show/7JCmtoeH53neYyZeOZ6ym5

Twitter: https://x.com/drmichaellevin

Blog: https://thoughtforms.life

The Levin Lab: https://drmichaellevin.org

Transcript

This transcript is automatically generated; we strive for accuracy, but errors in wording or speaker identification may occur. Please verify key details when needed.

Main Episode

[00:00] Martin Picard: So the point of departure to get to energy constraints and trade-offs for us was mitochondria. And as a student, I imaged mitochondria in all sorts of ways and realized that there's this diversity of morphologies of mitochondria and also functions and proteome and molecular composition. And here there's an example on the left, or there's mitochondria on the surface of a muscle cell just underneath the sarcolemma. And here you see the collagen fibers, and just the cell membrane was peeled off, you see those globular mitochondria. But if you break a cell in half and you look inside, you see these elongated tubular mitochondria that are really quite gorgeous. So here we're looking at the sarcomeres. So you're looking at kind of in between two sarcomeres. And then we realized that in some cases, the inner membranes of the mitochondria where respiration happens, ATP synthesis takes place, actually bend out of their thermodynamically favorable state, the flat sheet, to actually become aligned. So you see there's alignment between the different mitochondria, which is not random. And that was quite surprising, suggesting the mitochondria can actually exchange information with one another. If you ask Google what mitochondria look like, you get this. But if you dig a little deeper, and this is a slide from my PhD thesis, you find that there's this beautiful diversity of mitochondria. And the little 3D video on the right is a reconstruction of a little piece of muscle from a patient in the clinic that has mitochondrial diseases. And we found that when the mitochondria have mutant mitochondrial DNA, the electron transport chain can't really flow energy as easily as it should, the mitochondria kind of reach out for help and they send these tubular connections to each other. So those are called mitochondrial nanotunnels. And I think that changed how I think about mitochondria from little bean-shaped powerhouses to interconnected network. And the nanotunnels between mitochondria is similar to what we know now as the intercellular nanotubes, where mitochondria can be shuffled and other kind of cytoplasmic content. And it's also similar to what bacteria do to exchange plasmids and what plants do with stromules. So there seems to be kind of a conserved process there. If you're lucky enough to work in a lab with a nice confocal microscope, you can make the mitochondria fluorescent. I think most cell biologists haven't seen videos of living mitochondria in the wild. That's what they look like. You can see that they have this social life where you have mitochondria that don't interact, then they start to interact with one another. Then this third one here comes along. They're probably exchanging information and then this one says, "Okay, I've had enough," and then just leaves. So this kind of social behavior, and we know there's a life cycle as well for mitochondria. The old ones will bud a piece with damaged content that gets degraded, and then new mitochondria are born out of growing from the existing ones. And there's an old model we developed to try to link the biology of mitochondria with the abundance of energy, the metabolic state. Either you can be in positive energy balance, there's a lot of resources around more than what you need, and then you can be in negative energy balance when the system is starved. And what emerged from in vitro studies is that if you go through the schematic here, is the normal schematic of a day: you're a little hungry, you eat and you're a little full, and then you're waiting, you're hungry again. And then there are periods of oversupply like Christmas break where you overeat and you don't move, or undersupply where, you know, you're calorie restricting, for example. And the literature in cultured cells was showing if you starve cells a little bit, the mitochondria start to branch and fuse with one another. If you overfeed cells with too much glucose, too much fat, they become fragmented. And what we knew from manipulating the proteins regulating the fusion, the fission of mitochondria is that small fragmented mitochondria, if you prevent mitochondria from fusing, they start to accumulate damage much faster. And that leads to an erosion of mitochondrial bioenergetic capacity. If you promote mitochondrial fusion, it seems to preserve mitochondrial respiratory capacity. And for a cell with more fused elongated mitochondria, that promotes survival in the context of stressors, a stress rate that would normally kill a cell if the cell has fragmented mitochondria. If that same cell has elongated fused mitochondria, it might survive and be able to come out alive on the other side. So this led to a model that maybe that's why physical inactivity and high calorie intake, overeating, is bad for us because it fragments the mitochondria and maybe why physical activity and low calorie intake is good for us because it promotes the recycling of bad mitochondria, the fusing of the existing ones. And then if we bring this back into the cell, I realize also through imaging cells and mitochondria, very often they're lined up. You see at the bottom, this is an electron micrograph. The nucleus is in blue, the cytoplasm in green, the outside of the cell in pale green, and the mitochondria here in like orangish brown. And if you image this interface, the nucleus, the mito, the cytoplasm, the mitonuclear interface, and you can see here actually the little dark areas at the connection between the mitochondria, these are electron-dense contact sites, or what we call intermitochondrial junctions, and the cristae are aligning there.

[04:45] Martin Picard: So there's kind of this network of mitochondria that surrounds the nucleus and effectively filters information coming from the cytoplasm. And there's a new paper last week in Nature showing that mitochondria are physically tethered to nuclear pores in some cases. Here's a magnified version of this mitonuclear interface where you can see that mitochondria are just a few hundred nanometers diffusion distance away. This is like less than a minute, I think, for diffusing a small molecule, like a polar metabolite. And then, of course, in the nucleus is the epigenome and the malleable transcriptional machinery. So mitochondria can literally turn on, turn off several genes. And in one study where we perturb, we dialed up the mutation rate, what's called heteroplasmy in mitochondrial DNA, we found that over two thirds of genes in this cell line were regulated just by tweaking the mitochondria. So that suggests that a large number of genes in nucleus, of course, perhaps, are under the control of energy metabolism and signaling from the mitochondria. We wanted to know if this was true also in mice and large living creatures and mammals. So we took mice with normal healthy mitochondria, the wild type, and then green and yellow mice have mutations in the mitochondrial DNA, like some of the patients we see in the clinic. And then blue and red mice have other gene defects. ANT1 is an ADP ATP translocator. So it allows the ATP in mitochondria to leave, to go into cytoplasm, and ADP to come back inside the mitochondria to be re-phosphorylated by the OXPOS system. And the NNT is nicotinamide nucleotide transhydrogenase that uses the membrane potential that mitochondria generate by fluxing electrons to regenerate antioxidant defenses like peroxyreduxin and thyreduxin systems inside the mitochondrial matrix. So these are four different genetic ways of perturbing the biology of mitochondria. And then we ask, if you perturb the mitochondria, do you change how an animal responds to perceived or induced stress? And here we wanted to stay away from physical stress and just look at mental stress, which from the perspective of an organism, perceived or actual stress, in many cases, stimulate the same kind of responses. And one of those responses, the HP axis, the hypothalamic pituitary adrenal axis, leads to the release of steroid hormones, cortisol in humans, corticosterone in mice. And what you see here is that the mice with the normal mitochondria, wild type, black line, there's a beautiful increase in corticosterone during the stress. Here we put the mice into a small, a 50 ml falcon tube with the holes at the end so they can breathe. But when they figure out they're stuck there, they can't back out. We put a little Kim wipe with a piece of tape. Then they freak out and activate the HP axis, mount the stress response, neuroendocrine axis, very similar as what humans activate in response to socioevaluative stress like public speaking or where you feel you're being judged or you feel like your ego is threatened. So in that response here you see is magnified, doubled in mice with the energy exchange defect, the ANT1, and is decreased by 50%, half in mice with the intramitochondrial oxidative stress. So if you can extract a few parameters here, the magnitude of the response, the time of recovery and things like this, we extracted features of the stress response from the HP axis. We looked at cytokines, circulating amino acids, gene expression, the brain, and so on. So we extracted a total of 74 different stress response parameters across modalities. And what you find here is compared to the mice with the normal mitochondria, the different mitochondrial defects elicit not just a magnified overall stress response, they actually qualitatively change the stress response. And if you feed, condense those 74 features into two principal components with PCA, you find indeed that depending on which way you perturb the mitochondria, you tweak the stress response directionally or qualitatively. And here the proportion of explained variance in this two component model is about 75%. So I think that was some of the first data to show that mitochondria really can regulate physiology, not just in terms of magnitude or vulnerability, but tweaking the nature of the stress responses. And what we've concluded from the imaging studies and some other work and this physiological work is that mitochondria really are a dynamic, a social collective involved in energy transformation and signaling. And the question I was excited about asking when I opened the lab was what role do mitochondria and energy play in regulating complex cellular and human behaviors? And in particular, what happens in the context of aging? And I know that's something you've thought about. So I wanted to just see how this lands in the context of aging and what you've been thinking about in that direction. What seems clear is that as cells age, things change molecularly and they accumulate a lot of damage. Changes happen in the epigenome. There are mutations that accumulate. Pieces of mitochondrial DNA actually translocate inside the nucleus and become inserted. These are called numts, nuclear mitochondrial DNA insertions. We found people who have more numts in their brain die earlier.

[09:30] Martin Picard: So there might be some functional consequences there that seem to destabilize the nuclear genome, trigger retrotransposons and transposable elements. Anyway, there's a bunch of things that happen. Collectively, all of these drive stress responses, right? And as a result of mitochondria signaling some things and the intrinsic damage accumulation, aging cells start to secrete this senescence-associated secretory profile that costs a lot of energy. You need to turn on those genes, make those RNA, make the proteins, package those proteins, fold them, and then exocytose them. All of this costs energy. There's also mitochondria release cell-free mitochondrial DNA that triggers DNA sensors that cost energy. Transcriptional noise, some genes are supposed to be repressed. Some neuronal genes in a fibroblast or in a cardiomyocyte should not be expressed. With aging, you seem to have this ectopic gene expression that's induced, in part driven by epigenetic drift. There's more aging, senescent cells seem to make more mitochondria, likely to compensate for some of the impairments and the increased energetic cost. And in many cells, at least in vitro settings, there's a shift towards glycolysis, kind of a Warburg-like effect, which is a shift away from what was most optimal. The optimal metabolic strategy to subserve energetic needs in the young cells is no longer appropriate. So that shift requires that you increase the flux through some pathways, make more enzymes, and so on. So all of this leads to a state of hypermetabolism. And that happens quite clearly. We reviewed that literature a few years ago. It happens in aging cells, but when, as we'll see, you look in a whole organism, you don't really see that phenotype. So there's clearly an increase in energetic expenses that we can derive from first principles, but it also happens measurable if you look at this in cells, where damage accumulation leads to stress responses, and nothing is free in biology, so stress responses increase the cumulative energetic cost. And we think in this context, senescence and quiescence, when you have aging cells that go quiet and stop replicating, on the surface they look like they're not doing anything. But if you peel, you know, the surface of that phenotype, you find that there's a bunch of things that are activated. And if you look at secreted proteins, transcriptome, there's a lot of things that are up when the cell has actually down regulated its behavior. So aging senescent cells are clearly hypermetabolic, but the whole body is not. And this is some of the best data that's from a few years ago showing total energy expenditure that's adjusted here for body weight. Total energy expenditure from the moment of birth throughout the whole adult lifespan up to 97 years old, I think. What you see is kids are hypermetabolic. When babies come out, they don't burn too much energy. But within the first year, they become hypermetabolic. A lot of calories need to come in as they're growing their bodies. And then at some point, the kids age five to 10, they start to grow personality and learning about the world and patterning their brain. And so there's a lot of energetic costs. But you look at the adult lifespan, right? From 20 to 60, it's pretty much flat. But 20 to 50, 20 to 60, and beyond is where you have some cells that become senescent that start to burn more energy, become hypermetabolic. So there's cells that become hypermetabolic, but the whole-body energy expenditure either is flat or even declines in older years. And that's been shown in other studies as well. That's from the Baltimore longitudinal study of aging. You have age here on the x-axis, resting metabolic rate. How much energy does it cost to just stay alive? You're sitting down, not contracting a muscle. How much energy do you burn? And you see here the trend to a decline. And that's true in both men in blue here and in women in red. If you look at longitudinal data, so you get those patterns cross-sectionally and longitudinally. People become hypometabolic with age.

[14:15] Martin Picard: So if you bring those things together, that doesn't quite align. There's cellular hypermetabolism that happened, but there's whole-body hypometabolism. So is this some kind of energetic paradox, and how can we resolve this? So we want to know what's the link between energy expenditure, aging, and lifespan? And there's a bunch of work that's been done on this, on allometric scaling, linking metabolic rate and how long animals live, how quickly they accumulate damage. And there's kind of this concept that processes, energetic processes in the world that are very exothermic, happen within seconds. And then there are living creatures that live for centuries or millennia, like there are redwoods that live for about 2,000 years. And then the animals are kind of somewhere in between this. They kind of contain or slow down combustion process through metabolism, burning oxygen, making CO2. And then humans are kind of in between there. And small creatures tend to live shorter. They're more like fire, and big creatures are more like long-lived trees. So hypermetabolism, if you look at correlations, it looks like hypermetabolism is associated with lifespan. So is hypermetabolism driving the aging process is one question that a lot of people have been trying to ask. It's not that simple. It's not hypermetabolism, I think, that drives aging. And we can talk about what might actually be happening. But to study this in the dish, what we did was to take human cells, expose them to a stressor, something that doesn't damage the cell per se, but makes the cells anticipate stress and then mount a stress response. And what we use for this is dexamethasone, which is a glucocorticoid agonist, and it taps into this evolutionarily conserved stress axis that basically, through evolutionary history in humans and in other creatures, if cortisol is secreted in the blood, it's because either there's something dangerous happening right now, or you're anticipating something dangerous is going to happen. So our cells evolve to respond to that signal and then mount this preparatory or allostatic response. What we did here was take cells from three different people, healthy donors, and then you have the primary human fibroblasts. And then you look at how they divide, this is a Hayflick limit type experiment, over days in treatment up to 250 days, that's nine months, a full gestation period. And then you just monitor population doublings. And you can see at the beginning of the healthy cells, every 40 hours or so, there's a population doubling and there's a nice linear increase in the number of population doublings over time. And then that eventually slows down. And when you get to about nine months, cells no longer divide, right? So there is senescence and what we call, I guess, deep senescence. If you expose the same cells, the same person, but to dexamethasone, you do your lifespan study, you find that they slow down cell division right away, and they never reach the same population doubling as the other cell. You can do this in other different people. Roughly, the result is the same. We get about 20% reduction in the Hayflick limit, which, interestingly, 20% reduction is what you see in human studies of chronic life adversity and severe trauma and deprivation. So maybe there's a quantitative parallel there. Then the question we want to ask is how much energy does it cost, right? You have these cells that divide less so that they should be consuming less energy, but no, if you actually quantify total ATP turnover, and you can do this with the Seahorse longitudinally, right? So you do this in the same cell line, the same cell population week after week, you find that compared to the control condition where there's no dexamethasone, in the presence of this stress hormone, which again isn't damaging, but it tricks the cell into thinking there's something dangerous happening, cells are mobilizing energy. And you can see on some passages, on some weeks, you're like 200, 300%. So those cells are dividing more slowly, but they're burning two or three times more energy. So if you average this across a whole lifespan, you're at about 62% increase in energy expenditure. What effect does this have on aging? You can monitor aging biologically with the epigenetic clocks.

[19:00] Martin Picard: People have trained those clocks in human tissues. You take a few thousand people age 1 to 90 years old, and then you just do genome-wide DNA methylation profiling, and you can identify a few hundred positions in the genome that either go down or go up with age, take those into an elastic net model, make a prediction algorithm, and then you can predict the age of the person with three, five years accuracy, pretty accurate. We found that those clocks actually work in the dish, and you can use them longitudinally to track the rate of aging. If you take here the first derivative, that slope is the rate of aging for those cells. And you can see the three donors, which roughly age around the same rate in the presence of stress, that all three donors increase the rate of aging. This epigenetic age is by about 36%, and then a third way you can look at aging is telomere length. I'm not showing here the same result. And then here's the correlation between how much energy is this person's cells burning on the x-axis and then how much cell death do we find at every given passage? And what we found is that on passages where the cells are burning a lot of energy is where there's a lot of cell death, suggesting that there's a connection. Either the cell death process costs a lot of energy or the other way around, which is the hypothesis that I think is most consistent with other results. When the cell is undergoing a very strong stress response, it burns a lot of energy and then eventually that becomes unsustainable and cells die. So how much energy does it cost to worry about the future? Quite a bit. And we profile a number of things in these cells longitudinally to find that there's even those cells are dividing more slowly and they age faster and they divide for fewer passages, there's actually a lot more happening under the hood. And ultimately that leads to kind of the wear and tear, and it's consistent with the model of allostatic load and allostatic overload. So aging accelerates depending how you look at it by 10 to 40%. So what could be happening here? And this is where we start to think about energy constraints, and there might be kind of a pool of energy that needs to be distributed competitively between different processes. And that's well known to happen, you know, and people have studied this in life history theory, and birds that have access to less food around will have smaller clutches. And that's true also in other animals. So the physiology of the animal can kind of respond to how much food is available and change its reproductive rate. If you're a young woman and you train for long-distance events, you will lose your menses. Amenorrhea is not a symptom of dysfunction ovaries or some reproductive defect as far as we understand, the organism feeling that there's not enough energy around and trying to cut costs, and it costs energy to menstruate every month. And there's also well-documented immune growth tradeoffs. So in kids that live in Central America that are exposed to more pathogen in their food, they don't grow as quickly. And many of them don't reach the same height that their counterparts would in a condition where there are fewer parasites and immune challenges. So what's common here is that the function of the behaviors that happens at the cell, organ system, and behavioral levels compete for the same energy budget. And maybe there are limits to energy dissipation. There's a competing kind of theory around this that I'm not sure is as robust as the energy budget theory. So the bottom line here is that the energy budget comes from two things: the ability to supply energy through the vascular system and the ability to transform energy from the food you eat and the metabolic pathways, both glycolysis, anaerobic, and aerobic oxidative phosphorylation. And together, those contribute to cellular energy. ATP is one currency. There are other currencies as well. And then that leads to or yields a total energy budget. That's X number of units that's different across different species and it's different across different people. But if you look at what's true in humans, you can plot here how long an event is, right? How many hours something lasts. And then on the y-axis is metabolic rate. How much power can you develop over that period? The first dot on the top left is the 100 meters sprint. And you can see you can develop a lot of power if you're only going to push for 10 seconds. If you're going to run a marathon, which is close to the final point on the first segment there, you can't go as hard as when you run 100 meters, obviously. And if you're going to grow a human being for nine months, that's pregnancy.

[23:45] Martin Picard: You're going to be above resting energy expenditure for a normal human. But you can see that you're here on this line. And these are all world records, by the way, like this is the human body being pushed to the maximum. And what you see is that on the right is the red zone. It's not humanly energetically possible. And anything that's on the left is within budget. So what sets that constraint, right? And over a nine-month period, you can clearly eat more food. Metabolic rate is a little higher, but there's nowhere near exercising intensely, for which we know the human body can do. So those data suggest, and many other data suggest, there is a limit and the organism somehow there's maybe some kind of central pattern generator that can sense and control how much energy is expended and adjust behavior so that you don't go over budget and die. So one approach to start to explore this in humans and other systems is to think about how energy is partitioned across different processes in the body. And there's a classic terminology that involves resting energy expenditure and how much energy you expend to move, activity-related energy expenditure, and then how much energy you expend to digest the thermic effect of food. And those have been useful methodologically if you're trying to measure this in human beings, but they don't reflect what's happening under the hood, which really is what we're trying to understand. So the proposed terminology that is more biologically informed and I think behaviorally relevant is this. The energy budget, which is the vertical bar here, is compartmentalized into vital processes, which are absolutely required for sustaining waking life. There's stress-related processes, which is anything that comes above what's required to sustain life, required to face acute challenges, as well as chronic stressors and challenges. And then there's this third bucket, which is what we call GMR, growth, maintenance, and repair. And those processes are there to optimize functions, contribute to healing processes, and enhance the long-term functions and lifespan and health span, right? And the key idea is that there's a hierarchy of energy needs. Those three buckets cannot be prioritized the same. Otherwise, physiology wouldn't work. And the concept here is inspired by Maslow's hierarchy of human needs and Maslow's insight. There's some issues with this, but I think the core idea is important and valid, is that there's some things that need to be secured first before you can engage other processes. And Maslow's insight was, for a human being to fulfill their potential, first, they need to have their physiological needs met. Then they need to feel safe. Only then can they think about creating connections with other human beings and developing relationships. Only then, once that's set, you can think about developing your esteem, your skills, and so on. And then only then you can engage in the most frivolous things about being a human being, like think about consciousness and spirituality and these kind of things. So under stress, the core idea here, the top of the pyramid is the first piece to go when the **** hits the fan and you're kind of short on resources. So the same thing we propose happens at the level of cells and organisms. And so there's a hierarchy of energy needs where there are vital costs that the cell cannot trade off. Maintaining membrane potential and keeping metabolism going is not negotiable. Then there are stress responses, making your proteins do to engage in responses and adaptive processes, stress response, like the integrated stress response, signal transduction and so on. And once those energy needs are met, then the cell can kind of expend extra energy towards replicating, making DNA, RNA, repairing its genome, telomerase, antioxidants. And then if those needs are met, then the cell can kind of invest energy into things that will keep it healthy and happy over the long term, over the long run. But the top of the pyramid, again, are the most dispensable processes. And if a cell is in trouble because it's turning on massive stress responses because of DNA damage, because of some genotoxic stress, then of course it's going to stop replicating and it's going to stop investing in living a long, healthy life because it just cares about surviving tomorrow. And the same thing appears to happen as well at the level of physiology, their vital organs, their stress responses, HP axis and so on. And then there's maintaining very large muscle mass and a well-diversified immune system. You don't need that to survive. And then developing skills, healing vitality, and engaging with others and reproducing. These are things definitely that are frivolous if the goal is to survive tomorrow or next week.

[28:29] Martin Picard: So if you bring this into the time domain, now you have the stress, GMR, and vital, and you're looking at what happens over days to months. And what happens if there's a small stressor that comes by is you need to expend the stress cost, right? You need to expend energy to power a faster heartbeat, to release cortisol, and fire some more neural circuits, pathways in the brain. All of this costs energy. That energy needs to come from that budget. Over a short period, you can expand the budget a little bit. But remember, there's an evolutionary pressure that suppresses the expansion of energetic cost. So what happens then is the expansion of stress cost needs to dig into GMR, right? So the GMR processes are partially temporarily dispensable or flexible as well. Every cell can shut down those GMR processes. And then when the stressor disappears, now you can expand into the stress cost and GMR can do its thing and prevent long-term damage. If there's a severe stressor, now you can cut completely GMR cost, and in very severe cases, let's say sepsis, then the stress costs might even dig into vital costs. And that's when you would start to have vital organ failure and multi-organ failure and so on. And then when the stressor goes, then you might be expending a lot of energy recovering post-marathon, post-surgery, and then hopefully the GMR processes are good enough, and there's enough energy to support them, that you can completely eliminate the damage that has accumulated in the meantime. And sometimes you can't. I think this model is getting us to a point where we can start to measure. The question can you measure, and how do you know if the stress costs are expanded or contracted? And these are some initial ideas that we think reflect energetic stress and the stress portion of the energy budget fairly well. There's some growth, maintenance, repair-related things like growth hormone, insulin growth factor one, and repair enzyme at the cell level. There's some hormones that are anabolic. And some of those things, GMR, you know, they're well known, for example, to be suppressed in context of aging or in the context of severe mental stress, probably because of the trade-offs we've talked about. And then vital, I think, is the most challenging. How do you quantify how much energy an organism is spending just to stay alive is a tough one, but there might be some surrogate proxy. So that paper, I had it in press in previous slides, but it's out as of last month. And I'd love to have people's thoughts on this. What happens when the demand is chronic is that over minutes, again, the budget can expand. You can do a sprint and expand a lot more energy than your basal. But then over days, months, and decades, there's this kind of pressure to bring energy down. And we call this the dogma of efficiency. Somehow organisms have evolved to tend towards maximal efficiency. And the chronic suppression of growth, maintenance, and repair could be driving the exponential nature of the aging process. And as we age or as stressors become chronic, there seems to be this increase in vital costs. It just costs more energy to be alive and therefore you suppress. That might explain why there's decreased resilience, you know, what people call resilience or the ability to respond to challenges and mobilize energy where it's needed. These were just a few examples. I think it's best that we skip this and have more time for discussion. Question is, how is this happening between different organ systems and how are energetic states signal from one part of the body to another part of the body? And one way we've been thinking about this is like pain, nociception, right? There's a signaling cascade for nociception where there's an injury, that injury, tissue damage is transformed into an electrical signal by a nociceptor. And then that travels to the brain. The brain then becomes aware that there's something damaging or dangerous.

[33:14] Martin Picard: And then it does two things. It mobilizes response to conserve the integrity of the organism. And then it also mobilizes processes that release energy or make energy available for fight or flight, for example. So that's called the nociceptive cascade. We think there's something similar that we could call metaboception. Energy is so central to all of life that there have to be processes in place to sense your metabolic state and respond to it and make sure you're not running out of energy. We know there's a lot of these metaboceptive processes that connect peripheral physiology to the brain. GLP-1 is an example. The gut can make GLP-1 when there's food in the tube, and then that tells the brain there's food here, there's energy, stop eating. Leptin is another one. It comes from fat cells, and there's a few examples of this. But what about how can you sense how well energy is flowing in your mitochondria? And how can you sense the flux of energy and the energy demand relative to the capacity to transform energy, right? That delta. And we think a good way to think about this is think about this as mitoception. So the ability to interocept the state of energy deficiency or energy gap inside the cell. And in that context, every cell is a little agent, a little computing agent that is a metaboceptor. So every cell can make that computation. How do I have enough energy transformation capacity, mitochondria, other metabolic pathways relative to what I need right now? And if you're exposed to a challenge, a stressor, you're going to start to activate stress responses that cost it a lot of energy. So if the cell makes that assessment, either in the moment or predicts that it's going to run out of energy, the body is a cell collective. That cell needs to let other cells know that it's energetically in trouble. So cells can sense, we think a key trigger here is reductive stress, electrons basically that can't make it to oxygen as smoothly as they should. And then the NADH, NAD ratio is a sensor of reductive stress that initiates gene expression cascades, releasing cytokines. One of those cytokines is GDF-15, growth differentiation factor 15, which can travel to the blood, go to the brain, and then in the brain, there are beautiful mouse studies that show if you activate the neurons that sense GDF-15 in the area of post-trema, kind of the nausea, vomiting center of the brain, it does two things. It conserves energy. It triggers experiences of fatigue, social withdrawal, and kind of sickness behavior, but it also mobilizes energy. It triggers a HP axis, glucocorticoid release, norepinephrine, the sympathetic nervous system, which together increase lipolysis, gluconeogenesis. So it makes energy available for the body to basically rescue this energy deficient metaboceptor or cell. If you bring this together now in the context of aging, this leads to this concept of brain body energy conservation. And this is, I think, a plausible pathway that leads to phenotypic aging in humans and in other systems where if you're young, plenty of energy available relative to what the body needs, right? So the capacity versus demand. But as you age, senescent cells accumulate damage, become hypermetabolic. Those senescent cells signal their hypermetabolic state via GDF-15 and other cytokines, chemokines that reaches the brain. The brain perceives and interprets this as there's energy deficiency either right now or incoming. Let's save energy. And the subjective experience of this here is like fatigue or low energy, vitality, physical activity decreases, libido decreases, that costs energy too, appetite costs energy to digest. And then stressors might kind of compound this. But overall, the brain has a veto on so much of the body and the brain can, you know, modulate the immune system and the muscle mass and all sorts of anabolic, catabolic processes. So many of the manifestations of aging and frailty, like losing hair color, shrinking the brain, having less acute sensory capacity, hormonal deficiency, you know, quote unquote deficiency like tired hormone insulin and sensitivity, like all of those features could be understood as the brain trying to save energy and suppressing systems and organs that consume a lot of energy. So we've formalized this into this concept where you have hypermetabolic cells that signal the brain. The brain is kind of the energy broker and says, okay, something's running out of energy.

[37:59] Martin Picard: Let me save costs. And that's what we call phenotypic aging. So the brain moderates energy trade-offs and the organism. And to look at this and ask specifically, what effect do mitochondria have on these different processes? And if you change how well energy flows in the mitochondria, can you change the energetics of the whole body, of the whole system? So we developed this mitochondrial daily energy expenditure study, MDE, which was a sister study of the parent MISB study, the mitochondrial stress brain imaging epigenetic study. This ran over six years. It was a lot of work to do, but we've learned so much from that cohort. We recruited 70 people with healthy, normal spectrum mitochondria. And then 40 individuals with either a point mutation in the mitochondrial DNA. It's a point mutation in a transfer RNA that messes up protein translation of the 13 protein-coding genes in the mitochondria. And then people with large-scale deletions, they're missing a chunk of mitochondrial DNA, which takes out a bunch of transfer RNAs and mRNA coding genes. So the mitochondrial disease group are two different types of mitochondrial defects. So in humans, MISB was meant to be kind of the human translation of the mouse study that we did initially, where we procured the mitochondria, looked at stress responses. So MISB had this protocol of stress reactivity and collecting blood at multiple time points. And I'm not going to talk about this here, but there's a, on our website, there's a MISB page with a little summary video that talks about what we've discovered. And there's, I think, eight or nine papers on the MISB study that are out there. I'm going to focus on MDE, the sister study, where we recruited 20 people, 10 controls, 10 mitochondrial disease, and then we ask, what's happening to their energetic state? How much energy does it cost to live with mitochondrial disease? And for MDE, we brought people in again in the clinic, and then we took a basal metabolic rate measure in the morning, quantitative magnetic resonance scan to look at body composition very accurately. And then we dose them with doubly labeled water, which is a technique that allows us to measure metabolic rate over a whole week, up to 10 days. Then we put them into the MRI. We scanned the whole body to measure how big the muscles were, the volume of the kidneys, the liver, the heart and the brain. And then they met a physician to make sure they're healthy enough to go into the chamber. And then we collected urine, and that's a protocol. When they went at home, we collected urine every day for 10 days to look at their energy expenditure out in the wild. And there's a cool special technique that I won't talk about here. The coolest part of MDE was this: day number two, they come back and then we put them into this chamber, which is, you probably know of metabolic chambers for mice and metabolic cages. This is like the human metabolic cage. And I was the first participant in this. I spent 24 hours, and it's a very small bedroom-sized thing. There's a cot, a toilet and a sink and a TV. And you're in there with an IV line in your arm for 24 hours. And then every hour, the nurse draws a bit of blood unbeknownst to you, even during the night when you're fast asleep. The meals are exactly the same for everyone at exactly the same time. We collect saliva also during the whole day. And then the cool thing with this chamber is that it's mostly airtight and you can collect air from the chamber every minute to measure VO2 and VCO2. So you get a direct in and you measure the electron flux in the mitochondria, how much oxygen are they burning, how much CO2 are they producing. And then we measured accelerometry and movement, continuous glucose monitor and actigraph. And during the night, we have a little EEG system that lets us know if the person is really sleeping, are they in REM sleep or light or deep sleep. So here that, again, we did this in 10 healthy controls, 10 people with mitochondrial disease. I'm just gonna show one piece of data from that cohort, which is most amazing. Here's the data for energy expenditure. This is actual versus predicted, just a way of mean-centering the energy expenditure for each person based on their body weight, body composition and so on.

[42:44] Martin Picard: For every minute, this is a smooth average of minute-level energy expenditure across the whole day. So they come in at 9 a.m. and they're in the chamber. You see the three meals here, the bars. 00 p.m., we turn the lights off. So the blue segment is the lights-off period. 00 a.m., we turn the lights back on, turn on the little intercom to say good morning. And then we draw blood at four different time points and we can quantify the energetic cost of waking up in the morning. So it turns out you can see here how much energy people are spending. Every line is a person. The dark lines are the group averages. People with mitochondrial disease burn more energy across the waking hours. 00 PM here, there's a little spike. This is when we ask people to put the electrodes at the back of their head and to get ready for the night. So we engage with them. And it looks like engaging with them costs a significant amount of energy. And then they go to sleep. Look at this, this beautiful hypometabolic state, which I suspect is why we need to sleep. Why animals need to sleep is to go into this hypometabolic state, where the stress costs, right, and our stress GMR vital costs, the stress costs go down to almost zero. And then GMR can expand and heal. The body can heal during that period. So that's our working hypothesis. But even during the night, right, when people are unconscious, having a mitochondrial disease costs you more energy. And you can quantify this. It's quite robust and significant, even with 10 people per group. And what we found was that people with mitochondrial disease, they burn more energy, not because they move more. If anything, regardless of how you quantify this, people with mito disease move less. And we've known this just because they feel terrible when they move, so they avoid movement. So if you compute non-resting energy expenditure and you normalize this to movement, that's a measure of efficiency, right? How much energy do you need to do a given amount of movement? This is about double in mitochondrial disease. And the biggest physiological signal we found is parasympathetic activity. You can tap into this with heart rate variability. This is SNDD, it's same RMSSD or whatever, or the metric of heart rate variability. You see in normal healthy people, in controls, heart rate variability increases during the night. The first day is this chamber section here, then they go home and we continue to monitor heart rate continuously for 10 days. So you can see every day there's this beautiful oscillation. In mitochondrial disease, it's practically a flat line. So that suggests that the parasympathetic arm of the nervous system is hypoactive. And the parasympathetic arm is what puts the organism into this kind of rest and digest state. And that's true during the wake and during sleep in particular. So what we think is happening here is this: at the cell level, there's a mitochondrial defect that triggers stress responses in a nucleus, integrated stress response.

[47:29] Martin Picard: The nucleus tries to make more mitochondria biogenesis, that costs energy. And then there's compensatory responses that happen, increased glycolysis, for example, reductive stress, that costs energy. Then the cell tries to tell other cells that it's in trouble through releasing cytokines, metabokines. That costs energy as well. And so overall, there's an increased cost of having just a mitochondrial DNA mutation. And there's intracellular signaling that increases ATP consumption at the cell level. So the result here is decreased cellular efficiency. That is signal to the organism. And the state of hypermetabolism now drives increased energy expenditure, increased heart rate, increased minute ventilation. There's all of these physiological systems that are activated. And then there are biomarkers like lactate and GDF-15 that signal onto organ system to change physiology. That costs energy as well. And there's psychosocial factors, how you feel, whether you feel supported or the stress in your life can also increase your heart rate, increase stress responses, and contribute to the energy expenditure. So there's also physiological efficiency, which I show this decrease now we know in mitochondrial disease. So cumulatively, this can lead to accelerating biological aging, like we saw in the cells exposed to dexamethasone. And then that can lead also to system-level energy constraints. Like as we know, overtraining, if you overextend energy in one domain or exercise, then there's not enough energy for menstruating, for example. So some of the symptoms in mitochondrial disease could be driven not because of a mitochondrial defect that leads to ATP deficiency. It actually leads to an increase in ATP turnover that forces trade-offs. You shut down digestion, for example, you shut down vigor, so you have fatigue. So some of the fluctuation and symptoms and why many people end up dying of infections, because it seems like the immune system can really steal energy away from the vital organs when that's needed. And we have evidence that some hormones like FGF21 are involved in this. And there's a beautiful dynamic that were unsuspected in some of those hormones that were thought to be like flat and general, but they're actually quite dynamic. So overall, this is the picture. It's a bit of a mitocentric framework, but I think it reflects where energy is transformed initially from food to oxygen. And then that ripples out into the function of cells that talk to each other. And then cells talk to each other and make functioning organs. Organs talk to each other and make functioning people. And then people talk to each other and make functioning communities. And so there's a rippling of energy and information and patterning from low level within the cell to out into our behaviors. So there's information flow this way. There's also information flow the other way around. And I didn't talk about this, but it's clear now that our states of mind, as we know, can change brain activity and heart rate and things like that, change cellular biology, but it also can change the biology of mitochondria and energy transformation at that level. And I think there's a lot to be discovered there. So in summary, I showed you how starting from mitochondria, we became interested in how mitochondria regulate physiology and the stress responses, the aging process. It's clear now that if you increase stress responses that cost energy, and that is correlated at least with accelerated aging, likely energy trade-offs and competition drives human aging. And with this energy constraint model, I think we're in a position to start testing this. And the concept of metaboception, mitoception might help to bridge cell-level, subcellular-level energy dynamics to whole person behaviors. And we have initial evidence now that in mitochondrial disease, if your mitochondria can't flux energy properly, you don't go into energy conservation mode. You actually struggle and that costs energy, physiological and psychological struggle has an energetic cost, and that might be driven by some intra-organ, intercellular dynamics. This is our amazing team that's doing most of the work now, and we're exploring some of those questions at larger scale, and I think it's a really exciting time, including measuring mitochondria in the brain and asking how this relates to how people feel. We have amazing collaborators that make this possible and funders, including Baszucki Group. And for those of you who might be interested, we have a new Substack called the Science and Experience of Energy, where with my wife, Nirosha Murugan, we're bringing the science of energy together with the human experience and the stuff that makes life real and worth living for and challenging. And I think we're all human beings. Science is done by people, and the kind of things we decide to focus on and we think and make a difference in the world necessarily involves both the scientific process, but also the human experience. So if you're interested, please join us. There's a growing community there. So if we have some time, Mike, I'd love to hear your thoughts on this.

[52:14] Michael Levin: Super. Thanks so much. Unfortunately, I got about four minutes before I have to peel off. But that was really good, and I can just tell you a couple of things and we can discuss them in depth next time. One thing is that we did some simulations of whether, if given the choice for coordinating things—I mean, in my case, morphogenesis, but you can, I think, expand it to other things—whether evolution would prefer to use limited or unlimited pools of resources. And in simulation, one of the things we found, I can send you this paper, one of the things we found is that it actually prefers to use limited pools. Scarcity is good. And the reason it does it is because if it's an infinite pool and you take from the pool, I can't tell what you've done. Whereas if it's a finite pool and you take from it, I see it going down. So it's a mechanism of coordination. So it actually, and evolution notices this like immediately when you start to evolve over systems like this, they prefer to use it ends up preferring to use these kind of limited pools as a coordination mechanism. So that's kind of interesting. And there's all this, like we did a review of all this biological data where, kind of like some of the stuff you showed, where you do one thing, like for example, in food animals, there's a lot, you do something and then some other feature balloons out. And partly, I think it's this, I think actually it's being used as a coordination medium. So that's the first thing. Something else that I really want to talk to you about is the stress markers you're using. We've done a bunch of work now on stress as an indicator of basically the delta from target morphology. So my hypothesis now, and we've got a paper coming on this soon, I hope, is that basically you go through the stages of development, because every time you're different than your set point, it's stressful. And so you try to relieve that stress by undergoing morphogenetic change. Of course, by the time you do that, the bioelectric set point has moved on, so now you're wrong again. And so you've got this up, down, up, down, up, down kind of thing, which we've actually found. So we use a very particular stress marker, but I want to try yours and see if there's any difference, because I think you've got some different ones than what we're using, so that's super cool. All right, one other thing that I now realize that we have to start paying attention to, so I don't know if I've told you, but our anthrobots are actually younger than the cells they come from.

[54:43] Martin Picard: Yeah, you said that.

[54:44] Michael Levin: Right? And so my crazy hypothesis here is, I call it age evidencing. So the idea being that the age of these things, of anything, is probably, like many other features, a cognitive estimate. So by being made into an anthrobot, they're sort of getting evidence that they're embryos, and it conflicts with their priors that say, no, we've been around for a while. And so they roll back to some extent, but it's kind of like some cognitive dissonance. They don't roll back all the way because they still have the older memories. So what I'm now, and so we have a research program on trying to give them more evidence of the fact that they're young and whatever, but it now seems to me maybe we need to be talking to the mitochondria as well. And so we may want to start thinking about, and actually we have a collaboration with Nick Lane looking at how mitochondria talk to the rest of the cell. But we haven't, like we've done a lot of work communicating with cells and trying to do that, but not really with mitochondria yet. So it sounds to me like we should talk about that. We should talk about ways to communicate with them, and maybe that'll be a much more effective intervention. And I guess the final thing, the final thing is that I love this hypometabolic state thing that you saw, that you showed during sleep and so on. And one of the, super cool, one of the things that we wonder a lot about is whether xenobots and anthrobots sleep. And so we maybe need to do some metabolic profile. We started doing some anyway for a guy foundation project, but we need to do some better metabolic profiling and see if we can detect that kind of a thing, you know, they're actually sleeping.

[56:31] Martin Picard: Well, I have three or four ideas that might be relevant when there's a