Why do some giant sea creatures live hundreds of years, while small land animals such as insects may live only a matter of months? Why does a living thing’s ability to overcome wear and tear generally decline over time?

These and other questions about the nature of aging remain deeply mysterious even to those who have studied them for decades. But physicists think they might have some quantitative tools that can help.

A few months ago, some 160 scientists across physics, biology, and medicine met in Singapore for the first ever global conference on the physics of aging, or gerophysics. It’s a niche discipline whose first glimmerings arose in the 1960s but that is only now coming into its own. The scientists gathered in Singapore talked about how to develop a predictive and testable science of aging using certain concepts such as thermodynamics and network theory. Together, they intend to establish mathematically precise definitions of terms like aging and healthspan and to find simple mathematical principles that can explain the complex processes that contribute to longevity.

I spoke with Max Unfried, a research fellow at the National University of Singapore's Centre for Healthy Longevity, and one of the organizers of the gerophysics conference as well as a founder of the field itself about the maximum lifespan of humans, why bears and bats may unlock the secrets of successful aging, and the importance of thermodynamics to the future of medicine.

What does physics have to do with aging?

One thing is that biology cannot make its own rules. Everything has to obey the laws of physics. The fundamental processes that occur when a star dies are probably also relevant when a human or other organism dies. Currently gerontology is primarily described in terms of biology. You have your genes, you have your proteins, and these things change with age. If you’re a physicist, you might describe that process in different ways, perhaps in terms of thermodynamics.

Physics can also bring questions of systems stability or instability to the conversation. Does an unstable system increase mortality? If you combine this with the biological way of thinking, you might notice that genes, proteins, and lipids change in a linear fashion with age. But as these molecular components change in a linear way, a person’s mortality risk and rate of disease goes up exponentially. One question is, how can a bunch of linear processes create something exponential? That’s probably not a question biology by itself can answer, because you have emergent phenomena, new dynamics that need explanation. Some fundamental laws of physics might help.

Why do you think this field of gerophysics is only coming together now?

The first paper in the field was published in the 1960s, which attempted to use statistical physics to describe aging. Over the past 65 years, you had individual attempts to bring a little bit more of physics thinking into the field of aging biology and gerontology. But it never reached a critical mass. Some scientists in Germany and the United States did a little bit, and these efforts got attention, but the ideas never spread. Then, in the past 10 to 15 years, we slowly have more physicists coming into the aging space, people thinking more about network dynamics.

Three years ago, I started a workshop at one of the large aging meetings, and people liked it, so we started a conference. And now we finally have people that understand each other, a peer group. Not physicists, and not biologists, but gerophysicists. These are people that come from physics, engineering, and mathematics, who share a fundamental language of mathematical modeling and physics thinking and are applying it to aging biology. Within physics there are different subdisciplines: Some people come from a thermodynamic lens and think more about entropy, and other people come from a stability science perspective. Still other people study network science and want to understand how network connectivity is lost in aging and what that means for resilience. Now they can all talk to each other and discuss those ideas and get feedback.

A lot of predictions that gerophysics makes are probably not what biology would have expected, which probably has relevance for how we measure aging and how we target aging with treatments.

Read more: “The Longevity Skeptic”

Why would biologists not have expected some of the predictions that gerophysicists make?

There are a few areas where physics has surprising things to say about aging: maximum lifespan and stable versus unstable lifespan.

Let’s start with maximum lifespan. For humans, it’s around 120. It’s just like a barrier that can’t be crossed. But if you had 10 times more humans or 100 times more humans, would we actually see humans live until 200? We just don’t know. Biology can’t answer that question. But physics models like the stress response model collapse at around 120 or 150 years of age, which gives you an indication that this is probably the limit for human lifespan with our existing blueprints. Humans just won’t be able to live to 300, even if we had infinite humans. For most people, the max is 110, 115, and the people who live to 120 are extreme outliers.

The second thing is work from the Uri Alon group at the Weizmann Institute of Science, who modeled different animals according to their “stability,” and found they come in two broad categories: stable animals and unstable animals. Unstable animals decline more rapidly toward death. Stable animals live in a state of equilibrium for longer and die later.

Assume you have a U-shaped bowl that represents resilience. And in that bowl, you have marbles, and they roll around in there, which represents your daily fluctuations of blood pressure, heart rate variability, and glucose levels. In a stable animal, the marbles all roll around in there, but they don’t fly out of the bowl. Now, assume that aging erodes the bowl, making it flatter. Now, those marbles, as they change, are much more liable to exit. The flatter the bowl, the less stable it is, and the more likely the animal is to develop disease and die. Ecology has similar analogies for how ecosystems fail.

What’s an example of an animal that’s stable versus an animal that’s unstable?

Luckily, humans are fairly stable animals. That’s why we live a long time. Our diseases come only in late age. Most of them appear around 50 or 60. The problem is that most of the animals we use in the lab are unstable: mice, drosophila, C. elegans. That’s why mice are very short-lived. They usually all get cancer. This makes them good for cancer research, but not good for longevity or aging drug trials, because all the things that you modify in a mouse aren’t relevant to human aging. But the entire academic biomedical ecosystem depends on mice, so what the field has to do is find the shortest-lived stable animal that we can use to test our theories and drugs. That’s going to cost more money and require new infrastructure.

Lifespan stability has implications for drug targets. Returning to the bowl, our current drugs address the marbles, but physics might aim to treat the shape of the bowl instead. Targeting the marbles might give us up to 15 years of additional healthy life, which is pretty good. But it’s different from trying to increase maximum lifespan. To change maximum lifespan you have to target the bowl, which increases your resilience. This resilience is related to thermal fluctuations, entropic damage, random damage, whose root cause is life itself. It’s not possible to target the damage caused by thermal fluctuations because they’re random and uncorrelated. If we were to cure all those fluctuations in the marbles, but do nothing about this thermal damage, you’d get a 110-year-old without diseases who still died. They wouldn’t have any cancer. They won’t have Alzheimer’s. Their brain will work. They’ll be super fit for their age. But they’ll still be frail.

If aging obeys physical laws the way that falling objects obey gravity, does that mean that it’s inevitable in the same way?

Inevitable is a hard word. Most things age. You have a few creatures, those long-lived animals that show negligible senescence, where mortality doesn’t increase with age. Humans follow an exponential trajectory called Gompertz law, where your risk of dying doubles every eight years. A few creatures, like the naked mole-rat, seem to have a constant mortality rate. Trees have a constant mortality rate in most cases. There seems to be something that doesn’t increase mortality in these cases.

The issue is, all those organisms are living systems that are far from thermodynamic equilibrium. We’re not talking about normal thermodynamics—we’re talking about non-equilibrium thermodynamics in dissipative systems. They have to spend or dissipate energy to keep local order in the organism. If aging is fundamentally caused by this entropy, a universal tendency of everything to break down, then maybe aging itself is a kind of universal force, similar to gravity.

Then the argument is that nature found some tricks, some genetic blueprints, to extend life. And the question is, can nature build something much better than that, especially in living systems? Currently, the longest-lived animals are 200 to 400 years old, so it seems to be possible to live that long for some animals, but this comes with certain side effects, and almost all of the long-lived animals reside in cold water, which results in fewer thermal fluctuations and less entropy.

Maybe if you want to live to 200 years, you have to become a whale and find solitude in the deep ocean. Or maybe we can learn something from the long-lived animals. Maybe we can find genes or proteins that, if we have enough of them, can provide additional lifespan. Maybe we have to go into germline editing. It would be too late for people already alive, but we can edit the future in the germline to make newborns more stable across their lifetimes.

You mentioned one big question the field is trying to answer is how can a linear molecular process lead to a compounding death rate. What are the other big questions that the field is interested in answering?

Currently, we have 200 or 300 theories of aging. It’s telomeres one year. It’s reactive oxygen species the next year. It’s something else the year after that. In a way, all those theories have linear effects, but none of them fully explains aging. If you have hundreds of theories, nobody really understands it, and probably each of them is partly true. But physics can tell you how all those different linear aging theories fit together.

But what other open questions are there? Some things are far from settled. We currently have two popular models that describe the stability of mortality, and they’re approximations. I hope they’ll get better in the future. I also hope that stochastics will come more into the aging space. We’re exploring that. One area that’s really coming in more is from thermodynamics—especially the subfield of stochastic thermodynamics, which is very interesting. It’s not a mainstream field yet.

About 30 years ago or so, people in the physics community started thinking: “Okay, thermodynamics really deals with microscopic objects. It’s like all your gas atoms in a box.” Stochastic thermodynamics asks the questions of what happens to thermodynamics on a small scale where fluctuations matter. They started figuring out certain equations that give you ideas of entropy at a small scale, which is where a lot of our molecular damage lives. This field is currently heavily focused on thermodynamic computing. Silicon Valley is building companies on that, and many of those people work on quantum computing.

You could look at biology as just information being delivered between different proteins and molecules. Does the thermal noise make distorted signals, and how does this change with age? This has to be figured out, and this has to move into biology and aging biology.

I believe the future of medicine will be thermodynamic. This is very long term. I’m talking 50 to 100 years. Currently, biomedicine is very crude. It tries to modulate biology on the physiological level, but you take a drug and it hits multiple targets at once. We don’t even really understand which target it hits. We kind of hope it’s the one, but we need much more precision. Physics could help us deliver precise energetics to precise locations in the body. But we need much better tools to study a lot of those things, because we don’t have the instruments that could measure some of those things, especially in biological systems.

Another question is around the fact that there seems to be this plateau at old age. Very old people above 100 or 110 have a constant mortality rate. It’s a high mortality rate. You have a 50-50 chance of dying every year, compared to the doubling of your mortality risk every eight years before that. Maybe physics can help us explain why, as well as what really makes certain people age faster than others. There’s a lot of variability. Also, what impact did evolution have?

This recent paper about the impact of somatic mutations on lifespan, which was published a few weeks ago, got some traction. It tried to answer how much extra lifespan would we get if we eliminated all somatic mutations, which are changes in the DNA of a body cell that happen after conception. And at what age does it really become relevant? We should be able to rank what we should target and say, “Well, you know, if somatic mutations only start killing us at around age 150, maybe we can put them off a little bit?” We could focus instead on what kills us at age 110—or earlier.

Read more: “This Is Life at 400”

What kind of research do you think is most pressing right now in gerophysics?

From a drug development perspective, and a pragmatic perspective, the best way forward currently is to figure out what we can learn from longest-lived animals and try to see if we can implement it in humans. Let’s be inspired by nature. It’s probably an easier task than to try to develop something from scratch.

Is there one animal in particular you think is most promising in that regard?

I probably would put my hope into bats. They’re one of the largest groups of mammals, the second largest group after rodents. There are more than 1,500 species of bats. That’s the first thing. The second thing is that there are a lot of bats that live vastly longer than their body size should predict. A lot of bats are the size of a mouse or smaller, and some of them live up to 40 or 50 years. That’s quite impressive. And then, what we know is that they’re very good with pathogens. They’ve got enhanced immune systems that allow them to live with a heavy viral load.

The immune system plays a big role in human aging. If we could learn how bats control their immune systems so that they don’t spike into cytokine storms—maybe we could improve the human immune system. That probably helps with cancer and viruses and a lot of other things.

The naked mole-rat is also very fascinating. Similar story. It’s a very small animal, very long-lived, very cancer-resistant. And then, bears and other hibernating animals are very interesting. If a human were to do what a hibernating bear does, they’d have to go to the doctor: Half of the year, you binge eat, and the other half you just lay around and don’t eat at all. It’s like a seesaw between obesity and anorexia. If a human laid around like a bear, they’d get thrombosis. The bear wakes up, and they’re kind of good to go.

I’ve just started a new research organization on this subject called the Thalion Initiative. To study these things from a physics perspective, you need data. Right now we don’t have enough data for all of those animals where we could say, “Are they stable? Are they unstable? How fast do they age?”

What we’re trying to build is like a CERN for aging biology, a biomedical mega-project that focuses on aging biology, and especially on the things that nature already figured out. We want to build this super large dataset of over 200 different species. For each of the species, get many specimens across every age group, from young to old, and then measure everything that we currently are able to measure: the genes, transposomes, the methylomes, the proteomes, lipidomes, and inflammation markers. Then we could see molecular changes across evolution, across animals, and see what controls lifespan, what gives them stability.

We know that complex systems become better through stress and noise, so we probably should observe modularity and resilience in longer-lived species.

You just held the inaugural conference for the field earlier this year. One of the focuses of the conference was using artificial intelligence to screen compounds for longevity effects. Why is AI a useful tool in this space?

Biology is always in a very high-dimensional space, and what AI is really good for is compressing everything into a few variables. It can take all those features of biology and, for example, put all of the inflammatory molecules into a single effective variable, all the heart-related ones into another, which is helpful for some physics modeling.

But it also depends on data, and there’s a lot of unstructured data in biology. Some of it is good for simple modeling, but other data isn’t good enough for AI yet, so we need approaches that collect more data to make AI useful.

Could physics help us separate lifespan from healthspan or sickspan?

Yes, it can, and that’s probably the strongest thing it offers right now. The marbles are probably the healthspan, and the bowl is the maximum lifespan part. Realizing those are two different things is an important starting point. Because two different species can have vastly different maximum lifespans, but can have similar diseases. A human and a mouse both get cancer but the humans have a maximum lifespan 40 times longer. That probably tells you the biology is somewhat different. Maximum lifespan is probably subject to a bunch of other evolutionary constraints. It probably also tells you why, for instance, caloric restriction has a lot of benefits in mice but probably won’t have much benefit in humans.

It will be important to bring evolution back into the biology of aging, because while evolution is the central dogma of biology, it gets ignored when it comes to drug discovery. We have cancer therapies, and if you only give one cancer drug to a person, the cancer becomes resistant. If you throw different drugs at the cancer, it’s actually more likely to die because it cannot adapt to all of them. That’s basic evolutionary thinking, but something like that holds for aging as well.

Lead image: Inna / Adobe Stock