Let’s start with a stylized fact: almost every cell type in the human body is removed and replaced on a regular basis. The frequency of this turnover ranges from a few days (for many immune cells and cells in the gastrointestinal lining) to ten years (for fat, heart, and skeleton cells). Only a handful of tissues are believed to be non-renewing in humans - e.g. eggs, neurons, and the lens of the eye (and even out of those, neurons are debatable).

This means that the number of cells of any given type is determined by “homeostatic equilibrium” - the balance of cell removal and replacement. If an ulcer destroys a bunch of cells in your stomach lining, they’ll be replaced over a few days, and the number of stomach cells will return to roughly the same equilibrium level as before. If a healthy person receives a bunch of extra red blood cells in a transfusion, they’ll be broken down over a few months, and the number of blood cells will return to roughly the same equilibrium level as before.

As organisms age, we see a change in the homeostatic equilibrium level of many different cell types (and other parameters, like hormone and cytokine levels). In particular, a wide variety of symptoms of aging involve “depletion” (i.e. lower observed counts) of various cell types.

However, human aging happens on a very slow timescale, i.e. decades. Most cell counts equilibrate much faster - for instance, immune cell counts equilibrate on a scale of days to weeks. So, suppose we see a decrease in the count of certain immune cells with age - e.g. naive T cells. Could it be that naive T cells just wear out and die off with age? No - T cells are replaced every few weeks, so a change on a timescale of decades cannot be due to the cells themselves dying off. If the count of naive T cells falls on a timescale of decades, then either (a) the rate of new cell creation has decreased, or (b) the rate of old cell removal has increased (or both). Either of those would require some “upstream” change to cause the rate change.

More generally: in order for cell counts, or chemical concentrations, or any other physiological parameter to decrease/increase with age, at least one of the following must be true:

  • the timescale of turnover is on the order of decades (or longer)
  • rate of removal increases/decreases
  • rate of creation decreases/increases

If none of these is true, then any change is temporary - the cell count/concentration/whatever will return to the same level as before, determined by the removal and creation rates.

Of those three possibilities, notice that the second two - increase/decrease in production/removal rate - both imply some other upstream cause. Something else must have caused the rate change. Sooner or later, that chain of cause-and-effect needs to bottom out, and it can only bottom out in something which equilibrates on a timescale of decades or longer. (Feedback loops are possible, but if all the components equilibrate on a fast timescale then so will the loop.) Something somewhere in the system is out-of-equilibrium on a timescale of decades. We’ll call that thing (or things) a “root cause” of aging. It’s something which is not replaced on a timescale faster than decades, and it either accumulates or decumulates with age.

Now, the main criteria: a root cause of aging cannot be a higher or lower value of any parameter subject to homeostasis on a faster timescale than aging itself. Examples:

  • Most cell types turn over on timescales of days to months. “Depletion” of any of these cell types cannot be a root cause of aging; either their production rate has decreased or their removal rate has increased.
  • DNA damage (as opposed to mutation) is normally repaired on a timescale of hours - sometimes much faster, depending on type. “Accumulation” of DNA damage cannot be a root cause of aging; either the rate of new damage has increased or the repair rate has decreased.
  • DNA mutations cannot be repaired; from a cell’s perspective, the original information is lost. So mutations can accumulate in a non-equilibrium fashion, and are a plausible root cause under the homeostasis argument.

Note that the homeostasis argument does not mean the factors ruled out above are not links in the causal chain. For instance, there’s quite a bit of evidence that DNA damage does increase with age, and that this has important physiological effects. However, there must be changes further up the causal chain - some other long-term change in the organism’s state leads to faster production or slower repair of DNA damage. Conversely, the homeostasis argument does not imply that “plausible root causes” are the true root causes - for instance, although DNA mutations could accumulate in principle, cells with certain problematic mutations are believed to be cleared out by the immune system - so the number of cells with these mutations is in equilibrium on a fast timescale, and cannot be a root cause of aging.

For any particular factor which changes with age, key questions are:

  1. Is it subject to homeostasis?
  2. If so, on what timescale does it turn over?
  3. If it is subject to homeostasis on a timescale faster than aging, then what are the production and removal mechanisms, and what changes the production and removal rates with age?

These determine the applicability of the homeostasis argument. Typically, anything which can normally be fixed/replaced/undone by the body will be ruled out as a root cause of aging - the timescale of aging is very long compared to practically all other physiological processes. We then follow the causal chain upstream, in search of plausible root cause.

New Comment
25 comments, sorted by Click to highlight new comments since:

Forgive me if you've covered this already, but... Isn't it also possible that you have some shortscale homeostatic pair where neither rate changes, but were in fact always just a little miscalibrated relative to each other? Such that the equilibrium point shifts very very slowly over time?

Like, in Harry Potter canon you can cast reparo on an object as often as you break it, but it's just slightly less than a perfect repair. It's not a difference you'd notice until you'd repaired the object numerous times.

Great question, and it's a very subtle point.

In general, the "state variables" in biology are counts of identical things (or at least functionally-identical things). So for instance, paralleling your Harry Potter example, DNA is constantly breaking and being repaired. But the imperfections in that repair process aren't continuous - either the repair is perfect, or a mutation is introduced. Once a mutation is introduced (assuming it's in something functionally important), we have a functionally different molecule, so it needs to be reflected in a state variable: we need to have a state variable like "mutation count" or counts of specific kinds of mutations. That state variable will be what's out-of-steady-state on a long (in this case infinite) timescale.

This generalizes: if there are slight differences accumulating over time, then those will have a state variable (like mutation count), and that variable will be out-of-steady-state on a long timescale.


I wonder if perhaps something more environmental might also be playing a part. The protein toxins that are associated with Alzheimer's seem to build over time and the effectiveness some of the processes that work to clean them up may be negatively impacted they its presence. Seems the the sleep cycle (non-REM) results in a reverse flow type flushing of the brain that help clear this out. But the build up itself seem related to not getting that sleep needed.

So what about internal and external to the cells themselves? Could some elements or combination of things build up that we're just not looking at yet -- have not see it as connected to any of the processes?

I remember that first paper which found some sort of flush-out mechanism for the brain opening up during deep sleep. Do you have a link? I've been meaning to dig into that a bit more. I remember when I first saw it I thought it would be huge for understanding Alzheimers, but that was a while ago before I started seriously reading up on aging.

Anyway, environmental factors...

There's a qualitative general pattern that various kinds of physiological stress - exposure to radiation or harsh chemicals (including smoking), chronic infection, malnutrition, sleep deprivation, etc - tend to accelerate aging. These are results which I generally don't put too much faith in, at least individually - they're too flashy, so there's likely publication bias. That said, there is a plausible mechanism by which general physiological insults would accelerate aging. I'll save discussion of that for a later post, but David Sinclair published a (general-audience) book earlier this year which discusses it quite a bit.

One important thing to note, more related to the OP: whatever the root causes of aging are, i.e. the things which are out-of-equilibrium on long timescales, those do need to be internal to the organism. We would have noticed centuries ago if changing the environment could forestall aging long-term. That does not mean these factors need to be inside particular cells; for instance, extracellular aggregation of certain long-lived proteins is one plausible root cause (e.g. elastin deposits as a cause of wrinkles). It is very likely that we are still missing connections in the causal graph, so there could easily be things building up that we're not paying attention to yet. That said, there are relatively few things in the body which turn over on decade-plus timescales, so that severely limits the list of possible root causes.


Here is one link: https://www.sciencedaily.com/releases/2019/10/191031174650.htm . I was not able to find the one I was actually reading earlier (and apparently my poor sleep last night was not sufficient and I cannot remember how I found it....) but the link here seems to be referencing the study I was reading about.

BTW, when I mentioned "external" I was not thinking external to the organism (e.g., me) but rather external to the cells (or at least many of them) but within the confines of our body or organ.

[Just found it with a different seach. https://www.sciencenews.org/article/sleep-may-trigger-rhythmic-power-washing-brain ]

Aha, that study linked to the one I was thinking of: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3880190/


Alzheimer's, and several other age-related diseases, seem to be related to lifestyle, since they're rare enough in hunter-gatherer tribes that they can't be detected.

The kinds of age-related deaths that are common to all environments are mainly due to frailty, susceptibility to infectious diseases, and cancer.

Do you have a reference for that Alzheimer's thing? I'd be interested to read more.


Frailty seems a questionable cause in this context. Am in interpreting incorrectly perhaps?

I would think frailty while young might be a symptom of something that leads to death but how do we go from "sturdy" and so healthy and living well (in a functional sense) to old and frail and more likely to die?

The other two, seem more like lottery type cases, yes we all have a probability of contracting some infection or virus that our immune systems just cannot deal with so we die. We have a probability of cancer destroying critical systems. But that doesn't quite explain the whole aging story to me -- why the slow path to what we see physically rather than a sudden break? Or is this inference about how we should observe things missing something you perhaps have bundled into the three causes you mention above I perhaps I should understand why (if I were more knowledgeable on this area)?

I wasn't trying to describe the root causes of aging. I was trying to distinguish between diseases that are avoidable via lifestyle changes, and age-related diseases that are sufficiently determined by our genes that we'll need major new technology to avoid them. The latter include things that impair our immune system and repair mechanisms.

My best guess is that the root causes of aging involve some clock-like processes that have been actively selected for different metabolism at different ages. See Josh Mitteldorf's writings if you want more on that topic.

Only a handful of tissues are believed to be non-renewing in humans - e.g. eggs, neurons, and the lens of the eye

Wikipedia says: Un-repaired DNA damages accumulate in non-replicating cells, such as cells in the brains or muscles of adult mammals, and can cause aging.

Weakening of muscles seems to me like a plausible driver of aging.

Yeah, lots of sources say that various aging-related things "accumulate". If you dig into it, it turns out that by "accumulate" they usually mean "increases" - DNA damage does not actually stick around for long, it turns over very quickly (see e.g. here, or compare the damage frequencies to the steady state levels on the wikipedia page). But the steady-state level of DNA damage does increase with age (even in many types of replicating cells), so either the damage rate increases or the repair rate decreases.

Weakening of muscles is an interesting case. Muscle "cells" (myocytes) are really huge, centimeter-long super-cells containing hundreds of nuclei, and satellite cells (the first type of stem cell discovered) spin off new nuclei to add to the myocyte on an as-needed basis. So although the cell itself is technically long-lived, that doesn't mean the components - even including the DNA - aren't replaced. Loss of muscle mass (sarcopenia) is considered one of the main hallmarks of aging, but as far as I can tell, the gears of sarcopenia are still quite poorly understood. (A lot of people spent a lot of time and effort imaging neuromuscular junctions, but that seems to have been driven more by the desire to publish pretty pictures than by any actual evidence that NMJs play a causal role in sarcopenia.)


(Great post!)

It seems like another implication of this model would be that correlated shifts in multiple equilibria on the same timescales provides some evidence of common causes. E.g., DNA damage and cell turnover rates changing at the same time would give some evidence in favor of them being regulated by the same mechanism.

What about telomere shortening? Are there other things that slowly break apart as they're used (and not rejuvenated constantly) that could explain aging beyond a few slow changing cells?

Telomere shortening is an interesting case. (I'm going to give my current understanding here without trying to dig up references, so take it all with a grain of salt.)

It's clearly a plausible root cause - it's a change which could stick around on long enough timescales to account for aging. On the other hand, it is possible for telomeres to turn over: telomerase is active in stem cells, so telomere length should at least not be an issue for cell types which regularly turn over - the telomeres turn over with the cells, which are ultimately replaced from the stem cells. For long-lived cells, there's a stronger case that telomere shortening could be an issue.

Telomeres do get shorter with age, BUT they get shorter even in cell types which turn over regularly. That's a bit of a red flag - either the telomeres aren't being fully replaced by telomerase in the stem cells (in which case the stem cells ought to die a lot sooner), or some other mechanism is making them short besides accumulated loss over lifetime. The alternative mechanism which jumps out to me is: DNA damage, and oxidative damage in particular, has been observed to rapidly shorten telomeres. DNA damage and oxidative damage rates are generally observed to be much higher in aged cells of most types, so that would explain why telomeres are shorter in older organisms.

In terms of actual experiments, telomerase-boosters have been experimented with a fair bit, and my understanding is that they don't have much effect on age-related diseases (though of course there's the usual pile of low-N studies which find barely-significant and blatantly p-hacked results).

Other things will eventually be covered later in this sequence.

Is telomerasa active in all stem cells?

That's a tough question to answer; there's an awful lot of stem-cell types and telomerase activity is not necessarily binary. I wouldn't be shocked if some of the more-differentiated and/or slower-dividing types don't express it.

We can do a back-of-the-envelope estimate: wikipedia quotes typical telomere length of 11k bp (base pairs) at birth, and one replication eats up 20 bp. That's a limit of ~500 divisions, so any human stem cell which divides much faster than every ~(80 yrs)/500 = 60 days needs to express telomerase just to keep dividing throughout our lives. In practice, I'd expect that to be an extreme underestimate, since my understanding is that oxidative damage eats up telomeres considerably faster than replication does, especially for infrequently-dividing cells.

I was recently introduced to evolutionary theories of aging in the book Other Minds: The Octopus, the Sea, and the Deep Origins of Consciousness. See the "Aging" section of this lesswrong book review for an overview, or these wiki sections on antagonistic pleiotropy. The core concept is that there is selective pressure in favor of mutations which benefit an organism early in life but harms it late in life. This is because organisms die of other causes, so will have a lower chance of reaching the age where the mutation becomes harmful. A higher the rate of death from these other causes, the lower the natural lifespan of the organism.

If that is an accurate description, there would be many different issues that only emerge later in life. The lower rate of production might be a adaptive feature, which partially avoids the negative effects of these harmful-late-in-life mutations. Removing the root cause of lower production could then be harmful. I would expect there to be many distinct issues, and many distinct "root causes" that were evolved to optimize around those, and that the age of onset would generally be in line with its natural lifespan.

Just encountered this, also: https://nintil.com/longevity

Most cell types turn over on timescales of days to months. “Depletion” of any of these cell types cannot be a root cause of aging; either their production rate has decreased or their removal rate has increased.

In order for that to be a cause, those cell types would have to produce each other (and a decrease across the board, would lead to a decrease in the production rate).

Correct. However, even in that case, the overall level would still reach a steady-state on a timescale much faster than aging itself. More generally, feedback loops equilibrate on roughly the timescale of their slowest component. So, even if there's a feedback loop in the mix, it still can't be a root cause of aging unless one of the components is out-of-equilibrium on a timescale comparable to aging (i.e. decades or longer).

The root cause of aging is programmed aging. Bowhead Whales (mammals, genetics remarkably similar to a cow) live 230 years, Greenland Sharks live 430, arctic quahogs live over 500, various wacky plants live as long as religious memes.

Then I'd very much like to know what the program is, what its variables are, and in particular how the program state is stored.


"The root cause of aging is biology."