I’ve been commenting in various groups (including here) for a long time on why I don’t think aging is fundamentally programmed, even though gene regulation clearly strongly influences how aging progresses.
I recently wrote a longer piece that lays out this argument in more detail, including why cellular reprogramming alone cannot address all age-related damage. I’m posting it here in case it’s of interest.
This debate has been happening on twitter recently. I therefore asked three LLMs for a definition of programmed aging and posted the results here:
As people probably know Olafur and I disagree about these issues.
I think Aging is fundamentally programmed as part of an aging and development process driven by nuclear acetylation levels which are driven mainly by mitochondrial citrate export.
This is a link through to the discussion on twitter/x
Just to clarify for readers: the article isn’t arguing that gene regulation (including cellular reprogramming) is not important, but that regulatory explanations alone cannot account for accumulated non-cellular damage, which in turn places hard limits on what reprogramming can fix.
Thanks mate. I’ll read more carefully to address the points you make later. For now I’m more interested in the points you don’t make. To wit, you spend a lot of energy arguing against certain interpretations of programmed aging, but as an alternative you only hint at some nebulous “damage” (yes, this is a common term, and I hate it for imprecision). You’re clear that this refers to some subset of ongoing stochastic changes. Which ones (you refer to ECM changes, but as one in many), and why do you think they are sufficient to explain the phenomenon of aging?
You seem to be arguing some combination of
As to the first point, I’m inclined to agree that tissue level matrix regulation is central to aging. I’m skeptical that it’s stochastic in nature, for two reasons. First, in general, the repeated failure of all attempts to modify aging by preventing various forms of stochastic changes, from antioxidants to anti-crosslinking agents. Second, the fact that despite there being a host of identified proteins composing the ECM, no progeroid syndrome is associated with mutations of them, AFAIK. Surely some of the defective matrices would be more prone to “damage”?
The second point is surely not what you meant to argue, and just my misunderstanding the text. After all, different animals’ matrices change at different rates, so either the rate of accumulation of “damage” or the rate of its clearance is under genetic control. Could you elaborate on the meaning of paragraph 3, especially, in this context?
I agree that damage sounds a bit nebulous if it is’t clearly definied. But I don’t think it is nebulous. The types of damages I’m talking about are well defined chemical processes such as non-enzymatic glycation, cross-link formation, isomerization, racemization and deamidation. These are well known molecular changes that can be measured and are known to influence chemical reactivity.
Just to be clear, I don’t think these types of damages explain all aspects of aging fully themselves. Aging is influenced by many more things than those types of damages. My argument is that some of these damages cannot be fully prevented or repaired with genetic control and that limits what we can explain with programming.
Note that these damages are not meant as some theory of aging. They are meant as examples that puts limits on what can be explained by programming. Any complete theory of aging must not ignore the accumulation of changes that occur as a results of chemistry unfolding over time rather than from a program.
I get why you’re skeptical, but it’s important to point out that nobody has developed an effective therapy to slow down these stochastic changes. Effective anti-crosslinking interventions do not exist. There are some that have a minor effect (e.g. the ALT-711 drug that failed) but none that actually remove a significant part of them. The closest to an effective anti-crosslinking intervention we have is IMO to keep your blood glucose low, which is something we know is beneficial for prevention of some stochastic damages, and we also know that it slows down aging slightly. We also know that lower body temperature helps, which may be one reason why subclinical hypothyroidism is associated with longer lifespan in some studies. But to make a real dent would require reducing the body temperature much more than a degree or two, which obviously isn’t reasonable. So overall we kind of have nothing that puts a real dent in stochastic damages.
As far as antioxidants go, they fail for various reasons that don’t have so much to do with stochastic damages. There is no antioxidant that can massively reduce oxidative damage in the whole body of someone that is healthy and doesn’t suffer from a lot of oxidative stress to begin with. They aren’t that effective, and if they were, they would cause unintended harm because oxidative processes aren’t just harmful. They are also necessary for many cellular functions, so if an intervention existed that would prevent all oxidative stress, it would cause great harm to your body.
That’s good thinking and a reasonable question. I agree that the absence of clear ECM progeroid syndromes is interesting. However I don’t think it breaks the argument.
Just because there are no well-known progeroid syndromes clearly caused by ECM defects doesn’t mean that the ECM integrity is not relevant to aging. Many components of the ECM are critical to development so many mutations that result in faulty ECM would be lethal or would create connective tissue disorders rather than clear accelerated aging syndromes. Milder defects might cause smaller changes in the rate of aging that do not show up as clear progeroid syndromes.
Even if a defective matrix were somehow more susceptible to stochastic damages, the are some things that limit how much that would accelerate aging. Non-enzymatic processes such as glycation, racemization, and isomerization are constrained by chemistry, mainly by exposure time, molecular turnover and temperature. Even a defected matrix has these constrains.
For those reasons, I don’t think we need some ECM progeria syndrome to think the ECM is important and limiting for aging. The stochastic ECM damages are accumulating chemical modifications in long lived structures driven mainly by reaction kinietics and time rather than by geneticaly coded structural properties. Such kinds of effects are not something you would expect to show up as clearly as progeroid syndromes caused by defects in a single gene.
I have a simplistic sort of engineering approach to the question as to what damage cannot be repaired, recycled or prevented which is to fix what we can more readily fix and then see what remains to be fixed.
It happens that I think more can be fixed than most people think. I accept that nuclear DNA mutations cannot be fixed although non functioning cells can be cleared. It remains, however, that I don’t see any merit in spending a lot of time discussing where any arbitrary line can be drawn.
I think the phenotype of development and aging primarily results from changes in nuclear acetylation levels (when cells are functioning normally not when they are stressed). Hence I am to intervene on acetylation.
I accept that not everyone agrees with me and in our discussions on twitter recently there was some agreement that the burden of proof is on me to prove that my approach works.
This is similar to my view on programming in that even though I don’t think aging is fundamentally programmed, I do think programming is very important and am very supportive of research into reprogramming as a means to solve aging. I think that if we were to theoretically solve programming, we would then see more clearly what remains to be fixed to fully solve aging.
In the end we don’t at this stage have to agree on what is the best solution. We can simply agree that doing research is useful. I don’t agree with Aubrey de Grey’s view that people should be prevented from doing certain types of research - although he has pulled back a bit from this perspective.
I agree, with the caveat that unfortunately funding and attention is very limited, so it is important to have somewhat of a consensus on what is most important to work on or else we risk that a lot of resources and time go to waste. I wouldn’t go as far as to say people should be prevented from doing certain types of research though.
I don’t myself agree that a consensus is helpful. It would act to prevent novel research. The conservative nature of many funders already tends to limit what research is done.
That’s a good point. I see that as a bit of a separate problem related to risk taking. Many funders are very aversive to risk which has the negative effect of limiting novel ideas. This is probably even worse in academia where you need grants to do research.
Thanks for the thoughtful reply. It might be useful to introduce a philosophical point, dividing viewpoints on aging into a spectrum spanned by roughly 3 categories. First one has “geroscience denial” (pejorative reflecting my bias!) – those who don’t admit aging as a phenomenon worth naming. Rather, they see only a basket of diseases. Second comes the “weak geroscience” view. Here aging is seen as a basket of loosely related changes united by their apparent inevitability – not a unitary phenomenon, but one that deserves a common name. Finally one has “strong geroscience”, wherein the facets of aging are tightly coupled and proceed in a coordinated manner. I.e. it is seen as useful to view aging as a single phenomenon, rather than a term of convenience. Note that these are viewpoints, not scientific claims themselves, and it’s possible to agree on facts and differ in viewpoint, or vice versa.
I mention this because I suspect you fall more towards the middle camp in comparison to my position more toward the third.
As an example, many mammals can not regrow or repair teeth as they wear down due to use. In the wild this can be life limiting. But it’s not the type of thing that I refer to as “aging”. Similarly, I agree with your abstract point that some molecular level changes to the ECM, for example, can be both irreversible and theoretically life limiting (as in a barrier to immortality). I’m not convinced elastic fibers are one of these examples, but that’s not the core issue. Either way, the abstract point does not translate automatically to the concrete phenomenon of aging as we see it.
In another thread you write:
I’m bringing this in because it’s a stronger statement than in your article or this thread. I recognize that as a brief comment in a forum it cannot represent the nuance of your thinking. It’s a waypoint for discussion.
I don’t know why there are no 130 year humans – humans are already very long lived. I can say that the oldsters appear to die of old age, with old age appearing pretty similar across mammals. So let’s consider rats (which also die of old age).
There are no 5 year rats, despite enormous numbers observed. The same arguments would appear to imply that rat lifespan is not under genetic control, and due to stochastic chemical changes. But we know this to be false. There are plenty of 10 year squirrels – same size, higher metabolic rate, closely related, just longer lived. There are 30 year vampire bats in captivity – smaller, massively higher metabolic rate. There are even 40 year siberian bats in the wild, with presumably higher metabolic rate yet again when active, but they hibernate half the year. If the lifespan of rats was limited by unavoidable, unrepairable, stochastic changes to their matrix then that would seem to mean that their matrix was just unusually crappy in comparison to the others.
Which brings us back to to the question – why aren’t there progeroid mutations of ECM components? This isn’t rhetorical. I’m genuinely puzzled as to why this appears so. If we posit that rats have short lives due to crappy ECM, then ECM changes should exist that are nonlethal but affect the rate of accumulation of deleterious changes.