It is a bit dated but I always listen carefully to Aubrey, here commenting on the Information Theory of Aging and the eyesight results quoted in the paper I just posted. Right or wrong but always welcome his laser sharp analysis of results in this space:
"The information theory and the eyesight result are not quite so closely intertwined as you may be thinking. The information theory, in its broadest sense, is the claim that the main determinant of the age at which we start to exhibit the chronic progressive pathologies of late life is the lifelong accumulation of epigenetic noise. Stated like that, the information theory is neither new (it was first put forward in 1982 by Richard Cutler, who named it the dysdifferentiation theory) nor, in all probability, correct, any more than any other “theory of aging” is correct: the more likely state of affairs is that there is no such single main determinant, but rather that several types of lifelong accumulating damage contribute substantially but no contribution exceeds 50%. (You might get another impression from David’s recent book - but, well, he’s not the only scientist who makes certain choices about what to emphasise.) The next question is mechanistic, i.e., how does that epigenetic noise arise? David’s key idea is that the main mechanism is that the process of repairing genetic (as opposed to epigenetic) damage, such as double-strand breaks, causes an accumulation of epigenetic damage as a bystander phenomenon: first of all the epigenetic state (methylation, in particular) in the vicinity of the genetic lesion is lost when bases are replaced, and secondly the repair process entails the temporary redistribution of DNA protection proteins from elsewhere in the genome, thus exposing those other areas to accelerated damage. That’s a highly plausible and valid theory. So is the theory of Andrei Gudkov, that the DNA damage is mostly caused by retrotranspsons. Not only that, those two theories are linked, because retrotransposon activation is itself caused by epigenetic damage. But honestly the mechanism doesn’t much matter, as I’ll explain.
So, what about the (very impressive) eyesight result? It really tests (and validates) a different theory, namely that partially removing epigenetic noise is beneficial. Note the difference: there’s no reference to the mechanism by which the noise was created in the first place, nor to whether epigenetic noise contributes more to aging than all other types of damage combined. A lot of confusion arises when people fail to make this sort of distinction - when they conflate theories about how aging happens with theories about how it can be mitigated. Indeed, it could be argued that the main thing I did by introducing SENS was to highlight the importance of keeping the two distinct.
So far so good. But the thing is, the specific method of removal of the epigenetic noise is rather critical to what is really being shown. What’s actually being shown is a very much stronger result than I wrote above, namely that whacking the entire genome with a a stupefyingly indiscriminate and non-selective baseball bat called OSKM (or in David’s case only OSK) can do more good than harm if you do it just right. The point here is that such interventions don’t have a way to distinguish between epigenetic marks that are noise and marks that are signal, i.e. are acting to make the cell do what it’s supposed to do. All it can do is remove some proportion of those marks, of both types. So what we are showing when we do this and get a desirable result is that the differentiated epigenetic state is really robust and redundant, such that we can blitz quite a lot of it and the cell still knows what sort of cell it is. Plus, the fact that the cells become more regenerative means that the marks that define what sort of tissue the cell is part of are even more robust than the marks that narrow it down to its terminally differentiated state.
So, what’s not to like? Well, as so often in aging, that comes down to one word: cancer. David’s main reason for using OSK rather than OSKM is that M is c-Myc, an oncogene. The fudamental issue here is that every adult, let alone middle-aged or older adult, is chock full of cells that have already acquired most of the mutations needed to turn them into cancer cells, and the only reason they haven’t acquired all such mutations is that mutation is stochastic - a numbers game. And that can also be thought of in terms of robustness and redundancy. A pristine cell has layer upon layer of ways to stop itself from dividing more than it’s supposed to, and these mutations progressively degrade those defences and make them more and more fragile. Not the sort of thing you would want to hit with a baseball bat, right? And what’s worse, you won’t know you’ve turned a bunch of nearly-cancer cells nto actual cancer cells until they have gone on and divided enough to become detected, which usually means someone displaying symptoms. So we can’t infer much at all about the magnitude of this problem from short-term experiments. What I’m hoping for (and yeah, I’m trying my best to make it so) is that people like David, and Belmonte (the most prominent researcher in this area), and others, will take the cancer risk more seriously than their not-necessarily-all-that-far-sighted reviewers and investors may, and devise smart ways to quantify the cancer risk of such interventions as early as possible. This won’t be easy, but it’s vital. If we hide our heads in the sand and engage in oversimplistic overoptimism about this risk, we are very likely to spend a great deal of time and money on therapeutic dead ends. I’m certainly not saying that partial reprogramming (as this is called) in general is a dead end, but I’m very much saying that it will not fulfil its potential, or at least nowhere near as soon as it could, unless we focus squarely on the cancer side-effect every step of the way."
https://qr.ae/prY1I3
