Robert Bradbury's Grand Unified Theory of Aging

Robert Bradbury of the Aeiveos Research Library has been working on a grand unified theory of aging of late. He was kind enough to post a summary to the extropy-chat list, which is reproduced here with permission:

I've been working on this theory for the past year or so and thought I would present it to the list so it gets into general circulation. You could call this either "A Grand Unified Theory of Aging" or "The Mutation-Energy Catastrophe Theory of Aging." In it, I try to bring together a number of elements from a number of other aging theories to see if we can begin to reach a greater understanding.

First, lets assume the Free Radical theory of Aging which involves various aspects of Mitochondrial damage and aging are correct. This explains why caloric restriction works.

Second, lets assume you can't do too much about them because radicals and/or other pro-oxidants (e.g. nitric oxide) are being used as signal molecules. This assumption may be somewhat controversial.

Third, lets assume that the free radicals lead to DNA mutations (which is one way cancer develops) or worse leads to DNA double strand breaks. Radiation and perhaps toxic substances in food or the environment might contribute to this as well.

DNA double strand breaks are bad. There are 3 possible results:

(a) Repair the break via the homologus recombination pathway. This can lead to "gene conversion" where a masked defective gene gets copied such that it becomes dominant. So for example you may have a cell that can function well with one good and one bad p53 gene, if the bad p53 gene gets copied to where the good p53 gene once was you are in big trouble. The net result is an increased risk of cancer.

(b) Repair the break via the non-homologus end-joining pathway. This appears to involve perhaps the Artemis protein and/or the Werner's Syndrome protein both of which seem to be exonucleases. Bottom line your DNA gets chewed up and you get a microdeletion during the repair. Alternatively if you happen to have two double strand breaks at the same time the two chromosomes can get mispaired with the wrong chromosome. This leads to several types of cancer.

(a) and (b) are aspects of various "mutation" theories of aging.

(c) Avoid repair by the cell committing apoptosis. In this case you lose cells and if the cells are not replaced by stem cells - which may themselves have damage from (a) or (b) - then you suffer a gradual loss of function.

The above seems to explain much of aging and cancer, but now the problem gets worse. If (b) goes on for long enough you will gradually accumulate mutations in various (most probably different) genes in *ALL* cells. i.e. the genomic "program" that the cells require to operate properly is gradually corrupted in random ways.

So gene expression may become defective in many various ways in many cells. This incorporates the dysdifferentiation theory of aging and perhaps aspects of the neuroendocrine theory of aging. However if the mutations occur within genes rather than say regulatory regions now you will probably have a protein that will not fold properly. This will probably be detected and the protein will be degraded. But the lack of a sufficient quantity of these proteins will probably result in cellular signals to make more of them. But they or at least half of them will not fold properly either. Now both protein manufacture and many types of protein degradation require energy (ATP). So when the cells detect a decline in ATP (due to futile synthesis and degradation of proteins) they may attempt to increase energy production. This might be through making the mitochondria work harder or making more mitochondria. In either case the result of this will most probably be more free radicals which feeds back into the start of this whole process. So over time cells will "age" increasingly faster.

The net result is that you get an exponential decline in function (i.e. aging). So far I've only managed to imagine two solutions for this.

1. Develop better DNA repair processes that do not allow the genome to become corrupted.

2. Shift things to allow more apoptosis when DNA double strand breaks are detected but also increase the replacement rate by stem cells.

(1) is a reason to support the sequencing of the genomes of other long lived species -- to see if they have figured out better solutions to the problems outlined. (For example we know that Deinococcus radiodurans has better double strand break repair but we do not fully understand this yet or know if it can be applied to humans).

(2) is a reason to be very supportive of stem cell research.

As Robert Bradbury mentioned to me, there is a lot of background information behind this summary, unfortunately still in the rough notes stage. His attempt to shift genome sequencing priorities is related to this theory and its implications for research.

Comments

> Third, lets assume that the free radicals lead to DNA mutations (which
> is one way cancer develops) or worse leads to DNA double strand breaks.

The idea that DSBs are worse than cancer is the only thing up to this point that is dubious -- see below.

> DNA double strand breaks are bad. There are 3 possible results:

There's a fourth -- cell senescence. It's long been suspected and is now firmly demonstrated that the reason why mouse cells in culture have a very small Hayflick limit is because they respond to double strand breaks by senescing. Human cells are better at repairing (or preventing) DSBs so they only senesce when they get short telomeres. In vivo, we don't yet know for sure what happens because the best histochemical assay for human senescent cells doesn't work well in mice. There is some evidence that some senescent cells in vivo in humans are caused by short telomeres but some may also be caused by DSBs.

> The above seems to explain much of aging and cancer, but now the
> problem gets worse. If (b) goes on for long enough you will gradually
> accumulate mutations in various (most probably different) genes in
> *ALL* cells. i.e. the genomic "program" that the cells require to
> operate properly is gradually corrupted in random ways.

This sounds very logical, but it breaks down when you plug in numbers to compare mutations that cause cancer with mutations that don't. The key point is that cancer can kill you starting from just one cell, in contrast to non-cancer mutations which have to accumulate in a large proportion of cells (just as Robert points out). The only things that can be set against this are (1) that there are more genes mutations in which can make the cell dysfunctional than genes mutations in which can contribute to its becoming a cancer, and (2) that it takes several mutations to make a full-blown cancer. But (1) isn't a particularly big number -- probably a good 1% of genes are involved in cell cycle control at one level or another -- and (2) is undermined both by the "mutator mutation" phenomenon (the first relevant mutation may be one that globally increases genomic instability) and also by the amplification phenomenon (an early mutation that releases cell cycle control will cause the cell to become at least a few thousand cells before anything like angiogenesis inhibition kicks in, and it only takes any one of them to acquire subsequent mutations).

These facts, added to the fact that there are lots of ways for the cell to lose cell cycle control, mean that evolution has been forced to make our DNA maintenance and repair machinery far better than it needs to be in respect of non-cancer-related genes, just in order to be good enough in respect of cancer-related genes to stop us dying of cancer very young. This is brilliant for life extension, because it means that "all" we need to do to make somatic nuclear mutations irrelevent to our health for far more than a currently normal lifespan is to cure cancer really well. See

http://www.gen.cam.ac.uk/sens/SENS3.htm

for a summary of how I think we need to proceed.

The above logic applies just as well to epimutations (chances to histone and DNA methylation, etc) as to sequence changes.

> if the mutations occur within genes rather than say regulatory regions
> now you will probably have a protein that will not fold properly. This
> will probably be detected and the protein will be degraded. But the
> lack of a sufficient quantity of these proteins will probably result in
> cellular signals to make more of them. But they or at least half of
> them will not fold properly either. Now both protein manufacture and
> many types of protein degradation require energy (ATP). So when the
> cells detect a decline in ATP (due to futile synthesis and degradation
> of proteins) they may attempt to increase energy production.

This doesn't really work. The extra demand on the cell consists of making and degrading the proteins expressed by just one gene, possibly at twice the normal level. That won't change the cell's overall bioenergetics.

> The net result is that you get an exponential decline in function (i.e.
> aging). So far I've only managed to imagine two solutions for this.
>
> 1. Develop better DNA repair processes that do not allow the genome to
> become corrupted.

This is a decidedly mediocre solution since it only slows things down, rather than eliminating existing damage.

> 2. Shift things to allow more apoptosis when DNA double strand breaks
> are detected but also increase the replacement rate by stem cells.

This is a much better solution. It also applies to senescent cells, in the scenario that they actually matter at the tissue level despite being very rare; Judy Campisi is working on trying to push senescent cells to be more apoptosis-prone.

Aubrey de Grey

Posted by: Aubrey de Grey at April 4th, 2004 4:35 AM

Aubrey, I'll agree that double strand breaks are not "worse" than cancer because cancer can kill one before the DSB have a chance to cause many of the other aspects (I believe) of aging. But it isn't clear what fraction of DSB lead to cancer -- we do know they cause some fraction.

The only point about my emphasis on DNA repair is that we do not know how cells control (a) vs. (b), i.e. the homologous recombination pathway vs. the non-homologous end-joining pathway. It also appears that cells may sometimes switch between them in the middle of a repair operation(!). It looks like the HR pathway *might* work better if we had more copies of each gene (this may be how Deinococcus radiodurans gets away with a very high tolerance for double strand breaks). But shifting human cells to that as a default state seems likely to be difficult. What I would like to know is whether we can find a pathway in humans or other species that would allow end-joining of double strand breaks that does not include processes that create deletions in the genetic code?

Now I believe (though am not certain) that Rafal Smigrodzki may have once suggested to me that stem cells had very low resting metabolic rates (and so would have lower damage from free radicals over time). However that doesn't eliminate damage from radiation or perhaps toxic nutrient sources. So there are limits on how far you can push repair via promoting apoptosis and stem cell replacement.

Right now I don't think we understand the processes well enough to judge the advantages of specific paths. We do have enough information to begin to connect the dots and know what we should study next. But of course you have been doing that and I haven't been following it very closely the last couple of years for various reasons involving other priorities.

Robert

Posted by: Robert Bradbury at April 4th, 2004 6:18 PM

Aubrey: "evolution has been forced to make our DNA maintenance and repair machinery far better than it needs to be in respect of non-cancer-related genes, just in order to be good enough in respect of cancer-related genes to stop us dying of cancer very young"

That is an interesting interpretation with considerable scope for debate, but putting that aside - one is compelled to ask the question: what would the evolutionary disadvantage be (why has it been selected against) of having an even better genomic maintenance system?

Posted by: Harold Brenner at May 19th, 2005 8:19 AM

Harold, the problem is that repair of many types of DNA damage does not come free of charge. You have about 150 DNA repair genes involved in the 5 or so DNA repair strategies (everything from single base mutations to repair of double strand breaks). Repair does not come "cheap" in that you have to manufacture and recycle the proteins in the repair systems as well as manufacture the DNA bases required for replacement and dispose of the defective DNA bases which may be removed during repair. The DNA base manufacture (or import) & disposal processes are probably not particularly expensive but maintaining a pool of repair enzymes probably is. I'm guessing it works out to ~1-2% of the number of active genes within various cell types. Presumably most evolution takes place in an energy scarce situation so there are unlikely to be situations where systems are "over-engineered", e.g. keeping lots of repair proteins around in robust condition, "just-in-case", unless they are absolutely necessary for survival. This tends to be associated with whether the species are K-selected (for longevity) or R-selected (for rapid reproduction). In K-selected species there will be a tendency to push capabilities up, e.g. moving from 99.9% robust repair to 99.99% robust repair but you will rarely see something jumping to 105% robust repair (i.e. genome corruption never occurs or is never allowed to accumulate).

The evolutionary disadvantage of a better repair system would be less energy available for the purposes of survival and/or reproduction. Ultimately you would like to go back to ground zero and engineer an energy production system to replace the mitochondria and its enzyme complexes (that produce most of the ATP a cell requires) with an entirely different system which does not generate free radicals. But that is quite a non-trivial exercise. That would not eliminate the need for repair systems entirely however as one would still be left with endogenous and exogenous radiation sources which would damage the DNA (radioactive isotopes and cosmic rays respectively).

Posted by: Robert Bradbury at January 8th, 2006 9:17 AM

The evolutionary advantage in having a sub-optimal DNA repair is intimately involved in the process of evolution itself. If DNA repair becomes too efficient, it inhibits mutation which in turn decelerates evolution. Thus a highly efficient DNA repair system is a trait that would be selected against.

Extreme lifespan would require a genome that is resistant to damage for the duration of that lifespan. Consequently, extreme lifespan can come only at the cost of a slowing of evolutionary rate.

Posted by: Harold Brenner at June 2nd, 2007 11:36 AM

Post a comment; thoughtful, considered opinions are valued. Comments incorporating ad hominem attacks, advertising, and other forms of inappropriate behavior are likely to be deleted.

Note that there is a comment feed for those who like to keep up with conversations.