Caloric restriction slows brain aging

Aubrey de Grey ag24 at
Sun Jul 2 13:11:22 EST 2000

Lou Pagnucco wrote:

> Thanks for the very informative reply.
> I hope, though, that you are at least somewhat too pessimistic.

I don't consider my view pessimistic!  Microarray and related work is
easily overinterpretable, that's all.  Since it's in its infancy, that's
no surprise.

> IMHO (as a layman), I tend to agree that many of these changes in gene
> transcription are indeed compensatory.  However, I am inclined to think
> that many are causal given that modifications of single genes have been
> found which significantly increase maximum lifespan in mice, drosophila
> and worms.

That logic is seductive but very fragile.  I recommend Gordon Lithgow's
recent article in Nature 405:296-297.  To paraphrase: the best-studied
life-extending mutant in C. elegans shows no compensatory diminution of
fitness (such as lowered fertility) in the normal lab environment, so
evolutionary theory predicts that it would enjoy a selective advantage
relative to wild-type, which would make it become the wild-type (i.e.
take over the natural population) once it arose naturally.  It's not
very attractive to argue that the mutation simply never HAS arisen
naturally -- basically, anything one can induce in the lab by random
mutagenesis using a comparatively tiny number of flies is certain to
arise often in nature.  Lithgow resolved the paradox by growing mutant
worms in conditions that mimicked nature more faithfully than standard
lab conditions do: specifically, he alternated them between food-rich
and starvation conditions.  The long-lived mutants rapidly lost out in
coculture with wild-type worms.  Everyone predicts that the same thing
will be found with all the other long-lived mutants in C. elegans and
other species, once we find sufficiently natural conditions.  Since we
do not ourselves live in natural conditions, the implications of such
mutants for human longevity are less clear... but it would be extremely
surprising if similar-sized life-extensions were found from single-gene
mutations in long-lived species.  Most of my colleagues are doubtful
that November's mouse result (Nature 402:309) will repeat even in long-
lived mouse strains, let alone long-lived species.  (Ever the optimist,
I have a princely five dollars with George Martin that C57BL6 will be
longevised at least 20%....  I should add that I am unaware that such
an experiment is yet in progress.)

> those "non-genetic" changes appear to be
> down range results of genetic programs activated earlier.  Why else
> would such changes be so species-specific?

I'm not sure I understand you.  Which ones are species-specific?  The
RATE of such changes is species-specific, to be sure, but that doesn't
make them programmed.  For example, most such differences can plausibly
be attributed to variations in the rate of mitochondrial superoxide
production as a percentage of oxygen consumption; that's under genetic
control, to the extent that it can be affected by the precise shape of
some of the enzymes in the electron transport chain, which is of course
a consequence of their amino acid sequence, but one can't really call
oxidative damage a programmed process.

> there are eminent gerontologists who surmise that the increasing
> fraction of senescent and near-senescent fibroblasts does indeed have
> deleterious effects on the extra-cellular matrix which, in turn, has
> damaging effects on the cells which must rely on the ECM to maintain
> their proper state of differentiation.

Actually that view is losing ground at present.  Judy Campisi, who first
identified senescent gene expression in vivo, now suggests that the most
important secreted products of near-senescent are growth factors, which
could promote tumours.  (I believe she first set this out in J Am Geriatr
Soc 1997; 45(4):482.)  But dysdifferentiation of the sort you describe
could also potentially have that effect, yes.  However, my point was that
this process is uneven -- some cells of a given type are affected a lot
more than others.  The stochastic element may be due to random choice
of which cells divide and which don't during growth (since the cells we
are talking about here are skin cells, dermal fibroblasts); it may also,
to a large extent, be oxidative, since telomeres shorten as a result of
oxidative damage as well as incomplete end-replication.

> >Moreover, microarray data
> >give virtually no clue as to the relative importance of these various
> >primary processes, because the changes they reveal are compensations
> >for the cumulative effect of all the primary processes mixed together.
> Amen.  This will certainly require the most powerful of supercomputers.
> Such a complex multifactorial process may be too intricate for the
> unaided human brain to disentangle.

Hm...  A contrary sentiment, which I share, is that a lot of the work we
could do now to intervene in aging is being held back by reluctance to
invest effort in projects that are not based on detailed understanding
of the system being manipulated (i.e. the aging human).  Ultimately we
must appreciate that there is a trade-off here -- the longer we wait for
more data and better theory, the surer we are to succeed when we try, but
the surer we are not to succeed soon.  I am in favour of being ambitious
now and hoping to get lucky, even if quite a lot of effort is likely to
be wasted in the process.

Aubrey de Grey

More information about the Ageing mailing list