From: Aubrey de Grey (ag24@gen.cam.ac.uk)
Date: Thu May 29 2003 - 09:28:19 MDT
Rafal Smigrodzki wrote:
> ### But is 7-ketocholesterol the non-degradable molecular species?
> According to: Lyons MA. Brown AJ.Metabolism of an oxysterol,
> 7-ketocholesterol, by sterol 27-hydroxylase in HepG2 cells.Lipids.
> 36(7):701-11, 2001 Jul., it is well-metabolized in the liver.
Many thanks for this ref - I was unaware of this work. Either 7-KC is
a species that is degraded in the liver but not elsewhere (which must
not be ruled out, because the liver can afford to make toxic things
and then chuck them out into the bile, an option that macrophages do
not have), or, like native cholesterol, it accumulates as a secondary
side-effect of the accumulation of other stuff that must not be let
out into the cytosol. More on this latter possibility below.
> ### If the balance is tilted towards formation of foam cells (e.g. by
> ingestion of dietary poisons, like trans-fatty acids, or by genetic
> factors in familial hypercholesterolemias), you will have them early
> on, and you could have your first heart attack in grade school (or in
> your forties if you have only margarine going against you).
Right. My point was that the veiw of Lusis (and others) is that early
foam cell accumulation is the norm. But see below.
> We already discussed reversibility of atherosclerosis, I think. Since
> you can observe a bulk reduction in the atheromas in human carotids
> with some dietary interventions (vit. B supplementation) as measured by
> Doppler sonography, the majority of pathologically relevant material
> responsible for artery occlusion is apparently digestible and is
> removed if the equilibrium is forced back towards lipid removal. If
> there are truly indigestible molecular species present in the atheroma,
> they would form only a part of it, and would not have a pathological
> significance in the age range that humans currently achieve.
All true. I mentioned one aspect of this earlier -- that there will
always be a proportion of stuff in the lysosomal compartment that is
digestible (and, indeed, being digested), and that is the material to
which your logic correctly applies. But that presumably applies only
to intact cells, not to the acellular core. I couldn't find a ref to
your claim that the **majority** of pathologically relevant material
in late-stage (artery-occluding) plaques is apparently digestible --
please provide it.
However, there's another important point re foam cells. It's clear
that a lot of what's in foam cells is straight, native cholesterol or
cholesteryl ester. Why? The prevailing interpretation, I believe,
is that there is a modified cholesterol species present in the cells'
lysosomes which will get out if the normal cholesterol efflux system
proceeds, so that system is arrested (or at least slowed) to protect
the cell. So then, the question becomes whether pre-foam cells --
healthy arterial macrophages that have begun to accumulate stuff but
have not yet become engorged -- can be isolated and their lysosomal
contents analysed for such products, which would be present at higher
relative concentrations than in more advanced cells. Might be a hard
experiment; I think looking at the acellular core of advanced lesions
will be more efficient.
> ### Lusis mentions the early fatty streaks without qualifications
> - but AFAIK, there are large differences in the age of onset of
> atherosclerosis between ethic groups. The Japanese, who were the
> subject group in Homa's study, seem to develop lesions later
> (presumably due to lifestyle factors).
Right, but that's a difference concerning frank atherosclerosis, not
foam cells. But I haven't found literature on whether or not foam cell
accumulation occurs later in Japanese -- it would certainly be good to
know.
> Homa didn't do a random sampling of the aorta - all samples were
> taken from predefined locations.
My poor use of "random". I just meant that it may be that essentially
all 10-year-olds really do have foam cells in their aorta, but that
different people have them in different places, such that a sample of
people may give some with none in any of the regions examined. Also,
the paper pooled the observations from all the regions, so we don't
know whether "no lesion in x%" means none in any of the places for x%
of subjects of none in x% of samples.
> ### Chinnery specified the "maintenance of wild-type mechanism", yes,
> but AFAIK he didn't say that random fluctuations are irreversible and
> cannot push the total number of genomes back down (except when driven
> to homoplasmy).
He said that the total number can fall due to random fluctuations, but
not be "pushed" down. That's where the bias in directionality comes
from: there is no push either way when the number of wild-types is
sufficient, but there's an upward push when it isn't.
> His model applies to functionally neutral mutations (Am J Hum Genet
> 2001; 68: 80206.), with the maintenance of wild-type being an add-on
> to the hypothesis, meant to help explain MERRF-like phenomena.
Ah, sorry, sure, yes, I was talking about this "add-on" the whole time.
It's entirely possible that something like what Chinnery discusses is
indeed correct for neutral mutations, but I've only been discussing
loss-of-function ones. Apologies for the confusion.
> There are threshold effects for derangements of mito function by
> various mutations, ranging from 30 to 95% mutated genomes needed to
> produce a detectable pathology. The cell cannot read the quality of its
> mtDNA genomes directly (at least, so we believe), so it must use a
> biochemical sensor to decide when to push mtDNA replication higher.
> Since the biochemical derangements appear late because of the
> thresholds, even if the cell tries to maintain a certain constant level
> of mito function, it will only react with proliferation to advanced
> heteroplasmy, and in an inconsistent, tissue- and mutation-specific
> manner, just as seen in real life.
Ah, but it's not just as seen in real life: this is the whole problem.
We see two clear deviations from real life in his model: (1) his model
predicts lots of proliferation at modest heteroplasmy (say, under 50%),
and (2) it predicts (for parameters chosen to give the observed number
of COX-negative cells) a lot more heteroplasmic cells than homoplasmic
(or very nearly) mutant cells. Thresholds don't help re the former,
because in Chinnery's model the only reason there's a threshold is that
there's a maximum number of mt genomes in total (wt plus mutant). Re
the latter, Chinnery has even cited Brierley 1998 from the same lab in
support of his prediction, when it explicitly shows the opposite!!
Tissue- and mutation-specificity is hard to interpret, of course, since
it depends on mutation rate as well as on selective advantage or lack
of it.
> ### Is it in review? I can't find it on Medline.
Capps GJ, Samuels DC, Chinnery PF.
A model of the nuclear control of mitochondrial DNA replication.
J Theor Biol. 2003 Apr 21;221(4):565-83.
PMID: 12713941
> ### Well, one way or another, large neurons would be generally much
> more likely to be either COX-negative or dying, by more than an order
> of magnitude.
Maybe, but not observably so, for the reasons I gave.
> our data show that there is a significant difference
> between the frequency of mutations in the D-loop and in the protein-
> coding regions, by almost an order of magnitude.
Right -- in fact if you only got "almost an OOM" you saw less of a
difference than most have.
> If all the mutations we observe were due to degradation of the DNA,
> then we would expect that the frequency would be the same for both
> areas, unless there is a significant influence from e.g. SSB vs. MTFA
> interactions.
Plus, e.g., that the D-loop is where the mtDNA is attached to the inner
membrane (nasty place to be).
> Also, the size of the fragmented, highly-oxidized mtDNA in Suter's
> article is about .2 to .5 Mb. We amplified fragments of 1 kb to 4 kb,
> and gel-purified the DNA to exclude smaller fragments, in both cases
> finding the same overall mutation frequencies, indicating that the
> mutations are not from the oxidized mtDNA which would be excluded by
> our approach, especially with the longer PCR products.
Fair enough.
> Finally, most types of DNA damage seem to block PCR rather than
> introduce errors, which is the basis of some methods for adduct
> detection.
Really? I thought only a few types block PCR (eg thymine dimers).
But I'm not well up on that, so I take your word for it.
> BTW, I just read the abstract of your paper in Protoplasma - sounds
> intriguing, but the full text is not available online. Could you send
> me a copy?
I actually don't have full-text access, even though Cambridge does get
the journal in hard copy. If you (or anyone else reading!) can get it
(http://link.springer.de/link/service/journals/00709/tocs/t3221001.htm
and my article is the second in the issue), I would be most grateful!
Aubrey de Grey
This archive was generated by hypermail 2.1.5 : Thu May 29 2003 - 09:39:53 MDT