Hyperlipid by Petro Dobromylskyj
24 August 2019You need to get calories from somewhere, should it be from carbohydrate or fat?
Dave Speijer, an extremely insightful person if ever there was one, has a new paper out:
Can All Major ROS Forming Sites of the Respiratory Chain Be Activated By High FADH2/NADH Ratios?
the link to which I am extremely grateful to Bob for forwarding to me. This concept is purely from Dr Speijer. But I like it. A lot.
I'll start with an old doodle I produced about a decade ago depicting the front end of the electron transport chain. Matrix is at the top, cytoplasm at the bottom:
Electrons travel from NADH to Coenzyme Q, reducing it to QH2 which donates them to complex III being oxidised back to Q in the process.
The unlabelled blobs in the diagram are complex II (succinate dehydrogenase), electron transporting flavoprotein dehydrogenase and mitochondrial glycerophosphate dehydrogenase, all of which compete with complex I for CoQ as an electron acceptor. Given a high membrane voltage, a deeply reduced CoQ couple with most of the CoQ as QH2, then reverse electron transport back through complex I gives superoxide/H2O2 generation (provided the mitochondrial NAD pool is also highly reduced so unable to accept electrons (ie little NAD+, lots of NADH).
This is pretty straight forward and is the gist of the Protons thread. The extension of this is that, under high substrate availability, H2O2 from this process stops insulin facilitated caloric ingress to the cell.
Another major site of ROS generation in the ETC is complex III. Dave Speijer would hold that this is also triggered, like RET, by a deeply reduced CoQ couple. This is why.
So here is a stripped out version of the above doodle:
The problem here is that it's not that simple. Complex III does rather odd things with its two electrons. Electron bifurcation is a standard enzymic technique perfected very early on by proto-biology and it is exactly what happens here. One electron travels to cytochrome C (which only ever carries one electron at a time) and then on to complex IV and O2. The other does not:
The second electron is transferred backwards to another CoQ molecule and so partially reduces it to QH*, the radical semiquinone.
So (oxidised) Q is a necessary electron acceptor for complex III. Under high substrate availability and with most of the CoQ couple present as QH2 there will be very little Q available.
With one electron securely on cytochrome C, with any delay in the availability of Q, the second electron can be left sitting on one of the two haem groups along its route, easily available for donation to O2 to give superoxide, so adding its ROS signal to that of complex I. Both indicate that enough substrate is present and it is time to limit insulin signalling, to limit caloric ingress.
Nothing, absolutely nothing, about the construction of the ETC is random.
The general principle that a highly reduced CoQ couple is a signal to halt caloric ingress in to the cell applies to complex III just as much as to complex I.
When you want to resist insulin, you really want to resist insulin. Superoxide is then your friend.
If you fail to limit caloric ingress then eventually ROS from complex III are ideally placed to break the cardiolipins which anchor the water soluble cytochrome C to the inner mitochondrial membrane. Destroy these anchors and the ROS signal changes from"resist insulin" (good) to "perform apoptosis" (possibly not quite so good)...
And of course ROS in excess of physiological signalling are going to activate all sorts of inflammatory pathways.
Aside: The process of ROS generation is probably limited by pairing of complex IIIs (complex III is always a dimer in-vivo) for cooperation to use one Q to accept an electron from each of the pair. This should limit ROS generation as one Q will be available to two complex IIIs, twice that avaiable if they were each working alone. Nothing is random. End aside.
Altavista put up an excellent link to an article describing an AI algorithm capable of mining pubmed abstracts and coming up with gold. The exact opposite of modern meta-analysis (repeat after me: the meta-analysis of dross is dross). And the AI algorithm could use old data to predict recent discoveries! Sadly, the university I graduated from is currently teaching final year vet students that any publication over five years old is unimportant/can be ignored. The end of science. Alt was prompted to share his gem following this link put up by Hap:
C‐reactive protein promotes atherosclerosis by increasing LDL transcytosis across endothelial cells
Which is worth thinking about (and ignoring). Unless anyone thinks that LDL transcytosis across human umbilical vein derived endothelial cells (HUVECs) in culture resembles the process of arteriosclerosis, I think we can safely ignore this aspect of the modelling in the paper. Just ask yourself how severe is venous arteriosclerosis, with or without C-reactive protein. And the cells in the model die if you expose them to anything greater than 35mg/ld LDLc! But it brings up the apoE-/- mouse, a truly fascinating subject.
The apoE-/- mouse is also a model. They develop hyperlipidaemia and rapidly progressive "arteriosclerosis". Obviously, the apoE-/- lipid particles, which lack their apoE attachment protein, work their way (using active transcytosis of course) through the endothelium of blood vessels and cause cholesterol to accumulate in the subendothelial space and... Well. It's a pretty neat evolutionary dead end, if you believe it.
The apoE-/- model tells us more than anyone could ever want to know about apoE-/- mice. I've just spent some considerable time on Pubmed and SciHub trying to find out if there is any sort of full blown apoE-/- syndrome in humans. Not apoE2/E2 etc, more like no apoE at all. Zero. Zilch. Like the mice.
No luck finding it so far.
EDIT Yay, Adam found it
Effects of the absence of apolipoprotein e on lipoproteins, neurocognitive function, and retinal function
No suggestion of CVD. Just shows how important adding over 1% by weight of dried (oxidised) egg yolk to the diet to generate the "model" might be!
But apoE-/- mice are interesting in their own right. They really do accumulate lipids on their arterial linings in a manner exuberant enough to make a lipidologist wet their knickers. None of this messing about with the non-lipid intimal thickening so characteristic of real human arteriosclerosis. Lipids, lots of them, just "invade" the arterial walls and stick. Right there, pretty well on the surface.
Every now and then you trip over an interesting paper, in this case about why apoE-/- mice really have vascular problems. Here's one:
TLR2 Plays a Key Role in Platelet Hyperreactivity and Accelerated Thrombosis Associated with Hyperlipidemia
The paper is long and complex and very, very clever. As per usual. And less than five years old.
Here is the scene-setter from the discussion
"Patients with enhanced platelet reactivity are at increased risk for cardiovascular events.4, 37–39 Enhanced platelet reactivity is associated with chronic and acute inflammation, infections, diabetes, and a number of pathophysiological states related to dyslipidemia, including atherosclerosis, diabetes, and metabolic syndrome". My italics.
The mechanism appears to be through CD36, a multifunctional scavenger-type receptor present on most cells but here they are looking at platelets:
"Previously we have linked platelet hyperreactivity in dyslipidemia to accumulation in circulation of specific oxidized phospholipids, oxPC-CD36, which activate platelets via the scavenger receptor CD36"
The "previously" citation is to:
Phosphoproteomic Analysis of Platelets Activated by Pro-Thrombotic Oxidized Phospholipids and Thrombin
which introduces us to KODA-CP, or to put it more elegantly 9-keto-12-oxo-10-dodecenoic-phosphatidyl choline. This was just one of the more effective CD36 activators of the many lipid products present in oxidised lipoproteins.
So. Hyperlipidaemia facilitates the generation of KODA-CP which activates CD36, which activates TLR2, which makes platelets super sticky.
KODA-CP must have linoleic acid or arachidonic acid as part of its parent molecule.
The platelets stick. In apoE-/- mice enough of them stick to form massive aggregates on the arterial surface that look a bit like late stage lipid infiltrated arteriosclerosis plaques. It's a model.
BTW, platelets carry apoB labeled lipoproteins, among the many physiologically appropriate contents of their cytoplasmic granules (link below). Under the more normal generation of arterial intimal hyperplasia which precedes pathology I consider this lipid will simply be used for normal repair/hyperplasia processes. But there is nothing physiological about apoE-/- mice. They look like they should stick a ton of platelets to any damaged vascular wall, with more apoB labels than any-(mouse)-body knows what to do with. Given enough omega 6 PUFA to generate the KODA-CP.
Apart from the cardiologist derived omega 6 PUFA (another link below), did you notice the core involvement of the CD36 receptor? CD36 also facilitates free fatty acid uptake in to many cells. It is stored within cells and translocates to the cell surface, a bit like GLUT4 proteins, when needed. Stored in the cell, translocated when needed.
What controls CD36 translocation to the cell surface?
Insulin, of course (another post there).
Just thought you might like to know.
Let's just summarise. Lacking the apoE protein limits the utilisation of lipoproteins, much as having a fully non functional LDL receptor does in familial hypercholesterolaemia. This increases the concentration of lipoprotein particles in the circulation.
Applying the Dunning-Kruger effect to lipidology: Lots of LDLc particles = lots of invasion. QED. That's been it for the last 50 years. I don't thing many lipidologists every get past this obvious, unarguable, simple fallacy. Oh, also core to lipid "therapy" has been, and still is, giving corn oil to lower the LDLc count.
In reality the elevated lipoproteins are a marker of reduced utilisation and are associated with an increased residency time in the circulation. Given lipids based on palmitic, stearic or oleic acids I don't think that would matter.
Given lipids filled with linoleic acid, the essential precursor of KODA-CP, you will get a progressive rise in KODA-CP associated with increasing persistence of the lipoproteins. The more KODA-CP the more activation of platelets via CD36/TLR2 (and undoubtedly other pathways) and the stickier the platelets become.
Given the pathological intake of linoleic acid promoted by cardiologists and lipidologists working under their cholesterophobic hypothesis it seems perfectly possible that seed oils (and insulin) may well be drivers of the platelet adhesion which is core to the vascular damage in apoE-/- mice. Platelets even carry apoB100 labeled lipoproteins in their cytoplasmic granules which allows us to immuno-stain lipid accumulations with this LDLc implicating flag.
Given lipoproteins which lack apoE on their surface, accumulation of KODA-CP, hyperreactive platelets and a surfeit of insulin we are in a position to understand how the apoE-/- mouse works. Which is cool for those of us who like to understand things.
How much of this applies to actual human arteriosclerosis? Increasing platelet stickiness will amplify the normal response to arterial injury. I think this may be real. The rest is just a very extreme, rather bad model.
Most models, like this one, are usually useless.
Is it conceivable that cardiological dietary advice represents the exact opposite of the correct approach? That it would actively worsen the problem it is aiming to ameliorate?
Yep. But we knew that anyway.
Increasing linoleic acid in the diet is undoubtedly a facilitator of the generation of KODA-CP and the activation of the subsequent cascade goes a long way to explain the Sydney Diet Heart Study and the Minnesota Coronary Experiment. People died. From corn oil.
I'll stop now.
Some helpful links that didn't integrate neatly in to the text.
Effects of saturated and polyunsaturated fat diets on the chemical composition and metabolism of low density lipoproteins in man (1980, written on papyrus)
Apolipoprotein B release from activated human platelets (1986, probably on parchment, safe to ignore).
Dr Thomas "Just-take-the-statin" Dayspring writes on twitter:
"Any apoB-lipoprotein less than 70 nm in diameter can pass be pass thru endothelium - The LDLs are 20.5-25 nm. Remnants and IDLs are less than 70 nm and greater than 30 nm. The term small, dense LDL is way too simplistic - Big LDLs, like small LDs (less than 20.5 nm) if present in excess can invade artery"
This statement makes a prediction. It predicts that the accumulation of lipid in arteriosclerosis will be, initially, sub-endothelial.
As in this review of transcytosis (because passive "leakage" of LDLc down a concentration gradient across an endothelial cell layer is laughably impossible for particles over 6 nm across. No, that 6 nm is not a typo, according to the review). Not that "impossible" means anything to a lipidologist.
"During the initial stages of atherosclerosis, LDL particles are transported [transcytosed] across the EC [endothelial cell] barrier and accumulate in the subendothelial space".
So. All we need to do is a few post mortem examinations, find some poor people who had early arteriosclerosis at the time they died, and look for that lipid which will be sitting neatly under that single layer of endothelial cells lining their arteries. That is the prediction embedded in Dayspring's tweet.
If we go to this paper:
Early Human Atherosclerosis. Accumulation of Lipid and Proteoglycans in Intimal Thickenings Followed by Macrophage Infiltration
EDIT: from a link in Subbotin's excellent "Excessive intimal hyperplasia in human coronary arteries before intimal lipid depositions is the initiation of coronary atherosclerosis and constitutes a therapeutic target" END EDIT
we can find real images of real arteries from real people who died of non related causes while carrying different levels of arteriosclerosis:
Left side images are van Giesen stained for histology, central images are with Sudan IV for lipid and right hand are immunostained with anti-CD68 antibodies to show macrophages. The pairs of small black arrowheads indicate the level where the intima stops and the outer media layer begins. The vascular endothelium is a single cell layer at the top of each image.
Let's look at the circled image in detail. This is an example of early atherosclerosis from a real human with real early changes who died of non related causes. It's just the sort of place you might hope to catch an LDLc particle creeping between the cells of a single endothelial layer or freshly spat out after transcytosis by an endothelial cell. Lipid stains bright red:
Well, there's the lipid, deep, deep down at the junction of the intima and the media, right between the arrow heads...
There is none anywhere near the endothelial cell layer. If you believe that LDLc, as a result of a concentration gradient between artery and sub endothelial layers, "moves" or "invades" across that endothelial cell layer you have to explain how there is none at all in the sub-endothelial area and there is a progressive accumulation at the intima-muscularis junction. How does the lipid get from the top of the image to the deep spaces without any of it showing up in the lipid-free zone between the two?
"Beam me down, Scotty" is undoubtedly the most plausible explanation.
It is very, very hard to explain how utterly disreputable the lipid hypothesis is. All of this angst about increased LDLc and/or apoB counts on LC diets is based on the assumption that somewhere, somehow, cholesterol is the cause of heart disease. How LDLc "invades" (by active and controlled transcytosis!) the sub-endothelial space, disappears from there and then suddenly appears at over 200micrometres deeper, with none showing in the intervening zone requires a belief tenet which bears no resemblance to reality...
This was bollocks in the 1950s. My question is, as always, at what time did it stop being bollocks?
No one would reasonably doubt that the lipid deep down at the intima/media junction level comes from lipoproteins (though there are other plausible explanations). No one would doubt that loading the lipoproteins with with linoleic acid is likely to be a Bad Thing. No one would doubt that generating oxidative derivatives of the lipids in those lipoproteins might be a Bad Thing.
But trans-endothelial "invasion" is beyond belief.
This would suggest that all lipidologists are talking crap.
Nothing new there then.
This paper is interesting (and badly written):
A high carbohydrate diet does not induce hyperglycaemia in a mitochondrial glycerol-3-phosphate dehydrogenase-deficient mouse
It uses a mtG3Pdh knockout mouse, which is essentially a mouse which behaves as if it was on an enormous dose of metformin without all of that toxic blockade of complex I which gives a potentially lethal lactic acidosis at high dose rates. If you feed these mice standard crapinabag they are phenotypically normal. If you feed them a diet consisting some casein, a little PUFA to avoid EFA deficiency and the rest of the calories from pure sucrose they become rather interesting.
Eating pure sucrose does not make normal mice fat. It does make them insulin resistant and hyperinsulinaemic and, of course, insulin resistant adipocytes refuse to retain fat unless insulin action is facilitated by the oxidation of PUFA. Hence the normal body weight.
But the knockout mice actually become slim on sucrose. Here are the data, we can ignore the heterozygous (HET) groups:
They are slim because insulin levels are low. From the Protons perspective the function of the glycerophosphate shuttle in the pancreas is to drive enough reverse electron transport through complex I to trigger insulin release. Less RET, less insulin release, less fat storage, less hunger. Here is the isolated response of the perfused pancreas model to hyperglycaemia:
First phase insulin release is about a third of that in the normal mice. The mice are not diabetic because, in the absence of the glycerophosphate shuttle, RET to allow insulin signalling is generated by beta oxidation supplying electron transfering flavoprotein the CoQ couple via mtETFdh. Insulin signalling still happens but at the "cost" of increased lipid oxidation in the peripheral tissues.
What doesn't happen is sucrose induced insulin resistance. Again I consider this is triggered via the glycerophosphate shuttle causing RET at a level to shut down insulin signalling, which simply doesn't happen in the knockout mice. Lack of glycerophosphate shuttle also stops the generation of insulin-induced insulin resistance under conditions of high insulin concentrations coupled with energy replete cells.
Does anyone recall this figure from Metformin (01) post back in 2017?
Insulin was given at 90 minutes. At 150 minutes in the upper (non metformin-ed) rats insulin action starts to fail, at about the correct time for insulin-induced insulin resistance. By 180 minutes that upper trace, the non-metformin group, shows an upward trend in glucose as exogenous insulin levels are no longer high enough to overcome insulin-induced insulin resistance (the rats are DM T1 under insulin withdrawal).
At 180 minutes in the lower line showing metformin treated rats we can see the continued action of insulin being facilitated by the metformin because it blocks insulin-induced insulin resistance. It was mention in the comments to the post that, clinically, this effect of metformin might worsen the possibility of hypos in humans if combined with exogenous insulin usage. Potentially fatal hypos.
So what happens if you inject a sucrose treated mtG3Pdh knockout mouse with exogenous insulin to check their insulin sensitivity? Insulin sensitivity is preserved, to fatal effect:
All of the mice with the mtG3Pdh knockout died under exogenous insulin. This is exactly how I would expect metformin to behave in humans using insulin. A functional glycerophosphate shuttle allowed a sucrose diet to block this fatal sensitivity to exogenous insulin.
Obviously the mtG3Pdh mice have a normal complex I. Might they still develop lactic acidosis? Sadly the group didn't look at this (they had no idea back in 2003 that they had developed a meformin mimic model mouse). I do think there might be some elevation of systemic lactate despite a normal complex I.
In the absence of the glycerophosphate shuttle glycolysis is going to run directly to lactate to maintain redox balance. If glycolysis proceeds at a rate in excess of oxidation of the lactate within mitochondria (recall oxphos is slow compared to glycolysis) then some glycolytic lactate will spill outwards, though this is never likely to reach ICU-needing levels. No need to have a complex I blockade to generate mild lactic acidosis.
Does this metformin-ed like mouse have the exercise gains seen in human cyclists after popping 500mg of metformin pre-race?
That requires that we look at a different model.
Back in 2008 Noha Mesbha published her excellent PhD thesis
ANAEROBIC HALOPHILIC ALKALITHERMOPHILES: DIVERSITY AND PHYSIOLOGICAL ADAPTATIONS TO MULTIPLE EXTREME CONDITIONS
which introduced the world, via this paper
The halophilic alkalithermophile Natranaerobius thermophilus adapts to multiple environmental extremes using a large repertoire of Na+(K+)/H+ antiporters
to Natranaerobius thermophilus and its antiporter nt-Nha. Which gives every impression of being a stand alone Na/H+ antiporter very closely related to the invariably operon controlled MrpA protein (named shaA) of Clostridium tetani. As she says
"Gene nt-Nha had 35% identity to the shaA (mrpA) gene of Clostridium tetani. The Mrp proteins belong to the monovalent cation/proton antiporter-3 protein family....Sequence analysis of the regions surrounding gene nt-Nha, however, did not show that it was part of an operon. This indicates that gene nt-Nha does not encode a subunit of an Mrp system, but rather a mono-subunit antiporter".
All well and good.
Then in 2010 Morino et al published
Single Site Mutations in the Hetero-oligomeric Mrp Antiporter from Alkaliphilic Bacillus pseudofirmus OF4 That Affect Na+/H+ Antiport Activity, Sodium Exclusion, Individual Mrp Protein Levels, or Mrp Complex Formation
Although the whole paper was about B subtilis (and how none of its Mrp subunits worked in any incomplete combination to antiport anything) they did make this throw-away comment:
"A MrpA/MrpD homologue encoded by a “stand alone” gene from polyextremophilic Natranaerobius thermophilus was recently reported to exhibit Na+/H+ and K+/H+antiport activity in anaerobically grown E. coli KNabc (24)"
where (24) is Mesbha's PhD paper. Notice that at this stage Mesbha's
nt-Nha ~ shaA, very closely related at 35% conserved gene sequence,
has been changed by Morino in to
nt-Nha ~ An "MrpA/MrpD homologue".
This is a just about acceptable per se because MrpA and MrpD are homologous to each other and nt-Nha is closely related to MrpA (shaA) of C tetani. But Mesbha herself never mentions MrpD in her 2009 paper or in her PhD thesis. And "MrpA/MrpD" is open to mis-interpretation. So we have "definition-creep" here, where nt-Nha could be accidentally seen as some sort of composite of MrpA in combination with MrpD. Ouch.
So next we have the 2017 offering by Jasso-Chávez et al
Functional Role of MrpA in the MrpABCDEFG Na+/H+ Antiporter Complex from the Archaeon Methanosarcina acetivorans
where we have this bizarre statement
"On the other hand, a fused MrpA-MrpD homolog in the alkaliphilic Natranaerobius thermophilus displayed Na+/H+ antiport activity when produced in E. coli strain KNabc (5, 28)"
Ref (5) is Morino's paper on B subtilis Mrp, in which one rather misleading citation suggests that nt-Nha is an "MrpA/MrpD homologue". This has developed to the extent that nt-Nha has now "become" a fusion of two genes to give a rather mythical monster.
Ref (28) is just Mesbha's PhD paper in which nothing of the sort was suggested.
So the Jasso-Chávez paper is utterly flawed due to misinterpreting a poorly phrased statement and adding an erroneous modification so as to grossly mis-represent an initial very solid finding by Mesbha. The Jasso-Chávez discussion of nt-Nha can be distilled as:
"Send three and fourpence, we're going to a dance".
The chance of their understanding how nt-Nha or their very own archaeal MrpA subunit work as a stand-alone antiporter appears to be approximately zero.
BTW The folks who worked from the actual gene to model the nt-Nha protein structure suggest that
"The final model presents 13 transmembrane α-helices organized in a similar arrangement as the NuoL subunit".
You know the picture but here it is again 'cos I think it's lovely