Thursday, February 25, 2010

Arteriosclerosis in 1957

I was born in 1956, so I was an infant at the time these photographs were being taken. Each one records a personal tragedy. We should learn from them.

All illustrations are taken from Henry Moon's classic 1957 paper.

This is a normal coronary artery. It has been stained to emphasise elastic tissue. Note the continuous folded band of elastin with nothing visible between the elastic layer and the artery lumen (there is actually a single layer of epithelial cells here).

The thicker layer outside the thin black elastic layer is media and is made of muscle. The more granular layer outside the muscle layer is the adventitia. The slide comes from a 5 months gestation human foetus who died, without being born, in a tragic accident. Accidental death autopsies are where all of the pictures in this post come from. Non are cardiac patients.

Next is this picture of the coronary artery of an infant who was 4 days old. Note that most of the intima (dark red wavy layer) is normal and that all of the media is normal. But look at the lower right, there is an abnormal area of the elastic layer. It is shredded and there is fuzziness over the surface. This is the earliest stage of arteriosclerosis noted in this series.

If we next go on to look (slide below) at the coronary arteries from an infant of four months of age we can see a small section of normal intima, with intact elastic layer, on the upper left of the section of artery with normal muscle in the media outside it. Over the rest of the artery the elastic layer is grossly disrupted or absent, the intima is grossly thickened and the muscular media is still relatively normal, though it is a bit thickened at the lower area of the section:

And then we have this superb section actually through a branch of the coronary artery of a child of three years of age. There are some areas of elastic tissue intact but much of it is damaged and the intima has generally thickened wherever the elastic tissue is disrupted. The effect is most marked by the mouth of the branch. The smooth muscle layer (media) is still quite normal but somewhat thickened at the mouth of the branch too.

These last two high magnification images in one picture are stained to show the material of which the thickened intima is composed. In the lower image the changes extend in to the muscular media too.

Do you think it might be cholesterol?

You know, LDL-C, the stick-and-die stuff? Oddly enough if you do frozen sections and stain them with sudan red there are very occasional macrophages with a little lipid in their cytoplasm scattered thinly through this gunk.

But no, the wall to wall stuff stained with Prussian Blue is mucopolysaccharide, I think nowadays it's called glycosaminoglycans or GAG.

To me it's really weird how a cardiologist can think that LDL causes this, and that statins might stop it.

But then the world is a weird place!


It's Friday and we're heading south for the weekend so the blog will be quiet for a few days.

Tuesday, February 23, 2010

Saturated fat and sdLDL?

The post itself and the exchange of comments on Dr Davis' blog about genetic causes of sdLDL piqued the interest of many of us. You need to have read the post and comments to make sense of this particular post here. I'm not keen on decrying the concept of genetic sdLDL out of hand but, obviously, there is a great deal that can be thought about around the non genetic concept. My thoughts are down here as I don't want to go cluttering Dr D's post up with comments that are clearly mine, from my biased viewpoint, and very probably not congruent with those of Dr D. The situation is too interesting not to expand a little though...

I roughed out Stan and Ollie on Fitday, very crudely and making big guestimates.

Here they are:

Very crudely for Stan

I gave them both half a kilo of cabbage each, nuts as almonds, (slightly more for Stan as he is weight stable) and chicken, again slightly more for Stan to keep him weight stable. I also allowed Stan some cannola oil to make his total calories up to 2000kcal/d.

And here is a guess for Ollie

I assumed Ollie was burning 0.34lb/d of his own fat and entered this as lard. I allowed him a total (including the lard from his butt/belly) of 2,300kcal/d as he was carting a fair bit more weight around than Stan. Probably an underestimate.

Saturated fat "consumption" for Ollie (who near eliminated sdLDL) worked out as 67g/d, total fat was 83% of 2300kcal, MUFA 105g and PUFA 31g. Most of this fat was from belly-fat plus almonds.

Fat for Stan (who retained sdLDL) was from his diet only as he, quite correctly, lost very little weight over 6m. Results were saturated fat 17g/d, total fat 70% of 2000 calories eaten, MUFA 84g/d and PUFA 53g/d.

I find it hard to see that Stan's saturated fat is the reason for his sdLDL, unless one posits that 17g/d of saturated fat as chylomicrons (apoB48 labelled, gut produced) causes sdLDL in some way (sdLDL is apoB100 labelled and produced by the liver). While at the same time Ollie's 67g/d of saturated fat, available as FFAs after release by hormone sensitive lipase from his adipocytes, behaves differently to Stan's 17g/d saturated fat released from chylomicrons by lipoprotein lipase....

It's possible, but it seems implausible to me.

Dr Davis mentioned the tendency of this sub group of his patients to be borderline diabetic. This to me is far more interesting and makes my ears prick up. Both apoB100 and apoB48 have glycation-predisposed sites which, once glycated, inhibit their uptake by the LDL receptor (certainly for the apoB100 particles, probably something similar works for the apoB48 particle, I've not chased this).

If apoB100s are not taken up because they are glycated due to a borderline diabetes tendency we have a very plausible mechanism for "atherogenic" remnant particles both being formed and "remaining". It just needs a little sugar to get them there. And of course, some of us think sugar and arteriosclerosis might be linked anyway...


Sunday, February 21, 2010

Physiological insulin resistance: Dolphins

This article was sent to me by Stan, and it's interesting on many levels. At the most basic is the gross error in the description of the management of diabetes. This is what the article says:

"In diabetic people, chronic insulin resistance means having to carefully control blood glucose, usually with a diet low in sugar, to avoid a variety of medical complications."


Human diabetes is managed by a diet low in FAT. Ask any diabetologist.

The experience of Dr Dahlqvist encapsulates the monstrous medical approach to the use of low carbohydrate diets in diabetes.

I'm feeling a bit polite tonight for some reason so I won't mention what I think about low fat diets and diabetets. Perhaps I need a glass of wine.

This marine mammal researcher can see physiological insulin resistance in dolphins and see that it is PHYSIOLOGICAL. The difference between a healthy dolphin and a healthy human is minimal (can she see that too?). We humans "do" physiological insulin resistance. But she and her collaborators cannot see that there is a difference between physiological insulin resistance and breaking your liver by living on soda and bagels to get pathological insulin resistance... As she says:

"If we started feeding dolphins Twinkies, they would have diabetes."

Not true. Their insulin resistance would go as they switched on carbohydrate metabolism in their muscles. It would take several years of Twinkies to cause diabetes. Like humans. We're fine for the first few hundred/thousand Twinkies. Then we break.

EDIT: Being in the UK I hadn't realised how small Twinkies are. Let's say 100,000 or so to break your liver...

But ultimately we humans need Twinkies to survive. We must eat them to remain happy and feel part of normal society. Imagine a teenager saying no to a Twinkie, just because they are diabetic! No, we must help people to eat Twinkies while diabetic, so we MUST research the "fasting gene" which is abnormally activated in human diabetes. And develop a drug to turn it off, of course.

BTW the activator of the fasting gene will turn out to be palmitic acid. What other messenger would you use to suggest that there is a fasting state? So we're back to using Palmitofake and a continuous supply of Twinkies.

The title says "Dolphins have diabetes off switch"

No, they do not. There is no off switch for a broken liver. Unbroken dolphins are just behaving like unbroken humans. They turn off physiological insulin resistance when carbohydrate becomes available, even if that carbohydrate come from fish via gluconeogenesis. It's simple and it's NOT diabetes.



There is potentially a whole load more posts from this link, follow up depends on all sorts of things...

Wednesday, February 17, 2010

John Hawks on Paleo in NY

I was looking for this report by scrolling back through John Hawks' weblog (it wasn't there, it was in Thinking Meat).

It had come to mind because I had made a bolognaise sauce for tonight's supper and had seared the mince and onions, quite deliberately, because I like the taste of seared meat. I guess like sweetness because it was rare before sucrose became widely available. But burned meat? AGEs and ALEs? Why? And is it bad for me? With quite well defined hearth use at 750,000 years ago I think I'm OK to eat seared meat. I'm not sure why I like it. It's not quite the same as feeding gamma irradiated "food" to cats (their brains fall to pieces) and I suspect it's something humans have been doing for a long time. A lot longer than the 10,000 years of growing grains, which Hawks rightly suggests some of us may be better adapted to than others, even on this brief time scale.

So AGEs and ALEs, yes. Unless, of course, those out-of-Africa-a long-time-ago people had stainless steel cooking pots for boiling meat in, which we've just not found yet...

Scrolling back through Arrested Adaptation, in a vain effort to find what wasn't there, I stumbled on his take about the Paleo in NY article which was discussed in "our" zone of the bloggosphere when it came out. I enjoyed it a lot. Though he has some gripes and beefs about "modern paleo" (with a particular BIG down on Crossfit), his article was generally enjoyable and informed. Being one of those weirdo's with strange dietary habits, even if not paleo, it's a good to get a bit of fun poked at me occasionally, provided it's not malicious...

You do tend to forget how strange you are.

Well, I do anyway.


Cholesterol: Near miss in Edinburgh

Just another brief post. This group were very, very, very lucky. Crucifiction is supposed to be unpleasant.

What did they do? Well it was more of the usual stuff but, back in the early years of this century, it was still considered ethical to include a placebo group in a statin trial! Gasp, horror, call the ethics committee.

Anyway, 54 people, all with known heart disesae, were allowed to go for 24 months without the benefit of atrovastatin. You wouldn't do this in the states! But here in Scotland, well, a few more heart attacks and no one will notice.

So what does a greater than 50% reduction of LDL-C and a near 50% reduction in C reactive protein do for CAC score progression?

Diddley squat, as you would expect. It's all the usual quotes:

"statin treatment does not have a major effect on the rate of progression of coronary artery calcification"


"Serum low density lipoprotein concentrations were not correlated with the rate of progression of coronary calcification (r = 0.05, p = 0.62)"

All utterly tedious.

What grabbed me was the near miss. The CAC score progression in the atrovastatin group was not significantly different to the placebo group.

The p value was 0.18

What would have happened if the p had cracked the mystical p<0.05, say with bigger groups or longer follow up?

In this study atrovastatin "allowed" or "facilitated" progesssion by 26% per year, placebo by only 18% per year.

As I say, p=0.18

A near miss for the careers of all involved.


Oh, and in the full text you can find that there were actually 2 diabetic patients and 10 current smokers in the placebo group vs no diabetics and only 5 current smokers in the statin group. Equalising these might have given catastrophe.

Tuesday, February 16, 2010

Cholesterol: Peto seeing some light?

Even Sir Richard Peto (second author) is seeing the light. Sir P is famous for stating (loosely remembered by me) that no individual study of the role of cholesterol in CVD is particularly convincing, but the overall weight of evidence was. I'm stuck with reading study after study and realising they are crap. A big heap of crap is no more convincing than a small heap, to me. But then I'll never get a knighthood.

Back to the Oxford abstract:

"Given usual apoB, lower LDL-C (consistent with smaller LDL particles) was associated with higher risk (P < 0.0001)."

Translation: at a given number of apoB100 particles, the lower the measured LDL the HIGHER the risk.

"The ratio apoB/apoA(1) was substantially more informative about risk (chi(1)(2) = 550) than were commonly used measures such as LDL-C/HDL-C, total/HDL cholesterol, non-HDL cholesterol, and total cholesterol"

Translation: Most of what we have measured in the past is bollocks. Our new ratio is slightly less bollocks.

Notice they didn't mention HDL/trigs. And they still believe sdLDL is out to get you. And no one has pointed out to them that sugar is a great generator of sdLDL.

I've not bothered down loading the free full text. The abstract says exactly what you would expect a real view of the world to say. That's enough to take notice of without working too hard.


Lipoprotein(a) and ascorbate

OK, ascorbate and Lp(a).

There are some posts which are quite hard to write. I really like the "Lp(a) is a surrogate for ascorbate" hypothesis. It's neat and elegant, so the more information I read which undermines it, the more depressing I find it.

We humans are members of the sub order haplorrini of the order primates. Haplorrhini, that is humans and apes, Old World Monkeys and New World monkeys, have all lost function of that gene for the last enzyme in the formation of ascorbate from glucose.

Of course the crucial species are the tarsiers.

Tarsiers are quite interesting as they have been batted back and forth from the haplorrhini, ie us lot, to the prosimians, ie the rest of the primates. It's a very very close call genetically as to whether tarsiers are haplorrini or not.

There is one isolated report that tarsiers cannot make vitamin C. This should put them firmly in to the haplorrini but geneticists still argue the exact grouping.

It matters because the loss of ascorbate synthesis appears to have occurred in haplorrhini very soon after the split between our sub order and the rest of the primates.

Now, there is this concept that humans lost the ability to synthesise ascobate because a common ancestor to all of us haplorrini sat around all day eating melons and nectarines. Of course it's quite hard to say exactly what the ancestor of the tarsiers actually ate, but it's quite simple to say what current tarsiers eat. They eat insects, lizards and small snakes. I've seen at least one textbook entry which suggest that tarsiers are about as close to anything resembling the common ancestor of we haplorrhini as anything alive today.

Perhaps tarsiers use Lp(a) as an ascorbate replacement? No one has checked this, but what they have checked is whether the New World monkeys make Lp(a). They don't. Or, if they do, it is unrecognisable using human Lp(a) sensing antibodies. If tarsiers can make an Lp(a) like substance it puts them firmly down with the Old World monkeys and apes. They're certainly not an Old World monkey or an ape.

To summarise: Only Old World monkeys and apes make recognisable Lp(a). All haplorrhini have lost ascorbate. The haplorrhini are probably derived from an ascorbate-less near carnivore. New World monkeys are doing fine without Lp(a) or ascorbate. Modern tarsiers are doing fine as pure carnivores without ascorbate....

Of course the guinea pig is a fascinating little beast too. As far as I am aware, it does not make Lp(a) (Pauling and Rath appear to have made a mistake here), it doesn't make ascorbate and it doesn't eat much fruit either. OK, guinea pigs never ever, ever eat fruit in the wild. They live in the high Andes where the year round fruit availability is probably as good as it is in Antarctica.

They eat grass. And guinea pig pooh. That puts ascorbate intake down at very low, but not zero, levels. Like rabbits, they actually live primarily on volatiile fatty acids produced by hind gut fermentaion of cellulose. The minute you feed them cr@pinabag they will immediately become either hyperglycaemic, hyperinsulinaemic or both compared to living on short chain saturated fats. That's not exactly how they are designed to run and their ascorbate requirement rockets. BTW they become obese too. Feed them rabbit cr@pinabag and they die. It's a bit like feeding sailors or tarsiers on dried salt beef and ships biscuits, but probably doesn't taste as good. Doesn't have the weevils either.

Then there is that fruit bat which has also lost its ability to synthesise ascorbate because, well, it eats fruit. This sounds quite convincing as a reason for losing ascorbate until you realise that it seems as if all bats have lost their ability to synthesise ascorbate. Most ascorbate-less bats are pure carnivores. Insects are where it's at. Insects don't synthesise or supply much ascorbate. Fruitbats are the oddballs among bats and it's probably not why they don't make ascorbate. That's just a bat-thing.

Now, I hate to mention hedgehogs again but, here I go, we all know that they not only do make ascorbate, but they also make an Lp(a)-like substance too. As an aside to pointing out that this doesn't go along with Lp(a) as a surrogate for ascorbate, I'd just like to mention kringles III and IV to clear up a minor point:

Humans repeat kringle IV from plasminogen in their Lp(a). Hedgehogs repeat kringle III. How convergent is that? Not very. But in hedgehogs it is kringle III of plasminogen which binds to fibrin, in humans it is kringle IV. So hedgehog Lp(a) seems to be markedly convergent with human Lp(a). This makes me happy as it suggests Lp(a) is not some idiotic mistake only made by humans and their kin. It's worth evolving by non related species. That just leaves the big unknown of whether hedgehogs are also using Lp(a) to deliver oxidised lipids to where they are most needed (guess yes!).... And what we are both doing with those lipids. And their kringles in addition to binding to fibrin.

Anyway, the simple concept that humans and guinea pigs make Lp(a) as an ascorbate substitute seems to be so full of holes that I can't see it. Sigh.

Mega dosing on C in the attempt to put the clock back 40 million years and so reduce Lp(a) and heart disease does not hold water. Whether there is a pharmacological benefit of mega dosing ascorbate, along the lines of using niacin to mimic beta hydroxyburytate, seems possible. But if this is the case it's an accidental benefit derived from faulty logic...

Oh, a last comment on bats. None of them make ascorbate. Some, but not all, are VERY long lifespan species. Not in the league of Naked Mole rats of course, but some certainly live far longer than the same sized mouse.

Never mind mega dosing on ascorbate, that's without making any ascorbate and living on insects.


Tuesday, February 09, 2010

Lipoprotein(a): 7-ketocholesterol and cancer

This group did lots and lots and lots and lots of very, very, very clever things. So many, so complex and so clever that I'm not going to try and discuss them all. I was stuck with having to read them all because of the big lump of truth that they tripped over without apparently commenting and I had to look for any justification of the actual conclusions they came to! It's one of those papers where you can't tie the discussion to the results...

This is the executive summary to let me get to the holes in the study.

They found that oxLDL acts on cancer cells in tissue culture to induce two specific changes. The first was the induction of autophagy. This is the housekeeping process of clearing out the odds and sods of damaged and non functional proteins which accumulate with aging. This is a Good Thing, classically induced by ketosis. Good old beta hydroxybutyrate is the classic trigger. So, apparently, is oxLDL!

Obviously, if you are certain that oxLDL causes cancer, you have to consider that any Good Thing happening in bad cancer cells is a Bad Thing for the organism in general. There is no doubt that oxLDL does induce autophagy in cancer cells. This is purported to encourage super cancer cells to develop. These are what I was looking for in the results.

The second thing which oxLDL does is to induce apoptosis in cancer cells. Generally, death is considered to be a Bad Thing from a cancer's perspective. For the organism affected by cancer cells, apoptosis of those cancer cells might be considered a Good Thing.

I dunno.

Perhaps you could say that oxLDL is really Good all round, the cancer cells can die with the happy satisfaction that they had dusted the cupboard under the stairs and emptied the rubbish bins before they committed suicide and the organism is happy to lose the cancer?

There is an unspeakable amount of detail about cell pathways involved but the bottom line is free radicals, ROS. Some ROS equals autophagy, lots equals apoptosis.

OxLDL causes lots of ROS.

Now, if you ignore the discussion and introduction, both of which are exercises in cognitive dissonance, and go to the results you can play hunt the super squeaky clean autophagic cancer cells. Until you are blue in the face. They're not there because they are all either dead or looking VERY sick. No one mentioned finding sleek autophagous cancer cells. Just dead ones.

The action of oxLDL on cancer cells is to kill them. There was no other finding noted on the cellular level, using a microscope. OxLDL, especially the 7-ketocholesterol component, is a potent apoptosis inducer in cancer cells. Quite how this can be a Bad Thing, as expressed throughout this paper, is beyond me.

But there are a series of snags.

OxLDL also induces apoptosis in cultured human umbilical vein endothelial cells. This is undoubtedly a Bad Thing from the CVD point of view. Like Really Bad.

So perhaps oxLDL is really the cause of arteriosclerosis?

It's a nice idea but there really isn't a lot of oxLDL in the blood of any species. In humans it's bound to Lp(a) and in all other species it is present at the lower limits of detectbility. Let's just go back to the Bergman/Krauss/Tsimikas paper and recap:

"In individuals with low Lp[a] levels, there is a corresponding low level of OxPL/apoB, suggesting that in the absence of Lp[a], these OxPLs do not accumulate on plasma apoB-containing lipoproteins other than to a minor degree. A similar situation exists with most animals that we have studied (32, 33). For example, in mice with marked hypercholesterolemia, a situation in which OxPLs recognized by E06 are abundant in the arterial tissues (and probably elsewhere as well), the levels of OxPL/apoB in plasma are very low, and often just at the level of detection of our assay (33). In contrast, Lp[a]-transgenic mice have very high OxPL/apoB levels, even in a C57BL/6 background without obvious atherosclerosis (34). Presumably, this reflects the generation of such OxPLs as a component of normal physiological processes. Lp[a]-transgenic mice express both human apoB-100 and apo[a] and thus can form a true covalent Lp[a] similar to that found in humans (34). Mice expressing high levels of human apoB-100 alone, or apo[a] alone, nevertheless have very low levels of OxPL/apoB (34), suggesting the need for an intact Lp[a] to enable preferential binding of OxPL."

So when you take some native LDL, cook it with copper sulphate and then pour the resulting mess of goo on to cultured cells on a petridish you are really getting a handle on what is happening in the real world of human cardiovascular disease aren't you? Like yeah.

In the real world the oxidised lipids of oxLDL rapidly transfer to Lp(a) (in humans, apes and transgenic Lp[a] mice) or (in all other species) they are excreted, presumably by the liver. Either way, they don't hang around.

Lp(a) has a large sticky moiety which binds to fibrin and proteoglycans. It's not going to see either of these in healthy vascular tissue. Delivering 7-ketcholesterol to the damaged tissue around a cancer cell might be a bit of a magic bullet. It's hardly surprising the body makes lots of apo(a) when the liver picks up markers that you have cancer.

Delivering it to damaged vascular tissue, where fibrin or proteoglycans are exposed, might be good or bad. Probably good in view of a number of other oxidised lipid products present in Lp(a). But if the damage is on going and severe it seems a little unfair to blame the sticking plaster for the damage.

But I would maintain that circulating Lp(a) through healthy tissue has NOTHING in common with pouring copper-oxidised LDL on to cultured vascular smooth muscle cells or endothelial cells.

I suppose at some stage we have to talk about the natural arteriosclerosis which occurs in aged rats and rabbits, but there is so much more of interest about Lp(a) and so little research published to work from...


BTW Barry Groves has an interesting article by Wayne Martin on the role of Dipyridamole in cancer management. It has a nice description of cancers and fibrin which fits neatly with Lp(a) being a potential anticancer stratagem based on targeting blood clot with an apoptosis inducing agent.

Monday, February 08, 2010

Lipoprotein(a) and oxidised lipids: What's your mimium requirement?

Just briefly, from this paper:

The paper was written before the Lp(a) is oxLDL paper, so the authors are showing the two seperate plots, one is the amount of oxidised phospholipid per unit LDL and the other is Lp(a).

I like the similarities in the curves! Life should be logical.

From the practical point of view, at what level of Lp(a) do you need to hurry in making out your will? Look at graph B.

Obviously there is a "least risk" value for Lp(a), but if you had null genes for apo(a), and so come out with zero Lp(a), you would still be in exactly the same risk category as the sextiles 3 and 4 but wouldn't be worrying about it! So, in this study, using this Lp(a) assay, for people eating modern food, anything below 24mg/dl has the same risk as zero Lp(a). Perhaps lower if you are lucky.

Note also from graph A that having undetectable oxidised lipids per unit LDL in your blood DOUBLES your risk of incident CVD. That's it. No oxLDL [ie no Lp(a), whichever you please] DOUBLES your risk of heart disease.

EDIT: I guess "is observationally associated with" would be a better phrase. Don't want to sound like the AHA about cholesterol here!

And you thought your genes evolved apo(a) to kill you!



Sunday, February 07, 2010

Lipoprotein(a) is oxidised cholesterol

If you want someone with serious expertise on the separation of plasma lipids using an ultracentrifuge then Krauss, obviously, is your man. Author number nine out of twelve on this next paper. He was, obviously, their man too.

If you want to look at oxidised lipids in lipoproteins you need an antibody which locks on to oxidised lipids but not to undamaged lipids. This is called E06. It is quite specific and only binds to the phosphocholine residue of an oxidised phospholpid. It's probably the most commonly used antibody for detecting oxLDL. Or what people thought was oxLDL until this paper came out.

What did this group do? All sorts of things, most of them very clever in deed.

Let's scream through the results. First they took an antibody to Lp(a) and pulled the Lp(a) out of plasma with it. E06 reactivity was pulled out along side Lp(a).

They got Krauss to spin some lipids and found, lo and behold, E06 reactivity separated out ONLY with the Lp(a) fraction.

They took some LDL and oxidised it artificially with copper ions. Then they offered it a choice of Lp(a) or native LDL to share lipids with. Not only did E06 reactivity jump out of the oxLDL and in to Lp(a), it ONLY jumped in to the Lp(a). None jumped from copper oxidised LDL to native LDL. None would leave its Lp(a), even to go to another Lp(a).

They went on to check if Lp(a) is just susceptible to oxidation in its own right, by looking for malondialdehyde-lysine residues. It's not oxidised itself. It just collects oxidised lipids. The antibody for malondialdehyde-lysine is E14. It ignores Lp(a). However the E06 antibody to oxidised phospholipid not only recognised Lp(a) but also apo(a) alone, presumably both from sources where they have had access to oxidised lipids.

The implication from this is that while some of the oxidised phospholipids are in the lipid particle of the Lp(a), a big chunk are also bound to the apo(a) protein.

The group feel that apo(a) initially captures oxidised phospholipids from the aqueous plasma phase and they are then transfer over several hours to the lipid droplet.

EDIT: I miss read this, the lipid drop captures the oxidised phospholipds and then they get arranged on the apo(a) glycoprotein. This may have some relevance to apo(a) isoforms, Lp(a) levels and vascular injury.

They doubt that Lp(a) really goes around stealing oxidised lipids from oxLDL particles. What they suspect is happening is that whenever oxidised lipids are released from damaged tissues, Lp(a) is the mop which mops them up. They probably never get as far as native LDL.

As far as they are concerned Lp(a) IS oxLDL. And oxLDL IS Lp(a).

Do tissue damage, the liver makes a sponge for oxidised tissue lipids. Probably many more oxidised lipids than E06 recognises.

You have to wonder whether the liver senses free oxidised lipids in the bloodstream and makes apo(a) in response to them (almost certainly the case, because Lp(a) spikes after injuries such as percutaneous cardiac procedures, where everyone expects oxidised lipids to be mechanically released without dietary warning). Or whether, as in the Finland intervention, volunteers do something grossly stupid such as reducing the fat content of their diet. And the liver pre-empts...

Probably a bit of both.

So, Beth, you asked in the comments section of the last Lp(a) post:

What does Lp(a) actually do?

It preferentially accumulates oxidised lipids and binds them in a form where they cannot be immediately excreted from the plasma. It also puts a great big sticky label on them that allows them to firmly bind to damaged tissue.

Only Lp(a) does this.

Only in humans and related apes. Oh, and in mice genetically engineered with both human apoB100 and apo(a) in combination. Of course.


That too is an interesting question.


Friday, February 05, 2010

Lipoprotein(a) Bantu recap 2

More from the Lugalawa study in Tanzania.

We all get two copies of any particular gene, one from our mother and one from our father. The gene coding for the number of kringle IV repeats on apo(a) is exactly like this. Some people get apo(a) of the same size from both parents, others get a pair of different apo(a)s, one size from each parent.

If you are trying to compare the effect of apo(a) size on Lp(a) concentration, and the confounding effects of diet, it would be really helpful to have everyone in the study with just one size of apo(a) each. They're not actually that rare and in the Bantu study Pauletto found quite a number of individuals with only one apo(a) size who could be pair matched with someone similar from the neighbouring village.

The effect is shown here in Table 2

Just in case the table is not 100% clear on a casual glance (it wasn't to me!!) here is the section we need, just copied from the middle section of the table.

We can see that the median in the fishermen is half that of the vegetarians with a p value < 0.001

Just imagine a drug had been found to halve Lp(a) level. Eric Topol would wet himself!

But it gets more interesting. Pauletto went on to plot isoform size against Lp(a) concentration. Here are the data from the matched apo(a) size:

Look especially at apo(a) with less than 15 kringle IV repeats. Lp(a) is 55mg/dl in the vegetarians and less than 20 in the fishermen. Same genes. Now this is not statistically significant, but that comes from trying to compare 6 farmers with 4 fishermen!

So Pauletto cheated a little to bump his numbers up and went on to do the same plot but included anyone with more than 85% of their expressed apo(a) size in a given bracket. This gave more people and significant p values across all ranges.

I don't find the differences very biologically convincing at the right hand end of the plot with the long apo(a) kringle IV repeats. With long apo(a) you can eat what you like and still have low Lp(a), within these cultures. But it seems likely, if you believe Lp(a) furs up arteries, that the short apo(a) repeat farmers are in a lot more trouble than the similarly gifted fishermen... Lp(a) 60mg/dl vs 20mg/dl.

But it gets worse for the farmers! They are significantly more likely as a group to have short form apo(a). Here is the isoform distribution between the groups:

The bottom section clearly shows that the vegetarian farmers are more likely to have killer apo(a) isoforms, as well as making more Lp(a) than they should for a given apo(a) size. A double whammy.

Why should this be?

Pauletto thinks it's a founder effect:

"...even though the 2 populations are racially homogeneous, we have found a significant difference in apo(a) isoform size distribution between the inhabitants of the 2 villages, probably due to differences between the founders of the 2 villages."

This seems implausible to me. Founder effects would certainly be likely to persist in isolated populations but these populations intermarry commonly.

What seems far more likely is that there is an on-going positive selection pressure for short apo(a) isoforms in the vegetarian farmers. Intermarriage dilutes this effect but doesn't eliminate it. Of course you could argue the converse, that fishermen are selected to have longer apo(a) isoforms but this seems less likely when you consider what Lp(a) might actually be doing, other than keeping Eric Topol off of the streets.

You have to ask whether the farmers benefit from high Lp(a) levels.


Wednesday, February 03, 2010

Lipoprotein(a) Bantu recap

Just to recap from a few years ago, there are two well studied villages in Tanzania. One has a diet which is subsistence agriculture and so complex starch based, the other is also mostly starch based but supplemented with half a kilo of fish a day. Their daily food intake is approximately described in this Lancet paper:

"Daily energy intake was similar in the two populations (2196 kcal [9·19 MJ] in the fish-diet group vs 2109 kcal [8·82 MJ]). There was no difference in salt intake (4·4 vs 4·0 g daily). In the fish-diet group, 23% of energy intake was from fish with consumption of 300-600 g daily (three to four fish meals per day). Among the vegetarians, most energy was derived from complex carbohydrates (82% compared with 70% in the fish-diet group) such as maize and rice. The proportions of energy derived from protein were 11% and 18%, respectively, and those from fats 7% and 12%. "

The vegetarians are as close to an ideal version of a low fat vegetarian diet as you can get. I don't know much about Ornish's ideas but I'm guessing this comes as close to doing it "correctly" as you can. Exercise too!

The fishermen on the lake shore seem closer to the Kitavans in their macronutrient intake. Still high carb, but not quite up at the 82% of calories mark eaten by the vegetarian farmers...

Using blood pressure as a surrogate for CV health, the fish eaters appear to beat the complex carb group quite convincingly:

So, if you are on an extremly low fat vegetarian diet and your blood pressure isn't doing what it's supposed to do, don't blame yourself. Eat some animals.

Pauletto's group think it is specifically the omega three lipids which have the beneficial effects on BP, but I'm not so sure. If you eat a diet based on fish and seal alone, and virtually zero carbs, you still get this increase in blood pressure with age. Despite the omega three fatty acid intake being very high.

The comparison comes from Paal Røiri's 2005 discussion paper "Eskimo-kostholdets betydning for dødeligheten av hjerteog karsykdommer". We have this table of blood pressure changes with age, rising from just over 100mmHg systolic in childhood to around 150mmHg over the age of 60.

Either extreme does not appear to be ideal. It looks like the upper tolerable limit of unrefined complex carbs seems to be some where above 70% of calories but below 80%. Living down at 4% might not be perfect either. But I digress.

The striking difference between the two Bantu groups is in Lp(a) level. Obviously the vegetarians have higher Lp(a) levels than the fish eaters, as you would expect from their slowly rising blood pressure with age.

You can see that the vegetarians have a median value for Lp(a) of 27mg/dl where as the fishermen have a median of 14mg/dl.

The assumption is that the high Lp(a) is bad and is furring up their arteries and putting up their blood pressure as they age.

Probably genetic. Bad genes mean the farmers have short kringle IV repeats and so their high Lp(a) sets about giving them their just desserts.



Lipoprotein(a) Oxford abstract

This is Watkins, one of the senior authors of this paper on Lp(a) and heart disease:

"This is the most convincing evidence so far that this protein [Lp(a)] is directly part of the pathway that causes heart disease rather than a bystander. If we can target it through treatment, we might expect to lower the risk of disease," coauthor Dr Hugh Watkins (University of Oxford) told heartwire.

OK, we have known for quite a long time that Lp(a) is associated with heart disease. We also know that genetics has a big influence on the level of Lp(a) in the bloodstream. The basic rule is that if you have a small number of kringle IV repeats from your apo(a) genes you will have a large amount of Lp(a) in your blood stream. All of this is old news.

How many repeats of kringle IV you have is a genetic trait. The people at Oxford have used a very sophisticated search technique to locate a couple of changes, each involving a single base pair, in the region of the DNA which makes up the chromosome area coding for apo(a). Either of them affects the number of kringle IV repeats you have in your apo(a). Hence the amount of Lp(a) in your blood.

Finding two single point DNA variations which influence kringle IV repeats is a huge achievement. We now have a piece of genetic mechanism information. What puzzles me is how this is any different from simply counting the number of kringle IV repeats in apo(a) extracted from an ordinary blood sample, which we have been doing for years, without knowing exactly which change in the DNA affected the count.

What is assumed in this paper and the inferences drawn from it is that Lp(a) is intrinsically bad and that genes are the sole control of the level of Lp(a).

Both assumptions are probably wrong. We have to back to Tanzania to see what might actually influence Lp(a) in people with identical apo(a) kringle IV repeats. Hint, its not atrovastatin or ezetimbe.

But one thing came out of the abstract which is actually quite interesting. There is an association between these gene changes, "a small Lp(a) lipoprotein size" and heart disease.

I'm just wondering if a small Lp(a) particle is also a dense Lp(a) particle..........

Back to the Bantu next.


Tuesday, February 02, 2010

Lipoprotein(a) genetics press release

Richard over at Free the Animal discussed this press release, relating to this paper, delivered a day early from the NEJM, by Santa, to cardiologists all over the world.

The paper itself is a piece of trivia (next post) but it is useful personally as it has me back to thinking about Lp(a), arteriosclerosis and genetics. Before I start talking about the abstract there are a couple of laughs/cries to be had from the press release itself.

First is Eric Topol. When you read a technical paper and you see the word "Atkins™" in the discussion, you know you are sitting on a garbage heap. When you see Eric "rent a quote" Topol in a press release you know you are in for a laugh too. Just how funny Eric is requires that you read Malcolm Kendrick's superb essay on Treating to New Targets, the TNT study, which obviously went like a bomb. Or bombed.

You really must read the original essay if you haven't already, but always remember the statistical significance and the biological significance of being dead are two quite different things, unless you sell a statin of course. Eric is owned by statin pushers.

Now, for another giggle, down at the end of the press release there is the section which is as predictable as the musings from dear Eric. This time Kathiresan gets the idiot award:

"Finally," adds Kathiresan, "the genetic data suggest the hypothesis that lowering the plasma Lp(a) lipoprotein level by pharmacologic means will lower the risk of coronary disease."

That sounds perfectly reasonable, if you have Kathiresan's outlook, until you get to the follow on, added by the pharmaceutical company which owns Heartwire's tame author, Lisa Nainggolan:

"Another group of drugs in development that lower Lp(a) are the cholesteryl-ester transfer protein (CETP) inhibitors (such as anacetrapib), which also raises HDL..."

Now if anyone out there thinks that anacetrapib is going to turn out to be remotely different from torcetrapib or the genetic defects giving low CETP activity in Japanese or non-Japanese USA citizens, all of which raise HDL while killing people of heart disease, well you must be on a statin.

Equally, if you think a pure Lp(a) reducing pharmaceutical agent is going to stop heart disease, you're probably wrong. As Dr BG says, Lp(a) has some seriously good effects but you shouldn't run with scissors...

Now, reducing the need for the body to make Lp(a), that's a different approach to lowering it pharmacologically!