Insulin: Are you hungry? Part 2

Well, over the years I have made the occasional serious blooper on Hyperlipid.

Perhaps the worst of these, to my intense shame, is the acceptance of insulin as a satiety hormone. This is complete bollocks and, thankfully, some deleted-expletive person in obesity research has finally opened my eyes to this. The gift was from Dr Guyenet of course. This is how he convinced me that insulin is not a satiety hormone:


Let's feed some rats standard crapinabag and inject one group with nothing much, one with glargine insulin and another with detemir insulin. But here's the trick. Because we know that hypoglycaemia triggers overeating and the overeating causes weight gain, let's limit the insulin dose to one which does not cause hypoglycaemia... No overeating, active satiety hormone, weight loss...

Because we have been (mis)informed that insulin is a satiety hormone we would expect the insulin-injected rats should eat less, weigh less blah blah blah. What really happens? This does:



I've seen this paper cited as showing insulin can reduce weight gain. By Dr Guyenet no less. Who didn't mention the graphs. Which are core to the paper.

Technically this shows that insulin does bugger all to food intake and fat storage. This is hardly surprising as giving a sub hypoglycaemic dose of insulin will simply attempt to lower blood glucose which will be avoided by reduction of endogenously produced insulin. Total insulin will stay the same. There will be subtleties of peripheral administration vs portal secretion but I guess these are a bit too subtle for this study. There are also fascinating differences in duration of binding of detemir insulin to the insulin receptor vs other insulins. Not surprising as it has a socking great fatty acid tagged on one end but that's another set of stories.

Ah, but what about the effect of detemir insulin on limiting fat gain of rats fed toffeefudgecheescake, aka D12492?

We are talking Fig 3 parts b and e here:



Okay, we're now utterly convinced that insulin limits weight gain. Well, detemir insulin does. Of course glargine insulin doesn't, as Dr Guyenet forgot to mention when citing this paper. It produces a non significant increase in weight over vehicle treated controls.

Just for a giggle, consider changing the grams to kgs on graph e and imagine these rats as humans. Given a group size of 6-8 leading to statistically ns changes in fat mass, would you consider 5kg fat mass gain (a ns change) on glargine, without eating any extra, non significant? Biologically? In dress size? Tee hee.

Now let's look at section e in a little more detail.

The study is very, very carefully set up. The insulin and the D12492 were both started on the same day. It is utterly convincing (to me) that detemir insulin limits weight gain IN THE FIRST 7-10 DAYS of D12492 feeding. From day 10 onwards the fat mass does not change in the control, the detemir or the glargine groups. Not even a trend. A bit like the crapinabag groups demonstrated throughout. In fact, identical to the crapinabag groups. Where's your satiety Guyenet?

Now here's a thought experiment. Let's pretend that all rats were fed D12492 from day 0 to day 10 without injected insulin, so became equally obese with a fat mass of 65 grams, same as the controls on day 10. From day 10 onwards all groups then received their respective insulin or vehicle for four weeks.

Would the fat mass have changed from the 65g starting weight? Of course not, look at the last weeks on graph e. These people are not stupid, though they do like to give that impression.

From this study the follow on question has to be: What is the difference between detemir insulin and either endogenous insulin or glargine insulin during the first 7 days of feeding D12492 to rats?

We all recall from the paper from the Schwartz lab featuring the world's greatest mis-citation expert, that the first few days of sucrose/fat feeding produces an acute inflammatory lesion in the hypothalamus of rats which get fat on D12492. If I had to guess I would suggest detemir insulin limits this injury. How and why cannot be guessed at from this paper but needless to say groups working with gold thioglucose injury have considered what factors influenced hypothalamic injuries. That leads to far out speculation, so I'll limit this post to what Guyenet's citations really do show.

They show that physiological insulin does NOT suppress appetite. Are you surprised? Me neither.

Of course an increased dose of insulin might suppress appetite. But this would need a glucose infusion to maintain life, which would promote DNL in adipocytes and inhibit lipolysis. No hunger while you gain fat. You have to wonder what the point of the above study was, excepting it supports a grant maintaining position and is a self justification for a bizarre mindset.

I also notice Guyenet re-cited this crap. Doesn't he read Hyperlipid????? Giggle... That was a rhetorical question!

Less rhetorical is to ask whether he has actually read the Vanderweele paper at all, particularly Fig 4 of the paper and whether he has reverse engineered said Fig 4 to see the problems with the conclusions of the paper!

Finally he has cited a drug study using an insulin mimetic, not insulin. Well, bully for insulin mimetics. With an insulin mimetic you can mimic lethal doses of insulin without all that inconvenient death. The body does not produce lethal doses of insulin under physiological conditions. If you want to know about physiological doses of insulin within the CNS I can just quote this paper. I feel the authors are being just a teensy weensy bit over the top in their deprecatory attitude to the "centralinsulinisasatietyhormone" brigade. But I can understand why! Here's my fav quote:

"To reduce the likelihood of pharmacological effects of the insulin doses administered, we choose a dose of insulin that is more than 15,000–fold lower than those commonly used for ICV [third ventricle, CSF] insulin infusions"

That's about as rude as you get in Cell Metabolism! You can't use the word "pillock". Drug doses (pharmacologic) of insulin produce drug effects. If you give only physiological dose rates you get physiological effects! Now isn't that amazing?

Oh btw, at physiological levels brain insulin increases peripheral lipogenesis and decreases lipolysis. Did you think insulin would do the opposite through the brain compared to what it does in the periphery?

Duh.

A more believable scenario is that ATP generation within the brain using glucose metabolism, facilitated by insulin in those areas responsible for energy sensing, does occur. But this combination of glucose and insulin will also store fat, as it should, when it occurs post prandially. Which is exactly what excess energy sensing should signal. Insulin without the glucose is pharmacology, unless you suffer from reactive hypoglycaemia.

Peter

BTW I notice over on Woo's blog that there has been some discussion as to whether Dr Guyenet is just dumb or being very deliberately misleading, ie conspiring to mislead. I don't do orchestrated conspiracy theories. I don't really do the financial drive thing either, not for some body who is still as wet behind the ears as Dr Guyenet certainly is. No, for a junior post-doc it has to be:

He has the whole of the knowledge base of the Schwartz lab at his beck and call and the above three citations are the best dross that the Good Doctor can come up with... But still he believes! Stupid.

Happy Solstice

Well Happy Solstice, a day late, to all!

Bit of a bummer yesterday with working a late shift and needing a mega shop for the food for over the next week or so. After my bed time before the car was unloaded!

But I don't think we will run out of cream or butter. Phew!

All the best to all

Peter

Some of us eat a high fat diet

Just a brief post while I try to think of a simple descriptor for The Goof Doctor. Oops, typo there, I think I'll leave it. On to happier subjects:

A birthday card from a few weeks ago to my wife from her best friend Karla. They both understand...



(Card from here http://www.corrinarothwell.co.uk)

However, some people don't need help with butter, except getting it on to the spoon if it has just been self-served from the fridge and is rock solid!

Peter

More yawns on insulin and knockout mice

Ok, I've just got 10 minutes to spare. What shall we play?

Last Saturday was the pathology lab's Christmas Party. In my cracker I was graced by a small and extremely nasty quality plastic magnifying glass which, sadly, I discarded. It might have come in useful in searching for any intellectual understanding in the Good Doctor's latest demolition of the role of insulin in obesity. Actually, the loss of the hand lens is no real problem as I doubt whether even an electron microscope would allow us to find something which is not there.

Let's begin with TNFalpha knock out mice. First thing is that these are based on C57BL/6 mice and, as I have discussed on many occasions, these mice develop an hypothalamic injury which decreases sympathetic outflow to adipocytes and so increases their ability to store fat. You need some technique to make fat mice fat and C57BL/6 are your candidate critters. They are really useful in that they lose fat in to adipocytes, just don't imagine they have anything to do with human obesity. But as a tool, they're great.

Here are the diagrams we need, taken straight from the Good Dr's blog:



All very clear cut.

But what does it mean? Fasting insulin goes up when adipocytes become insulin resistant and so leak FFAs, especially palmitic and stearic acids. Adipocytes become insulin resistant when they become over stuffed with lipid. They don't want any more fat, so they refuse to respond to insulin. Remember that the role of insulin is the storage of dietary fat, plus a little DNL if dietary fat is very low. If adipocytes don't want to respond to insulin they will signal to the rest of the body to do the same. That's working on the basis that adipocytes control whole body insulin resistance based on the ratio of palmitic acid to palmitoleic acid they release, which is in turn dependent on their own insulin sensitivity and SCD1 activity.

Now I have yet to delve in to the mechanics of adipocyte distention induced insulin resistance but I can tell you something here and now for free. It involves the TNFalpha. If you knock out TNFalpha your adipocytes (and pretty much the rest of your body) cannot become insulin resistant. No one gets fat due to insulin failing to act. You become fat due to the action of insulin on adipocytes. When adipocytes refuse to listen to insulin you stop getting fatter, but become hyperglycaemic (unless you eat LC of course!).

So the difference between wild type high fat fed mice (plus sucrose of course, the "cookie dough" they "can't get enough of") and TNFa-/- HFD fed mice is that the wild type mice are sending the signal to the rest of the body that they are fat enough and would like to stop accepting any more calories, fat or glucose. A combination of hypothalamic injury, a sucrose rich diet and a pancreas of steel makes these wild type C57BL/6 mice continue to become obese because they need massive levels of insulin to maintain normoglycaemia. All because insulin resistant adipocytes are signalling that insulin resistance should be produced in insulin controlled cells throughout the body.

Summary: Insulin sensitive adipocytes distend in response to insulin. Insulin resistant adipocytes don't. Hyperglycaemia needs to be corrected. The large dose of insulin needed for this continues to drive obesity by force-enlarging insulin resistant adipocytes. With TNFalpha knocked out the adipocytes become "effortlessly" obese. They, and the rest of the body, will stay insulin sensitive and there will be "easily achieved" normoglycaemia. Healthy obese, but still very obese. Obese due to modest insulin acting on very insulin sensitive adipocytes.


Now, on to iNOS knock out mice. These are really interesting. Again, they are based on C57BL/6 freak mice. A useful model for basic physiology, so long as you have some concept of what is wrong with them. These mice have also had inducible nitric oxide synthetase knocked out ONLY from their muscles. Adipocytes are normal, or as normal as any C57BL/6 mouse can be. Here are the weights etc from the results:



Just look at those gluttonous food intakes and fat gains! Wow.

We can get a basic handle of what is going on from this set of graphs:



Graph a is worthy of the Good Doctor.  HFD fed Nos2-/- mice are not only enormously obese, but they manage it on a fasting insulin which is LOWER (admittedly not significantly so) than that of the WT mice fed crapinabag! Blooooodie hell!

I could spend days discussing graph b, but I'm already half way through my 10  minutes so let's leave it, fascinating though it is...

Graph c. This is the pay dirt. Here they injected insulin in to the various WT and KO mice and tracked the fall in blood glucose levels. The ruler-drawn straight line of black squares is the HFD Nos2-/- mice. These massively obese mice are the MOST insulin sensitive, whole body, of all groups, certainly at the 60 minute mark.

Their adipocytes aren't. These are insulin resistant. All that remains sensitive to insulin are the targeted knock out muscles. On insulin injection these mice simply pour glucose in to their Nos2-/- muscles because the muscles ignore the signals from the insulin resistant adipocytes. Their muscle cells are like a black hole in to which glucose pours. Now, at the risk of quoting the Good Dr yet again: Hypoglycaemia is a very, very potent driver of hunger. Eat, or die.

What's on the menu? Ah, a bowl of lard sweetened with sugar. Death is not an option, let's eat the lard to get the sugar. Good idea, stayin' alive. Now, we've used the sugar, what shall we do with the lard? Ah, as a brain injured C57BL/6 mouse we have adipocytes which are rather more willing to accept fat than a genuine wild type mouse might have. Bye bye fat, in to the adipocytes you go! But at least death due to hypoglycaemia is avoided, all be it at the cost of greater obesity!

If you have been following the protons thread you can see that linoleic acid, ie corn oil, is a mild mimic of TNFa-/- mice and of Nos2-/- mice. It's obesogenic while preserving insulin sensitivity. Your cardiologist made you fat.

Now, in my last 30 seconds: Why are adipocyte insulin receptor knock out (FIRKO) mice healthy and slim? Well you could ask the Good Dr for some sort of platitude, but, hey, that would be stupid.

No. Adipocytes control whole body insulin sensitivity. They see no insulin if they have had their insulin receptors knocked out. They sport minimal (zero?) GLUT4s on their surface. What fat they contain has been accumulated without the assistance of insulin. I think it is reasonable to assume they have some GLUT1s on their surface. Any glucose taken up will be available for lipid synthesis but, without insulin's action, there will be no insulin induced SCD1 desaturase activity. So palmitate it is and, in the absence of insulin's action, this will be freely released and should signal whole body insulin resistance. But it doesn't. It does exactly the same as the palmitic acid does in SCD1 knockout mice. Peroxisiomes. FIRKO mice eat more, weigh less and (probably) generate more heat than WT mice do. They are insulin sensitive everywhere except for their adipocytes. They behave exactly as SCD1-/- mice do but get there by a rather indirect route. Excess palmitate is burned in peroxisomes and the C8 end product in mitochondria.

Life is, in the end, logical. Having the correct tools helps. It must be awful to be wallowing in the mire of the Reward hypothesis.

Peter

Are humans just multicellular yeasts?

Again via the Brent Kearney link on Stan's site.

http://phys.org/news/2012-11-scientists-key-events-early-cellular.html

discussing this paper

http://www.nature.com/nature/journal/vaop/ncurrent/full/nature11654.html

I especially like the last sentence of these two paragraphs:

These results are just the newest chapter in a "decade-old interest," according to Gottschling. He and his colleagues have made several landmark discoveries in the past 10 years, including finding that aging yeast cells exhibit the same genomic instability seen in human cancer cells and proving that mitochondrial dysfunction causes that instability. Gottschling's team also has developed innovative tools to leverage the power of yeast as a model organism, including a technique called the Mother Enrichment Program that makes experiments more efficient by enabling researchers to generate large populations of aging yeast cells.

"It's worth using yeast to study complex things like aging because a lot of person-years of research have gone into understanding the fundamentals. The genetic and cell-biology tools available for studying yeast are unparalleled," Gottschling said. "Having the proper tools is like having new glasses; you can see things you never could before, and once you start to see new things, you can dissect them to understand how they work."

The full paper is a Letter to Nature. That means it is technically very dense, but very interesting.

Caloric restriction to a yeast is, essentially, glucose restriction. In organisms with a circulation, an hormone system and a background FFA supply, caloric restriction at the cellular level (where it matters) is achieved by resisting the insulin mediated facilitation of glucose access, ie glucose restriction. Using superoxide.

Palmitate please, with just a very little glucose.

Peter

Interesting times on Mars

http://boingboing.net/2012/11/20/big-news-from-mars-coming-soon.html

I picked this up through the link to Brent Kearney's blog on Stan's site. Olivine and water under a CO2 atmosphere, all present on Mars, are the precursor to biochemistry. I just wonder... Interesting times.

Peter

Protons: Physiological insulin resistance, addendum two

George put up the link to this paper, which allows you to tease information out about omega-3 PUFA as bulk calories vs lard as bulk calories. We are not talking about essential supplies of essential lipids here. We are looking at serious bulk calorie supply. This is quite, quite different.

Aside: The basic conclusion that FO is protective against endotoxin shock is fascinating but may be restricted to C57BL/6 mice. Pity everyone uses them. But it's interesting, and on file, never the less. End aside.

Here's the composition of the diets:



Fairly typical research diets, a little more sucrose than I would like but, well, everyone does it.

What about weight gains? Here they are:



Although body weights, at all time point, are not significantly different between groups this is just due to the initial group weights being different. If you look at weight gain rather than absolute weights, the lard group gains more weight than the fish oil group.

The insulin and glucose levels do support the idea that insulin sensitivity is controlled by the degree of unsaturation of the bulk lipid in the diet, ie PUFA diets increase insulin sensitivity. But there is no excess weight gain in the fish oil group. Why not?

C57BL/6 mice suffer an injury to their hypothalamus on exposure to a saturated fat based diet, especially if combined with sucrose. Omega-6 PUFA do not seem do this and I doubt omega-3 PUFA do either. I considered this back here.

So we are comparing obesity in an un-injured group carrying omega-3 enhanced insulin sensitive adipocytes versus a hypothalamic injured group carrying adipocytes which are partially resistant to insulin due to dietary palmitic acid and partially sensitive to insulin due to decreased sympathetic outflow from the hypothalamus to adipocytes. C57BL/6 mice are very special in their response to saturated fats.

This is a knotty problem to try and untangle. This paper helps a lot.

I just want to look at two of the control groups, both of which are C57BL/6 mice, both of which are exposed to a high palmtic acid diet and so both will have an hypothalamic injuy.

So we can have, among many, two groups of C57BL/6 mice, one fed a high fat diet to make it fat and the next fed a high fat diet to make it fat, but then add in a significant dose of omega-3 PUFA. Just to add some insult to injury. The first group gets a whammy. The second group gets a double whammy. Want to see the graph? Ok, ok, here it is:



First, strain your eyes to follow the open triangles. This is the high fat only control group. These are C57BL/6 freak mice with a brain injury triggered by palmitic acid. They have limited weight gain but, as they store palmitate without DNL, hence without desaturase activation, hence without palmitoleate generation, they develop metabolic syndrome. Visceral fat, fatty liver. Of course the group didn't measure either insulin or glucose (they are in obesity drug development), but these mice have metabolic syndrome and have lost the ability to get any fatter. They are in trouble. They don't actually weigh any more than un-injured C57BL/6 mice fed traditional crapinabag.

Now look at the open circles of ever increasing obesity. Fatties or fatties? This is what happens when you add fish oil to the diet of a palmitic acid injured C57BL/6 freak mouse. Impressive waistlines huh? Of course we don't get the insulin or glucose levels here either, but these mice do not have metabolic syndrome. They maintain the insulin sensitivity of their adipocytes, especially peripherally, and continue to become obese with sustained metabolic health. They will stay healthy until their adipocyte distension induced insulin resistance eventually over rides the insulin sensitising effects of the bulk fish oil.

We have a pair of models. Skinny-fat and obese-but-metabolically-slim. Both are explicable by looking at the basic effects of bulk lipid supply from the diet acting on the insulin signalling system within mitochondria.

Summary: These are palmitic acid injured C57BL/6 freak mice which have the added insult of having their adipocytes rendered extra insulin sensitive by the F:N ratio of a significant percentage of the fatty acids in the fish oil of their experimental diet. This postpones metabolic syndrome until they have become fat enough to develop it.

The F:N ratio concept delivers again.

Peter

BTW no one knows the omega-6 content of the fish oil is in this second study! The discussion mentions that there is zero omega-6 in the basic high fat diet, which has no added fish oil. Imagine running a Rimonabant study when you don't know the omega-6 content of the (high fat) diets. But this becomes irrelevant if you look at the basic metabolism at the molecular level. Either family of low F:N ratio PUFA will delay metabolic syndrome, while ever they assist weight gain. And you have to remember that C57BL/6 mice break by eating butter.

Protons: Physiological insulin resistance addendum

Edward sent me this paper. I think I did already have a copy on my hard drive but you can't really start to make head nor tail of what is really going on until you have a handle on F:N ratio and the physiological role for superoxide. I'd completely forgotten about the paper.

 For those people who think humans are in some way different from rats, here's Fig 1 from the paper on humans eating either a saturated fat ketogenic diet or a polyunsaturated fat ketogenic diet, just for 5 days:



Look at the glucose, lowest in the PUFA group, look at the ketones, highest in the PUFA group, look at the insulin sensitivity, waaaay higher in the PUFA group. Rat or human, makes no odds. PUFA fail to generate superoxide in mitochondria. Is this good or bad?

The whole point of a ketogenic diet (epilepsy excepted) is to induce starvation-appropriate physiological insulin resistance. What is the point of setting up a ketogenic diet which does not have the ability to convert from running on glucose to running on fat?

Aside: Why might anyone want to run their metabolism on FFAs? Superoxide. I want more mitochondria to supply spare ETC capacity, to minimise the sort of levels of free radicals which wipe out mitochondria when the pressure is on. Physiological superoxide signals for mitochondrial biogenesis, without all of that tedious exercise to do the same job on a mixed diet. End aside.

Now it is just possible to argue that chronically reduced insulin may render adipocytes immune to the insulin sensitising effects of PUFA. Maybe. The obese mice of the next Protons post are on a mixed PUFA-carb diet to assist their "ballooning" experience. High insulin plus insulin hypersensitive adipocytes gives obesity. Perhaps the combination of low insulin with distendable adipocytes is OK if the insulin levels are low enough. Volunteers queue over there please.

But we are still in a situation where FFAs are high, yet glucose can still enter cells. What does this do to cellular ATP status? The whole point of insulin resistance is to avoid cellular caloric overload, full stop. If you load liver cells with PUFA and add in glucose the end result, by whatever mechanism, is cirrhosis. Perhaps other tissues fare better. Am I about to try it?

No thank you.

Peter

And for a real giggle you can see exactly what sort of utter crap made up the high PUFA diet, just read Table 1.  "Imitation" if the first word of the first "food"!

Protons: Physiological insulin resistance

Woody emailed me a pdf of this venerable paper. I like it. Even though it was generated in-house by Proctor and Gamble!

The core findings are that if you eat a diet with around 60% of calories from fat, then you end up lighter if that diet is based on safflower oil (omega 6 PUFA) and heavier if it's based on lard. Here's the table:



Well, there you go!

Safflower oil is more ketogenic than lard. It also allows greater fat loss during 72 hours of starvation, 10.3g of fat loss from safflower fed rats vs 3.9g from lard fed animals, here we are:



Mmmmmmmm, linoleic acid, the elixir of life and Weight Watchers' friend. And I thought it was obesogenic!

So what does the small print say? Lots!

First thing to make my ears prick up is that during the whole study the rats were only allowed to eat for 2 hours a day, in a single solid block. That means they were force fasted for 22 hours a day. We've noted from Table 3 that safflower oil allows greater weight loss under fasting conditions. So is it any surprise that the safflower rats ended up at 195g vs the 222g of the lard fed rats after eight weeks of this intermittent fasting? I'll come back to why in a moment.

In contrast most modern papers show that linoleic acid, the primary component of safflower oil, is grossly and transgenerationally obesogenic...

But that's because there is a difference between ad lib feeding and 22h per day forced fasting!

The difference is in the glucose levels. Look at these graphs from Fig 1, especially the glucose levels in the middle top graph:



Now, far be it for me to put words in to the mouth of a lab rat, but which group of rats is the hungriest? Which group becomes most hypoglycaemic perhaps? Even Dr You-are-confused-man-Guyenet seems to have, in an aberrant moment of lucidity, a glimmering of perception that hypoglycaemia makes you hungry. Or should I say drives eating behaviour? Did I ever even mention gluttony? If you are force fasted for 22h you can't overeat. You can't even eat to your metabolic needs without accessing significant amounts of stored fat.

Under full starvation a rat lives off of its fat. If linoleic acid is what is being released from the adipocytes under fasting conditions it provides significantly less FADH2 relative to NADH in to the electron transport chain of all fat burning cells than the mix of fatty acids from the adipocytes of lard fed rats. ie there is less physiological insulin resistance. This failure means you fail to keep glucose levels normal during starvation. Let's rub that in: Failure to develop physiological insulin resistance during starvation results in hypoglycaemia and hunger.

Exactly the same will happen in any soy oil fed USA citizen. The end result will still be the failure to develop the essential physiological insulin resistance which is needed to keep blood glucose normal during fasting.

When the average soy oil fattened American is asleep they HAVE to, finally, stop snacking on carbohydrate crap, which is the only way they can maintain a decent blood glucose level. At this time blood glucose falls, simply because their muscles stay insulin sensitive and glucose falls in to them. The brain will not accept hypoglycaemia. Some time, in the middle of the night, there has to be a Refrigerator Raid.

And, OMG, they eat calories! And calories count! Did I ever mention gluttony? Or, perhaps, is the Refrigerator Raid simple physiology?

The 22h daily fasted rats have a locked refrigerator. However hungry they feel due to hypoglycaemia, they are not getting any extra food. But why is there extra weight loss under starvation? Insulin was tricky to measure from a rat in the 1970s, but I know that the safflower loaded then starved rats had the lowest insulin as they are both insulin sensitive and hypoglycaemic. Low insulin = more lipolysis = more ketones and more weight loss. Logical.

Now let's go a step further. Blood glucose is low. It's low because the F:N ratio of linoleic acid is low and that's what is being released from adipocytes. This is metabolism. Individual cell by individual cell, it's a metabolic phenomenon. Picked up by the brain as hypoglycaemia.

Hypoglycaemia must drive food intake to avoid cerebral catastrophe. Glucose is not a neurotransmitter per se. The hypoglycaemia has to be converted to a neurotransmitter based message to affect behaviour. Oddly enough, a derivative of linoleic acid is a neurotransmitter. Linoleic acid is the parent molecule of the endocannabinoids. Even the cleanest nosed obesity researcher has heard of the munchies. Wouldn't it be funny if eating linoleic acid messed with your blood glucose level and this hypoglycaemia triggered the production of endocanabinoids from, you guessed, the parent molecule (of the hypoglycaemia!), linoleic acid within the brain? The same molecule which triggers the hypoglycaemia, making the hunger essential, in the first place.

You could view it as linoleic acid ingestion is a signal that hypoglycaemia is in the offing. If random evolution was looking for a substance to "choose" as a neurotransmitter to deal with hypoglycaemia, it might just be simplest to modify the basic lipid which is most prone to trigger low blood glucose throughout the body in the first place... You could then end up thinking that the endocannabinoids "just happen" to control appetite as a primary function. Of course, once the control system is in place you are all set to mess with it by demonising saturated fat and pushing corn and soy oils.

Well, I find the concept hysterical. But then I find most brain centred obesity research pretty amusing.

Peter

Protons: SCD1 and the bomb

Our daughter has taken to watching DVDs. When we are not being tortured by Postman Pat (Special Delivery Serrrrrrviccccce, you know the tune) or the Mickey Mouse Clubhousssse we do at least get some amusement when she requests that blistering LC comedy "A Matter of Loaf and Death", by Nick Park.

Just how funny you can make a story about an obese cereal killer (no typo, the subtitles specify cereal killer, I said Park is funny!), murdering bakers as revenge for her obesity ("Are you ballooning?") has to be seen to be appreciated. It's a lot more amusing than Postman Pat.

One of the funniest scenes is where Gromit cannot get rid of Paella's bomb. It's a direct tribute to the 1960s Batman scene where "Some days you just can't get rid of a bomb".

You know, with the ducks

and the nuns. Park has kitten-enhanced the nuns



and included Yorkshire as the preferred site for bomb disposal. The Wars of the Roses are, apparently, over but not forgotten.

This post is about how physiology uses SCD1 to dispose of the metabolic bomb of hyperglycaemia in the presence elevated levels of palmitic acid.

It was pancreatic beta cells in culture which produced this picture:



I love this group because not only do they tell you in the methods section EXACTLY what glucose concentration they used in culture (5mmol/l vs 11-25mmol/l) without making you go back through three layers of references (to bury the 25mmol/l most groups use, but never discuss), but they also describe 11mmol/l as hyperglycaemia. That is, pathology.

This is Figure 4 from the same paper showing markers of apoptosis, superb:




Note the increase from palmitate to stearate. Note the complete protective effect of oleic acid and very modest toxic effect of linoleic acid. Aside: Note also the complete and total protection provided by limiting glucose to 5mmol/l, with any fatty acid, at any concentration. You have adipocytes leaking FFAs? Your best hope of keeping a functional pancreas is to limit your glucose to 5mmol/l. How? Answers on a postage stamp to...

It's also worth noting that stearic acid had to be reduced from the original 0.4mmol/l to 0.25mmol/l because at the higher concentration with high glucose they found exactly the same thing as Dave Lister did in Red Dwarf when Holly brought him out of stasis. Everyone is dead Dave. Everyone. Is. Dead. Dave. You have to U-tube the clip. Stearic acid at 0.4mmol/l with glucose at 25mmol/l, in cell culture, is utterly lethal to beta cells. If you pharmacologically block apoptosis the cells simply undergo the rather messy collapse of necrosis. This is a non survival insult.

Okay, okay, here's the clip:



The group went on to do quite intersting things with blockade of acyl-CoA synthetase and also with inhibition of fatty acid oxidation, which leads to all sorts of other threads which are, in part, where I have been wandering for the last few weeks. Far too much for this post.

So let's look at SCD1 knockout mice which have been rendered obese by also knocking out their leptin gene. Here we have rapid onset obesity due to adipocyte fat storage, free fatty acid leakage due to adipocyte insulin resistance and a complete inability to place a double bond in to palmitic acid or stearic acid. They have elevated FFAs and these are almost all saturated. This paper describes the study. It has to be noted that to obtain the FFA levels you have to reverse engineer Fig4 part A:



A ruler and calculator gives FFAs for the normal ob/ob mice as 0.32mmol/l and for the SCD1 k/o ob/ob mice as 0.56mmol/l. Any group which makes you reverse engineer in this way to get something as simple as FFA levels is, in my book, highly suspect. Does anyone think that 0.32mmol/l is quite low? Despite the greater obesity. Partly due to maintained insulin sensitivity in adipocytes (that's why they distend) while ever de novo lipogenesis produces palmitoleate using SCD1 and partly due to the higher levels of insulin production (normal beta cell mass) working on those insulin sensitive adipocytes... These mice are still in a slightly privileged position, metabolically, as they have yet to become obese enough for their SD1 equipped adipocytes to become seriously insulin resistant, they are still only six weeks old.

And here is the % of types of FFAs.



The column on the left is the one which represents about 0.32mmol/l of total FFAs and the column on the right is around 0.56mmol/l, as above. Glucose varies but fasting levels can be as high a 700mg/dl. So what happens to beta cells?

They divide up in to two types. The health ones and the dying ones.

The basic finding is that young ob/ob mice need either oleic or palmitoleic acid to maintain a functional beta cell mass. Exposure to high levels of glucose combined with palmitic and/or stearic acids induces apoptosis plus some necrosis in beta cells. Most non pancreatic tissues in the SCD1 knock out mice appear to be able to upregulate beta oxidation, especially in peroxisomes, of fatty acids which minimises both obesity and insulin resistance.

The beta cells of the pancreas do not appear to have this luxury.

They need to lower that F:N ratio with palmitoleate or oleate, otherwise they are left holding the bomb.

Peter

Dalcetrapib fails as it should

Still no time to post but this one liner is another gem. Dalcetrapib, another (yawn) HDL raising drug has bombed. Who is surprised?

Funnily enough Dr Nissen (Rentaquote-you-have-a-statin-deficiency) feels dacletrapib failed because it was too weedy to do any good, only a 30% increase in HDL.

Evacetrapib (Nissen's gravy train) and anacetrapib will REALLY work because they double HDL and slash LDL.

No they won't. They will do exactly what torcetrapib did as they are as potent as torcetrapib was. Body count will rise. Dalceptrapib killed no one as it is not potent enough to do so!

Peter

Skirting around leptin

I've been wanting to post about Wallace and Gromit, Batman and ob/ob SCD1 k/o mice for weeks now and it keeps not happening. Before we go there, just a word or two about leptin and weight gain. You can't work through anything relating to ob/ob mice (+/- SCD1 k/o) without having to, finally, sit down and read something about leptin. Or at least ob/ob mice...

To me the core question to ask is whether ob/ob mice are gaining weight because they have a brain disorder giving overeating or an adipocyte disorder storing calories. As always, there is an infinite supply of data suggesting a brain disorder, there's no denying leptin does things in the brain. The question is: Are ob/ob mice in caloric excess as they gain weight? ie Do they eat too much so gain weight or do they primarily lose calories in to their adipocytes and so have to eat more to just meet metabolic needs?

We have various rodent models of obesity which have the common feature of reducing the sympathetic nervous system drive to adipocytes, so failing to oppose insulin's lipogenic action. These animals gain fat even if you calorie restrict them. I'm thinking about hypothalamic ice-picks, various VMH neurotoxins or unprotected free radical generation. But the common thread is the loss of sympathetic nervous system driven lipolysis, facilitating insulin driven fat storage.

When confronted with the overwhelming literature on leptin it's hard to know where to start, especially when people are not asking the sorts of question which interest me, looking at the data from my very particular perspective.

I accept that ob/ob mice get fat. So too do brain injured rats and mice. Is there a common mechanism here? The smoking gun would be a period of enhanced insulin sensitivity in adipocytes, due to decreased sympathetic tone, which allows both fat gain and the preservation of insulin sensitivity in the early weeks, until adipocyte distension induced insulin resistance kicks in for the swelling adipocytes and systemic insulin resistance develops.

Of course the easy part, with absolute leptin deficiency, is asking whether leptin increases hypothalamic sympathetic nervous system outflow. That took about 30 seconds on pubmed and this was the 6th or 7th hit.

Leptin increases sympathetic drive. I think it's reasonable to conclude leptin deficiency does the converse and reduces sympathetic drive from the hypothalamus. So I'll take that as a yes. Don't forget I'm biased.

Sooooooo. Does leptin deficiency defend insulin sensitivity during rapid weight gain? As it should if the mechanism is enhanced lipid storage. And weight gain in young ob/ob mice is, well, rapid. To say the least. There will only be a very narrow window to pick up preserved insulin sensitivity before adipocyte insulin resistance and hyperinsulinaemia set in. By which time researchers have a usable model of established obesity.

I'm interested in what goes on before the model becomes "usable". We know that by rewarding volunteers to overeat we can spike their insulin levels massively within three days, probably faster. Does this happen with ob/ob mice as they over eat?

I've been working through a whole stack of papers on ob/ob mice, FFA levels, insulin levels, ketogenic diets... All the usual stuff. I ended up in a review by Lindström, giving this little gem of a quote:

"The muscle insulin resistance is not observed in very young [ob/ob] mice, but develops after 3–4 weeks [131]"

The abstract of ref 131 supports the concept of preserved insulin sensitivity, looking at muscle rather than adipocyte insulin resistance. The papers on palmitoleate as a lipokine suggest that muscle insulin resistance is controlled by adipocyte insulin resistance, via SCD1 and palmitoleate. The paper is rather nice because it is looking at ob/ob mice as ob/ob mice, not as some completely inappropriate model for hyperleptinaemic obese humans. It was, after all, 1980 when it was published.

So to summarise:

Danish volunteers who are paid to overeat spike insulin from 35pmol/l to 74pmol/l in just three days. Mice with zero leptin overeat massively, but do not show the same insulin spike. The insulin spike signifies insulin resistance, that characteristic antioxidant defence response to an excess of calories in the metabolic milieu. This does not happen with the early overeating phase of ob/ob mice. They are in metabolic caloric deficit, which they make up by eating enough to remain vaguely functional.

Absolute leptin deficiency appears to be a very harsh driver of fat storage. Losing this many calories makes you hungry. I guess some bit of the brain is involved in converting this state of actual calorie deficit in to a feeling of hunger, but that's not what interests me nearly as much as what is happening at the adipocyte level of calorie storage.

Peter

Now we can get on to Wallace and Gromit and knocking out SCD1 in ob/ob mice.

Protons: Love your superoxide (outside your brain)

Two off topic posts in a day! How come? I had the weirdest morning today. A three hour consulting session with only six appointments, all straight forward. Bloody hell, was I lucky for a Saturday! Can't blog at work so I had a quick browse to see what Nick Lane has been up to recently. He has a cracking article up (as a pdf) on heteroplasmy which rewards careful reading in its own right, but look at these two "throw away" quotes.

First on ROS, good old superoxide from reverse electron transport:

"ROS leak seems to optimize ATP synthesis by stimulating mitochondrial biogenesis (mtDNA copy number), an interpretation supported by the fact that antioxidants lower not only ROS leak but also mtDNA copy number and ATP synthesis. ROS leak, in effect, signals low capacity relative to demand, stimulating compensatory mitochondrial biogenesis".

How do we minimise mitochondrial biogenesis? By running metabolism on glucose of course, but don't forget the lack of superoxide generation when oxidising PUFA. But who needs mitochondria when you can lower LDL levels by swilling corn oil? Ah cardiology, you have a lot to answer for. Executive summary: Want mitochondria? Burn PALIMTATE.

And on cerebral metabolism:

"In the brain, where further mtDNA biogenesis is limited, neurons would then become compromised whenever energy demands were high, possibly causing acute cognitive and behavioral abnormalities".

The brain neurons are running on lactate under crapinabag conditions. We considered this before. No fatty acids. No glycerol 3 phosphate. No FADH2 input to the CoQ couple. No free radicals. No signal for mitochondrial biogenesis. No mitochondrial biogenesis. You could substitute ketones and maybe get a few mitochondria back if you were canny, but most medics aren't canny. What happens to the lactate supply for neurons when hyperglycaemia drops on to chronically elevated FFAs and triggers apoptosis in glial cells? I think we can attach various labels, depending on which neuronal cell types die first. Alzheimers seems a nice name for the commonest scenario.

Lovely pair of quotes. Glad I got the browse time. But don't ignore the heteroplasmy discussion at the core of the article, it's good stuff.

Peter

Look AHEAD trial stopped

Eat less, move more, have your heart attack on time!

With apologies for lack of any attention to the blog recently (which may be set to continue for some time) but this snippet just had to get passed on. This link from Karl:

http://www.theheart.org/article/1458351.do

More info here

http://www.nih.gov/news/health/oct2012/niddk-19.htm

And the glowing anticipation of success here from the planning stage:

https://www.lookaheadtrial.org/public/home.cfm

Pubmed gives a series of genuine success stories from the early days on all sorts of parameters. But the cardiovascular end points show how utterly useless these interventions are long term.

However the massive omission, from the quick look I've managed, is of any intention to report the all cause mortality. It seems very likely to me that more people died in the intervention group than in the usual care group, but p was > 0.05.

Call me a cynic, but I think they stopped the trial because they could see where that p number was heading. Has anyone seen a body count from anywhere in the trial?

Also, what might the outcome have been if the intervention group had been repeatedly bullied, harassed and indoctrinated to maintain a normoglycaemic, low grade ketogenic diet for 13.5 years? Say to an HbA1c of around 5%?

Ha ha ha bloody ha.

Peter

EDIT: Have started on the SCD1 k/o ob/ob mice. The thread WILL continue.

Protons: Zero fat

A bit speculative here, read with caution!

How do we lower free fatty acids? Obviously, with nicotinic acid. What does this do to insulin secretion in response to a glucose challenge? I'll just work through this figure from the same paper which gave us the insulinotropic effects of various FFAs a couple of posts ago.



Section A is very simple, it just shows that they succeeded in clamping glucose at just over 200mg/dl, about 12mmol/l, ie just in to supraphysiological levels.

Section B shows FFA levels, which they manipulated very carefully. All rats started at about 0.6mmol/l. Nicotinic acid lowered FFA levels to 0.1mmol/l. These are the black squares. Two other intervention groups were included. The white triangles had their lipolysis shut down using nicotinic acid but then had FFAs clamped back up again using a soyabean oil infusion (mostly omega 6 PUFA) and the black triangle group had an infusion of lard based lipids (a mix of lipids but with a significant palmitic acid content) to restore and hold FFAs at about 0.8mmol/l.

The nicotinic acid group, with FFAs of 0.1mmol/l, cannot secrete insulin in response to glucose. Flat line at the bottom of graph C.

The open squares are the control group. These rats show the normal response to an hyperglycaemic clamp. They reduce FFAs in response to the inhibition of lipolysis from secreted insulin, down to 0.2mmol/l. Insulin inhibits lipolysis. But the reduced FFAs also reduce insulin secretion. There is a balance struck with only a modest rise in insulin, sustained throughout the clamp. You can see this in section C, open squares.

The two lipid infused groups have clamped glucose and clamped FFAs. They secrete insulin in proportion to the amount of palmitate in the lipid infusion. A bit extra over control if you use low F:N ratio omega 6 PUFA, a ton extra when you include some palmitate. Section D is simply a summary of this.

Step by step at the mitochondrial level: The lower fatty acid supply results in decrease reduction of the CoQ couple in beta cells. This reduces the reverse electron transport and associated superoxide triggered by glucose as it feeds NADH in to complex I, so limits insulin secretion. You can virtually ablate the insulin response to glucose by eliminating beta cell fatty acid supply.

Now, nicotinic acid is one way of reducing FFAs. There have to be other, perhaps more physiological, methods. Maybe we could use insulin per se? From food perhaps? Let's try eating around 40g of carbohydrate and look at the Spanish study graph again. Insulin rises from 50pmol/l to 75pmol/l. This is enough to reduce FFAs from 0.5mmol/l to just over 0.1mmol/l. Look at the FFAs, especially the circles between 120 and 300 minutes:



Now (again, sorry!) look carefully at the insulin levels after the small carb load, bottom circles.



By 180 minutes insulin is actually lower than fasting, and FFAs are still well below fasting levels too. The rat model appears to hold in humans, not what the study was looking at, and a small effect. But I think the effect is real.

How about scaling this up to a massive dose of potato induced insulin and limiting dietary fat? Severely limiting dietary fat. And never mind pussy footing around at 40g of mixed carbs and protein. There is a limit to how low FFAs can be driven, and it seems safe to assume that a baked potato or three might just inhibit lipolysis maximally and keep it that low for rather a long time. But if you deprive beta cells of free fatty acids you blunt their ability to secrete insulin. Very, very high carbohydrate diets really ought to be able to inhibit lipolysis to the point where the knock on effect is the inhibition of insulin secretion, provided you don't supply exogenous fat. Look at the nicotinic acid treated rats...

Once you get FFA levels low enough to inhibit insulin secretion you will start to move in to the sort of territory where insulin secretion might be blunted enough to allow hyperglycaemia. But the feedback effect of reduced insulin levels is also the re commencement of lipolysis. This will restore enough FFAs to maintain functional insulin secretion and so avoid potential hyperglycaemia, which the body tries to avoid. Of course you have to throw in the increased insulin sensitivity of muscles deprived of exogenously supplied FFAs too.

So is it possible to eat an ad lib, calorie unrestricted diet based on near pure carbohydrate and lose weight? Working from the premise that lowered insulin is a pre requisite for hunger free weight loss, as I always do, the answer is possibly yes. We all remember Chris Voight on his all potato diet (plus 20ml of olive oil, low in palmitate, per day) who lost a great deal of weight over a few weeks, the rate of weight loss accelerating as the weeks progressed? I had a think about it here, well before I had any inkling as to what might be happening in the electron transport chain.

We need to know what the interaction of insulin and FFAs was during this particular n=1 self experiment, and we don't. The rats suggest to me that insulin levels were initially raised post prandially and FFAs were not then available from peripheral adipocytes. Assuming the fall in lipolysis persisted in to the post-absorptive period (the primary function of insulin, especially at low levels, is the inhibition of lipolysis rather than facilitation of glucose diffusion, we've all read Zierler and Rabinowitz) we have a method for limiting insulin secretion late post prandially using reduced free fatty acid levels.

As an aside I personally wonder it might be the ectopic lipid supplies typically found in muscle, liver and visceral adipocytes which might still be available for metabolism by the tissues when exogenous supplies are shut down.  It reminds me of how metformin most likely depletes ectopic lipid to improve insulin sensitivity, despite having complex I inhibition as its primary action. You need lipid from somewhere. So reducing FFA supply by inhibiting systemic lipolysis may well be a route to lower fasting insulin levels. Especially if you are not far in to metabolic syndrome.

Once ectopic lipid becomes depleted then lipolysis would accelerate in peripheral adipocytes as systemic insulin resistance falls and fasting insulin levels too, which might be what was reported as progressively increasing weight loss by Chris Voight. Insulin levels would be low, especially during fasting, and appetite low at the same time due to hypoinsulinaemia facilitated lipolysis, much as appetite is low under LC induced hypoinsulinaemic eating. There is more than one way to skin a.... Oops let's not complete that phrase!

What would happen to a healthy person under these conditions, long term, is anyone's guess. Chis Voight gave up after a few weeks when weight loss became alarmingly rapid. But we know from the crucial study by the vegan apologist Barnard that, for diabetic people at least, that a long term, whole food, low sucrose and low fat diet is a complete disaster, once the initial weight loss ceases.

This is playing with fire (possibly near literally, at the mitochondrial level) if you are a diabetic. Please don't go there.

But the physiology of weight loss on ultra low fat diets is basically comprehensible, especially once you look at lipids and superoxide at the ETC level, and what the body needs to function effectively. Running your metabolism on pure glucose would induce, theoretically, an infinite glucose sensitivity and low fasting insulin. If we do reductio ad absurdum you would end up with no fat stores and experience death from hypoglycaemia if you ever depleted your glycogen stores. Mitochondria like (saturated) fatty acids. Fatty acids keep them in control.

I think someone in obesity research used Chris Voight's experience to support some cock and bull story about food reward and a set point of body fat. We can wait for the recant on that one, if you could care less about it. The biochemistry is, as always, the fascinating stuff.

Peter

Never forget, never forgive

For recreational purposes only, from here:



How toxic is palmitate at any concentration from zero to 0.4mmol/l if glucose is held at 5mmol/l? It's not.

How toxic is glucose at 25mmol/l in the presence of increasing palmitate? Glucose at this concentration becomes progressively more toxic within increasing but still physiological levels of palmitate.

Glucose at 25mmol/l is UTTERLY non physiological.

There is potentially a huge amount to discuss from this paper, one day, maybe. Even though they never delve in to the ETC, which they should do. But anyway, someone commented that papers using 5mmol/l glucose in the culture medium were rare. This one is a hen's tooth.

Back to thread next.

Peter

Protons: The pancreas

We've seen the concept of superoxide being used to produce insulin resistance as a means of limiting (glucose derived) energy input in to cells which really don't want it. Superoxide appears to be the primary marker of energy excess at the cellular level.

We know from isolated mitochondrial preparations that superoxide is physiologically produced by reverse electron transport through complex I and is driven, gently, by succinic acid alone working through complex II. Far more is produced when the NADH level is high as well as having a reduced CoQ couple through FADH2 input, be that from complex II or from fatty acid oxidation products. Macroscopically fat and glucose together should produce enough superoxide to show as cellular insulin resistance, rejecting glucose from the cell, while allowing continued fatty acid oxidation. That's simple and logical.

But if you are building an energy sensor, it would be a bit dumb to restrict access to the very energy molecules which you are trying to look at to judge overall energy status, especially when energy status is high: You need to decide when to store calories...

The beta cells appear to use both fatty acids and glucose to generate superoxide, but instead of signaling beta cell insulin resistance, they signal insulin secretion. Several lines of evidence fit in with this.

You can get succinic acid itself directly in to beta cells by providing it as a methyl or ethyl ester. As a metabolic fuel source this acts as a near pure complex II substrate, pushing electrons in to the ETC through the FADH2 of succinate dehydrogenase to reduce the CoQ couple and set the scene for reverse electron transport and superoxide production, especially when NADH from glucose metabolism rises. In a commonly used model of functional beta cells, succinic acid methyl ester is a marked insulin secretion potentiator, especially at higher glucose concentrations. Glucose supplies NADH, succinate supplies FADH2, they clash at the CoQ couple and the generation of superoxide signals that there is a ton of energy available. Better store it. Better secrete insulin.

Succinic acid methyl ester drives complex II. This drives insulin secretion in response to glucose. But it's a drug. There is nothing physiological about this drug. So shall we go a little more physiological?

To recap from previous posts: Superoxide generation is directly proportional to the ratio of FADH2 generated to the amount of NADH generated for a given substrate, the F:N ratio.

Here's a nice graph of insulin secretion stimulated in response to 12.5mmol glucose on a background of assorted free fatty acids from an isolated pancreas preparation:



If you can't be bothered to work out the F:N ratios (shame on you), here they are added to the graph:



Please excuse the C8 value; as we all know, MCTs are shunted directly to the liver via the portal vein. They do not seem to feature too prominently in pancreatic superoxide generation and insulin secretion. It would take a ton of reading to see why and how they are handled differently to longer chain fatty acids. For the time being let's stay looking at C16 and longer as these make a much tidier story...

So, for the four longer fatty acids tested, the amount of insulin secreted is remarkably closely associated with the F:N ratio of the fatty acid available.

Does this work in people?

Of course it does. Remember the Spanish study? I lo0ked at it in some detail here.

In particular look at this graph:



From the top downwards we have butter, high palmitic acid seed oil, refined olive oil and a mix of fish and vegetable oils as the white triangles. It is very clear that the insulin secretion here is in direct proportion to the saturation and length of the fatty acids in the meal, in an intact group of volunteers..

Aside: Obviously, there is a glaring error in the graph. All of the curves except the control use 800kcal of total food, of which 40g is carbohydrate/protein and the bulk is fat. The graph is missing a group where 800kcal was supplied as pure carbohydrate. We can all imagine where this much bulk glucose would have put the insulin curve, needless to say there is absolutely no way it would fit on to the presented graph. We would need a much taller vertical axis, which would show the mixed meals in their true context!

But the principle, that insulin secretion at a given level of glucose is elevated in direct proportion to the F:N ratio of the background fat, holds perfectly well in this carefully contrived human study.

BTW, lucky for me they didn't include a coconut oil group!

The obvious conclusion from this finding is that to lower insulin maximally we should, taking as given that replacing carbohydrate with fat is the biggest step by far, all go for vegetable oil with some fish oil. Not butter. But in the original post on the Spanish paper I went on to discuss what appeared to be happening to the lipid from the meal. It could stay in the bloodstream and be used for metabolism, as the butter did, or it could be cleared rapidly in to adipocytes allowing metabolism to return to being glucose based. At the cost of expanding the adipocyte stores of fat.

The high PUFA meal really was rapidly stored as fat in adipocytes. The F:N based explanation is because we are supplying a low F:N ratio fat and so not generating insulin resistance phyiologically; we are allowing lipid easily in to adipocytes because the lipid does not generate adipocyte insulin resistance. We are going back to glucose metabolism as rapidly as possible. PUFA facillitates fat storage and glucose based metabolism. All is fine until you can't get any fatter. Butter limits fat storage and runs metabolism of palmitic and stearic acids. Those high PUFA-fed mice generate obesity when fed their high PUFA diets from pre-conception onwards:



In the butter group there is some excess insulin. Does this matter if no one (cellularly) is listening to it?

I next want to look at the flip side, the reduction of supply of free fatty acids to the pancreas. This was done in the same paper. You can certainly do this in intact rats (and humans if you so wish). Then we might get back to the fat mice.

I think that had better be another post as this one is getting overly long and it's light enough to let the chickens out.

Peter

Protons: Linoleic acid in the hypothalamus

Hi all, just getting my head above water now that we have two or three locums at work to cover some of the (rather difficult) gaps in the rota!



Before we look at the fat mouse study which wins the prize for most miserly hoarding of data, I just wanted to put up a brief post, based on that paper, about breaking your hypothalamus with a high fat diet. Just to re emphasis: This is NOT what happens to a human after 7 days on a high fat diet.

Remember Schwartz's rats? Put them on a high fat diet and this happens to food intake:



Note the very sudden and dramatic spike in the intake of food, shown by the red line which I've added to emphasise the abrupt change from baseline chow consumption. We can ignore the red oval for this post. What happens in the VMH neurons of these rats?

This is what happens:



The dark brown staining cells on the right are dying, the rats have been eating "cookie dough", which they "can't get enough of", for seven days. The nice healthy cells in the left hand photomicrograph are from rats on crapinabag. The basic idea appears to be that feeding rats a bit of fat and sugar makes them eat so much, starting in just one day, that by seven days their VMH is killed by over indulgence. You eat too much, you kill your brain. Simple. This is, of course, absolute bollocks.

At the risk of repetition, we can produce exactly the same lesions in the VMH with MSG or gold thioglucose (or an ice pick if you must be crude and don't want nice pictures). This injury results in fat gain which must be compensated for by overeating. Rats will gain weight more slowly if they are on low fat diets than on high fat diets because of the effects of increased de novo lipogenesis which I've discussed in previous posts.

Want pretty pictures from GTG injured rats? Here's some random immuno from a random paper, there's a lot of it around, only black and white though:



Gold thioglucose on the right, arrow marks the injury area. And I just noticed the same pics, also in black and white, from fat injured rats from elsewhere in the Schwartz paper (mirror imaged compared to the GTG pics, random choice of which side of brain got sectioned!) after just a week on their high fat diet:



So we can produce the pretty black stains of dying cells with gold thioglucose (or MSG if we looked at neonatal immuno) but this injury preceeds the loss of calories in to adipocytes and subsequent "hyperphagia". THE INJURY COMES FIRST.

Let's really look at the bizarre idea that non-forced "overeating" causes subsequent damages your VMH. This is how it works for over eating by a gold thioglucose injected rat, no yummie high fat diet needed: It simply decides to over eat crapinabag because this has suddenly and randomly become delicious and so it becomes obese. We all know overeating CAUSES the VMH injury in fat fed rodents. So how do GTG injured rats get the injury first and over eat secondarily? Gold thioglucose obese rodents might SEEM to have a chemical lesion causing obesity but clearly they get fat first, travel back in time (squeezing in to a time machine as obese chrononaughts) and retrospectively force the researchers to give them the injection of GTG to obtain the lesion in their VMH which they are going to produce in the future by eating too much crapinabag. Got that? You've all watched Back to the Future. I watched parts I and II but never managed part III. It's simple time travel. Ditto MSG and ice-pick (ouch!) obese rodents. Self inflicted injuries using time travel.

Or we could abandon such stupidity and say that high fat diets injure the VMH first and this injury increases fat storage by decreasing sympathetic tone to adipocytes, as it does.

And I suspect it's superoxide, generated by a high F:N ratio (classically derived from palmitic acid at an F:N ratio of 0.47) in POMC neurons, which probably does the damage. You all know POMC neurons, the ones in the VMH with both gluokinase to sense (via metablism) glucose and CD36 to monitor FFA status (via metabolism again). No lactate for the energy status sensing neurons of the VMH...

So the question is, as always, what happens to the VMH of a C57BL/6 mouse (bred to get fat on a high fat diet) when put on a high fat diet which does NOT generate superoxide in POMC neurons? You can do this.

No one has done the necessary immuno staining under these conditions to get the pretty pictures of dying (or non dying) cells, as far as I know. But it's easy to look at the weight gains, which are a reasonable surrogate for POMC injury. Schwartz again using rats:



Not the most lucid graph, but it gives the basic idea. The control weight gains on the left are comparable to the weight gains shown for day 14.

Now, here is what happens if you take a C57BL/6 mouse and put it on to 35% of calories from fat if you keep the F:N ratio of that fat well below 0.47, using omega 6 PUFA with an F:N ratio of 0.42, as the primary source of fat:



Ignore the top two lines (for now) and look at the weight gain of the mice in the bottom two lines. One group weaned on to crapinabag, the other weaned on to 35% of calories from fat, but a fat with a low F:N ratio. There is zero, zilch, nil difference in weight gain over three weeks. There is no excess weight because there is no VMH injury. No one generates significant superoxide from a low F:N ratio fat like linoleic acid. That appears to include the POMC neurons of C57BL/6 mice.

C57BL/6 mice (and Long Evans rats) are specifically bred to get fat on palmitic acid (sometimes plus fructose) based diets. They fail to deal with the absolutely normal levels of superoxide produced in POMC neurons in the VMH which are crucial to energy status sensing. They do not have the luxury of developing insulin resistance as their job is to monitor both glucose and fatty acid levels. They are not allowed to run on lactate with an F:N ratio of 0.2 the way much of the brain does. They take whatever plasma gives them and do their best to cope with it. Or, in the case of rodents bred to become fat on high fat diets, not cope with it.

Before we go looking at the linoleic acid paper a bit more carefully I think it's worth trying to look at energy sensing rather more peripherally than the POMC neurons of the VMH. Then we can come back to the fat mice and try to think about what's going on using the meagre data available. Because it's quite interesting.

Peter

Guess the weight of the mouse competition

Here's the Brownie points quiz.

What is the weight of either mouse?

If you find the answer, in this study, please let we know!

Peter

BTW you are allowed/obliged to scour results, discussion and all supplementary data. Please.

Protons: SCD1 knockout mice

Let's peep inside an adipocyte belonging to a mouse which has had its stearoyl-CoA desaturase gene deleted.

It's busy making lipid, being an adipocyte. Two carbons, four carbons, six, eight, ten, twelve, fourteen and hey, there's the sixteen for palmitic acid. Now, how much glucose and insulin is there around? Ah, lots. Need to signal this with palmitoleate. In goes the double bond to prove it... Oops. No SCD1. Hmmmm. We now have a ton of palmitic acid and no chance to convert any of it to palmitoleic acid. Tricky.

Is the adipocyte going to become insulin resistant? Unless it runs on glucose and never uses any lipid this seems likely. Will the adipocyte stay small? It should do, it's insulin resistant, so won't store fat. Should it export saturated fat as FFAs? Yes. Should we have a slim but insulin resistant mouse? On chow it should become hyperinsulinaemic. Well, you might expect so.

But that's not what happens. The mice stay slim alright, but have excellent insulin sensitivity. Like really, really good insulin sensitivity. You can even feed them on toffee fudge cheesecake and they stay fairly slim and very insulin sensitive.

The SCD1 deleted mice also eat more despite being slimmer than WT mice when on chow, ie they are in CICO-denial:

"On average, the SCD1−/− mice consumed 25% more food than wild-type mice (4.1 g/day vs. 5.6 g/day; n = 9, P < 0.05). Nonetheless, they were leaner and accumulated less fat in their adipose tissue"

Huh. Bloody gym sneaks again. Even while they are asleep:

"The SCD1−/− mice exhibited consistently higher rates of oxygen consumption (had higher metabolic rates) than their wild-type littermates throughout the day and night (Fig. 3A). After adjusting for allometric scaling and gender, the effect of the knockout allele was highly significant (P = 0.00019, multiple ANOVA, Fig. 3B)."

These animals have a hugely increased metabolic rate. The brown adipose tissue "looks normal". That's not the answer.

They are also ketogenic during fasting (daytime is sleep time but they don't really go to the gym while they are asleep). Fasting BHB was 4.4mg/dl. For those watching their ketone meter at home, eat your heart out. They do this even when living on toffee fudge cheesecake.

OK. Utter basics:

What is the F:N ratio within the mitochondria of these mice? Is it:

a) <0.45
b) <0.45
c) <0.45
d) <0.45
e) <0.45
f) Huh????
g) >0.48 (trick answer, don't choose this one!)

The mice are insulin sensitive. They do not have undiluted palmitic acid oxidation going on in their mitochondria. This would produce a ton of superoxide and severe insulin resistance. We know that their mitochondrial F:N ratio must be low. Their metabolism is ketogenic. What fats produce ketones on a high carbohydrate diet? Those MCTs from coconuts and breast milk do. Where do you get C8 caprylic acid from if you are an SCD1 knockout mouse on a low fat diet?

From your peroxisiomes.

Mice with palmitic acid on tap and no ability to lower the F:N ratio by desaturation simply oxidise it in peroxisiomes, FADH2 free, to C8 which is ketogenic, has a low F:N ratio and they produce a lot of heat in the process.

In the words of the paper:

"Northern blot analysis also supports changes in fatty acid oxidation and lipid biosynthesis. Probes for acyl–CoA oxidase (ACO), very long chain acyl–CoA dehydrogenase (VLCAD), and carnitine palmitoyltransferase-1 (CPT-1) indicate increases in β-oxidation"

My emphasis. VLCAD is the main one in peroxisomes, as well as being present in mitochondria. The authors do not come up with any comprehensive explanation of what is going on. The F:N ratio delivers.

I think I mentioned some time ago the explanatory ability of the F:N ratio is awesome. It just goes on.



I was going to leave it there, back to work next week so blogging will diminish, but here is some idle rambling which followed on from this post.

Now here's the question. If some guy like me set out to maintain the lowest practical insulin level (which will minimise SCD1 activation) and bases his diet on the very longest chain, most fully saturated fat practical, would you expect me to activate my peroxisomes? Might the result be that I might stay slim and be cold tolerant?

When we moved in to our current house I unpacked the scales after they had spent nearly a year in a box provided by Pickfords. I was 63.8kg after a year of not checking anything, down by about a kilo from Glasgow. But I was getting a great deal of hill walking in Scotland and probably had more muscle. I forgot about the scales for another year but dug them out recently. Down to 62.8kgs. I eat a huge amount of palmitic acid. I generate enough superoxide to maintain the needed physiological insulin resistance to eat LCHF and I suspect I might have quite active peroxisomes.

I still run a dawn phenomemon FBG of around 5.5mmol/l, if I get up early enough to check it beforehand it's about 4.3mmol/l, once 3.9mmol/l. Random BG through the day vary from 3.3mmol/l after a half day of walking to and from the beach while the car was being MOTed to 6ishmmol/l post prandial if I had parsnip chips (yum) with my high fat beef burgers. Yes, I pour the cooking fat over the chips. A big carb load will get me above 7.0mmol/l easily but only for a couple of hours. I try not to do this too often.

Posting-wise I have no idea what time will allow next week but beta cell failure in SCD1 knockout ob/ob-ve mice tells us interesting things about cells which have minimal antioxidant defences and are deprived of palmitoleic and oleic acids.

Peter