Wednesday, November 27, 2013

Protons (30) Uncoupling and metabolic rate in insulin resistance

I wanted to look at insulin resistance, uncoupling and metabolic rate. If we just review the effect of an intravenous bolus of palmitic acid on an anaesthetised rat we can see that the bolus produces a period of increased oxygen consumption, ie an increased metabolic rate, in a then-uncoupled healthy rat. From Curi again:

This next graphic is looking at glucose oxidation, rather than oxygen consumption, from the same paper.

The left hand graph shows what happens under basal metabolism, ie glucose is being taken up by isolated muscle cells without help from insulin facilitated GLUT4 translocation. Fatty acids uncouple, ATP levels fall, there is an increase in glucose oxidation to compensate, ATP levels are corrected. Presumably some of the fatty acids get metabolised too.

On the right is the effect under supra maximal insulin. We have no idea of dose response curve to insulin from this study, it describes either using none or using 10mU/ml. That's mU, not microU! At a concentration of 10mU/ml insulin will overcome any suggestion of insulin resistance, short of a knockout model. Insulin supports glucose oxidation, uncoupling increases this. The set up is not designed to look at insulin resistance as might be related to that uncoupling. But the effects on overall metabolic rate, at 5.6mmol of glucose, are quite clear cut.

What if that elevated FFA level was maintained long term?

As adipocytes become progressively more resistant to the the anti-lipolytic effect of insulin (why is another whole ball game), plasma free fatty acids rise even under levels of insulin which should be suppressing them. Unless these free fatty acids are converted to CoA derivatives they are will uncouple respiration. This should reduce delta psi and increase metabolic rate.

A reduced delta psi will not support reverse electron flow through complex I. The essential insulin induced pulse of superoxide, converted to H2O2, will not occur. There will be fasting insulin resistance. This paper spelled it out. In brief:

The insulin signalling cascade is tonically restrained by a phosphatase which deactivates the insulin/receptor complex (which activates itself by auto-phosphorylation) as soon as the process tries to get started. For insulin to signal you need to have an elevated delta psi which allows a nanomolar pulse of H2O2 to cripple this phosphatase and so allow the insulin/receptor complex to get signalling.

With reduced delta psi this isn't going to happen. We have enhanced insulin resistance of starvation until the time when food arrives.

You can increase delta psi, of course, to get insulin working. Increasing delta psi requires increased electrons in to the respiratory chain through complex I. In the blunted insulin signalling situation of low delta psi we can't use GLUT4 transporters but we can, given high enough plasma glucose levels, get sufficient glucose in to the cell to generate enough of a delta psi to allow the pulse of H2O2 and its downstream effects to occur. So overcome the inability of insulin to signal in the presence of FFAs.

The cost is that an elevated blood glucose level is needed. This is what we look at with the ratio of fasting glucose to fasting insulin, the HOMA score. An elevated HOMA score is a marker of UNCOUPLING at the delta psi level in mitochondria. It should be associated with an increased metabolic rate in proportion to the degree of fatty acid induced uncoupling.

Does this happen in real life? These data are from the Pima but the pattern is generic in insulin resistant states:

Note first that the changes in FFA concentrations are statistically non significant, but that the trend is nicely upwards with insulin levels

The authors comment:

"Alternatively, FFA may contribute to increased RMR via stimulation of mitochondrial uncoupling proteins (UCPs) (53,54). As early as 1976, Himms-Hagen (53) suggested that FFAs stimulate UCP-1. More recently, elevated FFA concentrations were reported to stimulate UCP-3 expression in rats (54)"

Bear in mind that as people progress from NGT through IGT to diabetes there are not only progressive changes in blood lipid and glucose levels but also progressive damage to mitochondria per se, which makes comparing a healthy rat to a metabolically challenged human being slightly dubious.

But the changes make sense.

If we look at the post prandial situation we have, after a carbohydrate containing meal, the combination of chronically elevated FFAs with acutely elevated glucose. There will be a high delta psi as soon as blood glucose rises high enough (supra physiological) to allow non-GLUT4 uptake to elevate delta psi high enough for insulin signalling. At this time point the mitochondria allow insulin to function. We can now translocate GLUT4s to the cell surface and start to lower postprandial hyperglycaemia by pouring metabolites through glycolysis or in to glycogen stores.

At the same time, among insulin's many diverse functions, the activation of free fatty acids to their CoA form increases, with a view to anabolic or storage functions. I would presume a different CoA ligase directs the activated fatty acids to oxidation. There is quite a group of LCFA CoA ligases.

At this point we lose the freedom of free fatty acids, they are ligated to CoA and suddenly become effective inhibitors of uncoupling rather than facilitators. We are now set up for the post prandial state. As soon as we have access to glycolysis combined with beta oxidation we have the possibility to generate marked reverse electron flow through complex I and re-inhibit the action of insulin using much greater generation of H2O2 than is needed for insulin's initial activation.

Metabolically, the mitochondria do this at this time to stop caloric overload in any individual cell, by diverting excess calories to storage in adipocytes. From the NADH:FADH2 ratio, long chain saturated fats do this best. Monounsaturates appear to be designed to allow a normal combination of glucose and fatty acid oxidation and PUFA fail to generate adequate insulin resistance to protect an individual cell (including adipocytes) from an overload of metabolic substrate. In the immediate post prandial state saturated fats can best protect cells from an absorptive excess of metabolites. These fatty acids are already present in supraphysiological levels, whatever has been eaten.

The elevated insulin needed to maintain post prandial normoglycaemia will, if the adipocytes are able to respond at all, divert fat from the circulation and into those adipocytes. This is the simple ability of insulin, and cellular resistance to insulin, to limit metabolic injury when calories are present in excess of immediate needs.


Under inappropriately elevated FFAs there is an initial failure of insulin's action due to depressed delta psi during fasting and a subsequent post-meal failure of insulin's action due to inappropriately elevated fatty acid metabolism through electron transport flavoprotein dehydrogenase's FADH2 when combined with high delta psi.

To me, this looks very much like impaired metabolic flexibility, reduced to the level of delta psi and superoxide. I like it.


Sunday, November 24, 2013

Death by dogma

It's amazing what you can find on the internet when you click on a link. I stumbled over this recently:

Premature aging in mice activates a systemic metabolic response involving autophagy induction

I picked up this gem of a paper from, of all places, a blog with somewhat limited enthusiasm for ketogenic eating. I'll just go through the results section of the paper, giving a staccato summary of each paragraph, because you have to be sure of exactly what a group have found before you consider whether you agree with their conclusions. These mice lack the ability to form prelamin A correctly, instead they form progerin, and they age very rapidly. Here we go with the results section:

A mutation which damages nuclear architecture and causes premature aging also increases autophagy.

The abnormal protein formed (that prelamin A precursor known as progerin) appears to be the cause of premature aging and to be associated with increased autophagy (in this model).

Other models, XPF and CSB/XPA, of rapidly aging mice (both with defective DNA repair processes) do the same thing but without accumulating progerin, especially they increase basal autophagy. So this upregulated autophagy is common to several models of premature ageing, not just the prelamin A model.

mTOR signalling is switched off. Really switched off.

The PI3K-Akt pathway, which usually activates mTOR, is not the explanation.

AMPK is switched on. Really switched on.

Stopping the response to DNA damage (p53 knockout) does not stop enhanced AMKP activity. So we are not looking at extra autophagy to recycle damaged DNA.

Next we get on to glucose. The five hour starved level of glucose is low, around 38% of control value. Insulin is low too.

In the liver things are strange.

Phosphoenolpyruvate carboxykinase and glucose-6-phosphatase are up-regulated, both are important for gluconeogenesis.

Puryvate kinase (a glycolysis regulator) is not up-regulated. So where is the glucose going if it's not going to glycolysis?

Glucose from gluconeogenesis appears to end up in the liver as glycogen granules, without needing glycogen synthase to be up regulated. This glycogen can be accessed if needed.

At the same time fatty acid producing genes are up regulated. Glucose is being converted to fatty acids. Genes associated with fatty acid oxidation are up regulated too. And a fatty liver develops. Very interesting.

Pyruvate dehydrogenase kinase-4, key for switching from glucose to fat burning, is strikingly upregulated. These mice burn fat. They reject glucose. And they die of precocious aging!

All of the "good" markers indicating longevity in many models are fantastic in these mice. The end product is early death.

Metabolically, everything appears to come down to PGC 1-alpha. It's production is very upregulated. This cofactor appears to responsible for the switch from glucose to fat based metabolism.

End of results summary.

This is where the paper stops.

On the basis of these findings a concern expressed in the discussion is that elevated PGC 1-alpha drives autophagy, which is initially adaptive but might become maladaptive when chronically activated. This is a potentially valid concern (there is an autophagy triggered form of cell death distinct from apoptosis and necrosis) but we have to bear in mind that there is zero data to support this specific concern provided in the paper. That's all of the results section summarised up above.

The authors are well aware that the reason for rapid aging is the genetic defect in nuclear architecture formation. This leads, indirectly, to genomic instability which immediately puts this model in to the same category as other premature aging models such as XPF and CSB/XPA, both of which have defects in DNA repair, also as mentioned above.

So why do the cells of these animals go in to a state of AMPK driven, mTOR inhibition dependent, persistent autophagy?

They do this because they have a severe ATP deficit. High levels of AMP per unit ATP drive AMPK.

Why is there insufficient ATP?

Let's have a read at Nick Lane's essay Mitonuclear match: optimizing fitness and fertility over generations drives aging within generations. Here's the quote:

"Oxidative phosphorylation involves the transfer of electrons through a succession of redox centres in respiratory chains, from an electron donor such as NADH, to a terminal acceptor such as oxygen. A slowing of electron transfer means that respiratory complexes become more highly reduced, which increases their reactivity with oxygen, corresponding to a rise in free-radical leak [25]. A slowing of electron transfer is the most likely outcome of any mismatch between mitochondrial and nuclear genomes. The reason relates to the mechanism of electron transfer. If the gap between adjacent redox centres in respiratory chains is increased by just 1 A, electron transfer by quantum tunnelling slows down by an order of magnitude [26] and free-radical leak should rise accordingly. Given that hydrogen bonds and Van der Waal’s forces act over distances in the range of 1–2 A, it is likely that any changes to optimal subunit interactions would disrupt the distance between redox centres by more than 1 A, slowing electron flow and increasing free-radical leak. Thus, a rise in free-radical leak is the predicted outcome of virtually any subunit mismatch, and indeed has been reported [27, 28]"

Question: How well might the electron transport chain function in the mitochondria of a cell which has a permanent defect in its ability to repair nuclear DNA damage?

Answer: Three separate mouse models, engineered to have defective DNA repair, have chronically activated AMPK.

I think these mice are "starving" with full stomachs because they cannot generate ATP. They have progressive un-repaired changes to the nuclear DNA coding for (amongst many things) electron transport chain components which means these proteins simply don't fit the proteins from mtDNA.

Soooooooo. Essentially all of the changes in these mice could be traced back to inadequate ATP generation, with the added bonus of excess ROS generation to further damage the "unrepairable" DNA.

The extended autophagy is a total red herring, sort of.

I picked this paper up as a FB link "liked" by Chris Highcock. The linked blogger is so impressed by the lethal effect of autophagy that he appears to consider that diet-induced autophagy is a "worthless dogma" and avoids it, not liking dogma. That's fine by me, we all have our quirks.

Unfortunately for this autophago-phobic concept, and possibly for the poor chap himself, it turns out that if you take a prelamin A mutant mouse and treat it with rapamycin you will actually promote even MORE autophagy. So, if autophagy is the cause of premature aging, things should get worse. Yes? The actual result is that:

"Here, we report the discovery of rapamycin as a novel inhibitor of progerin [defective prelamin A], which dramatically and selectively decreases protein levels through a mechanism involving autophagic degradation. Rapamycin treatment of progeria cells lowers progerin, as well as wild-type prelamin A levels, and rescues the chromatin phenotype of cultured fibroblasts..."

or just have the title:

Autophagic degradation of farnesylated prelamin A as a therapeutic approach to lamin-linked progeria.

As an approach, it works. Further increasing autophagy rescues prelamin A mutant mouse cells. RIP autophagy as a "cause" of accelerated aging.

I like Chris a lot, he's a really nice chap. Luckily the sort of hill walking he does (still love your Pentland pictures on FB, if you read this Chris) he will be regularly and repeatedly activating AMPK with subsequently increased autophagy. Exercise does this. So does the "misery" of ketogenic eating.

Actively changing your diet to avoid autophagy, based on a progerin model which can actually be rescued by increasing autophagy, just might be a booboo. Imitate with caution.


BTW It's sort of nice looking at technical papers which (superficially) challenge my low carbohydrate biased perspective. Knowing you are in a correct paradigm makes dissecting them a rather relaxed process when you start form a position where the world makes sense. Perhaps this is dogma.

Thursday, November 14, 2013

Ketoacidotic death in lactation!

Laura forwarded me the link to this epic study:

A gestational ketogenic diet alters maternal metabolic status as well as offspring physiological growth and brain structure in the neonatal mouse.

It has a very very clear message. This is the first paragraph of the results section:

"Lactation in KD [ketogenic diet fed] Dams Leads to Fatal Ketoacidosis. In the initial trial phase, 4 KD dams remained with their biological litters post parturition, and during lactation. However, lactation appeared to cause severe physiological strain to all KD dams. Soon after parturition they started exhibiting signs of high stress, as was also apparent by them cannibalizing their litter. Measurements of their blood glucose and ketone revealed elevated values (Glucose: ∼20 mmol/L and Ketone: ∼6 mmol/L), indicating development of ketoacidosis. This ketoacidosis rapidly progressed to a fatal metabolic state, leading to the death of the KD dams within a few days post parturition"

I hope that is clear. No beating about the bush.

These folks are cutting edge. They know how to get impressive results.

Take home message: If you wish to survive lactation do not, UNDER ANY CIRCUMSTANCES, base your diet on CRISCO. Do not do it.

My wife seems to have survived two ketogenic lactations in pretty good shape. Daniel's school is well in to spelling test this term. He is the only kid in his class to have made no mistakes so far. It doesn't look as if we have broken his brain. No Crisco.

Of course Sussman et al blame the ketosis. They keep very quiet about the Crisco:

TD.96355 Ketogenic Diet: "Vegetable Shortening, hydrogenated (Crisco) 586.4g/kg"

Want to die? Always ask for Crisco by name. Crooks.


Tuesday, November 12, 2013

Protons (29) Uncoupling with fatty acids

Trying to work out exactly what is happening in a given functional component of the inner mitochondrial membrane is not the most simple of undertakings. For anyone who wishes to bend their brain a little you could do worse than working through Federenko's paper on the likely structure/function of UCP1 in brown adipose tissue. I think the original pdf was sent to me by Alex, some time ago. Although it is looking at UCP1 I think we can reasonably assume other UCPs work in much the same manner. This set of line doodles summarises the paper:

Here's the explanatory text:

(A) The simplest mechanism of steady H+ IUCP1 induced by LCFAs. UCP1 operates as a symporter that transports one LCFA and one H+ per the transport cycle. First, the LCFA anion binds to UCP1 on the cytosolic side at the bottom of a hypothetical cavity (1). H+ binding to UCP1 occurs only after the LCFA anion binds to UCP1 (1). The H+ and the LCFA are translocated by UCP1 upon conformational change, and H+ is released on the opposite side of the IMM, whereas the LCFA anion stays associated with UCP1 due to the hydrophobic interactions established by its carbon tail (2). The LCFA anion then returns to initiate another H+ translocation cycle (3). Charge is translocated only in step 3 when the LCFA anion returns without the H+.

All nice and simple, compared to how complex it was to work all of this out. In plain English: The (removable) fatty acid flip-flops back and forth, passing a proton each time. UCPs absolutely MUST have access to long chain fatty acids to function. The LCFAs must be free, ie NOT be activated. Sticking a large CoA moiety on to the red dot (which is the carboxylic end of the LCFA) will completely inhibit any activity in UCP1 as there is then no negatively charged binding site on the fatty acid to allow it to ferry protons, repeatedly, down their concentration gradient back in to the mitochondrial matrix.

So fatty acids may just be slightly important to uncoupling. Uncoupling may be a Good Thing or the converse, depending on your point of view or your carbohydrate intake.

This leads us on to Curi's group and their 2006 attempt to look for the effect of FFAs on uncoupling in tissue culture, in isolated mitochondria and even in whole live rats. When you see a line in the methods which specifically states "in Krebs–Ringer bicarbonate buffer containing 5.6 mM glucose" you know you are in for a treat. You have to have read as much cell culture literature as I have to realise how precious a line like this is. All you normally get is the type of medium, probably the brand name and the supplier, but oops, forgot to mention it has, or might have, or might not have, 25mmol of glucose in it. Or some other random amount. Duh. The easy way to tell (when not specified) is that if saturated fats cause apoptosis in cell culture, be certain that glucose is >20mmol.

Anyhoo. Here we go:

This is from first generation harvested and cultured skeletal muscle cells. The fluorescence rises with inner mitochondrial membrane potential, delta psi. Antimycin A blocks ATPase so hyperpolarises the membrane, CCCP is a chemical uncoupler and so reduces it. These are just tests to show the prep works. The caprylic acid is useless (it works well on isolate mitochondria but the cytoplasm probably stops MCTs reaching the mitochondria without adding CoA while they are on their way there). Palmitic acid and linoleic acid uncouple, like a mild version of CCCP.

You can show exactly the same effect in chronically cultured myotubules, just another cellular prep.

This is our first look at isolated mitochondria. This is a graph of the same delta psi sensitive fluorescent dye plotted against time. The graph is upside down because, in real life, delta psi is negative. You add SMM (skeletal muscle mitochondria) to the observation medium. They fire up and delta psi increases. This is the curve dropping downwards from the SMM arrow. In the middle of the 2 minute bar, palmitic acid is added and delta psi decreases (upward flick in the upside down graph) marked by the PA arrow. The bottom line is what happens if you have added guanidine di phosphate, GDP, an inhibitor of UCPs. At the end they also added CCCP to fully uncouple the mitochondrial membrane and show that delta psi collapses.

Note, you can only show the inhibitory effect of GDP when the CoQ couple is oxidised. A reduced CoQ couple stops GDP's inhibitory effect on uncoupling. This gives a mass of contradictory studies in the literature, needless to say. The control of uncoupling as a fascinating area, perhaps another post.

And this graph is looking at the rise in oxygen consumption induced by uncoupling using assorted fatty acids. Again, mitochondria, so caprylic acid works, though never as well as the longer fatty acids. Note the lack of GDP effect, the CoQ couple must be reduced in this prep. BSA (bovine serum albumin) eliminates the uncoupling as it "hoovers-up" all available FFAs.

Does this uncoupling happen in intact animals? Yes. A stack of caveats, but yes. Ratties were bolused with intravenous palmitate or alcohol (As the control??!!?? Alcohol is not inert, even if it doesn't uncouple) and then oxygen consumption was measured:

Palmitic acid uncouples.

Right, that will do for now. The plan is to look at FFAs and insulin resistance next, within this framework.

Peter, chronically uncoupled.