Insulin resistance (12) Lipopolysaccharide

I've mentioned this paper by Nowotni from Shulman's lab on several occasions

Mechanisms Underlying the Onset of Oral Lipid–Induced Skeletal Muscle Insulin Resistance in Humans

It is a great resource on many levels, provided you completely ignore any of the paper's conclusions.

The most entertaining quote is this one from the control intervention period which involved the iv injection of a very small dose of lipopolysaccharide to induce insulin resistance. As the authors say:

"Only endotoxin administration led to mild flulike symptoms with a maximum after 3 h, while other interventions had no side effects."

What they didn't report was that the volunteers, after the endotoxin administration, came up with any spontaneous comment along the lines of:

"I could murder a steak now"

Almost certainly because it didn't happen.

I conclude that either a few hours of insulin resistance doesn't make you hungry or Nowotni is keeping shtum about the hunger generating effects of insulin resistance.

I struggle to identify inflammatory insulin resistance as a trigger for obesity. Even if it is chronically present. The last serious viral infection I had was probably in the late 1980s. If I recall correctly I lost 4kg in a weekend and only part of that was fluid loss. I wasn't hungry. I also went through a period of chronic inflammatory problems while I lived through mycoplasmal arthritis in the months soon after qualifying. I got through an awful lot of paracetamol before the medics got the diagnosis and subsequently I got through an awful lot of oxytetracycline at high dose to get it sorted out. No weight gain from inflammation though.

The endotoxin in the study caused the same degree of insulin resistance as ~900kcal of oral or iv soybean oil. But it provided no caloric input, so this insulin resistance is apparently not an obvious satiety signal.

And it's not pathology either.

Let's actually do some thinking about the advantages of insulin resistance during acute infection.

It's probably most illustrative to initially consider the insulin resistance of fasting. While this is often the situation during infection it is also a normal feature of simply not eating overnight. It is essential for survival, especially if fasting has to be extended.

The primary advantage of this aspect of insulin resistance is to spare glucose for the brain. While our brain can run perfectly well on ketones, these take time to develop and the default setting in modern agricultural society is to run your brain on glucose and leave daily ketosis to oddballs like myself and a few others.

So the ROS generation from FFA oxidation under fasting spares glucose for the brain, useful because the brain preferentially utilises glucose in modern times. Muscles and related tissues are very happy to run on fat and so resist insulin.

There is another part of physiology which is highly dependent on adequate calories from glucose for its optimal function and that is the immune system.

The case is laid out here

Metabolic syndrome and robustness tradeoffs

and here

Insulin resistance and inflammation in an evolutionary perspective: the contribution of cytokine genotype/phenotype to thriftiness

Obviously both sets of authors have zero understanding of the causes of metabolic syndrome but have successfully linked the insulin resistance of starvation to maintaining brain function and the insulin resistance of infection to maintaining optimal immune function.

The second paper also throws in the insulin resistance of exercise, which I hadn't thought of, but makes sense.

Please, please, please bear in mind that NONE of these folks understand the cause of metabolic syndrome from the Protons perspective and so have no idea of why the previously adaptive trait of insulin resistance should now appear to be maladaptive. Modern life is suggested to allow "pathological" insulin resistance through a deficiency of mammoth-killing exercise and/or via a deficiency of starvation on bad hunting days. So they say.

Stop sniggering at the back!

So what *should* happen in infection is that adipocytes and hepatocytes release both FFAs and glucose to optimise immune function. Flooding the immune system with caloric substrate is pro-survival in infection or trauma.

At the same time all non immune essential cells should limit their caloric uptake, especially that of glucose, to their needs and no more. Predominantly from FFAs.

During the early evolution of organisms which have a recognisable immune system and separate organ systems for caloric storage all that would be needed is for evidence of tissue damage to trigger caloric substrate release to facilitate effective pathogen attack and/or tissue repair.

The simplest signal for this would be fragments of damaged tissue, ie 4-HNE, or pieces of damaged pathogen, ie LPS (fragments of the cell wall from gram negative bacteria). Or similar substances. A basic system would have the energy storage tissues "seeing" evidence of  distant damage with a resultant generation of ROS in response. Those organisms which failed to do this failed to survive.

Because metabolic substrate is being flooded in to the circulation, use of the glucose component by tissues which can preferentially use fatty acids with their ROS generation should be limited. The ROS from FFA oxidation will automatically control caloric ingress to optimal for each insulin sensitive cell. The excess substrate, especially glucose to drive NADPH derived ROS generation through NOX enzymes, is then available to attack the pathogen and improve survival.

I suppose the next step to improve would be for immune cells involved in active duty at the site of damage to send a more reliable surrogate messenger and clearly tell the adipocytes/hepatocytes to respond decisively and release substrate without waiting to see how the untidy message of fragments of damage debris might work. So cytokines were developed as long distance messengers. I would hazard a guess that they might be derived from stabilised sections of proteins used in defence of integrity. Of course the same signalling molecules could be used to tell non involved insulin sensitive cells that a flood of emergency calories is on its way and better get resisting insulin now, in anticipation.

The next move is to have a sub population of immune cells actually sitting among adipocytes and hepatocytes, listening for cytokines, ready to act as local amplifiers of incoming signals. Their function would be to signal, via local cytokines and ROS, to adipocytes/hepatocytes to resist insulin, spill caloric substrate and gear up for battle.

There is nothing wrong with doing any of this. We can, with this perspective, regard the insulin resistance induced by inflammatory cytokines as the warning that a caloric flood is on its way, that that caloric flood is destined for the immune system, and it is now essential to enter excess calorie mode, ie resist insulin, in anticipation. I think of these higher level signals as a form of looking in to the future.

Although it might not be obvious initially, the sudden release of metabolic substrate in infection provides a surplus of energy availability at the cellular level despite anorexia. Exactly as fasting does. Running your metabolism on FFAs fundamentally generates ROS without needing insulin facilitated glucose uptake. Fasting, once metabolism is running of FFAs and ketones, is an energy replete situation at both the individual cell level and the central nervous system level. For modern glucose oxidisers the first three days might be a bit rough.

What *is* wrong is for adipocytes (and, presumably, hepatocytes) to accumulate so much lipid that they die by catastrophic failure secondary to over distension. At which point all hell breaks lose within the local tissue to form highly inflammatory crown-like structures and to signal over long distances via cytokines that a disaster is happening. The pathologically augmented inflammatory signalling in adipose tissues is not causative of obesity. It is a response to individual adipocyte obesity. This is most obvious in the most insulin sensitive adipocytes, ie visceral/omental adipocytes. Which become "inflammed" first because they distend soonest, ie because they are the *most* insulin sensitive.

The modern maladaptive inflammatory problems of obesity all stem from the failure to limit insulin mediated distention of adipocytes (and probably hepatocytes).

It's not gluttony and sloth.

It's not failure to find food for a week or failure to attack a mammoth.

Linoleic acid is, of course, the primary problem which breaks the system.

Peter

PS There is a follow on in this paper

Effect of a prolonged low-dose lipopolysaccharide infusion on feed intake and metabolism in heifers

which suggests that in modern organisms that the release of FFAs and glucose under endotoxin exposure is now mediated by active lipolysis via sympathetic drive and lipolytic hormones. Not really surprising. Eukaryote evolution has had a couple of billion years to produce functional heifers.

Insulin resistance (11) NAC might cure your diabetes

The next scenario is to look at the effect of NAC on metabolic syndrome with or without marked hyperinsulinaemia. We're looking at this section of the curve roughed out in this previous post, mostly for completeness before getting on to other aspects.

So this is our region of interest:

which we can re label as showing combined base load FFA oxidation ROS plus insulin mediated ROS generation via elevated delta psi, which limits further caloric ingress but without damage by virtue of the ROS signal being limited to an hypothetical equivalent to 0.3mM H2O2:

Now let's add in the situation where, despite ROS being high and insulin signalling being blunted, calories continue to enter the cell. Where might this caloric substrate be coming from?

Life is simplest if we think about an hyperinsulinaemic normoglycaemic clamp. If we just infuse insulin to give a high post prandial level, classically ~1000pmol/l and then infuse glucose to maintain 5mmol/l we can supply almost all of an individual cell's needs using glucose once steady state has been achieved.

Ideally the 1000pmol/l of insulin will suppress plasma FFAs to around 100μmol/l. These 100μmol/l will supply a base load of ROS from which glucose will top up energy supply until the ROS from FFA oxidation plus the ROS from glucose oxidation reaches the equivalent of H2O2 at 0.3mM on the graph, which resist further ingress of glucose via fully physiological insulin resistance.

In metabolic syndrome the same thing happens but the fixed 1000pmol/l of insulin cannot suppress the FFA levels to the preferred 100μmol/l because distended adipocytes are releasing FFAs independent of insulin while, under clamp conditions, insulin is fixed at ~1000pmol/l. Now the base load ROS generation from FFA oxidation will be higher. The amount of glucose required under fixed insulin levels will be lower and our subject will be labelled as insulin resistant, ie "unwell" or pathologically insulin resistant.

But individual cells are merely responding to having a higher base load oxidation of FFAs. There is nothing wrong with them. Drop the FFAs with acipimox and, given a little time, insulin sensitivity would return to normal. Except the subject's adipocytes would get even bigger and release extra "pathological" FFAs as soon as the drug wears off.

The only pathology is with adipocyte diameter. This is caused by inadequate FADH2 supplied by linoleic acid in the Protons hypothesis.

The situation in DMT2 is even worse in vivo, rather than under exogenous infused insulin, because there is a failure to produce adequate insulin. This leads a failure of suppression of FFA release from adipocytes, so an higher base load of ROS is invariably generated. There is a concurrent inability to suppress glucose release from hepatocytes, so the small amount of glucose needed to "top up" ROS generation to that 0.3mM H2O2 equivalent is minimal. The higher the plasma glucose level rises the less insulin is needed to allow a cell's energetic needs to be met.

If a cell is having all of its metabolic needs met by a combination of FFAs and "insulin facilitation free" glucose ingress then the 0.3mM ROS signal to resist insulin could easily be exceeded. Insulin resistance will become profound.

Let's imagine the level gets up to the equivalent of 1.0mM H2O2 in cell culture. This will fully resist insulin's action but fail to limit the caloric ingress, because this is happening independent of insulin facilitation.

ROS generation will be marked and profoundly damaging.

At this stage the nature of the fatty acids generating elevated base load ROS is immaterial. In fact saturated fats, with their optimal reverse electron transfer abilities, will be *worse* than PUFA in augmenting ROS generation. The downstream problems of ROS damage (4-HNE and its related compounds) will obviously be worse in an high PUFA environment which will be present because the primary pathology (adipocyte distention) is only possible in an high PUFA environment (pax fully hydrogenated coconut oil or butter induced obesity, which are real and different).

In to this soup of problems let's drop enough NAC to drop ROS levels by ~1.0mM of H2O2 equivalents, indiscriminately:

There may still be 1.0mM of H2O2 equivalent of ROS being produced, but they disappear in to the NAC sump.

Suddenly we are working to the left of the peak of insulin's action and, lo and behold, we have restored insulin sensitivity and everything looks much better. Bear in mind that it's not!

So. Can NAC "cure" your diabetes?

Of course. And not.

Can it "cause" diabetes, if you are currently metabolically healthy?

Of course. And not.

It's all quite straight forward, given the Protons perspective.

Peter

Insulin resistance (10) NAC might make you diabetic

Here is the next paper

Complexity of NAC Action as an Antidiabetic Agent: Opposing Effects of Oxidative and Reductive Stress on Insulin Secretion and Insulin Signaling

Amongst the many, many things that this group did there is this section from Fig 9. They injected a set of mice with a dose of insulin typical for an insulin tolerance test. Fifteen minutes later they euthanised them and extracted muscle tissue for assessment of insulin signalling as represented by the phosphorylation of AKT. Half the mice had been on NAC in their drinking water, half hadn't. All were eating chow.

Here are the Western Blots

and here is the darkness of the blots converted to numerical form:

Both groups of mice received exactly the same dose of insulin, 0.75iu/kg, high enough to seriously kick the insulin signalling cascade but still with a fairly low incidence of serious hypoglycaemia. The intention is to get a near maximal insulin response *without* lethal hypoglycaemia.

If we go to the doodles from the last post this is what that dose is trying to achieve:

The 0.75iu/kg will produce *exactly* the same ROS signal in all of the mice, on or off of NAC. But in those mice on NAC a large proportion of the signalling ROS are simply mopped up by the NAC. So instead of getting the phosphorylation of AKT we expect we get much less:

The dose of insulin is identical,  the ROS signal is identical but the NAC non specifically reduces those ROS which carry the onward signal.

The insulin effect is mediated through ROS, NAC limits it.

At low levels on insulin the small ROS signal would be completely wiped out.

At higher levels of exposure insulin will actually signal and should increase its downstream target activation, but not as effectively as it should. We're looking at the three points circled in red here if Fig 9B :

which are responsible for a significant increase in the AUC calculated for glucose.

The initial rise in glucose is unaffected by NAC, which might be surprising if you don't view it from the Protons perspective.

We are sightly stuck because we don't know the insulin exposure at times 15 and 30 minutes, which may not be comparable between the two groups.

We can guesstimate the insulin levels in the control group using the data from this paper for after a six hour fast. It looks like this, added on to the above. No units included for the insulin, it's the pattern I'm interested in:

Now, an OGTT is designed to elicit a maximal insulin response within the physiological range. A maximal physiological response will allow peak utilisation of glucose and, at some time point when combined with elevated plasma glucose, will produce insulin-induced insulin resistance as soon as the the mitochondrial delta psi rises high enough.

The signal to resist is, obviously, ROS.

A significant proportion of which disappears in to the soup of NAC so insulin-induced insulin resistance is blunted, so glucose continues to be disposed of at peak insulin. We're thinking about events somewhere within the time points circled in red:

If you choose your dose of NAC very carefully you can generate the above curves. This is NOT normality, this is balancing pharmacology against differing aspects of physiology.

How well this happens depends on your dose of NAC and probably on the metabolic state of your mice. Not all control groups are as free of metabolic disease as you might like.

Hence the results in JustPeachy's nice paper:

The Antioxidant N-Acetylcysteine Does Not Improve Glucose Tolerance or β-Cell Function in Type 2 Diabetes

I can't find a paper where NAC actually precipitated DMT2 but I had a vague recall that vitamin B3, a favourite of the orthomolecular practitioners, could occasionally precipitate glucose intolerance.

It turns out you need a massive study to pick up the small effect demonstrated in the second hour of the mouse OGTT above using the effective antioxidant B3 rather than NAC but I did find this bias-confirming meta-analysis (you know, one, two, skip a few, 99, 100):

Niacin therapy and the risk of new-onset diabetes: a meta-analysis of randomised controlled trials

Personally I avoid antioxidants. A few decent ROS is my preferred approach.

Peter