Things keep getting in the way of the next post, which is roughed out but needs tidying. I've also been meaning to post on the Somogyi overswing effect in diabetes treatment for some time, so here is a minor diversion down that route, mostly because it's very illuminating.
The Somogyi effect is quite common in those unstable diabetic dogs which tend to get shunted in my direction at work. Any clinician will recognise the effect. A dog is given 8iu of lente insulin at 7am with a meal of utter crap (ultra low fat Chappie usually). Blood glucose spikes to 22mmol/l by 9am from the carb load then falls progressively until about mid day, as the slow onset insulin struggles with the hyperglycaemia. It looks like there is a nadir at about 6mmol/l around mid day. The 1pm reading is unexpectedly high at 30mmol/l. No food, no behavioural signs. Just sudden hyperglycaemia. The Somogyi overswing. This fades slowly to around 15mmol/l by the next meal time at 7pm. The cycle repeats.
Management (if you can't change anything else) is to reduce the dose rate of insulin, which stops that sudden surge in blood glucose at 1pm. Somogyi attributed the effect to a reflex release of glucose from the liver to prevent catastrophic hypoglycaemia in response to insulin overdose. Most clinicians seem to still think in these terms.
Logical but incorrect. The advent of continuous glucose meters has pretty well disposed of the "hidden hypo" explanation and people are now looking at the effects of hyperinsulinaemia per se. The sudden rise in blood glucose appears to be associated with progressively rising or even peak levels of insulin in the blood.
Let's have a think about what is happening. Under insulin deficiency conditions glucose can still be used as a fuel, in a somewhat unregulated manner, using concentration driven supply through GLUT1, independent of insulin. Hyperglycaemia is essential for this. It's not good. The poorly regulated glucose supply generates free radicals in the electron transport chain. Superoxide is the main one and this appears to be the key to causing insulin resistance. Hyperglycaemia causes insulin resistance. This is not controversial, as far as I am aware.
As the insulin kicks in we have a period where glucose levels are falling so GLUT1 transport is decreasing and insulin regulated GLUT4 transport is increasing. Initially excess glucose above cellular needs diverts to glycogen and the respiratory chain is kept happy by insulin. As insulin levels continue to rise above physiological needs we end up with a situation where insulin is putting a ton of GLUT4s out, far more than are needed. This happens because we have inadvertently injected a supraphysiological dose of insulin.
All those excess GLUT4s allow glucose molecules to pour in to the cells. You might as well have hyperglycaemia and GLUT1 mediated oversupply, as far as the respiratory chain is concerned. Glucose in excess of the cell needs generates superoxide. Superoxide triggers, as an antioxidant defence mechanism, insulin resistance. With thanks to Dr Guyenet. Again. It is difficult to emphasise how good this paper is.
Somogyi overswing is likely to be caused by acute onset insulin resistance occurring as a direct result of excess glucose uptake in to cells due to supraphysiological insulin concentrations.
The temporal association with hypoglycaemia, which misled Somogyi, comes from the time course of switching sources of glucose oversupply. The hypoglycaemia is not causative, it is just common for it to occur at around the same time that insulin/GLUT4s oversupply substrate to the mitochondria and they say no to it, using insulin resistance.
Let's summarise. This is very, very important:
Excess insulin causes insulin resistance
This is just day to day internal medicine. You have to pay the mortgage somehow.
If anyone is interested there is a rather nice discussion paper here, it's pay per view and doesn't say much more than is in the abstract but it has a nice set of references. I have access to a great Athens account. All the comments on insulinomas ring so true to clinical life too.
It's also interesting to go back to the controversies around the Somogyi effect, you can read Somogyi's ideas here and the continuous glucose monitoring evidence here. All very fascinating stuff (well it is to me!) but what does it have to do with shooting fish in a barrel?
Question: Who are the Un-dead?
Which can be rephrased as: Can we control the Somogyi effect?
If we take the average bodybuilder from a few years ago and watch him self-inject with insulin for its anabolic effects and then forget to eat the carb load needed to balance it, we can see the acute effects of insulin overdose. Insulin rises very rapidly from the regular insulin used and every GLUT4 receptor in his body pops on to every cell surface which uses them. There is a free fall of glucose from plasma in to the cells, blood glucose plummets and the chap ends up in A&E or, quite possibly, in a mortuary. There is no time for the massive cellular caloric overload from over-translocation of GLUT4s to generate enough insulin resistance to stop the hypoglycaemia. Glucose pours out of the bloodstream until it drops to levels low enough to kill the brain. Sad but true. Somogyi effect is too late, too little. Insulin overdosed bodybuilders are not the Un-dead.
So who really are the Un-dead?
What if you give insulin as a constant rate infusion, initially at a low rate and gradually crank it up?
Think it through. Progressively increasing insulin levels allow progressively greater amounts of glucose in to cells. If the cellular glucose supply is greater than cellular needs there is increased generation of superoxide by the respiratory chain which signals the cell to become resistant to insulin. A balance is achieved. Increase the insulin CRI, overcome the insulin resistance, generate more superoxide, generate greater insulin resistance, achieve a balance. Do it again. And again. More. Again. How high can you get plasma insulin by playing this sort of game? Here's the table we need:
Okay, they stopped at a total of 6iu/24h/per rat. They could possibly have gone higher but hell, we have here a set of rats with a mean insulin level of 588.9microIU/ml. No, that is not a typo. The SEM was 89.7. Anyone like to guess how high the highest insulin level measured was? Quite high perhaps?
These are the Un-dead. They walk around, without any genetic modification, with an insulin level which, if achieved acutely, would have put them rapidly in to a clinical waste bag. They are very, very, very, (repeat ad nauseam) insulin resistant, otherwise they would look like the bodybuilder in the mortuary.
WARNING: There is a black box paradox warning about the paper providing Table 1. I'll stick an addendum on the end of the post.
Soooooooo. They are, undoubtedly, hyperinsulinaemic. Are they fat? Of course not. Why should they be fat? They are the Un-dead. If they were remotely sensitive to insulin they would be not be the Un-dead, quite the contrary. But insulin induced insulin resistance does not spare adipocytes. These have mitochondria and generate superoxide. They too will ignore insulin, to a level determined by their mitochondrial superoxide production.
Here's a bit of an aside: The process is physiological. It involves a careful titration of cellular insulin resistance to the cellular energy needs. This is no blanket insulin blocking drug. The responsiveness to insulin is carefully adjusted to just allow enough glucose in to cells to meet their needs. This applies to adipocytes as well as well as to muscle cells. With the number of GLUT4s being translocated by the residual insulin sensitivity, in an environment of 588microIU/ml of insulin, you don't need much of a blood glucose level to supply glucose needs. Table 1 suggests the body settles to a plasma glucose of about 71mg/dl, as opposed to 148mg/dl in the control rats. Metabolism is still largely glucose based, with some responsiveness to insulin preserved despite the need for resistance to survive at 588microIU/ml. Transplanting tissues to a petri-dish allows you to pick up this responsiveness. Free fatty acid release from adipocytes is not significantly inhibited because the adipocytes are insulin resistant to a level where they maintain normal function. Weight gain is similar to that of control rats.
And another BTW. The process is cellular. Bugger the hypothalamus.
A nail in someone's coffin?
Apparently these rats are a nail in the coffin of the insulin hypothesis of obesity.
The actual coffin nail [nb if the link comes up with a failed log-in just refresh the page] is a pay per view article in a journal not covered by Pubmed and I'm unwilling to shell out $40 for it. Perhaps I could ask The Good Doctor for a copy. Fortunately the information on CRI rodent models is freely available in the paper which provided Table 1 above. What is crashingly obvious is the utter lack of understanding of insulin induced insulin resistance by people who are fixated on insulin as a satiety hormone.
This might have been acceptable in 1980 when the physiology of insulin resistance was completely unknown. But to see this explanation promoted by the same obesity researcher who provides us with the concept of insulin resistance as a cellular antioxidant defence mechanism, mediated through superoxide, is utterly depressing. We are, after all, talking about a complete failure to understand the basic physiology of insulin resistance, with the key paper sitting as a free download from Pubmed.
Does the Good Doctor not understand his own citations or is he stuck with terminal cognitive dissonance?
Or perhaps he's just utterly confused.
I feel the coffin nail is misplaced.
OK the paradox: In the paper providing Table 1 the rats have a blood insulin level of 588microIU/ml with physiological blood glucose levels. BUT isolated muscle and fat cells taken from these rats are highly insulin sensitive, more so than those from the control rats. How is this possible? I can imagine the Good Doctor or some other idiot shouting that the rats aren't insulin resistant at all, because the paper clearly shows their tissues are extra insulin sensitive, ergo the insulin hypothesis of obesity is wrong. Peter is misquoting a paper, you know what I'm like!!!! Gotta read all those papers cited, the Good Doctor knows how few people follow the links.
But the rats are definitely Un-dead.
If you culture adipocytes at consistently supra maximal insulin levels they behave exactly as the whole rats do. So if you pull out a muscle or fat cell from an Un-dead rat, having made it an un-Un-dead rat by decapitation, how long will the insulin resistance last? This is probably determined by the elimination half life of superoxide. Which is, err, not very long... Actually, it's probably determined by the cellular redox state providing the superoxide, which should last at least a few seconds after decapitation.