Liver and insulin (not a cooking recipe...)

I'm umm-ing and ah-ing about posting this at all. In the end I'm going to hit post. It's up for shredding! Peter


The function of insulin is the inhibition of lipolysis. I cannot argue with this.

There is a widely held belief that insulin is also necessary for the cellular uptake of glucose. This is incorrect.

I hit on this paper as an accidental result of the Atkins and methylglyoxal searching. It grabbed my attention because it reminded me of a paper I had read many years ago (on vacation, I used to take British Journal of Anaesthesia on vacation!) which was probably this one. And this is the one where they got type one diabetics to skip their insulin and be studied in the hyperglycaemic an-insulinaemic state (see below).

This is my summary of some of the main concepts carried in the papers.


I started off with simple analogies to baths, bathwater, flows etc. Unless you have a very, very strange plumbing system, this doesn't work. Back to metabolism.

Life is simpler if you are fasting.

If you have 5mmol/l of glucose in your blood, you cannot get more than 5mmol/l inside your cells. There are no pumps for glucose, it follows a concentration gradient. If your cells are using large amounts of glucose there will be a bigger concentration gradient and so more glucose will flow through the GLUTs, but perhaps not enough. You might need more "holes" to let glucose through. Enter insulin, more GLUT4s, more flow, sustained metabolism. Still no pumping and still blood glucose is 5mmol/l because whatever the cells take is being replaced. From the liver (we're fasting). Nowhere else for it to come from.

Let us say there is no insulin. There will be a basal number of GLUT4s and a few other GLUTs, which will allow glucose to flow. How much? Not enough. Not enough if the blood glucose is 5mmol/l. But what about with a blood glucose of 30mmol/l?

Would a blood glucose of 30mmol/l force enough blood glucose through the few GLUTs that are present without the help of insulin and its extra GLUT4s?

Well, apparently that's a pretty easy question to answer using tritiated glucose and the answer is yes. With a blood glucose high enough you do not need insulin to allow as much glucose to be used as would be used when blood glucose is 5mmol/l in the presence of insulin. This is fact.

You can read the paper about the type 1 diabetics who volunteered to withdraw their insulin and were studied in the hyperglycaemic an-insulinaemic state. They burn glucose.

So, if insulin is not essential for glucose based metabolism, what is its primary function?

Insulin allows the pancreas to talk to the liver. The liver controls, under the influence of insulin, how much glucose it adds to that teaspoonful of glucose which is normally present in the total blood volume.

This is core. As core as insulin's inhibition of lipolysis.

In a normal person 85% of the glucose from a carbohydrate meal never makes it past the liver. Under conditions where a bulk supply of rapid uptake glucose is unavailable, I doubt that any glucose gets past the liver. The pancreas knows about dietary glucose, the liver knows. It's their secret from the rest of the body. The liver rations out the glucose.

Diabetics, type 1 or 2, are not hyperglycaemic because they cannot use glucose. They are hyperglycaemic because their liver can no longer hang on to its glucose hoard. The liver's inability to be influenced by insulin is central to diabetes.

So the aim, in diabetes management, should be the control of leakage of glucose out of the liver. You can actually force a fructose damaged, insulin resistant liver to listen to insulin in exactly the same way as you can replace pancreatic insulin in type 1 diabetes. Use exogenous insulin. But it's hard.

Your liver does not listen to your subcutis, which is where injected insulin comes from. It listens to your pancreas. The pancreas secretes insulin in to the portal vein which has a blood flow of about a 1000ml a minute. A minute's worth of secreted insulin will be carried in 1000ml of blood. From the liver that minute's worth of insulin enters the systemic circulation and mixes with the cardiac output, which at rest is about 5000ml/min, so is clearly diluted. Peripherally measured insulin is always less than what the liver "hears" when it listens to the pancreas. What was in 1000ml/min of blood is now in 5000ml/min of blood. Of course insulin recirculates so systemic concentration won't be as low as one fifth of portal concentration.

So the normal liver should be seeing more insulin than is detected in peripheral blood. Adipocytes and muscle cells only see peripheral blood. When you inject insulin under your skin it is carried by the full cardiac output and will be delivered in dilute form to the liver compared to what should have happened if the same amount had come from the pancreas. Muscles will get the full hit.

A high carbohydrate diet, coupled with industrial doses of peripheral insulin, is doomed to fail. You cannot effectively inhibit glucose release from the liver without hitting the peripheral tissues with a relative overdose of insulin. This opens the GLUT4 floodgates in to the peripheral cells in the process of trying to stop glucose release from the liver. This sort of balancing act, high insulin, high glucose throughput, has to rely on hyperglycaemia to keep you safe from hypoglycaemia. Hello ADA.

EDIT: The papers discuss ketones as producing blockade of peripheral glucose metabolism, as we know they do. Palmitic acid is my idea.

The only sensible solution is to make the peripheral tissues as resistant to insulin as possible (so minimal extra GLUT4s pop up as a result of exogenous insulin) and then supply enough exogenous insulin to inhibit hepatic glucose release, delivering it through the hepatic artery as well as the portal vein. Two things allow this to work. Palmitic acid and ketones. Palmitic acid can be delivered through chylomicrons or VLDLs, it causes insulin resistance, thank goodness, so makes exogenous insulin less effective on muscles (where most of the GLUT4s are). It's worth noting that MCTs (I would guess through ketones) raise peripheral insulin resistance but still dip blood glucose (ie they allow the liver to listen to insulin). Ketones are made by the liver. I can't see ketones making the liver insulin resistant. They will inhibit peripheral glycolysis independent of glucose uptake. They effectively replace the insulin/glucose combination in metabolism. Insulin can then be given to the liver to inhibit glucose release without turning the muscles in to a glucose sump. Most tissues outside of the brain and a few other places can run perfectly well on the palmitic acid.

Low carbing is the solution, aiming for mild ketosis, intense physiological insulin resistance and minimal insulin doses (aimed at the liver). Bernstein is the guru, Kwasniewski has a slightly more relaxed approach. Both seem correct to me. Neither shuns saturated fat. But it doesn't seem as simple as balancing insulin against carb intake (though this is an excellent, and probably the only practical, rule of thumb). I would expect it to work better on a palmitic acid based diet than one using any unsaturated fatty acid for bulk calories. Using PUFA and oleic acid (to keep LDL down, dontcha-no) won't hack it in providing the physiological insulin resistance that should be helpful for tight glucose control.

As an afterthought; it is under low insulin levels that the liver ships out VLDLs too. More palmitic acid to the periphery.

We're back very close to low carbohydrate eating mimicking starvation, without the weight loss.

Peter