Hepatic insulin resistance through caloric overload

It seems from Dr Lustig's commentary that one specific method of developing insulin resistance is by increasing fatty acid acyl-CoA moieties within a cell. Acyl-CoA is a single fatty acid molecule joined to a CoA group and represents an "activated" fatty acid, ready to do things metabolically. In both muscle and liver it appears to be these activated lipids, rather than stored triglycerides, which are the metabolic signal, via JNK1 and serine phosphorylation of IRS1, which is used to down regulates the activity of insulin on glucose control.

This paper suggests high levels of free fatty acids are taken up by the liver and inhibit it's response to insulin. My feeling is that the FFAs can come from hepatic lipase (as above), from overstuffed adipocytes leaking FFAs or even from dietary intake, as lipoprotein lipase spills diet derived FFAs from chylomicrons in to plasma as well as in to adipocytes.

Reducing FFA delivery to the liver by inhibiting hepatic lipoprotein lipase does nothing to get rid of hepatic lipid droplets (they have to go out as VLDLs) but the decrease in FFA delivery lowers Acyl-CoA and allows normal liver response to insulin.

Is it possible to overload the delivery of FFAs to the liver without in-situ generation of acyl-CoA from either fructose or alcohol?

Apparently yes. You can do it by diet. It's not easy, but if anyone wants to try it here's the technique. It works in dogs anyway.

This is a post I've had around for a few days while the little palmitate storm blew over:

There is a key concept in veterinary medicine which states that cats are not small dogs. No one would argue with this, especially those who associate with the superior species.

However, dogs might well be reasonably viewed as small humans, it makes a great deal more sense than considering mice to be very, very small humans.

So, if someone tells you that they fed a high fat diet to a group of dogs, restricting their caloric intake to weight stability, and that they all developed virtually complete hepatic insulin resistance within a few weeks, you might just have to sit up and take notice. Especially as the fat was cooked bacon grease, provided by the university canteen. Don't ask. The dogs didn't get the bacon (as far as I can tell). This is that study, free full text.

Anyway. They initially fed these poor dogs a "can a day" of an Hills "prescription" diet (that's the sum total of the methods info, except the macronutrient ratio in the can) and enough dry diet (some random food made by Wayne Dog Food) to maintain weight stability. They did this for two weeks then took away some of the dry cr*p in a bag and replaced it with bacon grease. About 2g/kg bodyweight of bacon grease. They kept it almost isocaloric with those first two weeks of eating traditional dog food. The idea was weight stability. The agenda was to prove that, under isocaloric and weight stable conditions, fat was bad, bad, bad. Replacing as little as 8% of mixed calories with bacon grease will cause total hepatic insulin resistance.

Like wow!

So what does this group consider a caloric intake to maintain weight stability in a 27kg mongrel while it is eating cr@p in a bag? Are you sitting down?


Cr@p in a bag: Total calories 3,885kcal/d

For "less cr@p in a bag but plus 2g/kg bacon grease": more like 3,945kcal/d


This is for a 27kg dog sitting in a cage.

Go on, read that again; 3,945kcal/d. I'm not joking.

OK, so the first question is whether these dogs were weight stable. Ha ha ha ha. The amazing thing is that they only gained about 2kg during the study period. Make that 3kg if you include the weight gained in the pre study "weight stability" period.



Aside: look at how they wangled the weight stability. Pre study admission time until study week zero, about 1kg weight gain. Not statistically significant. From study week zero to week 12 there was a 1.9kg gain which fluctuated in and out of statistical significance WHEN COMPARED TO STUDY WEEK ZERO. Had they compared the on-going weights to the weight at the time they first got their hands on these pooches (minus 2 weeks), virtually all weights from about week 4 would have been significantly up on the enrollment weight. Funny that. Back to the dogs:

3.0 kg in 14 weeks is >10kg per year. In three years, at this feeding rate, the dogs would weigh >60kg. Some weight stability!

Also, there is no control group. I would love to see what 3,885kcal/d of cr@p in a bag would do to a dog's weight in 12 weeks. Waltham's daily energy requirement for a 27kg "typical" adult dog is 1300 kcal/d and for an active dog 1480 kcal/d. Personally I doubt that chronically catheterised laboratory mongrels are getting a huge amount of exercise.

So this is another study where the introduction and discussion are utterly divorced from the methods and the results (and from reality). It's worth just flicking through the methods and, in your mind's eye, look at how much money was used on these dogs. A clinical MRI was around about £1000 a shot in the UK Home Counties in 2009.

For all this money spent, is there anything of interest in the study?

Fortunately yes, lots.

The first thing is that if, like me, you eat somewhere in excess of 2g per kilogram bodyweight of dietary fat every day DO NOT, under any circumstances, add 3000 kcal of carbohydrate to it. If you do this you will develop virtually complete hepatic insulin resistance within a few months. You will also get very very very very fat. Not in a week, but certainly in a couple of years. Thank goodness for this study, saved my liver.

Second is that you will not immediately develop peripheral insulin resistance. This will take significantly longer to develop. That's interesting. The liver is the initial site of injury in caloric overload, just as it is from fructose poisoning, or alcohol too for that matter. I might have guessed at muscle/fat for caloric overload.

Third is how would Garry Taubes view the achievement of getting a group of medium sized dogs to consume 4000 kcal/d? The equivalent of how do you get a 64kg human to consume 10.000 kcal/d? Challenging.

I would guess a mass of uncoupling proteins and elevated insulin to cover hepatic glucose leakage...

Anyway, if anyone has personally managed to consume 10,000kcal per day for a few years I'd love to know how you are getting on. Foie gras?

Peter

BTW the really scary features of this paper are that it got through it's grant proposal, it got through scrutineering and it spawned another, even more expensive, project using the same model which also got approved, completed and published. As my wife says, the peer review process is awful, but no one can think of anything less bad so far. Fortunately the group are wasting USA tax payer's dollars rather than my pounds sterling. Phew.

Addendum: What's the physiology behind the pathology? Well a dog never eats carbs in the wild, beyond the gut contents of herbivores. It usually takes in a massive caloric load of fat. It needs insulin to store that fat, so fat intake ought to make the liver a little insulin resistant, leak a little glucose and then it's up to the pancreas to sort out the glucose, taking the lipids along with it in to fat cells. This is normality. Adding massive carbs to massive fat will simply break a perfectly adaptive system... That's my take. Don't do it!