EDIT There is a follow on post to this one including some clinical data on the hypoglycaemia episodes. I'll put a link in here now it's up. END EDIT
Remi and Ken both pointed me toward this paper:
A Selective Sweep on a Deleterious Mutation in CPT1A in Arctic Populations
The paper itself is largely an account of the detective work involved in pinning down a specific mutation which has been positively selected for in a Siberian population living in the Arctic. The same mutation is also present in non related groups inhabiting the Arctic areas of northern America. The mutated gene is very common and frequently homozygous. It puts a leucine in the place of a proline in CPT-1a, the core enzyme for getting long chain fatty acids in to mitochondria. Putting a leucine where there should be a proline means the protein is basically f*cked. The mutation is linked, not surprisingly, to failure to generate ketones in infancy and can be associated with profound hypoglycaemia, potentially causing sudden death.
From the evolutionary point of view we have here a mutation which is significantly lethal at well below reproductive age, so it should have been weeded out because affected individuals are less likely to live long enough to pass on the gene. But it has been highly positively selected for in several populations, the common factors being cold climate and minimal access to dietary carbohydrate. It's a paradox.
Following a link in the paper gives us this abstract, with this snippet:
"Investigation of seven patients from three families suspected of a fatty acid oxidation defect showed mean CPT-I enzyme activity of 5.9 ± 4.9 percent of normal controls"
A value 6% with an SD of 5% suggests to me that some of these people may well have a CPT-1a function very close to zero. How common is the mutation?
"We screened 422 consecutive newborns from the region of one of the Inuit families for this variant; 294 were homozygous, 103 heterozygous, and only 25 homozygous normal; thus the frequency of this variant allele is 0.81"
I think "very common" is a reasonable description.
How dangerous is it?
"Three of the seven patients and two cousins had hypoketotic hypoglycemia attributable to CPT-Ia deficiency"
The next thing we can do is google CPT-1a deficiency and have a look what needs to be done to stay alive if you carry this gene.
Clearly, if you can't transport LCFAs in to your mitochondria, you should run your metabolism on glucose/pyruvate and avoid the dysfunctional fatty acid transporter. This means raw corn starch, as we have seen used (probably wrongly) for glycogen storage diseases. Properly cooked starches are too short acting to reliably keep a child alive all through the night. They aren't safe enough.
Of course MCT oils have a role too. A CPT-1a defect has no effect on MCT metabolism so these can be used either directly by tissues or indirectly via liver/glial produced ketones.
LCFAs, unable to be metabolised, accumulate in the tissues as a storage disease. The advice is to avoid them as far as possible.
So the archetypical CPT-1a defect tolerant environment would seem to be a person sitting on a South Seas Island beach by a pile of coconuts chewing on a raw yam, with copious flatus night and day.
But it's not.
The CPT-1a defect evolved in multiple non related populations where both starch and MCT were very notable by their near-complete absence. It's an Arctic selected gene. No starches. No coconuts.
Let's take a speculative look at what is going on.
Living on a very low carbohydrate diet is associated with chronically elevated free fatty acids, chronically low levels of insulin and an ignorance of glucose. i.e. the body ignores glucose. Synthesise what glucose is needed but, beyond that, who cares?
Living in a sea of free fatty acids, which are taken up in to cells in a largely concentration dependent manner, allows an increased gradient to push FFA-CoA at any residual function in CPT-1a. It would appear, from the evolutionary perspective of Arctic inhabitants, that near ketogenic levels of FFAs are adequate even if you have the proline to leucine substitution at amino acid 479 in CPT-1a. You can do enough beta oxidation to cope.
Of course, the minute you lower free fatty acids, perhaps to the level of a post prandial starchivore, beta oxidation is going to grind to a halt without the concentration gradient effect. This is pathological. The temporary fix of substrate level ATP synthesis and related pyruvate supply to the mitochondria is fine for a while, but any reactive hypoglycaemia is going to be potentially fatal, especially if you are asleep or food deprived at the time. We know that insulin suppresses lipolysis at levels which don't budge GLUT4s. When insulin has suppressed lipolysis and blood glucose is low, FFAs might be fatally limited.
If you have the mutation but you never do the starchivore thing your FFAs are high 24/7, whether you have just chewed on a lump of seal blubber or not. No paper in the reference list appears to have looked at the FFA levels of children with this mutation on a mixed diet, let alone on the ancestral fat based diet of the polar regions. Given sustained very high levels of FFAs, you might even make some ketones.
If free fatty acids are high and there is no insulin to divert them in to storage, all of the nasty storage diseases associated with CPT-1a dysfunction might well disappear. This is the situation where the mutation allows carriers to thrive.
I think elevated free fatty acids, without elevated insulin, is a recipe for the tolerance of this mutation.
But the mutation is not just tolerated. This is no neutral mutation, it is positively advantageous. The prevalence of the mutated gene is far from random. Why is it beneficial?
This is not quite so simple.
Uncoupling is one component. Uncoupling respiration generates heat. There might just be a positive advantage to running your metabolism fairly uncoupled in a very low temperature environment. Elevated FFAs are completely essential to uncoupling and heat generation. Limiting fatty acid removal from the cytoplasm to the mitochondria might be a facilitator of uncoupling. It's FFAs on the cytosolic side of UCPs which facilitate proton translocation. Having a higher level of cytoplasmic FFAs at a given level of plasma FFAs might give an advantage over the normal level of uncoupling seen under near ketogenic diet conditions.
The second possibility is that, once you have established high enough levels of FFAs to push through the CPT-1a bottle neck, you simply run at this level flat out, all the time. One of the features of the CPT-1a from the modified gene is that it fails to be inhibited by malonyl-CoA. Even with limited CPT-1a activity there must be times at which ATP synthesis exceeds metabolic requirements and fatty acid transport ought to slow. There is no longer any brake to be applied to FFA transport if excess acetyl-CoA, exported to form malonyl-CoA in the cytoplasm, fails to inhibit CPT-1a . Oversupply of ATP within the matrix is likely to provide optimal uncoupling conditions, in excess of those from a ketogenic diet with regulated fatty acid uptake. That would be my guess. If it's cold enough, this might make the difference between survival or not. It keeps you warm, especially when you are asleep and the TCA should be quiescent.
Flicking through other references in the paper it does appear that indigenous Siberian people do have an elevated resting metabolic rate. In fat free mass it is 17% above calculated values i.e. they are uncoupled.
Finally, adults are not affected by the hypoglycaemia syndrome. My presumption is that, after puberty, they are sufficiently insulin resistant to have adequate FFAs present to maintain relatively normal mitochondrial function. It's the children who need their ancestral diet.
People with glycogen storage diseases die of hypoglycaemia (amongst other problems). We know that a deeply ketogenic diet both protects from hypoglycaemia and sets the body up to run perfectly well without any dietary glucose, which might be lost to glycogen stored permanently in the liver/muscles. There is every justification for giving the finger to cornstarch here and the folks suggesting a modification of ketogenic eating appear to be on fairly safe biochemical ground.
For the P497L mutation everything from the evolutionary perspective suggest that a very high FFA inducing diet may be equally efficacious. But the risks associated with failure, from the occasional safe starch meal or unsafe birthday cake at a party, carries the potential for catastrophe once insulin puts free fatty acids in to free fall.
BTW: You just have to wonder if any other CPT-1 mutations might behave in a similar manner to the P497L change in the Arctic... Could it be bye-bye time for cornstarch?