Rapeseed oil for weight loss (3) Canola oil vs butter round two

Now to get down to the nitty gritty of the first anomaly in the paper

A highly saturated fat-rich diet is more obesogenic than diets with lower saturated fat content

The oddity is the blip downwards of weight in canola fed rats, highlighted by the red oval on the graph:

which needs to be read in conjunction with the food intakes in this graph, also highlighted by a red oval:

I've copied the food intake graph and laid it over the weight graph to make it a bit clearer:

Because the red graph is in grams of food I've annotated the important energy intakes next to the related data points. Obviously the switch from chow to high fat diet (fine orange line) massively alters the caloric intake despite the weight eaten being the same on the day before and after the switch.

Interminable aside: if any paper gives you the caloric density of any food used you are assured that the group are idiots, even if they are very, very clever idiots. They believe you can fool a rat's hypothalamus in to over-consuming calories by increasing the caloric density of the food. Or fool it in to under-consuming calories by diluting them down with sawdust or the equivalent (usually cellulose). As if evolution is that gullible. Do not, under any circumstances, accept any of such a group's ideas about satiety, food intake or bodyweight. Ever. End aside.

Anyhoo. On day one the high fat fed rats increased their energy intake from around 227kJ (55kcal) to 355kJ (85kcal), a rise of ~56%. I find this impressive and in the same league as the Schwartz lab rats on D12492, who increased their caloric intake by ~87% on day one in this venerable graph:

In Protons terms the canola oil fed rats still used 227kJ for running their metabolism but "lost" 128kJ in to their adipocytes via the enhanced insulin signalling of linoleic acid (LA) and alpha linolenic acid (ALA) due to their low F:N ratios. The rats needed 227kJ, they lost 128kJ so needed to replace those 128kJ, therefore they ate 227+128 = 355kJ. They stopped at 355kJ because that's when they stopped being hungry. Overall weight did not change because 227kJ of chow weighs the same as 355kJ of the canola oil diet (Hmmm, chance or pilot study????). 128kJ is about 3g of fat gain.

Simple.

What happened next is much more exciting. On day two of the canola oil diet the rats weren't hungry. They only ate 155kJ that day. The weight of food eaten dropped by ~13g, and so did the weight (~10g) of the rats, food is heavy, they ate less, they weighed less. But they still needed those 227kJ to run their metabolism. They ate until they weren't hungry, that's what they do. Why weren't they hungry?

They must have accessed the full 227kJ that they needed, otherwise they would still have been hungry. They obtained the missing 72kJ from their adipocytes. Without "trying", without being hungry. Remember, they always eat to satiety.

The logical explanation is that there was an acute drop in insulin signalling in their adipocytes which released 72kJ worth of FFAs, which meant that there was no need to eat the 72kJ's worth of food needed to reach the full 227kJ to run metabolism. 72kJ is the energy available from about 2g of fat.

From the Protons perspective acute loss of insulin signalling is a standard effect of reduced mitochondrial membrane potential caused by uncoupling. PUFA uncouple. At high intakes this over rides the obesogenic effect of their reduced F:N ratio.

The blip of weight down then back up again mostly reflects food weights with a lesser contribution from adipose weights, but the latter adipose changes are the factor responsible for the hunger changes which drive the food intake changes.

Uncoupling explains the blip. The only thing which surprised me was how rapidly the effect kicked in. I would have expected that about a week would be about the time scale but here it was 48h. Those are the data. I have to revise my views in the light of it. You don't often get given a gift-horse (of daily food intakes and daily body weight changes) to consider examining the mouth of!

The following day the rats increased their food intake. Under steady growth the rats still needed the standard 227kJ/d plus some extra for uncoupling losses. Adipocyte size does not continue to shrink because the uncoupling using biological uncoupling proteins is ATP dependent, low intra mitochondrial ATP reduces the effectiveness of uncoupling proteins. This is distinctly different from chemical uncoupling to which there is no upper limit and so these can give extreme weight loss or even death.

The rats still needed 227kJ/d for metabolism and 62kJ/d for uncoupling, hence 289kJ/d. This happened pretty consistently throughout the rest of the high fat period. They stayed slim but not too slim.  Despite eating around 20% extra in calories compared to the chow fed mice.

Peter

There is another very strange/interesting feature on the food intake graph. I'll see when I can find time to describe it.

Rapeseed oil for weight loss (2) and butter for obesity round one

This is the next paper. These people are good. Really good. There is almost nothing amateurish in this paper:

A highly saturated fat-rich diet is more obesogenic than diets with lower saturated fat content

and here is the graph which sums it up

I'm not saying these people aren't stacking the deck and I'm not for a moment suggesting that they understand anything about obesity. But they deliver in spades to support the current (anti saturated fat) narrative.

I'm going to speculate from here onwards about how they managed it.

The above graph did not materialise out of thin air.

They made their own foods in their own lab. For the high fat diet they "chose" 67% of calories as fat, For the low fat period they "chose" 27% of energy as fat.

Why?

Because, obviously, markedly different values would not have generated the above graph. They could, and I suspect they did, formulate almost any diet macro ratio they wanted to. It's called a pilot study and and I think you should seriously consider that they did one, and kept quiet about it.

Very few people using a rodent diet induced obesity model would consider 27% of calories from fat to be low in fat. Their chow was ~11% of calories as fat and very low in PUFA (Diet 5075, which included beef tallow as a partial fat source. It's not made anymore but you can guess it resembles 5001 but with less PUFA. Probably. But who knows?). My presumption for butter is that medium and short chain fatty acids are, by the Protons hypothesis, insulin sensitising. At low dose rates they are largely diverted from gut to liver and never reach peripheral adipocytes. So they are not obesogenic. They increase hepatic insulin sensitivity which lowers the penetration of both glucose and insulin to the systemic circulation and, certainly in the case of MCT oil, are slimming. My guess is that much less than 27% of calories in the "low fat" period would have produced active weight loss. So you simply do not use less than 27% of calories as fat from butter, if you want obesity in your model. 

If you supply enough insulin sensitising shorter chain fatty acids that they penetrate past the liver to peripheral adipocytes they will continue to work their insulin sensitising effect but this time on adipocytes rather than hepatocytes. With enhanced adipocyte insulin signalling there is enhanced fat gain. This appears to be optimal at 67% of calories from butter. Much more than 67% would tend towards ketogenically low levels of insulin exposure and the insulin sensitising effect would become irrelevant. Again, a guess, but much over or under 67% would have generated less obesity. It's a model, it's designed to make a point.

In the canola oil arm the high fat diet provide ~20% of calories as LA, verging on an uncoupling dose rate. In addition to this there are 10% of calories as ALA, a significantly better activator of UCP1 and UCP2 than LA. This is clearly a factor in the high fat group and there are oddities on the graph and elsewhere in the paper which do suggest that this is the case. Another post.

What would have happened if they had lowered the canola oil to a real low fat level, say 4% of calories from PUFA? LA at under 4% of energy is not usually obesogenic, though it should tend that way via the Protons F:N ratio, absenting uncoupling effects. Some fat storage is beneficial. And ALA would accentuate this effect, so long as levels were high enough for the Protons effect to occur but low enough to avoid uncoupling effects. So very low dose canola oil might well be obesogenic and was sensibly avoided.

You don't write your experimental protocol until you are damned sure it is going to deliver. The required result here is to show that "saturated" fat rich diets are obesogenic and "PUFA" rich diets are not.

They succeeded. As I say, they are good. Here's the start of their discussion:

"The present study tested canola, lard, and butter, respectively, low, moderate, and rich sources of SFA, widely consumed in the human diet, in an animal model of dietary obesity. As predicted, results confirmed the hypothesis that an SFA-rich diet is more obesogenic than diets with lower SFA content."

They obtained the result they predicted. Probably by dint of a lot of hard work.

Peter

Have you thought about electron transporting flavoprotein dehydrogenase and its substrate electron transporting flavoprotein?

This is not really a post. I just want somewhere to place a concept, outside of the original post in which it is embedded, where it is more easily recognisable and findable.

It's worth noting that this paper:

The CoQH2/CoQ Ratio Serves as a Sensor of Respiratory Chain Efficiency

describes the process of remodelling the ETC in great detail, especially down regulation of complex I when there is an high input from other sources to the CoQ couple. It's mediated by reverse electron transport and it happens fast.

The basic TLDR is that if you take fat adapted mitochondria they will be using mtETFdh to generate a significant proportion of their maximal oxygen consumption for ATP generation. This means that complexes I and II will be down regulated, so supplying electrons to these complexes cannot match the oxygen consumption which would be generated if mtETFdh was maximally active. We have no available direct supplies of electron transporting flavoprotein to supply FADH2 in the way that beta oxidation does. "Dysfunction" is actually an artefact of not inputting adequate electrons to the CoQ couple via mtETFdh.

This applied both to studies on high fat diets and studies on fasting. It implies extreme caution if one is to decide that high PUFA diets, when high in overall fat, do actually cause *any* mitochondrial dysfunction, if only tested using inputs from glutamate/malate or succinate.

So this has major implications as a generic "how to read a paper" factor.

The insight is based on the oxygen consumptions in this paper where "disrupted bioenergetics" are claimed.

Rapeseed oil‑rich diet alters hepatic mitochondrial membrane lipid composition and disrupts bioenergetics

I wrote this in the post about the above paper:

There is nothing wrong with these mitochondria. Bioenergetic are *not* disrupted, as suggested by the title of the paper. Let's dig deeper.


What is happening is that the study is taking mitochondria from fat-adapted rats and feeding them on either a complex I input or a complex II input. Fatty acids, even LA, make significant use of electron transporting flavoprotein (ETF) dehydrogenase as their input to the CoQ couple. Mitochondria adapt their electron transport chains to the substrates available. If mitochondria from rats fed 40% of calories from fat are significantly dependent on mtETFdh for input to the CoQ couple, and have down regulated both complexes I and II, then feeding the preparation on substrates specifically aimed at complex I or II will obviously produce sub-maximal oxygen consumption. Which is what happens under either state 3 respiration or FCCP uncoupling.


Under the "tickover" conditions of state 4 respiration the uncoupling from PUFA shows clearly.


Obviously, to restore visibly normal mitochondrial function, what's needed is a supply of reduced ETF to use as a substrate for mtETFdh. As supplied by beta oxidation. Sadly you can't just buy reduced electron transporting flavoprotein from Sigma Aldridge, so you end up with artifactual mitochondrial "dysfunction".

Another aside: that is exactly what is happening here too

Prolonged Fasting Identifies Skeletal Muscle Mitochondrial Dysfunction as Consequence Rather Than Cause of Human Insulin Resistance

discussed here

Will fasting destroy your mitochondria? (No).

It *appears* as if mitochondria adapted to high input from mtETFdh are dysfunctional if you fail to supply them with adequate electron transporting flavoprotein! The study did try to get around this by using octanoyl-carnitine (50μmol/l) as a lipid input to generate reduced ETF but clearly even 50μmol/l of palmitate will provide twice the ETF of 50μmol/l octanoate and the chaps in the study were running total FFAs of up to 3000μmol/l, not 50μmol/l, at the time that the muscle biopsies were taken. Utilising 3000μmol/l of FFAs provides a lot of ETF. So "dysfunction" is really an experimental artefact induced by lack of metabolic substrate for mtETFdh (secondary to using an homeopathic level of octanoate in this case or no mtETFdh substrate at all in most studies).

That is all.

Peter