Of mice and men (3) Sucrose

There is a lot of variability in the linoleic acid content of lard. Good studies get round this by measuring LA using gas chromatography. From this paper we have Chinese lard containing 10% LA and from this paper we have Brazilian lard containing 17% LA. I've seen values cited as high as 30% but they have never been quite as convincing as the results table in a publication, though I see no reason to doubt them per se.

From the Protons/ROS perspective even 8% of total calories as linoleic acid is usually enough to get to be obesogenic, particularly if a little extra soybean oil is added as an "essential fatty acid source" (no sniggering at the back), which nicely explains most obesity paradoxes.

I have been revisiting various high and low fat diets in the aftermath of the Speakman paper which was so confusing as regards D12451 and its specifically marketed control diet D12451B.

Here are the compositions of both because Research Diets no longer lists D12451B as a standard diet:

So we have D12451B, a slimming diet, used as the control diet in the paper discussed in the last post. That will relate to the LA content which I calculate out at 3.2% of calories. The obesogenic D12451 comes out at 6.7% of calories as LA if we used Chinese lard, 9.5% using Brazilian lard and goodness knows how much using lard from the USA or UK.

The sucrose values are fascinating. D12451B provides 35% of calories as sucrose plus 35% as starch/maltodextrin and is slimming. D12451 provides 17% of calories as sucrose and is obesogenic. So this long established, slightly old fashioned pair of diets speak strongly against sucrose being the driver of obesity, they are more compatible with it being linoleic acid.

However I find the sucrose cannot be ignored. You can design your study to show sucrose as good, bad or mystifying. Let's think about sucrose.

I used to really rather like this group, which I've mentioned before :

Ingestion of a moderate high-sucrose diet results in glucose intolerance with reduced liver glucokinase activity and impaired glucagon-like peptide-1 secretion

In this first paper they either fed mice on CE-2 standard Japanese laboratory rodent chow. It's specified at around 6% LA acid of total calories so is borderline high, has no sucrose and is clearly functional as a lab chow. They had two intervention groups, for one they diluted CE-2 down with starch and for the other they diluted CE-2 down with sucrose. As they say:

"the latter two diets were prepared by the addition of corn starch or sucrose, respectively, to CE‐2 (Table 1)."

Here is how the diets turned out:

This simple action diluted the protein, fat and vitamin/mineral content as was clearly documented. I don't really buy in to the protein leverage hypothesis so I'm not surprised that all groups ended up at exactly the same bodyweight despite reducing protein calories for two of the groups.

What I do tend to take note of is the fat dilution. The fat is all of the same composition, around half of the 12% total fat calories is LA, giving 6% for CE-2. Diluting its 12% of total fat calories to 7.7% reduces LA from 6% of total calories to 3.7% of total calories for either intervention. Neither the high starch nor high sucrose diet produced any excess weight gain, much as you would expect for a either diet with LA at 3.7% of calories.

Now, their second paper is MUCH more interesting. Here it is:

Chronic high-sucrose diet increases fibroblast growth factor 21 production and energy expenditure in mice

and here are the diets:

There is no suggestion of grinding up CE-2 with extra starch or sucrose this time. These are custom diets of unspecified origin/composition (so I'm going off these people now). I'm going to assume NC (normal chow) is still CE-2 and that the 12% fat in each diet is still soybean oil.

The protein is reduced for both intervention diets (just as it was diluted in the previous study), carbohydrate sources and quantities are kept unchanged.

But the fat is NOT reduced/diluted. If we assume that the now maintained 12% of calories is still mostly soybean oil we are now at 6% of calories as LA in the intervention groups, up from 3.7% in the previous experiments. Back to borderline obesogenic. 

This time their very high starch diet is grossly obesogenic, their chow (lower starch, possibly less refined) is not and the high sucrose diet is positively slimming. And these sucrose fed mice are definitely not "skinny-fat", see the rest of the paper re insulin sensitivity. Okay, here's the insulin tolerance test, it's from the first paper, the ITT in the second paper is similar but doesn't drop quite as low because there is a little more fat in the diet. I'm impressed by the drop to 20% so I like this particular image!

You can't use an OGTT or IPGTT as a high sucrose diet down regulates hepatic glucokinase so a glucose tolerance test would (and does) generate systemic hyperglycaemic for a short period. And here are the weights:

All of the mice this time are on 6% LA (assuming I'm correct re composition). The grey circles are mice on a very high starch diet, they get fat. The open circles are on chow. We could describe them as also being on a 6% LA diet with modest carbohydrate restriction cf the high starch fed mice on the same LA percentage. They don't get particularly fat. In the presence of 6% LA even modest carbohydrate reduction appears to be slimming.

Then we have the diet with 38.5% of calories from sucrose as the black circles. These mice are slim, absolutely not insulin resistant and they happen to have hot brown adipose tissue secondary to uncoupled mitochondria through generation of FGF-21 secondary to fructose ingestion.

I do not have the mechanism for this. It will undoubtedly be driven by ROS generation and this will more likely be driven by NOX enzymes than mitochondrial reverse electron transport. I am also suspicious that modest sucrose ingestion might drive obesity as in the Surwit diets and D12451 but that a very high sucrose diet might uncouple respiration and prevent obesity and metabolic syndrome. Why?

Consider the parallels with PUFA diets with between around 8% and 30% of calories from LA. These fail to limit insulin signalling and allow excess calories in to a cell. At these modest levels of LA the inappropriate excess caloric ingress can be stored by the inappropriate transfer in to triglycerides. Both ingress and storage are both allowed and achieved by un-resisted insulin signalling.

At very high levels of LA, moving from over 30% of calories from LA to over 45%, then limitations become apparent to this stratagem of "diversion to storage". A more potent mechanism for ameliorating the pathological ingress of excess calories is is to uncouple respiration. A concept I explored here for PUFA.

What if sucrose does the same? Moderate levels, say 17%, allow fructose to enter cells without insulin mediated control. At this level I would expect fructose to generate enough ROS to limit insulin action by just the correct amount to down regulate glucose ingress to compensate for the energy from metabolised fructose.

If sucrose is very high, say 38.5% of calories, perhaps the unregulated fructose ingress cannot successfully offset enough of insulin's action on glucose ingress to balance the books. It looks like the solution in response to un-manageable caloric ingress might be to uncouple respiration so as to off-load the excess calories as heat.

Parallels between dose response to LA and fructose? Medium doses are dealt with by storage, high doses require uncoupling.

Some Surwit diets and D12451 combine maximally problematic levels of both fructose and linoleic acid.

These are things I'm thinking about at the moment.

Peter