By George Henderson and Grant Schofield (….gets technical, mainly due to George)
There’s been a bit written about NZO mice and how they are used to understand the physiology of diabetes, and how dietary compositon works amongst it all. We wrote about it here. What are we to make of all of this moving forward?
Animal experiments are useful when a question cannot be answered by tests performed on humans for ethical or practical reasons. For example, investigations into the causes of beta-cell failure in diabetes involve damaging the pancreas and examining it after death. Obviously this is neither practical, nor ethical in humans!
Other questions are answered by human experiments easily and often – does the LCHF diet lead to weight loss or weight gain, does it improve blood sugar and insulin levels?
It is not necessary to use animals to answer this question, because the experiments that answer it are generally beneficial to their subjects.
We know a great deal about what happens when humans, including both humans with diabetes and humans at risk of diabetes, consume particular diets in experiments.
Data from animal experiments should first be interpreted in ways that are compatible with, and if possible explain, what is known to happen in human experiments and in observed clinical practice.
The University of Melbourne’s LCHF diet study in New Zealand Obese mice (Lamont et al 2016), which made such a big splash a few weeks ago, was poorly controlled, and the results certainly didn’t support the wild claims (that an LCHF Paleo diet will cause rapid weight gain and the progression of diabetes) made in press releases and interviews by Professor Sol Andrikopoulos, the president of the Australian Diabetes Society, nor did they support the interpretation put on them in the paper itself.
Implying that it’s not safe to rely on 20 or so well-designed LCHF RCTs in humans, and much other relevant evidence, because NZO mice behave differently from humans (as they are designed to), fails to meet basic standards of scientific procedure and medical prudence.
However, Lamont et al (2016) cited several other studies in NZO mice, and one of those seems especially informative and well-designed.
“Previous studies in NZO mice have shown that if carbohydrates are completely removed from the diet, hyperglycemia and progression to diabetes can be avoided, but only while the mice remain carbohydrate-free. However, outside of a tightly controlled animal study, achieving and maintaining a diet that has absolutely no carbohydrates will be practically impossible. In previous studies, when NZO mice were later exposed to dietary carbohydrates, diabetes developed rapidly and was perhaps even more severe than what was observed in mice maintained on a normal rodent chow diet.”
The reference is mainly to this study;
Kluth et al. (2011) Dissociation of lipotoxicity and glucotoxicity in a mouse model of obesity associated diabetes: role of forkhead box O1 (FOXO1) in glucose-induced beta cell failure.
Conclusions/interpretation: The dietary regimen dissociates the effects of obesity (lipotoxicity) from those of hyperglycaemia (glucotoxicity) in NZO mice. Obese NZO mice are unable to compensate for the carbohydrate challenge by increasing insulin secretion or synthesising adequate amounts of insulin. In response to the hyperglycaemia, FOXO1 is dephosphorylated, leading to reduced levels of beta cell-specific transcription factors and to apoptosis of the cells.
In the mice that became obese (lipotoxicity) on the carb-free diet, re-feeding carbohydrate (glucotoxicity) rapidly produced beta-cell loss. This is interpreted as follows:
“It is generally accepted that beta cell malfunction is caused by the following scenario. Obesity induces ectopic fat accumulation in the pancreas, thereby causing apoptosis of beta cells (‘lipotoxicity’)… However, lipotoxicity does not appear to be sufficient for the destruction of the beta cell. Carbohydrate-restricted diets fully prevented beta cell destruction in both NZO and db/db mice despite an extreme insulin resistance and a marked inflammatory state of adipose tissue. This finding is consistent with a previously suggested scenario in which postprandial hyperglycaemia (‘glucotoxicity’) plays an essential role in the pathogenesis of islet cell failure. Hyperglycaemia produces glucotoxicity for the beta cell through oxidative stress caused by formation of reactive oxygen species.”
This is more sober language than seen in Lamont et al, and the diets were properly controlled, with both diets containing the same amount of casein (20%) and the same fats (palm oil with 0.5% each of safflower and linseed oil).
Kluth et al (2011) tells us about two separate mechanisms that are sequentially causal for beta-cell loss (this is called a “two hit” hypothesis). The first is ectopic fat accumulation; this is the mechanism that has been explored in humans by Roy Taylor’s team in Newcastle. Ectopic fat accumulation induces insulin resistance in both the liver and pancreas, meaning that a higher production of insulin is required to manage carbohydrate meals, and post-prandial glucose increases over time. Reductions in ectopic fat in the pancreas result in reversal of type 2 diabetes pathology in humans on calorie restricted diets; this has also been shown with regard to reductions in liver fat in people put on a low-carbohydrate diet in David Unwin’s practice (it is almost certain that liver and pancreatic fat track together in obesity and weight loss).
The NZO mouse becomes obese on a high-fat, carbohydrate-free diet, because it has a unique mutation in the gene encoding phosphatidylcholine transfer protein, which produces deficiencies in phosphatidylcholine metabolism. Wikipedia states that “PCTP is produced in all tissues in the body at various levels. The protein is expressed at high levels in tissues engaged in high metabolism, notably including the liver and macrophages. No human patients with defects in PCTP have been described to date.”
Thus the NZO mouse can shed more light on the effects of ectopic obesity than on its cause, which is clearly different to the causes of human obesity.
The NZO mouse does not have unique defects in genes encoding glucose metabolism. Thus its response to post-prandial hyperglycaemia is more likely to resemble the human.
This is the second part of the two-hit mechanism for beta-cell loss in Kluth et al (2011); the rise in blood sugar after feeding a higher carbohydrate diet to mice which already suffer from a high ectopic fat load produces oxidative stress (an excess of free radicals which overwhelms antioxidant defenses and peroxidises lipids in the mitochondrial membrane), triggering apoptosis, or cell death.
If we extrapolate from the NZO mouse study of Kluth et al to generalise about type 2 diabetes in humans, we arrive at the view that a diet (or indeed a medication regime) that produces obesity or ectopic fat accumulation, and hence pathological insulin resistance, will eventually sensitize the beta cells of the pancreas to a toxic effect of post-prandial hyperglycaemia, particularly in persons with a strong genetic risk for type 2 diabetes. The effects of weight loss in reversing diabetes are mainly due to a reduction in ectopic fat.[3,4,5]
RCTs to date show that the LCHF diet produces a reduction in body fat and improvement in glycaemic control at least comparable to other diets, and superior to low fat or recommended diabetes diets. Studies that specifically look at ectopic fat, such as the measurement of fatty liver markers in the pilot study of Unwin et al (2015), show improvement on a very low carbohydrate diet.[6,7]
A higher carbohydrate diet in persons with type 2 diabetes is less effective for body fat reduction and produces greater post-prandial hyperglycaemia.
Thus the LCHF diet, properly formulated, combines a reduced risk of ectopic fat accumulation with a very low risk of post-prandial hyperglycaemia. This explains why such a diet can prevent the progression of prediabetes to diabetes, as in Maekawa et al (2014).
It is plain that other diets can work, especially if ectopic fat accumulation (lipotoxicity) isn’t so advanced that initial post-prandial hyperglycaemia will rapidly result in glucotoxicity. If an individual finds a Mediterranean or low-GI diet (or a higher-carb Paleo diet) more satiating than a LCHF diet and is able to reduce calories and body weight better on it, then they shouldn’t be discouraged from following such a diet. But the question can be stated as, what should be the default diet for diabetes and pre-diabetes, that is, what is the first diet therapy that should be tried in the majority of cases?
It is plain that this should be the diet most likely to reduce body weight and ectopic fat, based on RCT evidence, and also the diet that will produce the lowest risk of post-prandial hyperglycaemia, based on feeding studies.
And that is at present the LCHF diet. Something else may come along and prove itself better – science always leaves room for new possibilities. But for now, it’s LCHF, and because food quality and nutrient density is also important, it’ll tend to be a more Paleo or Primal or Mediterranean version of LCHF (as in What The Fat?), and not the pure-fat-and-casein diet fed to the NZO mice (who essentially lived on nothing but low-carb cheesecake).
What is the worst diet for diabetes? Well clearly a processed food diet of deep fried food, processed grains, and sweetened drinks is the very worst, but standard diabetes diets don’t seem to be very good either. Diets that aim at 30-35% fat, 10% or less saturated fat, and that, though recommending whole grains, fruits and vegetables, have a permissive attitude to refined grains and “treat” foods because “you need carbohydrates for energy” and because it’s socially desirable for people with diabetes to “eat like everybody else” seem to be the worst possible medical diabetes diets. These diets are really the human equivalent of the LCHFD in Lamont et al’s NZO mice.
They are also the dietary guidelines for the rest of us.
 Lamont BJ, Waters MF, Andrikopoulos S. A low-carbohydrate high-fat diet increases weight gain and does not improve glucose tolerance, insulin secretion or β-cell mass in NZO mice. Nutr Diabetes. 2016;15(6:)e194. doi: 10.1038/nutd.2016.2.
 Kluth O, Mirhashemi F, Scherneck S et al. Dissociation of lipotoxicity and glucotoxicity in a mouse model of obesity associated diabetes: role of forkhead box O1 (FOXO1) in glucose-induced beta cell failure. Diabetologia. 2011;54:605–616.
 Sattar and Gill. Type 2 diabetes as a disease of ectopic fat? BMC Medicine. 2014;12:123
 Lim L, Hollingsworth KG, Aribisala, BS et al. Reversal of type 2 diabetes: normalisation of beta cell function in association with decreased pancreas and liver triacylglycerol. Diabetologia. 2011;54(10):2506-2514
 Steven S, Hollingsworth KG, Small PK et al. Weight Loss Decreases Excess Pancreatic Triacylglycerol Specifically in Type 2 Diabetes. Diabetes Care. 2016;39(1):158-165.
 Unwin DJ, Cuthbertson DJ, Feinman R, Sprung VS. A pilot study to explore the role of a low-carbohydrate intervention to improve GGT levels and HbA1c. 2015. Diabesity in Practice. 2015;4:102–8.
 Browning JD, Baker JA, Rogers T, Davis J, Satapati S, Burgess SC. Short-term weight loss and hepatic triglyceride reduction: evidence of a metabolic advantage with dietary carbohydrate restriction. Am J Clin Nutr. 2011;93:1048–52.
 Maekawa S, Kawahara T, Nomura R et al. Retrospective study on the efficacy of a low-carbohydrate diet for impaired glucose tolerance. Diabetes Metab Syndr Obes. 2014; 7: 195–201.