A number of people have asked for my comments on the recent headlines claiming that “high-fat diets cause diabetes,” based on a recent paper published in Nature Medicine (1):
Oshtsubo K, Chen MZ, Olefsky JM, Marth JD. Pathway to diabetes through attenuation of pancreatic beta cell glycoslyation and glucose transport. Nat Med. 2011; Aug 14 [Epub ahead of print].
Robb Wolf and Denise Minger have already critiqued this study, but I have a few things to add about the potential for our good friend glutathione to protect us from diabetes.
It’s Not About the Diet . . .
For those who are concerned that this paper might indicate that traditional, nutrient-dense fatty foods are bad for us, I would echo what Denise has already written. “When it comes to studies like this one,” our approach should always be to “white out the headline” and “read with an open but critical mind.” In this case, we can safely ignore the media headlines and observe that even the title of the paper says nothing about high-fat diets. The authors plant a few token sentences in the paper about the ability of a “high-fat or Western-style diet” to cause obesity and thus predispose someone to diabetes, but the paper is fundamentally about mechanism, not diet.
The investigators fed mice either a “high-fat” or “standard” diet. Both diets contained about 12 percent of calories as maltodextrin, which is a string of an average of ten glucose molecules hitched together like links on a chain, and is recommended for use in these diets to aid in pelleting and to reduce heat damage during the pelleting process. Both diets used casein for protein, a small amount of soybean oil to provide essential fatty acids, and a collection of purified vitamins and minerals. Both of the diets contained hydrogenated coconut oil and sucrose. The only difference between the two diets is that the “high-fat” diet contained an extra 47 percent of calories from hydrogenated coconut oil and the “standard” diet contained an extra 47 percent of calories from sucrose. Since fat packs in more calories per gram than sucrose, the high-fat diet was also 37 percent richer in calories.
It should be evident by now that neither of these diets contains any food.
In “They Did the Same Things to the Lab Rats That They Did to Us,” I discussed the emergence of these purified diets in the 1970s, when they first began to replace the cereal-based (i.e., “food”-based) rodent diets of yesteryear. The animals consuming the new diets developed fatty liver, excessive bleeding, kidney calcification, and greater vulnerability to stress, toxins, and carcinogens. Leading scientists have revised and improved the diets to curb their worst effects, but they still don’t contain any food.
Purified diets are advantageous to scientists because they make it easy to control for single variables and to make comparisons between different studies using the same diets. At the same time, the studies become irrelevant to the dietary choices faced every day by human beings. When we decide whether to choose a fat-rich egg yolk or a sugar-rich orange, for example, our choice has a profound impact on our intake of choline, vitamin C, and dozens or hundreds of other chemicals. Rodents consuming purified diets don’t face these effects because the diets contain a standardized amount of each vitamin and mineral and lack hundreds of other substances found in natural foods.
Even if this were not the case — that is, even if we humans were to eat nothing but casein, sugar, refined fats, and multivitamins — we would still be faced with the fact that we are not and never will be mice.
Quite often we find that certain disease processes exhibit remarkable similarities across species but that the dietary factors that can induce those diseases do not. For example, hypercholesterolemia produces atherosclerosis in baboons, monkeys, cats, mice, chickens, parrots, chimpanzees, pigs, dogs, pigeons, goats, rats, guinea pigs, and hamsters, but some of these animals such as rats and dogs are remarkably resistant to the effects of dietary cholesterol. This would suggest that there might be something universal about hypercholesterolemia that is worth studying, but also that it would be profoundly foolish to generalize dietary factors from rabbits to humans if we can’t generalize them to rats or dogs.
As Denise Minger already pointed out, and as we will see in more detail below, the effect of high-fat refined diets on obesity is difficult to generalize even from one strain of mice to another; it would be all the more profoundly foolish, then, to generalize the effects from mice to humans.
This study provided no evidence that diets made from casein, sugar, and hydrogenated coconut oil cause diabetes in humans, but it did show that some of the mechanisms they observed in mice were active in human cells, so let’s take a look and see if we can learn anything that might be of interest to those of us who walk on two feet.
. . . It’s About the Mechanism
The authors concluded from their work that the following scenario is likely to be involved in the development of diabetes:
- Dietary and genetic factors lead to obesity and insulin resistance.
- These conditions elevate concentrations of free fatty acids, which in turn cause insulin resistance to worsen.
- In the pancreas, elevated free fatty acids decrease the ability of two proteins called FOXA2 and HNF1A to travel to the nucleus and bind to DNA.
- When these proteins fail to bind to DNA, the cell fails to make an enzyme involved in processing glucose transporters and stabilizing them at the cell surface.
- The loss of glucose transporters at the cell surface impairs the ability of the pancreas to sense glucose levels in the blood and adequately stabilize these levels. As a result, we lose glucose tolerance and type 2 diabetes begins to emerge.
The investigators obtained pancreatic cells from a very small sample of humans (six healthy donors and two diabetics) and found some evidence to support the belief that this basic mechanism is active in humans just as it is in mice: the cells taken from diabetics had lower binding of the same proteins to DNA, lower production of the same enzyme, and 80-90 percent loss of glucose transporters at the cell surface. Moreover, the researchers were able to reproduce this mechanism in healthy human cells by incubating them with palmitic acid, a free fatty acid.
These results suggest that we may be able to learn something useful to our understanding of human diabetes from these mouse experiments.
What’s With Those Funny Mice?
The authors of this study uncovered a little hint that oxidative stress is critically involved in the aspects of the disease process they studied and that the master cellular antioxidant, glutathione, may offer complete protection against the loss of glucose tolerance.
First, let’s consider the strain of mice they used, C57BL/6J. This strain is often called “B6” and the “J” indicates these particular mice are a substrain from Jackson Labs, so I’ll call them “B6/J” from hereon out.
The diets these researchers used were named “Surwit diets” after Professor Richard Surwit of Duke University, who had helped Research Diets formulate the diets in 1992. Surwit had shown in the 1990s that B6/J mice are uniquely vulnerable to obesity and diabetes when fed high-fat diets, and that the inclusion of sugar in these diets has little or no effect.
Surwit’s team published several papers comparing the effects of different diets in B6/J mice to the effects of the same diet in A/J mice, which are highly vulnerable to hearing loss and cancer but resistant to obesity, diabetes, and atherosclerosis. I’ve blogged about these studies before. Here is what happened when the team fed these mice high-fat or low-fat diets with or without sugar for four months (2):
If you need to enlarge this or any of the subsequent pictures, you can do so by clicking on it.
The fat in this study was hydrogenated coconut oil. We can see that the high-fat diet increased fat mass in B6/J mice to a much greater degree than it did in A/J mice. Sugar increased adipose mass even further in B6/J mice fed high-fat diets, but reduced adipose mass in B6/J mice fed low-fat diets. It had no effect in A/J mice at all. We should keep in mind that this is absolute fat mass, and not fat mass as a percent of body weight.
The Surwit team repeated the study a few years later, but excluded maltodextrin from the diet and expressed fat mass as a percent of body weight. Here we see similar effects of fat, but no effect of sugar at all (3):
In the earlier study (2), Surwit’s team also measured glucose and insulin levels. The vulnerability of B6/J mice to diabetes was even more apparent than their vulnerability to obesity:
Here we see that a high-fat diet increased fasting glucose and dramatically increased fasting insulin in B6/J mice, while sugar had no effect in these mice. At the same time, neither fat nor sugar had any effect on either variable in A/J mice at all.
Why would B6/J mice be so vulnerable to the effects of a high-fat diet? Well, it turns out that both strains of mice have a higher caloric intake on this diet, so it’s not that.
There are likely numerous genetic differences between these strains of mice, but one particularly interesting difference was discovered in 2006 (4). It turns out that B6/J mice — but not other sub-strains B6 mice or non-B6 mice — have a deletion in the gene that codes for a critical enzyme in glutathione metabolism. As a result of the deletion, the mitochondria of these mice are unable to regenerate a particular form of the B vitamin niacin, and thus have a seriously impaired ability to recycle the master cellular antioxidant glutathione under conditions of oxidative stress.
Glutathione To the Rescue?
Now, how might this relate to the fat-and-diabetes study we’re trying to pick apart at the moment? Well, these researchers also used B6/J mice. And when they induced the disease process in healthy cells by incubating them with palmitate, they showed they could block the effect entirely if they also incubated the cells with N-acetylcysteine (NAC):
The black bars show that palmitate reduced the ability of these two critical proteins to bind to DNA, while the gray bars show that palmitate no longer had this effect if the cells were simultaneously given NAC.
NAC is best known as a precursor to glutathione, and is used for this purpose not only in experimental science but even in clinical practice. Depletion of glutathione is essential to the mechanism of acetaminophen (Tylenol) toxicity, for example, which accounts for half of all cases of acute liver failure in the United States and Great Britain. Administration of NAC is a highly effective remedy in the first ten hours of overdose, and it is believed to act at least in part by restoring glutathione levels (5).
The authors of this study did not measure glutathione levels, but the hypothesis that glutathione is protective is consistent with a study I wrote about in a post back in January, “Eating Fat and Diabetes.” In that study (6), high-fat diets depleted glutathione and impaired insulin sensitivity and glucose tolerance in rats and mice, but treating the rats with a mitochondrial antioxidant and genetically engineering the mice to make lots of the antioxidant enzyme catalase both reversed these effects. Catalase is an enzyme that converts hydrogen peroxide to water.
How Does This All Fit Together?
When mitochondria are overloaded with more energy than they can handle, they begin making increasing amounts of the free radical superoxide. Superoxide carries out important signaling roles. Among them, it directs excess energy into fat synthesis. But it can also wreak havoc on the cell by forming oxidants that can damage vulnerable proteins, lipids, and other important molecules. Thus, a manganese-dependent enzyme called superoxide dismutase converts it into hydrogen peroxide. Hydrogen peroxide can also damage important molecules, but increasing evidence suggests it also regulates the activity of hundreds of proteins by controlling several “redox switches,” including glutathione.
An interesting picture begins to emerge as a working hypothesis:
- When the mitochondria’s capacity to burn lipids and fats in order to make ATP is overloaded, it makes signals such as superoxide that will redirect incoming energy to be stored as fat.
- Superoxide also generates hydrogen peroxide, which oxidizes glutathione and thereby flips a “redox switch” controlling a multitude of proteins. These proteins may then help the cell stop responding to insulin in order to minimize energy overload.
- This is a desperate attempt of the cell to protect itself from oxidants that would otherwise destroy its basic machinery, and has the unfortunate consequence of increasing glucose and other forms of energy in the blood, and thus contributing to the metabolic abnormalities we associate with diabetes.
- Supporting the cell’s antioxidant defense network helps it to handle more energy and thereby protects against this entire process. Thus, providing NAC to cells, synthetic mitochondrial antioxidants to rats, or extra catalase to mice all seem protect against the development of diabetes-like features in the face of energy overload.
In my view, the major causative factor in this pathogenic process is energy overload. This, however, does not simply mean “excess calories.” It means that calories are supplied in excess of our body’s capacity to burn them. This capacity is increased by exercise, optimal thyroid status, lack of infection and inflammation, and a variety of vitamins and minerals involved in energy metabolism and antioxidant defense.
If this is correct, does a “high-fat diet” cause diabetes? The obvious question that must follow is “which high-fat diet?” An anti-inflammatory, invigorating, nutrient-dense diet likely protects against diabetes regardless of whether it is low or high in fat.
For more on glutathione, see “The Biochemical Magic of Raw Milk and Other Raw Foods: Glutathione.”
Read more about the author, Chris Masterjohn, PhD, here.🖨️ Print post
Stan (Heretic) says
(reposting after WordPress page script crash – delete if duplicated!)
What you wrote makes a lot of sense and corroborates certain facts related to a high fat low carb nutrition from my own experience.
I have always wondered why is a high fat diet so much less forgiving against overeating than a mixed diet (like SAD) or a high carb low fat diet.
Symptoms of overeating on a HF diet are immediate (within a few hours) and unpleasant – nausea and a headache. Good think about it is that HF diets are intrinsically very hard to overeat because the way our digestive system works, a HF meal usually produces a strong and reliable satiety feedback within a short time, in most people. Those people that have this mechanism broken – may and do experience problems with any high fat diet. (Note: a good article about satiety, fat and carbs is linked here )
“The animals consuming them (cereal-based diets) developed fatty liver, excessive bleeding, kidney calcification, and greater vulnerability to stress, toxins, and carcinogens.”
You’d think that would tell them something about wheat! Guess not.
Great post, Chris.
Do you know of a similar study in which the fat used was a natural fat (not hydrogenated)? I’m wondering whether the fat being hydrogenated puts extra stress on the mitochondria during its metabolism and if this excessive stress might have an additive/synergistic effect to the energy overload.
Christopher Masterjohn says
It was the animals consuming the purified diets that developed these problems. I’ll fix the sentence to make that clearer. “The animals consuming the new diets developed…”
Virtually any fat will work, but fish oil seems protective. The most common fat now is lard, though I suppose how ‘natural’ it is by the time it is processed into the diet, and considering where it probably came from, might be somewhat ambiguous. On the whole, high-fat diets seem to be a stress for lab rodents, though their natural diet is arguably based on cereal grains. There is also quite a bit of variability in the response, so there’s probably some publication bias.
Jack Kronk says
I take a cold processed pure whey protein isolate from BlueBonnet about 5-6 days per week. It’s got only 2 ingredients: whey and vanilla. Regarding glutathione, does the intact of the 3 key aminos convert to glut successfully.. and if so, might this be a strong anti-oxidant factor with respect to LDL oxidation, or am I completely out in left field with that idea?… as in… antioxidant intake does not actually affect the oxidation of LDL particles.
Christopher Masterjohn says
The information about glutathione-boosting is actually from low-temperature processed undenatured whey protein. So yes, it is active if it is the good stuff. I would judge that by asking the company for measurements of the sensitive proteins that appear to act as precursors that I discussed in the “biochemical magic” article. I think the activity of truly raw proteins as found in raw milk is likely to be even higher but it hasn’t been directly tested. Yes, it makes sense that boosting glutathione status would help protect against LDL oxidation.
Jack Kronk says
Thanks for the reply. I had a bit of a computer black out but I’m back now. I foolishly clicked “Ok” on a pop up to authorize a certain program to ‘modify’ my settings. I guess I didn’t realize that ‘modify’ meant to promptly crash Windows 7 and destroy my pc’s ability start up… Fun times.
Anywhoo… I am wondering if I have too much antioxidant power in my diet. Is that possible. I know oxLDL is a potential risk for heart attack, but isn’t it true that if the LDL is never oxidated that it stays in the blood longer and if more LDL particles are created by eating fat, LDL raises by default? Does this mean than an non-oxidative diet by nature can cause higher LDL?
Jack Kronk says
Chris – My sincere apologies for double posting (and with a diff topic too) but I cannot for the life of me post on the correct LDL receptor article (yoru brand new one). It keeps blocking me. This is intended post for that article)….
Chris – Jack Kronk here. (from PaleoHacks) Thanks for this write up… a topic I’m quite interested in, obviously. It seems likely that I suffer from poor LDL receptor activity. Could it be that’s what’s causing my HDL to be so low at 40, or is it the reverse of that… that my HDL being so low contributes to the cause of the poor LDL receptor activity, causing mostly small dense LDL?
Also, what’s the best way for me to test for proper/poor thyroid function? I am finding that testing for certain things is not readily available and it’s pretty frustrating. I live in San Diego so I’d guess that I probably have several good options available to me. Do you know of a certian lab that can test me for Thyroid health? Should I also test for ApoE?
Christopher Masterjohn says
I’ll post this over on the thyroid/LDL post. If you still can’t post there, you can bring any replies back here and I can copy and paste them.
Jack Kronk says
I see your reply on your Daily Lipid site. I still can’t post over there and for some reason, my formatting is coming out super skinny here and I can’t change it. sorry 😮
I haven’t gotten the results yet, but I just got tested Friday for:
iron, magnesium, potassium, selenium, calcium
full thyroid “cascade panel” (but I don’t know what that means)
Hopefully the results can shed some light into where I might need to make some additional changes.
Chris Masterjohn says
Thanks Jack, let me know when you get the results in. Hopefully we can get you posted on the Daily Lipid!!!
Great stuff. I think it’s very interesting that the glucose transporter needs to be glycosylated to stay at the cell surface. Glycosyltransferases are activated by manganese, all of them as far as I know. Kind of suggests a high-fat diet isn’t going to give you diabetes unless you have manganese deficiency.
My understanding of ‘energy overload’ is similar to yours with one difference. Palmitate apparently causes oxidative stress by activating NADPH oxidase. I mention this because as you know, I’m always going on about copper and manganese being the missing links in explanations of modern disease, and it’s copper that detoxifies the superoxide produced by NADPH oxidase. Manganese does it in mitochondria.
Christopher Masterjohn says
Thanks! In this study, the vitamin and mineral content was normalized to the caloric load, so if the animals ate more food, they ate more nutrients. However, it could still be that manganese was insufficient to deal with the extra energy. Nevertheless, I think this is a more regulated process than that — I think the cell is deliberately becoming insulin resistant. Time will tell as the mechanisms are mapped out.
That’s a good point about NADPH oxidase, but I don’t think it’s an either or. This recent study showed that H2O2 activates NADPH oxidase in primary B6 mouse islets:
…and the the rat/mouse/antioxidant/catalase study I cited in the blog post used a synthetic antioxidant they claimed was specific to mitochondria to abolish the negative effects of the high-fat diet in rats.
So I think it is quite plausible to say that energy overload often strikes the mitochondria first and that superoxide and H2O2 then initiate a signaling cascade that involves NADPH oxidase and iNOS. But I doubt it’s as simple as hitting the mitochondria first. I just think that evidence to date suggests that mitochondria are a major source of ROS.
John Cameron says
With regard to the “depletion of glutathione by high fat diets” it is noted that in a low fat (20%) diet,the insulin response to dairy fat is about half the insulin response to other fats with the same polyunsaturated:saturated fat ratio. The same study (pubmed 14652359)found that the response of the appetite suppressant CCK to dairy fat was 50 to 70% higher than the response to non dairy fats. Obviously, response to dairy fat can be greatly different to that of other fats.
There are many studies that have found that an inverse association between risk of type 2 diabetes and intake of dairy fat. One study of particular interest by Harvard (pubmed 21173413) measured circulating levels of trans-plamitoleic acid, a fat that is distinctly related to intake of whole fat milk and is not synthesized by humans so it can be used to estimate whole fat milk intake. It was found that risk of new onset type 2 diabetes in quintiles 4 and 5 of circulating trans-palmitoleic acid versus quintile 1, had a relative risk (RR) of 0.41 and 0.38. In other words, the risk of diabetes was about 60% lower in the two upper quintiles of whole fat dairy intake! Insulin resistance was 16.7% lower and C-reactive protein levels 13,8% lower in those who consumed the most dairy fat compared to those who consumed the least.
Another study (pubmed 20424220)found that those in the highest quartile of vitamin K2 intake had a 20% reduction in risk of type 2 diabetes. This study found that Crp levels were inversely related to K2 intake and speculated that the reduced inflammation could contribute to reduced risk of diabetes.
I eat a lot of raw milk cheese made from milk of pasture fed cows that has been made at temperatures no higher than 102 degrees F. I am curious whether glutathione present in raw milk survives the cheese making process and would appreciate you opinion.
Chris Masterjohn says
I’m familiar with those studies on dairy; thanks for posting them. Milk is very poor in glutathione. It boosts glutathione because it is rich in certain whey proteins that provide precursors to endogenous synthesis. I doubt these are present to a large degree in cheese because cheese-making, as far as I understand it, tends to discard the whey proteins. But I think what we would really want to see is an experimental study to test the effect of cheese-feeding on tissue glutathione concentrations.
Hi Chris, yes the different ROS systems all talk to each other. This seems to be to do with coordination of metabolic oscillations, so that ‘energy overload’ would equate to disruption of the oscillations. I remember a diabetes researcher called Steve O’Rahilly telling me many years ago that he could get insulin resistance in his rats any time he liked just by preventing oscillatory secretion of insulin.
This may have some profound implications. According to Ron Rosedale, if you infuse insulin into an artery you get atherosclerosis, and this is why he thinks insulin is bad and you shouldn’t eat carbs. But presumably, the insulin was infused continuously.
I have wondered for a long time whether the reason cells do metabolic oscillations might be at least partly so they can coordinate their normal functions with cycles of maintenance and repair. Could help answer thorny questions like, how do you repair your heart if it never stops beating?
Chris Masterjohn says
I always found that statement of Rosedale’s pretty empty. The concept of oscillations here is quite interesting though, and likely to b equate illuminative.
Glutathione = another reason to EAT YOUR BROCCOLI!
Chris Masterjohn says
Thanks, Wenchypo. I love broccoli.
Hey Chris, just reading through your old posts. Awesome as I remember them. I have done some research on lipid overload and insulin resistance and, surprise surprise, it’s the state of the machinery not the fact that one ate fat that is the problem. But I have only looked at the ability to burn more fat, so this is another great addition to the picture.
I would really like to see a post on that aspect of fatty acids and insulin resistance/untimely apoptosis. Like with PPARs and AMPK and UPC3 and leptin, vitamins and minerals, and all of that stuff controlling lipid metabolism to prevent overload. Of course that’s a big topic, but it’s one that people should know, because there are so many “high fat diet kills rat” scenarios that are reversed by nutrients like omega-3s, resistant starch, quercetin, etc. Exercise too.
Basically we want to keep inflammation in the body low and nutrition high, as usual. Omega-3s are interesting because they’re PPAR binding ligands and that enables long-chain SFAs (the evil ones) to be metabolized better. I’m sure you already know some of this. One of these days, maybe.
Chris Masterjohn says
Thanks for the suggestion Stabby! I don’t think we want to keep inflammation low — I think we want to provide what is necessary to have a robust inflammatory response when needed, and a quick and effective resolution to it.
Exactly, that’s what I meant. I’m not in favor of swallowing a bottle of ibuprofen 🙂
I believe that in one of your posts that you commented negatively on the Life Extension vitamin K supplement (K1-1000mcg, K2,MK4-1000mcg, K2,MK7-1000mcg)principally because of the very high K1 amount. I wondered why you were so negative about the K1 supplementation.
Chris Masterjohn says
There is some unpublished data mentioned in a collaborative review by some vitamin K experts indicating that it may promote periodontitis at supplemental intakes of 1 mg/day, which indicates to me that it could be pro-inflammatory or pro-oxidant in certain contexts — I don’t know why that would be the case, but I suppose it could be metabolized more rapidly into menadione, which is a pro-oxidant. In any case, the fact of the matter is that there are no high-dose trials with it verifying its safety, whereas there are with K2, up to 45 mg/day. So, there is no clear evidence that K1 is harmful, but there the evidence is not as clear as for K2 that high doses are safe. So I think it would be best to use a low dose of K1 and a higher dose of K2 if one is going to supplement with high doses, and I would prefer to err on the side of safety by saying let’s keep total K1 intake below 1 mg. Also, I had several anecdotes of people contacting me to tell me they got some problems from taking the Life Extension supplement and did not have problems with other K2 supplements. So I wish Life Extension would lower the amount of K1, especially given that many people might take more than one capsule. In any case, it is not a strong case I can make, it is a weak one, but I prefer to be cautious where it seems potentially warranted.