Ketomyths II

In the first post of these series I reviewed the evidence regarding mucin deficiency and glucose restriction. Its time to move on into a more complex topic: vitamin C and scurvy. The best article written on low carbohydrate diets and scurvy comes again from Dr. Jaminet. I will use some of his statements to explain my points of view on the subject. Before I start, I have to mention that ketogenic diets for epilepsy are not comparable to a paleoish ketogenic diet. It is too high in fat, too low in protein and the main source of fat are heart healthy vegetable oils high in n-6 PUFA. Most of the time patients are also water-restricted to increase blood ketone levels. Having said this, I will not discuss any example which uses an anti-epileptic ketogenic diet in the article (except the one on selenium deficiency, which is relevant for the post).

First, it is important to understand that ascorbate is asborbed in the small intestine modulated by glucose (1):

Na+ dependent glucose transport via SGLT1 receptors in enterocytes modulates the asborption of ascorbate, while transport of dehydroascorbic acid (DHA) is by facilitated difussion. As glucose content in the lumen and in the enterocyte increases, ascorbate transport is reduced. In layman's terms: ascorbate absorption is reduced by high glucose, so if you eat more glucose you will need to eat more ascorbate. This means, glucose increases your minimal requirement for vitamin C in your diet. 

But once ascorbate and DHA enter the bloodstream, they must be transported inside cells. 

"DHAA can be recycled back into vitamin C, but only inside cells. In order to enter cells, DHAA needs to be transported by glucose transporters. GLUT1, GLUT3, and GLUT4 are the three human DHAA transporters; GLUT1 does most of the work."

 Structurally, DHAA and ascorbic acid are similar to glucose:

Ascorbic acid, reduced form (left) and DHA, oxidized form (right)

This could be one of the reasons for the sharing of membrane transporters. The affinity of GLUT4 for DHA is low, so the main transporters are GLUT1 and GLUT3 (2), that is, glucose and DHA are both competitive substrates for some glucose receptors. High glycemia reduces the transport of DHA into cells. 

Jaminet states:

"Glucose transporters are activated by insulin. Thus, DHAA import into cells is increased by insulin, leading to more effective recycling of vitamin C [8]"

The reference takes us to a study done on osteoblastic cells (3). When these cells were incubated with insulin, the uptake of DHA and the intracellular concentration of ascorbate from DHA increased. But if you read look into the results, you will find: 

"Transport of [¹⁴C]DHAA was inhibited by D-glucose and 2-deoxyglucose, both of which are substrates of GLUT"

What? But isnt insulin, which is stimulated by glucose, supposed to increase DHA transport? Yes, but here is where interpretation problems arise. When we study cells in vitro we isolate the variable we want to study. This helps us understand the exact mechanism of action of certain molecule or pathway (because there is no other molecule/pathway influencing it) but this also creates a problem: our cells are not isolated. In fact, they are under many different convergent signals which might have sinergistic or antagonistic effects on one cellular process. If we focus only on the proximal cause (insulin increases DHA uptake) we miss the big picture for understanding metabolic processes (why insulin increases DHA uptake?). From my point of view, this is an evolutionary acquired mechanism. Vitamin C transport/recycling modulated by insulin/glucose can be illustrated using a different scenario: muscle protein balance. Insulin's role in different tissues is mainly inhibitory. In skeletal muscle, insulin inhibits muscle protein breakdown. But has a passive role on muscle protein synthesis, which is stimulated by plasma amino acid availability. On the other hand, the most powerful stimulus for insulin secretion, hyperglycemia, increases proteolysis (4). Hence, insulin counteracts the effect of hyperglycemia (until a threshold is reached) on muscle protein balance. This effect is also illustrated by glucose induced inflammation. Hyperglycemia is inflammatory (5, 6, 7) and contrary to popular beliefs, insulin is anti-inflammatory (8, 9). My guess is that glucose is the main stimulus for insulin secretion because it is inflammatory. Just like in skeletal muscle, insulin is released to prevent damage from glucose. 

Getting back to the topic, these examples help to understand why insulin might increase DHA transport and recycling. When there is a higher level of glucose competing with DHA, insulin's role is to increase the number of GLUT receptors for achieveing a normal intracelular DHA level for proper conversion to ascrobic acid (AA). 

There are other vitamin C transport systems used by cells, namely SVCT1 and SVCT2, two sodium dependent transporters for AA, which are regulated among others, by AA plasma concentrations. This transport is not regulated by insulin. So we have two ways by which vitamin C is transported into cells (10):

AA is transported inside cells by Na+ dependent transporters coupled to a Na+/K+ATPase, while DHA is transported through GLUT proteins. DHA is rapidly reduced to AA, which can excit the cell by different uncharacterized mechanisms (beyond the scope of this post).

Remember that DHA is the oxidized form of AA, which explains its antioxidant nature. AA is oxidized to the ascorbyl free radical (transfering of one electron to a metabolic oxidant) which is further oxidized, losing a second electron, producing finally DHA (this is illustrated in the second figure in the post, although simplified and the ascorbyl free radical is not shown). If there are more oxidants, higher levels of AA are needed to reduce them so there is an increase in the DHA/AA plasma ratio, as observed in diabetics. Regarding this issue, Jaminet says:

"Confirming the role of insulin in promoting vitamin C recycling, Type I diabetics (who lack insulin) have lower blood levels of vitamin C, higher blood levels of DHAA, increased urinary loss of vitamin C metabolites, and greater need for dietary vitamin C. [9, 10]"

The first mistake is comparing no insulin (such as in type I diabetics) with low insulin. But why do type I diabetics have lower levels of AA and higher levels of DHA? Is it because a defective recycling via lower DHA cell uptake? Maybe. But I dont think it is the only cause. Hyperglycemia, as mentioned earlier, produces oxidative stress, increasing the formation of reactive oxygen species which react with AA, oxidizing it to form ultimately DHA. Because there is a defective transport of DHA into cells (because of competitive inhibition by glucose and abscence of insulin), the balance is shifted towards DHA. The main source of mitochondrial ROS seems to be complex III during the oxidation of complex I substrates (NADH dehydrogenase) (11). Without going into details (maybe on a different post), complete (aerobic) glucose oxidation produces a ratio of NADH to FADH of 5:1. This means that mitochondrial energy metabolism relies more on complex I than complex II, increasing the production of ROS. This scenario is also observed in type II diabetics (12, 13), who in fact have the opposite (chronic hyperinsulinemia), but glucoregulatory alterations. Despite having high insulin levels, they have low plasma vitamin C. Increased ROS production both by systemic inflammation, hyperglycemia, carbohydrate-based mitochondrial energy production and defective cellular transport are the likely causes of vitamin C alterations observed in these patients. In uncontrolled TIDM, abscence of insulin makes things worse. This is by no means applicable to ketogenic diets.

"Dehydroascorbate, the fully oxidized form of vitamin C, is reduced spontaneously by glutathione, as well as enzymatically in reactions using glutathione or NADPH. [11]"

Then, Jaminet states: 

"Glutathione is recycled by the enzyme glutathione peroxidase, a selenium-containing enzyme whose abundance is sensitive to selenium status. One difficulty with zero-carb diets is that they seem to deplete selenium levels."

For this assupmtion, he references a study on a sudden cardiac death on a ketogenic diet for seizure control (14). From the cited study:

"Selenium is an essential nutrient in the human diet. Sources of selenium include cereals, meats, and fish [8]. However, depending on the intake of those foods and the source of cereals (e.g., soil rich or poor in selenium), selenium levels in humans are variable. Patients on the ketogenic diet have little cereal intake, and only moderate fish and meat intake, and thus are predisposed to low selenium levels."

Even I said I wasn't going to refer to seizure-control ketogenic diets, its worth mentioning that according to the last recommendations of the International Ketogenic Diet Study Group (15), the "normal" diet used in this cases is 90% fat and 10% of carbohydrates and protein combined. They state that "Calories are typically restricted to 80%–90% of the daily recommendations for age". Not very similar to popular paleo-keto diets. To achieve this incredibly low level of protein, they have to restrict important sources of Se, specially those eaten in abundance by low carbers like meat and fish, being the former one of the foods with best Se bioavailability (16). The recommended Se intake seems to be around 40ug/day (17) and muscle meats, on average, have 0.3-0.4mg Se/kg (organ meats such as liver and kidney concentrate more Se, from 4 to 16 fold the amount on muscle). Remember that 1000ug = 1mg, so 0.4mg = 400ug. If you eat 100g of meat a day (excluding other animal sources) you are eating 40ug of Se, the recommended intake (for reference, 100g equals to 3.5oz/0.22lb).  

Jaminet concludes:

"So here we have a second mechanism contributing to the development of scurvy on a zero-carb diet. The diet produces a selenium deficiency, which produces a glutathione deficiency, which prevents DHAA from being recycled into vitamin C, which leads to DHAA degradation and permanent loss of vitamin C."

As we have seen, using a clinical case based on a ketogenic diet for seizure control leads to a flawed conclusion. 

Finally, as said by one of the PHD readers in the comment section, ketogenic diets have shown to increase GSH levels (18).

The hypothesis proposed by Jaminet is not supported by facts, only by assumptions based on misinterpretation of the existing data. Proper ketogenic diets dont produce vitamin C, glutathione or selenium deficiency. Loss of glucoregulation does.

Quick on starch, ketosis and non-oxidative glucose disposal

Post-exercise ketosis is a phenomenon rarely discussed these days. But there is some considerable number of research done on this subject. A very interesting study done in the 80's tested the effects of alanine, glucose and starch ingestion on starvation ketosis and post-exercise ketosis (1).

"The effect on the blood ketone body concentration of a 100 g oral dose of either alanine, glucose or starch was studied in forty-four healthy men. Twenty of the subjects were highly trained long-distance runners who underwent 'glycogen stripping' as previously described by us (Koeslag et al. 1980). Twelve non-athletic subjects fasted for 65 h before the test, and twelve were studied on a normal day after a normal breakfast."

Trained athletes were tested after running 2 hours. They ate a low carbohydrate diet for 48h prior to the experiment. The starvation group was composed of 12 non trained subjects, fasted for 65 hours. 

Results

The arrow shows the point in which the different solutions were ingested.

"The ingestion of glucose or alanine after exercise caused the mean blood ketone body concentration to fall to less than 0 5 mmol/l in 2 h. The fall was more prompt, less variable and longer lasting after 100 g alanine was ingested than after 100 g glucose. At 15.00 the mean blood ketone body concentration was rising again in the subjects who had taken glucose, but not in those who had taken alanine. The difference between the mean blood ketone body concentrations of the two groups at 15.00 is statistically significant (P < 0-01).

(...) Starch ingestion caused the blood ketone body concentrations to fall, but to a lesser extent than after alanine or glucose ingestion. After starch ingestion, as after glucose, the mean blood ketone body concentration was rising again at 16.00 (Fig. 1), thus reaffirming the evanescence ofthe antiketogenic effects of carbohydrate administration."

So the antiketogenic effect was alanine > glucose > starch. We are talking about 100g of each, after either 65h of fasting (almost 3 days) and running 22km plus a low carbohydrate diet. This is by no means a "typical" scenario. But it shows us that starch is less antiketogenic than glucose. Evolutionary reasons perhaps? 

Extrapolating the findings to real life situations, there shouldnt be much of a problem about getting back to ketosis consuming carbohydrates only post workout, specially if you follow a very low carbohydrate diet, fast daily and train in a fasted state. This is the winning combo for minimizing glucose oxidation and maximizing non-oxidative glucose disposal after a glucose load. 

Studies have shown that adaptation to a high fat-low carbohydrate diet produces a shift in glucose metabolism, reducing glucose oxidation and increasing glycogen synthesis and glucose storage. A similar metabolic response is triggered by short term fasting and resistance training. 

Further reading on high fat diets and glucose metabolism:

Dietary fat content alters insulin-mediated glucose metabolism in healthy men.

High-fat/low-carbohydrate diet reduces insulin-stimulated carbohydrate oxidation but stimulates nonoxidative glucose disposal in humans: An important role for skeletal muscle pyruvate dehydrogenase kinase 4.

Low-carbohydrate diet alters intracellular glucose metabolism but not overall glucose disposal in exercise-trained subjects.

Carbohydrate refeeding after a high-fat diet rapidly reverses the adaptive increase in human skeletal muscle PDH kinase activity.

Koeslag JH, Noakes TD, & Sloan AW (1982). The effects of alanine, glucose and starch ingestion on the ketosis produced by exercise and by starvation. The Journal of physiology, 325, 363-76 PMID: 7050344

Ketomyths

I recently started a discussion with Dr. Harris, author of PaNu. If you have read my posts, you will see that I refer to his website many times because I share almost all of his ideas. But one critical topic in which we differ is ketosis. I find it very interesting, considering that Peter also avoids ketosis permanently, and both of them are the two bloggers who I respect the most.

The answer posted by Dr. Harris lead me to find many ketomyths in the "paleosphere" and "blogosphere". Although most don't think ketosis is dangerous, the common perception is that "more is not better", that is, being in ketosis 24/7 is not better than being only for short periods of time. According to Harris, there is no need to if you aren't sick. The main argument is that ketosis is not the natural state of the body and represents a metabolic stress. Keep in mind I never said that we should be in ketosis 24/7, I just said that ketosis should be the baseline physiological state (so most of the time you are in, but not necessarily 24/7*). 

Initially, I was going to address the points made by Dr. Harris in his post but I found one ketomyth that has spread all over the internet which people use as an argument against glucose deprivation. Before I go on I must say that zero carb diets are only theoretically possible, unless you eat only yolks and oil. 

It all started with the incidence of gastrointestinal cancers in long time OD's that were reported on Peter's blog. People started speculating on the origin of this issue trying to understand why the "Optimal" diet was not that optimal. The Jaminet's, authors of the Perfect Health Diet, posted an article attributing mucin deficiency as the cause. I think this can be a likely cause but only in OD dieters. The OD is characterized by restricting both protein and glucose. But somehow people misinterpreted this and started using it as an argument against all ketogenic diets. 

But what exactly determines GI mucin production? Mucin biosynthesis is highly sensitive to protein malnutrition, specially threonine deficiency (1,2,3). Other aminoacids like serine, proline and cysteine can also promote mucin synthesis (4). When intestinal epithelial goblet cells are deprived of glucose, butyrate modulates MUC gene expression and becomes the main regulator of mucin synthesis (5). There is no evidence that the abscence of dietary glucose affects mucin synthesis. So any deficiency of mucin in ODs would occur because of a low protein intake, not because of a glucose deficiency (call it a secondary-glucose deficiency if you wish). 

Thr, Ser, Pro and Cys are all glucogenic aminoacids. The demand for these amino acids increases when eating a very low carbohydrate diet because gluconeogenesis is increased to maintain a normal blood glucose level. Eating a diet low in these amino acids could compromise mucin production because they are redirected to glucose production instead of other biosynthetic pathways. 

Dr. Jaminet goes on and explain:

Throughout my 2 years on this zero-carb diet, I had dry eyes and dry mouth. My eyes were bloodshot and irritated, and I had to give up wearing contact lenses. Through repeated experiments, I established that two factors contributed to the dry eyes – vitamin C deficiency and glucose deficiency. After I solved the vitamin C issue, I did perhaps 50 experiments over the following few years, increasing carbs which made the dry eyes go away and reducing them which made them immediately come back. This established unequivocally that it was a glucose deficiency alone that caused the dry eyes.

I find this last line specially interesting, considering that bOHB has shown to be protective in different types of dry eye models (6, 7), although used as eye drop. 

What about Vitamin C? Thats the topic for the next post. 

I am not trying to put down Dr. Jaminet's work. On the contrary, I find it highly valuable, specially the emphasis he has on micronutrition. I also like n=1 experiments. But I think glucose deficiency is a misnomer. Think about protein deficiency and dehydration instead. These problems should not arise if you eat a proper amount of protein, drink enough water during the day and eat a fairly good amount of sodium (is salt paleo?) and potassium.