Dietary fat, CCK and the cholinergic antiinflammatory pathway

Background

One of the most important antiinflammatory pathways is mediated by the vagus nerve. Vagal stimulation increases the release of acetylcholine, which then interacts with the alpha 7 subunit of the nicotinic receptor on macrophages, activating the Jak2-STAT3 signaling pathway (1). This interaction inhibits macrophage activation and thus reduces inflammation. This pathway is known as the cholinergic antiinflammatory pathway (CAP)*. Stimulation of CAP has been shown to reduce TNF-a, IL-1b, IL-6 and IL-18 levels during endotoxemia (2) and NF-kB activation (3). The autonomic nervous system controls inflammation by the adrenergic pro-inflammatory pathway and the CAP (4):

Nature Reviews Immunology 8, 776-787

Although there are different ways to stimulate the CAP, most research has been done either using electrical stimulation (or acetylcholine receptor agonists) or vagotomy in laboratory animals. As nutrition is a determinant factor contributing or reducing inflammation, it seems plausible to speculate about the role of different nutrients modulating the CAP.  

Luyer et al (5) tested the ability of dietary fat to modulate inflammation, by stimulating the release of cholecystokinin (CCK). They induced hemorragic shock in Sprague-Dawley rats, in order to increase proinflammatory cytokines such as IL-6 and TNF-a. They fed them either a low fat or high fat enteral nutrition, or fasting. Additionally, high-fat fed rats were vagotomized (VGX) or sham vagotomized (Sham). Nutritional composition of both diets was as follows, as percentage of total energy:

High fat: 6.9% protein, 40.9% carbohydrate, 52.2% fat.

Low fat: 6.9% protein, 75.4% carbohydrate, 16.7% fat.

Proteins were derived from lean milk and carbohydrates from a mixture of sucrose and corn starch. The lipid source was vegetable oil, cointaining 8.1% SFA, 58.9% MUFA (57.4% oleic acid), 28.2% PUFA (23% linoleic acid). The amount of n-3 and n-6 in the high fat nutrition was less than 5% of the total fat content. They found that:

- High-fat enteral nutrition reduced hemorragic shock-induced TNF-a and IL-6 in Sham rats, compared to low-fat and fasted controls. Vagotomy nearly abolished the fat-induced reduction in these proinflammatory cytokines (TNF-a: 205 +/-11 pg/ml [VGX] vs. 5 +/-1 pg/ml [Sham]; IL-6: 80 +/-5 pg/ml [VGX] vs. 19 +/-9 pg/ml [Sham]).

Intestinal barrier function was assessed by a. bacterial translocation to distant organs, b. leakage of horseradish peroxidase (HRP) in isolated ileum-segments and c. plasma endotoxin levels. Increased intestinal permeability and impairment of gut barrier function is observed after induction of hemorragic shock. According to the reduction in proinflammatory cytokines, the high-fat nutrition reduced endotoxemia, permeability of ileum segments for HRP and bacterial translocation to distant organs, compared to fasted and low-fat Sham rats. Vagotomy reversed the protection of the high-fat diet, elevating plasma endotoxin levels (from 12 +/- 2 pg/ml to 28 +/-1 pg/ml), increasing leakage of HRP (from 1.1 +/- 0.7ug/ml to 2.3 +/- 0.5ug/ml) and bacterial translocation (from 16 CFU/g tissue to 328 CFU/g). 

Using CCK-A and CCK-B receptor antagonists (or vehicle) in high-fat fed Sham rats, they showed that the protection from a high fat nutrition was mediated by CCK, as inhibition of CCK-A and CCK-B enhanced plasma TNF-a and IL-6 after hemorragic shock induction, as well as endotoxemia, HRP permeability and bacterial translocation to distant organs, compared to vehicle animals. Moreover, administration of chlorisondamine abrogated the inhibitory effects of a high-fat nutrition on the parameters previously evaluated, suggesting that inhibition of inflammation was mediated by stimulation of nicotinic receptors by way of efferent vagal fibers. The authors proposed the following model for explaining their observations:

The author's description is as follows:

"Ingestion of high amounts of fat induces release of cholecystokinin (CCK) that binds to CCK-A and CCK-B receptors (CCK-r) located centrally or on peripheral vagal afferents. Activation of CCK-receptors triggers vagal efferents leading to an increase of acetylcholine (Ach), the principal parasympathetic neurotransmitter. Release of inflammatory cytokines such as TNF-a and IL-6 after activation of Toll-like receptors by bacterial products is inhibited by way of binding of acetylcholine to a-7 nicotinic (a7-nAch) receptors."

The findings of this study are remarkable, but as far as I can see, there was not much practical application after it was published. It isnt surprising. The same authors have published studies showing that a high-fat enteral nutrition protects the liver from the remote effects of hemorragic shock (high-fat treated animals had minimal liver injury, no evidence of mtDNA damage and significantly lower expression of stress proteins) (6) and exposure to bacterial DNA (7), and have proposed the utilization of a high-fat enteral nutrition after sever trauma to attenuate the inflammatory response (8) and to treat inflammatory conditions (9).

Short after the publication of this paper, Tracey (10) expanded the findings of Luyer et al., proposing fat-induced activation of the CAP for the treatment of inflammatory diseases:

But controlled trials using this information is lacking. How can dietary fat, which is supposed to be inflammatory, can be anti-inflammatory? I guess that most nutrition researchers just ignore the awkward. Even recent papers dealing with the subject and the hypothesis of vagal nerve stimulation for treatment of inflammatory diseases do not mention anything about a high fat diet increasing the CAP. Moreover, you can find gems like this one, from a paper published in Medical Hypotheses recently by Undurti N. Das (11), called "Vagus nerve stimulation as a strategy to prevent and manage metabolic syndrome":

"It is proposed that consumption of energy dense food leads to acute raise in plasma glucose levels that triggers increased production of IL-6, TNF-a, and IL-18 by peripheral leukocytes, monocytes and macrophages [11]. Simultaneously, gut produces cholecystokinin that, in turn, enhances vagal tone and induces the release of acetylcholine [39]. Acetylcholine and the acute raise in plasma glucose levels trigger the release of insulin from pancreatic b cells that decrease plasma glucose levels and inhibit IL-6, TNF-a, and IL-18 secretion and thus, homeostasis is restored. However, this regulatory system quickly fades in the face of continued ingestion of a fat-rich and/or energy-dense diet. Thus, fat-rich (especially saturated fat rich) and energy-dense foods promote insulin resistance, obesity, type 2 diabetes mellitus and the metabolic syndrome, in part, by impairing nutrient-sensing systems that exist in the gut, liver and hypothalamus that are originally designed to limit food intake and enhance insulin sensitivity."

I find the statement "fat-rich (especially saturated fat rich) (...) foods promote..." misleading to say the least. In this regard, CCK could be a previously uncharacterized indirect antiinflammatory molecule. At the moment, this is highly speculative because the overall inflammation balance depends on several factors and not only stimulation of the CAP. 

* Or "inflammatory reflex".

Luyer MD, Greve JW, Hadfoune M, Jacobs JA, Dejong CH, & Buurman WA (2005). Nutritional stimulation of cholecystokinin receptors inhibits inflammation via the vagus nerve. The Journal of experimental medicine, 202 (8), 1023-9 PMID: 16216887

ChREBP: The forgotten factor

The discussion between Taubes and Guyenet has revived the long-life debate about the influence of carbohydrates/insulin in the pathogenesis of obesity and MetSyn. I think both hypotheses (carbohydrate excess and food reward) are complementary to each other. The problems associated with Taubes hypothesis are obvious and had been reviewed elsewhere. The food reward hypothesis, on the contrary, is very well documented in the literature and most researchers have shown consistent results. The question is, can food reward be a dominant factor in obesity in the absence of sugar? Generally, people get obese consuming high energy diets both high in sugar and fat. If we exclude sugar from the equation, would then the same people get obese? The food reward associated with junk food would be the same in the absence of sugar (assuming that there is sugar-free junk food)? Is high sugar (and probably veggie oils) the ultimate cause of leptin resistance and food reward? Is palatability only relevant when talking about high fat-high sugar foods? 

I think this is why maybe the two hypotheses are complementary. I have yet to see any evidence of someone getting obese eating a low carbohydrate high fat diet. Sure, you can gain weight with a low carbohydrate diet, and you can regain the weight lost after a period of calorie restriction eating a low carbohydrate diet. But can you get obese in the absence of sugar? We will never see any formal evidence of this theory, but I think is a valid speculation.

Anyways, on to the topic of the post. As a nerd as I am, I like to look at physiological processes deep inside. When discussing about the role of sugar/carbohydrates in obesity, the transcription factor carbohydrate response element-binding protein (ChREBP) is hardly mentioned. This transcription factor binds to the carbohydrate response element (ChoRE) located in the promoter of target genes and stimulates transcription. ChREBP is a member of the basic helix-loop-helix/leucine zipper (bHLH/ZIP) family of transcription factors and its expression is ubiquitous, being most abundant in lipogenic organs such as liver, brown and white adipose tissues, small intestine, kidney and muscle. 

Insulin and glucose both coordinate the transcription of key enzymes involved in de novo lipogenesis and glycolysis, the former by activation of SREBP1c and LXR. Both regulate different pathways which are integrated in the response to a high carbohydrate load (in this case, in the hepatocyte):

Copyright © 2011, The American Society for Clinical Investigation.

At low blood glucose concentrations, ChREBP is located in the cytoplasm, phosphorylated at the Ser196 residue. When blood glucose levels rise, glucose enters the hepatocyte and is phosphorylated by glucokinase and then converted to xylulose-5-phosphate (Xu-5-P) in the hexose monophosphate shunt (HMS). Xu-5-P activates protein phosphatase 2A delta (PP2A delta) and dephosphorylates ChREBP*, which can then enter the nucleus and stimulate transcription by dimerizing with Mlx. On the other hand, insulin stimulates transcription of both ChREBP and SREBP1c. Although Xu-5-P has been proposed as the key regulator of ChREBP, glucose itself can activate ChREBP through its GRACE (glucose response activation conserved element) domain, by unkown mechanisms. ChREBP activity seems to be controlled mainly at the post-transcriptional level. 

Inhibition of ChREBP is mediated by phosphorylation of Ser196 (inactivating nuclear import) and Thr666 (preventing DNA binding) by PKA and AMPK. 

Target enzymes regulated by ChREBP include Liver Type Pyruvate Kinase (L-PK), the NADPH supply system (glucose-6-phosphate dehydrogenase, transketolase, malic enzyme, etc.), glucose 6 phosphatase (G6P), Acetyl CoA Carboxylase (ACC) and Fatty Acid Synthase (FAS). 

Studies with ChREBP knockout mice

Iizuka et al (1) tested the importance of ChREBP for induction of several enzymes implicated in glycolysis, fatty acid synthesis and de novo lipogenesis. ChREBP -/- mice had slightly elevated glucose and insulin levels, deposited a large amount of glycogen in liver (but not in muscle), had almost half the plasma FFA of wild type (WT) mice and less adipose tissue, when fed a standard diet. Compared to WT mice, the level of LPK mRNA in ChREBP -/- mice was only 27% of that measured in age-matched WT. ACL, ACC1 and FAS mRNA levels were also lower in KO mice. Levels of mRNA for malic enzyme showed the greatest reduction (59%). When mutant mice were fed a high-sucrose diet, plasma FFA were markedly reduced and mice experienced progressive hypothermia, culminating in death in less than 1 week (>50% of the ChREBP -/- mice). When fed a high-fructose diet, they became moribund in a few days. This last observation was explained by low levels of fructokinase and triose kinase. 

To induce glycolysis and lipogenesis, they fed the mice a high starch diet. This diet increased levels of blood glucose compared to the standard diet, in both WT and KO mice. Plasma insulin in ChREBP -/- mice fed the high starch diet was significantly higher than any other group. They were moderately insulin-resistant (assessed by glucose tolerance tests) and had 40% greater liver weights than WT (from increased glycogen storage). Despite having increased glucose and insulin levels, ChREBP -/- mice fed the high starch diet showed reductions in liver mRNA for ACL, ACC1, FAS, malic enzyme, SCD-1 and LCE. This resulted in hepatic fatty acid synthesis rates that were 65% lower compared to WT. LPK remained lower in ChREBP -/- mice even when fed the high starch diet, and Glut-2 mRNA was <10% of that measured in WT. Glucose 6-Pase and PEPCK were also reduced**. 

Inhibition of ChREBP expression in ob/ob mice in vivo using a RNA-interference technique improves hepatic steatosis by decreasing lipogenic rates, leading to decreased levels of triglycerides and NEFA, and improving insulin signaling in liver, skeletal muscle and white adipose tissue (2). Using a double mutant model (ob/ob ChREBP -/-) (leptin deficient-ChREBP deficient), Iizuka et al (3) observed that inactivation of ChREBP expression reduced fat synthesis and obesity (body weight was very similar between ChREBP -/-, WT and ob/ob ChREBP -/-), and improved glucose tolerance and appetite control in ob/ob mice. Thus, deletion of ChREBP was able to override some phenotypic characteristics of leptin-deficient mice. According to this, it has been suggested that reducing ChREBP might protect against beta-cell dysfunction in type 2 diabetes because it inhibits the expression of Pdx-1, MafA, GcK and insulin (4), as well as PPARa (5).

Pleiotropic properties of ChREBP

Although ChREBP is a key regulator of glycolysis, fatty acid synthesis, gluconeogenesis and de novo lipogenesis, it seems that it has other important roles. Yun-Seung et al (6) tried to identify ChREBP target genes and gene expression patterns using ChIP-seq. They treated HepG2 cells with 25mM glucose for 8 hours. They found 783 target genes involved in different pathways. The most enriched pathway was lipid metabolism, followed by gluconeogenesis, as suspected.  Nevertheless, there were other target genes that are associated with diverse functions, such as protein dimerization, embryonic development, among others. 

One very intesting study was published by Tong et al (7). They investigated the role of ChREBP in cancer cell proliferation and metabolism using HCT116 colorectal cancer cells and HepG2 hebatoblastoma cells. It was seen that these cells require ChREBP to maintain their proliferative state and is rapidly upregulated upon growth factor stimulation. Moreover, comparison between HCT116 cells transfected with ChREBP siRNA and without transfection showed that inhibition of ChREBP caused a reduction in glucose uptake and lactate production, and increased oxygen consumption rates. This reflects increased mitochondrial respiration and decreased aerobic glycolysis. Transfected HCT116 cells also showed a reduction in glucose flux through the pentose phosphate pathway and de novo lipid biosynthesis. These observations were confirmed by 13C NMR. RNA microarray analysis of transfected cells showed an increase in the expression of p21, MDM2 and TIGAR, all of which are p53-dependent targets. Although the level of total p53 was the same in ChREBP deficient and non-deficient cells, the level of p53 that was phosphorylated in Ser-15 increased as ChREBP expression declined, effect that was explained by increased ROS concentrations. Supression of ChREBP resulted also in G1 and G2/M arrest. The authors finally showed that when injected to nude mice, ChREBP knockdown cells formed smaller tumors in vivo compared with control cells. 

Relevance to humans

Using animal studies to propose novel hypotheses can be fun. However, we cannot extrapolate findings directly. This is important when discussing metabolic pathways and the effect of different diets. For instance, mice have a basal metabolic rate that is 7 times greater than humans, so a 40% calorie restriction in mice mimics a therapeutic fasting in humans (8).There is evidence that mRNA levels of several lipogenic enzymes are different between rats and humans. In general, the lipogenic capacity of adipose tissue is lower in humans than rats, which may be explained by decreased ChREBP mRNA expression and lower abundance of SREBP1c (9). But it is interesting to note that expression of ChREBP differs in adipose tissue and liver. In obese subjects, hepatic ChREBP expression increases while decreasing in adipose tissue, and there is a significant increase of ChREBP mRNA and protein levels during preadipocyte differentiation (10). Differences between lean and obese subjects are shown below (A) in liver, omental and subcutaneous fat, as well as expression versus lean liver (considered as 1) (B).

We can see that lean people have less levels of ChREBP mRNA in liver, both have higher levels in both omental and subcutaneous adipose tissue. Because we know that we cant make conclusive statements only with mRNA levels (see here), we must look at protein levels to see the full picture:

Western Blot analysis of hepatic tissues of both lean and obese subjects (A) revealed the pressence of ChREBP (95 kDa) in obese but not in lean samples. Comparison to b-actin is shown in B. Protein concentration could not be detected in adipose tissue samples (indicating low absolute values and correlation between ChREBP mRNA and protein abundance). If there is a relevance for ChREBP and obesity-MetSyn in humans, we must look then at the liver, which is central to the disease. 

We have seen in studies mentioned above that inhibition of ChREBP reduces hepatic steatosis, confirming the role of lipogenesis and de novo lipogenesis in the process. It is often stated that de novo lipogenesis is very low in humans and has no physiological relevance. I differ. It might be not relevant to adiposity directly, but indirectly by promoting hepatic insulin resistance and steatosis (11). Maybe this is why very low carbohydrate diets work very well treating NAFLD (12,13), as hepatic DNL is increased with a high carbohydrate-low fat diet (14). This should also amielorate hepatic insulin resistance and subsequently hyperglycemia, not only by decreasing dietary glucose, but by restoring hepatic insulin signaling. A high fat diet should also reduce ChREBP liver activity, thereby reducing hepatic steatosis (15).

Summing up

There is not much human research with ChREBP but until now, the evidence supports its importance in the pathogenesis of obesity and MetSyn. For me, it seems that a chronic high sugar diet could influence development of obesity and MetSyn by chronic stimulation of ChREBP. This creates an obsogenic environment in which lipids begin to accumulate in the liver, leading to loss of hepatic insulin signaling. ChREBP overexpression in beta-cells could also contribute to insulin resistance and diabetes onset. These effects could be further exacerbated by excessive fructose consumption, as fructose has been shown to increase ChREBP DNA binding (16). This might explain why some high carbohydrate diets (ie. paleoish) do not cause obesity or MetSyn, compared with high sugar diets. Gut flora's effect on lipid metabolism and obesity might also include upregulating hepatic ChREBP, as shown with conventionalization of germ-free mice with normal microbiota from conventionally raised animals (17).

In conclusion, ChREBP might be the connection between high carbohydrate/sugar diets and obesity/MetSyn, as it promotes lipogenesis. This transcription factor directly links glucose/fructose to metabolic dysregulation and probably other "kind" of diseases, like cancer. The ability to improve the metabolic state of ob/ob mice suggests that a dietary treatment which reduces ChREBP activity should be the default treatment (at least initially) to MetSyn and leptin resistance. Especially when fructose seems to be the bioactive compound behind leptin resistance in a Western-type diet (18). Moreover, this could explain why maybe a high carbohydrate diet can predispose to obesity. But as always, more research is needed. 

* Recent research points out to G6P as the relevant molecule for ChREBP activation.

** Which suggests that increased liver glycogen was not by gluconeogenesis. 

Del Pozo CH, Vesperinas-García G, Rubio MA, Corripio-Sánchez R, Torres-García AJ, Obregon MJ, & Calvo RM (2011). ChREBP expression in the liver, adipose tissue and differentiated preadipocytes in human obesity. Biochimica et biophysica acta PMID: 21840420


Iizuka K, & Horikawa Y (2008). ChREBP: a glucose-activated transcription factor involved in the development of metabolic syndrome. Endocrine journal, 55 (4), 617-24 PMID: 18490833

Advanced cancer and the ketogenic diet

A new pilot trial about cancer-ketogenic diet has recently been published (1). This german group evaluated specifically the feasibility of a KD and its influence on the quality of life of patients with advanced metastatic tumors. The full text is free, so anyone can go and check the details and methodology used. The nutritional intervention was basically an ad libitum KD (<70g/CHO/day), plus extra omega-3 fatty acids. 

Dietary guidelines for the patients were (Table 2):

1. Avoid all types of bread, cake, processed snacks, sweets, potatoes, pasta,
   rice, polenta, vegetables rich in starch (corn, beans, peas) and cereals.
2. Be aware of hidden sources of CHO in sugar sweetened drinks, candy,
   chewing gum with sugar, milk and milk products, lunch meat and some
   cheeses as well as in most “low fat” products.
3. Fruits are rich in CHO, therefore always calculate the amount and select
   those which are low in CHO.
4. Vegetables are often rich in CHO - but mainly in dietary fiber, therefore
   calculate the usable CHO only.
5. If possible, prefer cold-water fish and meat from grazing cattle as protein
   sources, because of their preferable fatty acid pattern.
6. Vegetables and the few fruits allowed should be grown organic
7. As nibbles, select oil-rich nuts (walnuts, brazil nuts, macadamia nuts) and
   seeds (sunflower), and only occasionally chocolate with very high cacao
   content (min. 85%).

This guidelines look very paleoish to me (whatever that means). Some things I found interesting are points 5 and 6. Not seen emphasized commonly. Additionally, patients were told to drink two liquid meals as snacks. Components of this shake were provided to the patients and included: 250ml of highly fermented yogurt-drink, 8ml vegetable oil mixture and 10g of protein preparation. Ingredients of the components can be seen in Table 3. 

Because of problems with compliance to the diet, from 16 initial participants, only 5 remained until the end of the study (31%). Two patients dropped out during the first week, one because of inability to adhere to the diet and the other because of personal problems. Two patients died from their malignant disease during the study, one patient dropped out because he suffered excessive weight loss and weakness, one patient quit because he felt he wasnt able to stick to the dietary guidelines, one because resuming chemotherapy and four due to progress of their advanced cancer situation. The compliance problem is common. Adopting a ketogenic diet involves a lifestyle change. Something even advanced cancer patients can't do. This reminds me of a study in which some cancer patients wouldnt adopt a KD because that meant "giving up the candies and ice cream", despite the fact that it could improve their condition. To complicate things further, the acceptance of the diet varied greatly. One patient said that after 3 days on the diet it was not feasible at all and stopped the diet. Two patients rate feasibility as "very good", seven patients rate it "good", three "moderate" and one "poor". This was after 2 weeks of dieting so included the 16 initial participants.

Quality of life was measured by the EORTC QLQ-C30 questionnaire. Global scores remained relatively stable during the evaluation time. Physical and role functioning worsened slightly over time and constipation was reported by most patients. Because of the advanced cancer stage of the patients, fatigue, pain or dyspnoea increased over time. Nevertheless, emotional functioning increased slightly and insomina improved. 

Of those who completed the whole 12 weeks of dieting, 60% reached a stable ketonuria, predominantly being 1.5-4.0mM. Among blood parameters, only some patients had available data. Overall, CRP levels increased slighlty over time, considering the initial values were high. Two patients initially had elevated glucose, which returned to normal. In other patients, cholesterol levels were "normalized" (meant by reduced to conventionally accepted levels), as well as triglycerides in one patient and ALT in other patient. Total leukocyte count significantly increased during the intervention (even though one patient with initial low leukocyte counts showed a further reduction). 

Patients lost an average of 2kg. Progress of the disease occured in 5 patients who then discontinued the diet, while 5 patients who adhered to the diet had stable disease progression.

Overall, the percentage of days in ketosis (>0.5mmol/l) was not correlated with the results of the study. We can see in table 4 that for example, patient 6 reached 97% of days in ketosis, but because of impaired food intake only completed 6 weeks and showed progress in the disease. On the other hand, patient 16 reached 100% days in ketosis, completed the trial and showed no progress in the disease. Both patients 5 and 11 only reached 25% of days in ketosis, but completed the trial and maintained their condition. 

The limitations of this study were:

- Patients had advanced stage cancer. While a KD might help preventing and/or treating some cancers, there is no much left to do when the disease is too severe.

- Most patients werent from the author's hospital. Blood samples and laboratory parameters had to be provided by their family doctors or local oncologists.

- Short sample and short intervention time. 

Being fair, at the time the study was done (2007) guidelines to apply a KD for cancer treatment were scarce. The only premise was that reducing carbohydrates (hence sugar and cancer's fuel) would reduce progression of tumors. Since then, there is more information available which suggest how to implement the KD for these patients. Overall, evidence suggest that the diet should be not only ketogenic, but calorie restricted. This is for achieving low blood glucose levels and increased KB. In the study reviewed in this post, calories were ad libitum and with a carbohydrate intake limited to 70g/day. Glucose should be ideally around 55-65mg/dl and KB 4-7mM. Checking the study data, most patients had much higer BG (mean 93) and only mild ketosis. Therapeutic fasting is another valuable tool, but harder to comply with. Another factor to take into account is the cancer phenotype. A restricted KD should be more efficient in predominantely glucose-consuming tumors, which can be assessed using some phenotypic markers. Serum LDH levels, for instance, have been shown to be correlated with activation of HIF related genes (2), which include glycolytic enzymes (3). Or using the more conventional FDG-PET. Finally, utilization of gluconeogenesis and glycolysis inhibitors (ie. 2-Deoxyglucose or metformin) with the KD has also been proposed (4,5).

Recent evidence suggests that this metabolic therapy is promising. One case report (6) has shown a rapid regression of glioblastoma multiforme in an old patient using the guidelines proposed by Seyfried et al (7). This patient started the metabolic therapy with a water-only fast, switching then to a restricted KD which delivered 600kcal/day for 14 days. Dexamethasone was also eliminated (because high dosage steroid medication increases gluconeogenesis and blood glucose levels, while enhancing apoptosis resistance in tumor cells). Because of development of mild hiperuricemia, the KD was changed for a non-ketogenic calorie-restricted diet which also delivered 600kcal/day. Aside from the complete regression in such a short time (2-2.5 months), the most surprising finding in my opinion was the recurrence of the tumor after discontinuing the metabolic therapy, which strongly suggests that the therapy itself was the most infulential factor in cancer regression.

It seems that controlling blood glucose levels is the more important part of the metabolic therapy. Achieving blood glucose levels of 55-65mg/dl and 4-7mM of ketone bodies has been termed as "the zone of metabolic management":

This resembles the results from the study on the GB patient:

Although target blood glucose levels were not achieved, the reduction observed was sufficient for controlling disease progression. The other difference was the method of detection of ketosis, urinary ketones, which doesnt always correlate to blood ketone levels. 

Schmidt M, Pfetzer N, Schwab M, Strauss I, & Kammerer U (2011). Effects of a ketogenic diet on the quality of life in 16 patients with advanced cancer: A pilot trial. Nutrition & metabolism, 8 (1) PMID: 21794124

Of chimps and men

This is another post not related to ketosis (I have one almost finished which should be published shortly). 

The reason for this post comes after the "evidence" presented by Don Matesz in his last post. I think that my answer got a little short and I think this topic is very important for understanding certain aspects of nutrition, especially when talking about evolutionary nutrition. First of all, I have to mention that humans can adapt to almost every diet. Obviously I have my own reasons for believing that a high fat diet is better than other diets, but I don't dismiss the idea that different people can get the same results using different dieting approaches. 

Ok, back to the main topic. Don made the following assertion:

"Ninety-eight percent of the human genome is identical to the nearest primate relative, chimpanzees, who eat a 95 percent plant diet.  Recent hunter-gatherers consume up to 20 times more meat than chimpanzees on a percent energy basis, a substantial deviation from the primate baseline."

This leads us to comparative genomics. What is comparative genomics?

"Comparative genomics is the analysis and comparison of genomes from different species. The purpose is to gain a better understanding of how species have evolved and to determine the function of genes and noncoding regions of the genome."

But ultimately, the genome tells us nothing about how genes are transcribed and/or translated which is what in the end is physiologically relevant. The fact that two genomes from different species have the same genes doesnt mean that a. they are equally expressed, b. they are equally translated or c. they yield the same final product (protein or RNA). The collection of all transcripts present in a cell, at a given time and condition is called the transcriptome. The set of expressed proteins in a given cell type, time and condition is called the proteome. Further, another area of research is the metabolome. 

Copyright © 2002, Bios Scientific Publishers.

The genome is divided in coding regions (genes) and non-coding regions (cis-regulatory elements and junk DNA). Briefly, transcription of a gene is performed by RNA polymerase II, which is the polymerase that synthesizes mainly mRNAs, which are then translated into functional proteins. This is a dynamic and highly regulated process, which means that the merely presence of a gene in a genome gives us no information about the importance of this gene in a particular cell type at a given time, or for that matter, for an organism. Gene expression can be regulated in multiple ways. For most eukaryotic protein-coding genes, the binding of RNApolII and general transcription factors depend on the binding of several proteins (activators and repressors) to transcription-control elements, such as promoters, promoter-proximal elements and enhancers. The binding of transcription factors to these elements in turn depend on the energy status of the cell, as well as activation of different signaling pathways activated by different stimuli (ie. hormones). Another important aspect of this process of activation/repression (or on/off status of a particular gene) is its location within the chromosome (heterochromatin vs. euchromatin) and histone modifications (methylation, acetylation and phosphorylation being the most characterized). The condensation of chromatin regulates the binding of proteins and RNApolII to DNA, thereby regulating transcription initiation, which is the main mechanism in the control of gene expression. 

Another level of gene expression control are post-transcriptional modifications. After the transcription of a gene, the initial product (primary transcript) has to be processed to yield a mature RNA. This gives us high variability in the final product given that a particular gene can yield different proteins by alternative splicing (the process of removal of introns and splicing of exons). 

Schematic representation of alternative splicing. Exons are colored while introns are represented by lines. (Source: designmatrix.wordpress.com)

During and after the translation of mRNAs, proteins can be further modified (post-translational modifications). A detailed review of these processes can be found here. 

The overall effect of these modifications explain why the number of actual proteins greatly exceeds the number of predicted proteins by DNA and RNA analysis. 

But what is the relationship between the transcriptome and the proteome? One can assume that the transcriptome is directly related to the proteome, the amount and specificity of mRNAs present in the cell determines the amount and specificity of proteins being synthesized. But, as complex as multicellular organism are, this is not the case. Recent integrative analysis of the transcriptome and proteome has shown that there is, in general, a weak correlation between them; and mRNA levels could not be consistently relied upon to predict protein abundance (1,2). The amount of active protein is dependent on factors such as its location, half-life, post-translational modifications or interaction with other proteins to become effective. These might be some of the reasons for the differences observed between the transcriptome and proteome. 

So if we want to compare two species for making physiological assumptions, we must look into the transcriptome and proteome, and not only the genome. Applying this to the comparison of humans and chimpanzees, there is evidence that differences in gene regulation is a major cause of the difference in phenotype between human and chimpanzee (3). 

A very long and detailed review on human-chimpanzee comparisons has been written by Kehrer-Sawatzki and Cooper (4). I will copy some relevant excerpts for this discussion (my bolds).

"It is always possible that measured mRNA levels may not reflect the actual levels of functionally active proteins synthesized by the respective genes [Gygi et al., 1999; Preuss et al., 2004]. Interestingly, parallel patterns of gene expression differences and protein divergence have been observed in comparisons of human and chimpanzee transcriptomes and proteomes [Bustamante et al., 2005; Khaitovich et al., 2005a; Nielsen, 2006; Gilad et al., 2006]. These parallel patterns indicate that significant gene expression differences correlate with extensive divergence of the encoded proteins."

"The central question remains as to the molecular basis (and its ultimate causes) of the observed gene expression divergence between humans and chimpanzees. The processes involved are likely to include the sequence divergence of regulatory regions, differences in the control of transcriptional initiation, RNA processing and translation, the modification of chromatin structure, and potentially also differences in DNA methylation. (...)

Just as gene expression diversity between individual humans is known to be influenced by multiple cis- and trans-acting factors [Morley et al., 2004; Stranger et al., 2005], so it is likely that multiple genetic differences will be found to influence expression divergence between humans and chimpanzees. Particularly interesting in the context of the expression divergence between these higher primates is the finding that transcription factors have evolved rather rapidly in the human lineage by comparison with other human proteins [Bustamante et al., 2005]. The accelerated evolution of transcription factors is indicative of how a relatively small number of genetic changes in key locations may pleiotropically affect the expression patterns of a myriad of different genes by influencing their transcription factor binding capacity. We may therefore surmise that the accelerated evolution of transcription factors in the human lineage has in all likelihood had a major impact on gene expression divergence between human and chimpanzee."

"Alternative splicing contributes very significantly to increased transcriptome and proteome diversity as well as to interspecies divergence [Boue et al., 2003]. An estimated 60% of all human genes employ alternative splicing [Mironov et al., 1999; Lander et al., 2001]. Dual coding regions in particular are characterized by alternative splicing; the same exon sequences are shared by different transcripts encoding distinct amino acid sequences in different reading frames. At least 7% of all alternatively spliced genes in the human genome contain multiple coding regions [Liang and Landweber, 2006].(...) Interestingly, 4% of human dual coding regions are not conserved in chimpanzees, as indicated by the presence of stop codons that disrupt one of the two reading frames [Liang and Landweber, 2006]. (...) When human and chimpanzee exon flanking regions were compared, a significant reduction in the nucleotide substitution rate of the flanking regions of alternatively spliced exons was observed, independent of the site-by-site variation in mutability due to different CpG contexts. Thus, increased purifying selection appears to have impacted on exon flanking regions and thus, in all likelihood, the regulation of alternative splicing in humans and chimpanzees [Xing et al., 2006]."

Other mechanisms such as gene duplications, insertions/deletions, chromatin condensation, genome rearrangements, etc. are also important but discussing all of them would make this post extremely large. Nevertheless, the difference in gene expression patterns between humans and chimpanzees has been reported in several studies (5,6,7). This pattern also seems to be influenced by diet (8), which could influence phenotypic differences between humans and chimpanzees.Thus, it seems that cis-regulation is a major mechanism by which humans and chimpanzees differ (9).

Summing up

I think is pretty clear now how invalid is to use genome comparisons as evidence of the superiority of a given diet. Although this is a valuable tool for making phylogenetic and evolutionary assumptions, as well as for identifying orthologous genes, it does not help predicting gene expression dynamics. Moreover, using the idea of parsimony (nature is lazy), mutations driving species diversification should not be observed primarily in protein-coding regions, but in cis-regulatory elements which control gene expression. This in turn influences levels and expression of transcription factors for specific genes, like the differences observed between humans and chimpanzees. Finally, differential alternative splicing can yield completely different proteins from the same gene, thus contributing to phenotypic differences between humans and chimpanzees. Overall, these processes show how two divergent species can have a great sequence identity but a completely different phenotype, and thus require different diets. 

Addendum

A very interesting review on diet and gene expression among humans and non-human primates can be found here.

Kehrer-Sawatzki H, & Cooper DN (2007). Understanding the recent evolution of the human genome: insights from human-chimpanzee genome comparisons. Human mutation, 28 (2), 99-130 PMID: 17024666