Nutritional immunotherapy: dietary fatty acids

Most people familiarized with a paleo-type lifestyle already know that a healthy diet should include good fats, rich in saturated (SFAs) and monounsaturated fatty acids (MUFAs). This profile is characteristic of animal fats. Conversely, intake of polyunsaturated fatty acids (PUFAs) should be limited. These recommendations are in accordance with the biological, chemical and evolutionary aspects of fatty acids in the human diet. We might see this dietary fatty acid profile as a "natural" one. 

As I stated in my last post, I support the notion that some inflammatory/autoimmune disorders deserve a different approach, owing to the body's unnatural state. This deviation from normality implies that we cannot expect the same response to some nutritional components in diseased people. 

Background

For understanding the basis of my recommendations, we must review a little bit about Toll-like receptors (TLRs) and the classic cellular inflammatory cascade, the NFkB pathway. TLRs are pattern recognition receptors which bind pathogen associated molecular patterns (PAMPs). PAMPs are conserved molecular motifs found in a broad range of pathogens that are recognized by receptors mediating an innate-type immune response (like TLRs). PAMPs include LPS, lipoproteins, peptidoglycan, lipoteichoic acid and other molecules capable of binding to these receptors and trigger a response. There are different type of TLRs, with different cellular localizations and associated intracellular pathways, but most converge in the activation of NFkB. The final effect of ligand binding and protein signaling is the expression of inflammatory genes. 

TLRs have evolved not only a function in immunity, but recent evidence suggests a pivotal role in metabolism. 

TLR4/MD-2 binds to LPS

LPS, the classic TLR ligand, binds and activates signaling through TLR4 (1). This interaction is mediated by the lipid moiety, Lipid A. The toxicity of LPS is thought to be due to the interaction of TLR4 with Lipid A, and the shape and conformation of this lipid may determine the toxicity of a given pathogen (LPS are not created equal) (2). The differences arise from the number and length of fatty acid chains. 

LPS general structure. Supplemental Material: Annu. Rev. Biochem. 2011. 80:917-941. Link.

Typically, 4 to 7 lipid chains with 12 to 14 carbons of length are anchored to the glucosamine backbone. 

For being able to recognize LPS, TLR4 binds MD-2 and forms the complex responsible for the interaction. Only one third of MD-2 is involved in the interaction with TLR4 and the remaining part is available for interaction with other ligands. The presence of LPS is necessary for TLR4/MD-2 dimerization (3). 

TLR4/MD-2 complex. Annu. Rev. Biochem. 2011. 80:917-941.

Structural studies have proposed that certain factors determine the endotoxic activity of lipid A. The number of lipid chains seems to be the most important: Six lipid chains of 12 to 14 carbons each promote an optimal inflammatory activity. Because of this, the deletion of two lipid chains can destroy agonistic activity, making some LPS TLR4 antagonists (4). The SFAs linked to the Lipid A backbone are 3-O-acylated by lauric acid, myristic acid or palmitic acid, which seem to be important for the immunological effects of LPS, given that their loss abolishes LPS endotoxic properties and produces an antagonist of Lipid A (5, 6). The logical inference of these facts is that these SFAs are important for the binding and/or LPS-induced TLR4 activation.

The role for TLR4 in metabolic abnormalities has been corroborated using animal models. Mice with a loss-of-function mutation in TLR4 or TLR4-null mice are protected against the development of diet-induced obesity and insulin resistance (7, 8, 9). Deletion of CD14, a TLR2/TLR4 co-receptor, attenuates the cardiovascular and metabolic complications of obesity (10). TLR4 has been involved in other metabolic complications which are less understood, like the progression of simple steatosis to non-alcoholic steatohepatitis (NASH) associated with obesity (11). 

From the above evidence, researchers thought they had found the link between dietary fat and inflammation. Nevertheless, Erridge and Samani (12) found that the in vitro evidence showing a direct activation of TLRs (they tested TLR2, TLR4 and TLR5) was caused by contamination of fatty-acid-free BSA (used to present SFAs to cells) with LPS and lipopeptide (although some authors show that BSA alone is not sufficient to activate TLR4, see below). Others have suggested that SFAs activate TLR4 signaling indirectly, promoting TLR4 dimerization and association of TLR4 with MD-2 and downstream adaptor proteins (TRIF and MyD88) into lipid rafts (13). This latter explanation seems to be more plausible.

If SFAs promote inflammation and seem to act in part through TLR4 in in vivo studies, then maybe they are acting as carriers of another molecules which bind and activate TLR4. This seems to be the case. As Peter has blogged briefly before, Ghoshal et al. (14) demonstrated that chylomicron formation in the gut promotes LPS absorption. However, this effect is not exclusive to SFA, but to long chain fatty acids (LCFA) as they need chylomicrons for absorption/transport, in contrast to short- and medium-chain fatty acids. 

Does this makes LCFAs inherently unhealthy? No. Appropriate TLR stimulation is important of adequate innate responses to pathogens and for maturation and development during childhood. Excessive uptake of LPS, promoted either by calorie excess and/or overgrowth of gram-negative gut bacteria, seem to contribute to chronic endotoxemia and disease. Endotoxin overload is particularly problematic in adults with inflammation and/or immune related diseases. 

Association between postprandial lipemia and inflammation

Fisher-Wellman & Bloomer found that isocalorically, high-fat meals promote a stronger postprandial oxidative stress than carbohydrate and/or protein meals in healthy subjects (13). However, they used heavy whipping cream as the fat source, dextrose powder as carbohydrate and casein-whey protein powder as the protein source. This helps isolating variables (macronutrients) but doesn't help much for assessing the effect of a mixed meal composed of real food. Moreover, they didn't control for calories, as the "lipid meal" contained more calories than any other meal:

Fisher-Wellman & Bloomer, 2010

In fact, the "protein meal" included more fat than the "lipid meal" (98g/38% vs. 93g/34%). The carbohydrate percentage was the same between the carbohydrate and the lipid meal, but the absolute amount was higher in the latter because of the calorie content. If dietary fat were indeed to blame, it can be expected that the higher the fat content, the higher the measures of oxidative stress. But the meal with higher fat content was the mixed meal, which didn't produce significantly different results than the carbohydrate or the protein meal.  

Another measure of inflammatory changes induced by specific macronutrients or meals is the activation of transcription factors and proteins involved in cellular inflammatory pathways, such as NFkB. For example, glucose ingestion (75g in 300ml water) stimulated nuclear transport of NFkB, reduced IKB-alpha protein levels and increased the activity and expression of IKK-alpha and IKK-beta in mononuclear cells (15). This was paralleled by an increased expression of TNF-alpha and activation of NADPH oxidase. Similar results have been found after a 900kcal mixed meal (81g carbohydrate, 51g fat and 32g protein) (16). This is to be expected as energy intake will increase mitochondrial respiration, stimulating mitochondrial ROS production. Some ROS are capable of activating NFkB, so any increase in intracellular ROS can increase NFkB-mediated signaling (17). 

Ingestion of 300kcal of cream or glucose stimulated NFkB binding, expression of SOCS3, TNF-alpha and IL-1beta in mononuclear cells of healthy subjects, but only cream increased plasma LPS and TLR4 expression (18). This was not seen with ingestion of orange juice, probably due to the increase in uric acid associated with fructose ingestion. These results are in accordance with chylomicron-mediated transport of LPS through the gut. As expected, high FFA plus high glucose amplify the inflammatory response (19).

Hypertriglyceridemia increases endotoxemia

The human body must be able to cope with acute increases in LPS in plasma, attenuating the inflammatory response induced by fat ingestion. Clemente-Postigo, et al. (20) showed that in morbidly obese subjects, endotoxin increases were strongly correlated to the difference between baseline and postprandial triglyceride levels. They also found that baseline triglyceride level was the best variable that predicted basal LPS level in serum. In this regard, very low carbohydrate diets have shown to reduce baseline triglyceride levels and postprandial lipemia (21, 22). In the metabolically healthy, the immune system is capable of attenuating postprandial endotoxemia (as with inflammation induced by any meal). The inflammatory nature of absorption, digestion and metabolism of macronutrients must be coupled with an anti-inflammatory period, such as fasting (depending on the inflammatory load of the diet, an overnight fast might work). By this way, the body's inflammatory balance is maintained in a healthy range. 

Effects of dietary PUFAs on immune cells

It is important to note that, although LCFA (which means they stimulate chylomicron formation and thus, LPS transport), MUFAs and PUFAs exhibit different effects than long chain SFAs. This could be related to their effects on membrane fatty acid composition. Cellular membranes are highly structured, and subtle variations in the unsaturation of phospholipids can have diverse but important molecular consequences. It is well known by now that dietary fatty acids alter the composition of membrane lipids, as they are incorporated. In immune cells, this is extremely important for the overall response to a certain stimulus, from phagocytosis against a pathogen to secretion of cytokines for proliferation and clonal expansion. Fatty acids incorporated to immune cell membranes act through different mechanisms:

• Altering the composition of lipid rafts. This, in turn, influences protein-protein interactions, as well as coupling ligand-receptor interaction with scaffold and intracellular signaling proteins.
• Producing intermediate molecules, such as prostaglandins. The final effect is difficult to assess, but there seem to be clear differences between the action of metabolites produced from omega 6 (O6) vs. omega 3 (O3).
• Altering membrane permeability. A higher unsaturation index (that is, the degree of unsaturation of phospholipid chains) renders a more fluid membrane, being the opposite true for a low unsaturation index. Increasing the proportion of SFA in cell membranes decreases permeability because unsaturated fatty acids chains form a "kink", increasing the degrees of freedom of the molecules and its physicochemical characteristics (both individually and as a group). 
• Providing energy.

Animal and some human studies have shown that altering the amount of either O6 or O3 can affect the function of immune cells (23; references therein):

• Increased dietary intake of EPA (2.7g/day) has shown to reduce PGE2 production (a metabolite of arachidonic acid) in human mononuclear cells (MNCs). 
• Fish oil ingestion has shown to increase the production of 5-series leukotrienes, products derived from EPA.
• EPA/DHA or fish oil also induces the production of resolvins, which have anti-inflammatory properties.
• Increasing membrane permeability by increasing the unsaturation index might increase phagocytosis by MNCs. The phagocytic index of neutrophils and monocytes has shown to be negatively correlated with palmitic acid content, but positively correlated with the content of PUFAs, specifically O3. In healthy humans, 1.5g/day of EPA+DHA for 6 months increased the phagocytic activity in monocytes and neutrophils by 200% and 40%, respectively. 
• Arachidonic acid, EPA and DHA have shown to inhbit T-cell proliferation and IL-2 production in vitro. This has been replicated in animal models with fish oil and/or EPA/DHA in high doses. O3 might also affect the composition (and hence function) of lipid rafts, as treatment of T-cells with O3 displaces acylated proteins anchored to the inner lipid leaflet from lipid rafts, but not GPI-anchored proteins. This displacement (probably as a direct consequence of incorporation of EPA and DHA into membranes) affects the intracellular signaling pathway associated with the protein being displaced, such as LAT. 
• Increasing the amount of dietary fish oil in rats causes a reduction in MHC II expression on dendritic cells, as well as levels of CD2, CD11a and CD18. Arachidonic acid and DHA, by slowing the transit of new MHC I molecules from the endoplasmic reticulum to Golgi, have shown to decrease surface MHC I expression, decreasing cytotoxic T-cell mediated lysis of target cells enriched in these fatty acids. 

The acute effect of increasing doses of animal O3 is a reduction in arachidonic acid-derived inflammatory metabolites, increases in membrane permeability and anti-inflammatory molecules derived from EPA/DHA, as well as reduction in T-cell activation and antigenic stimulation. O3 also have direct effects: inhibition of LPS or lipopeptide-stimulated COX2 expression and LPS-induced NFkB activation (24, 25). Interestingly, there is evidence that the anti-inflammatory effects seen for O3 are dependent on their oxidation. Oxidized EPA, but not unoxidized EPA, inhibits NFkB activation and expression of inflammatory molecules in a PPARa dependent manner, as well as chemotaxis (26, 27, 28). Oxidized, but not unoxidized DHA, inhibits polychlorinated biphenyl-induced NFkB activation and MCP-1 expression, effects probably mediated by its oxidation products (A4/J4 neuroprostanes) (29). Thus, it seems that contrary to what is believed, oxidation of O3 PUFA is necessary to mediate their beneficial biological effects. 

The effects of MUFAs (mainly oleate) have not been studied in detail as with PUFAs. In contrast to palmitate and stearate, oleate do not seems to induce TLR2/4 activation in monocytes (19) (in this case, the authors used a BSA-only control, showing no activation of TLR). This makes sense, as oleate is the main FFA in human circulation (30). However, oleate has a strong inflammatory effect on human islet cells, increasing the levels of IL-1beta mRNA, IL-6 mRNA and IL-8 mRNA compared to palmitate and stearate, effect which was amplified by high glucose levels (31). In contrast with the latter two, the expression of IL-1Ra (antagonist of IL-1) was lower with oleate. The authors suggested that oleate-mediated islet inflammation could be hormetic (which makes complete sense).

Oleate levels in circulation are determined by oral intake as well as de novo synthesis from SFA. This process is mediated by stearoyl-CoA desaturases (SCD), specially SCD-1 in humans (32). This enzyme catalizes the introduction of a single double bond at the delta9, 10 position of long chain acyl-CoAs, preferentially to stearoyl-CoA and palmitoyl-CoA. Over-stimulation of SCD-1 increases the synthesis of MUFAs (like palmitoleoyl-CoA and oleoyl-CoA), affecting intermediary metabolism and promoting obesity, pathological insulin resistance, hypertriglyceridemia and hepatic steatosis (33). SCD-1 expression is induced by SREBP-1, LXR, and inhibited by PPARbeta/delta and PPARgamma. Accordingly, SCD-1 expression and activity is increased with high carbohydrate diets (34, 35), because insulin activates SREBP-1 and glucose (actually glucose-6-phosphate and/or xylulose-5-phosphate) activates ChREBP, which increases SCD-1 expression (36). However, it is important to interpret this data with caution, as lipogenic/lipolytic enzymes in rodents are more active than humans. Nevertheless, a high carbohydrate diet can contribute to the pool of MUFA, thereby influencing the secretion and expression of pro-inflammatory cytokines.

In contrast, EPA has shown to decrease the level of SCD-1 mRNA and SREBP-1c mRNA in Hep G2 cells (37) and omega 3 status is important for controlling the activity and expression of SCD-1 in rats (38). 

Albumin binds fatty acids and LPS

In vitro, one of the most inflammatory fatty acids is lauric acid, which activates NFkB, partially mediated by the TLR4-MyD88/PI3K/Akt pathway, while DHA inhibits this effect (39). SFAs released by adipocytes (mainly palmitate) are also able to activate TLR4 in macrophages, activating NFkB by a mechanism shared partially with LPS (40). It seems that activation of inflammatory genes in different immune cells is related to chain length (41). This suggests that in addition to promoting dimerization and organization of TLR4 with adaptor and co-stimulatory molecules into lipid rafts, some SFA could indeed activate TLR4 independently. What is really interesting is that free fatty acids travel in the bloodstream bound to albumin (and the levels of individual fatty acids correlate with those found free in plasma) (42), and recently, analysis by surface plasmon resonance found that albumin not only binds to LPS, but also modulates its interaction with TLR4 and MD-2, and thus, controls the inflammatory response to a given endotoxin load (results not published)*. So we have a situation in which increasing the dose of O3 PUFA might increase the relative proportion of O3 bound to albumin, thus inhibiting the interaction of palmitic or stearic acid with LPS, or at least, ameliorating it. On the other hand, we can decrase the proportion of saturated fatty acids in circulation and bound to albumin by dietary means (43, 44). Ultimately, the balance between lipolysis and oxidation determines the level of free fatty acids in circulation. A high lipolytic environment uncoupled to mitochondrial oxidation contributes to lipotoxicity and inflammation. This also holds true for LPS-induced inflammation. The perfect balance between hydrolysis of stored fatty acids and oxidation is achieved under fasting conditions. 

TLR4 is modulated by specific fatty acids. Saturated fatty acids (SFAs) activate TLR4 either by direct interaction or by promoting its dimerization and association with MD-2 into lipid rafts. Conversely, omega 3 polyunsaturated fatty acids (PUFAs) inhibit this effect, and reduces SFA-induced TLR4 activation. Both O3 PUFAs as well as O6 PUFAs are able to modulate the immune response by affecting membrane phospholipid composition. Increased levels of arachidonic acid (ARA) derived from O6 metabolism, as well as its metabolites are mainly pro-inflammatory, while those derived from O3 PUFAs are anti-inflammatory. Increasing the proportion of unsaturated fatty acids also changes membrane fluidity, affecting immune functions such as phagocytosis. Activation of TLR4 promotes the association of adaptor molecules such as TIRAP, MyD88 and IRAK. IRAK activates TRAF6, which interacts with TAK1, leading to the phosphorylation (denoted by the yellow P) of IKK/NEMO and its activation. This complex is able to phosphorylate IkB, thus letting transcriptionally active NFkB migrate into the nucleus. Once activated, NFkB increases the expression of pro-inflammatory genes, such as TNFa, IL-1 and IL-6. See text for details.

Summary

The composition of different fatty acids in the diet modulate endotoxemia. From the available evidence, there is consistent research which shows that:

• Saturated fatty acids (SFAs) activate TLR4 and the downstream signaling pathway, ultimately leading to the activation of NFkB, which increases the expression of pro-inflammatory molecules (TNFa, IL-6, etc.).
• SFAs might contribute directly (by interacting with LPS and/or TLR4-MD-2) or indirectly (by reorganizing lipid rafts). In either case, an increase in the level of SFA promotes the activation of this pathway. 
• The activation of TLR4 has been shown to be important for the onset and development of metabolic diseases such as obesity, diabetes and non-alcoholic hepatic steatosis.
• LPS uptake is mediated through chylomicrons and is promoted by a loss of barrier function of the small intestine.
• The level of endotoxemia correlates with baseline and post-prandial triglyceride levels.
• O3 PUFAs (EPA and DHA) have shown an inhibitory effect on LPS and LPS plus SFA-induced TLR4 activation. 
• The oxidation of O3 PUFAs seems to be necessary for their anti-inflammatory effects.
• The level of SFA in the bloodstream is controlled by diet as well as the cellular energy status. 
• MUFAs, in most studies, seem to be neutral. However, there is some evidence linking excess oleate and SCD-1 activity to cellular dysfunction, particularly beta-cell abnormalities. EPA has opposite effects and reduces SCD-1 expression.
• Albumin binds both fatty acids and LPS, and modulates the inflammatory response to a given LPS load. The relative proportion of individual fatty acids bound to albumin might influence the binding of LPS to TLR4, thus affecting the activation of the downstream signaling pathway. 

Practical recommendations

For people with autoimmune and/or inflammatory problems, I recommend the following measures to be taken with respect to fatty acids in the diet:

1. Reduce and control the amount of O6 PUFA, specially from vegetable sources (linoleic acid).
2. Control the amount of SFAs. Consumption of dairy fat seems to be protective against endotoxemia (45). Ghee might be a better option than butter. Better to avoid protein-rich dairy.
3. Increase the amount of marine EPA and DHA (O3 PUFAs). This should work best increasing the consumption of marine foods, but might be a problem for those with leaky gut given the presence of some metals in seafood. Individual tolerance must be assessed. If very sensitive, start with dietary supplements. A high dose (3-5g/day of EPA + DHA) might work first, and the those should be lowered afterwards (46). The higher the baseline triglyceride levels, the higher the dose. Additionally, the worse the inflammatory/immune status, the higher and longer the supplementation. This can be assessed using traditional blood markers (C-reactive protein, etc.) and symptoms. It has been shown that the effects of O3 supplementation are influenced by the O3 status of the subject (47). High inflammatory markers and/or symptoms might reflect O3 status.
4. Avoid industrial trans-fatty acids.

* Work was presented in the Inmunoperu 2012 conference. More information when available. 

News and some links

Unfortunately, I will not be able to attend to AHS this year, due to some unexpected financial and academic (damn single molecules!) problems. I was looking forward to meet most of the “paleosphere” and discuss about science with very bright people. I hope I can make it in 2013.

I have been very busy working on the lab, teaching biochemistry and cell biology in Med School and doing a post-graduate diploma in basic and clinical immunology. Because of this, I have had little time for reading carefully studies which are not related to my thesis, and in my short spare time I wanted to let my brain “rest” a little bit, although I have read some (but not as much as I would like to). Accordingly, I haven’t had any time to update the blog. Nonetheless, I have an almost finished post on dietary fats and immune function, which may raise some controversy in the “low carb” community. I want to finish and publish this post once for all (it has been almost finished for almost one month now). I might add some information that I was keeping for my AHS talk (see bottom of the post).

In the meantime, and to make something productive out of this post, here are some studies I have found interesting lately (not all related to the topic of my blog, by the way).

Myocardial Infarction Triggers Chronic Cardiac Autoimmunity in Type 1 Diabetes

Inducing myocardial infarction (MI) in NOD mice (non-obese diabetic mice) triggers the development of severe post-infarction autoimmune syndrome characterized by lymphocyte infiltration to the myocardium, infarct expansion, and autoantibody and Th1 type response to cardiac alpha-myosin. 83% of a sample of post-myocardial infarction T1D patients showed positive tests for autoantibody against cardiac proteins. 

Post-infarcted myocardium from NOD mice showed extensive lymphocytic infiltrates resembling those from  pancreas islets,  and extended into the non-infarcted myocardium (left image). Normal mice (B6) showed no infiltration nor expansion, but normal infarct healing and scar formation (right image).   © 2012 American Association for the Advancement of Science. All Rights Reserved.

Maternal western diet causes inflammatory milk and TLR2/4-dependent neonatal toxicity

Wild-type (WT) female C57B6J mice were fed with normal chow or chow supplemented with extra cholesterol, fat and sucrose (Western diet, WD) for 2 weeks before breeding with WT male mice, and were maintained on chow or WD during pregnancy and lactation. Pups from mice fed a WD exhibited alopecia* (see figure), increased body weight, thicker lipid layer in the skin (increasing the lipid/skin weight ratio), skin leukocyte infiltration and increased expression of inflammatory genes (IL-1b, TNFa, COX-2, iNOS, MCP-1) in the skin, intestines and liver.

WD-milk promotes alopecia.  Copyright © 2012 by Cold Spring Harbor Laboratory Press 

These deleterious effects were related to the composition of the WD, which contained more saturated fatty acids (SFA) than the chow diet (48.9% vs. 17.6%) **, as well as to the composition of the milk from WD fed mice, which was 24% higher in fat than control. Milk triglycerides from WD-milk had a higher percentage of long chain fatty acids (elongation ratio (C18:0/C10:0) was increased 1.9-fold) and saturated fatty acids (saturation ratio (C18:0/C18:2) was 2.8-fold higher). 

Furthermore, the inflammatory effects induced by the increase in long-chain saturated fatty acids in maternal milk was in part mediated through the production of ceramides and glucosylceramides, as well as lactosylceramide, which was capable of inducing the expression of the aforementioned inflammatory genes in RAW264.7 cells (mouse macrophage cell line). Deletion of TLR2/4 abolished or decreased some of the toxic effects observed, suggesting that TLR-signaling is involved (deletion of adaptor molecule Myd88 also reduced toxicity). Germ-free mice showed an increased toxicity for WD milk, suggesting a protective effect of the normal gut microbiota. 

Time-Restricted Feeding without Reducing Caloric Intake Prevents Metabolic Diseases in Mice Fed a High-Fat Diet

Fasting improves metabolic alterations induced by high-fat diets in mice. Copyright © 2012 Elsevier Inc. All rights reserved.

When fed a high-fat diet, C57BL/6J mice develop inflammatory symptoms and ultimately, metabolic disease. What Du, et al. wanted to know if circadian metabolic oscillations influence the response of mice to a HFD. They used a restricted feeding regimen, which has shown to reset some circadian rhythms. They found that, although eating the same calories overall, mice eating all of the calories from the HFD during an 8-hour window do not developed the symptoms exhibited by mice eating a HFD ad libitum (see figure). In other words, eating the same unhealthy diet at the same level of calories, but restricting it to an 8-hour window, practically abolishes all the deleterious effects***. This effects were mediated by preventing the abnormalities in the circadian rhythm induced by high-fat feeding.

Interactions between commensal fungi and the C-type lectin receptor Dectin-1 influence colitis

Iliev, et al. show that although bacteria are a very important (and the major component) for the gut microbiota, there are also other organisms which play an essential role in the host's health. This is the case for some fungi, which interact with the host through dectin-1 receptors, inducing a Th17-type immune response. Mice lacking dectin-1 are susceptible to DSS-induced colitis, and show an abnormal inflammatory response. Inducing colitis altered the mycobiome, and treatment with fluconazole (antifungal) ameliorated inflammation. Analysis of the mouse fungal microbiome revealed that 65.2% of the sequences identified corresponded to Candida tropicalis, an opportunistic pathogen. Their results showed that dectin-1 restricts its body localization to the intestinal lumen (so there is no invasion of inflamed tissues). Colitis also increased the proportion of Candida and Trichosporon (opportunistic pathogens) and reduced the levels of Saccharomyces (non-pathogenic). Supporting their hypothesis, they found an haplotype in the CLEC7A gene (human dectin-1 gene) which was correlated with severe forms of ulcerative colitis, but not with non-severe manifestations.

Fungi levels are an important part of the murine gut microbiota. This probably occurs also in humans.   © 2012 American Association for the Advancement of Science. All Rights Reserved.

A Single-Molecule Hershey-Chase Experiment

The mechanism by which phages transfer their DNA to bacteria has been a key question for decades. Using a single-molecule system with two different strains of lambda phage (differing in genome length), Van Valen, et al. found that the time it takes for lambda phage to transfer its DNA to a single E.coli cell takes 5 minutes (with high cell-to-cell variability and sometimes showing long pauses), compared to what it is  seen in vitro, where the process takes roughly 10 seconds. The in vivo velocity of DNA ejection seemed to be determined by the amount of DNA ejected, not by the amount of DNA left in the viral capsid (which is what has been seen in vitro). The method utilized by the authors is theoretically simple (see figure). They stained viral DNA while still in the capsid with a cyanide dye (SYTOX orange), after which excessive dye is washed out. The stained phages are then briefly bound to bacterial cells, which are pipetted into a flow chamber. After washing again with buffer, the sample is imaged with time-lapse bright-field and fluorescence microscopy. The ejection of DNA is measured by comparing the fluorescence intensity inside the phage capsid to that of the bacterial cell: a loss of fluorescence inside the capsid with a concomitant increase in the cell represents DNA translocation.

Copyright © 2012 Elsevier Inc. All rights reserved. 

This study is important for several reasons. First, it supports the notion that biological process must be seen at a single-molecule level for understanding them. When we do an experiment in bulk, we have millions of molecules at the time interacting. Any measure of a given parameter is just an average of what is being seen. But many times, we fail to see the behaviour of outliers, or in other case, we can have individual differences which are obscured by the number of molecules being analyzed in a given moment and a given space. What is more realistic is to measure single molecules and average repeated experiments. In this way, we a. reduce individual variability obscured by the average, b. observe how molecules behave inside cells (interactions are between single molecules) and c. reduce the influence of other co-solutes and molecules which might influence the process being observed. 

Fluorescence imaging of the process of DNA ejection to a single E.coli cell. Copyright © 2012 Elsevier Inc. All rights reserved. 

Second, this finding challenges what was thought about this process based on in vitro evidence: that the amount of DNA inside the capsid was the driving force for DNA ejection. This was the most logical hypothesis, given that before ejection, the repulsive forces experienced by the tight packing of the DNA would promote its ejection. Lastly, the method utilized its a novel approach that can be expanded to other systems. 

* Hair was recovered after weaning.
** It was also lower in PUFA (12.2% in WD vs. 61.8% in chow), but much higher in sucrose (152.8g/kg in WD vs. 0g/kg in chow). 
*** Nevertheless, some inflammatory cytokines were still increased compared to control.


PD. I wanted to make public my thanks to the AHS organizers for their help with paper work and answering any question I had immediately. I feel very bad for cancelling my talk, specially after their help. Also, I want to thank to everyone who supported my talk me via donations. Please, if anyone who donated wants a refund, send me an email. For anyone who was expecting my talk, I will compensate it with more information on my blog. First, I will finish and post my second post on my nutrition immunotherapy protocol, and second, I will add a new page to the blog with schematic and didactic diagrams made for my presentation. They will be available for free to everyone. 




Iliev ID, Funari VA, Taylor KD, Nguyen Q, Reyes CN, Strom SP, Brown J, Becker CA, Fleshner PR, Dubinsky M, Rotter JI, Wang HL, McGovern DP, Brown GD, & Underhill DM (2012). Interactions between commensal fungi and the C-type lectin receptor Dectin-1 influence colitis. Science (New York, N.Y.), 336 (6086), 1314-7 PMID: 22674328

Van Valen D, Wu D, Chen YJ, Tuson H, Wiggins P, & Phillips R (2012). A Single-Molecule Hershey-Chase Experiment. Current biology : CB PMID: 22727695

Hatori M, Vollmers C, Zarrinpar A, Ditacchio L, Bushong EA, Gill S, Leblanc M, Chaix A, Joens M, Fitzpatrick JA, Ellisman MH, & Panda S (2012). Time-Restricted Feeding without Reducing Caloric Intake Prevents Metabolic Diseases in Mice Fed a High-Fat Diet. Cell metabolism, 15 (6), 848-60 PMID: 22608008

Du Y, Yang M, Lee S, Behrendt CL, Hooper LV, Saghatelian A, & Wan Y (2012). Maternal western diet causes inflammatory milk and TLR2/4-dependent neonatal toxicity. Genes & development, 26 (12), 1306-11 PMID: 22713870

Gottumukkala RV, Lv H, Cornivelli L, Wagers AJ, Kwong RY, Bronson R, Stewart GC, Schulze PC, Chutkow W, Wolpert HA, Lee RT, & Lipes MA (2012). Myocardial infarction triggers chronic cardiac autoimmunity in type 1 diabetes. Science translational medicine, 4 (138) PMID: 22700956

Nutritional immunotherapy: An overview

In previous posts (1,2), I have briefly reviewed the importance of adipocytes and molecules secreted by the adipose tissue in immunity. Overall, increases in adiposity alterate the inflammatory balance, shifting towards a pro-inflammatory state, which contributes to the development of obesity-associated diseases. 

Given that metabolic and immune pathways are interconnected, and that metabolism controls function in immune cells, it is possible to modulate the immune system through nutrition. This is what I call "Nutritional Immunotherapy", which simply means targeting the immune system with nutritional tools for treating inflammatory and autoimmune diseases. 

Nutrition acts both directly and indirectly on the immune system, as shown in the following diagram:

Relationship between nutrition and the immune system. See text for details.

1. Nutrition influences the composition and function of adipose tissue (AT): Energy intake regulates fat mass, affecting the function and differentiation of pre- and mature adipocytes (direct effect). The concentration of specific fatty acids in AT is proportional to their abundance in the diet (indirect effect). The dietary fatty acid profile also affects membrane lipid composition on other cell types, which regulates cell functioning.  
2. Adipose tissue is an immune organ: There are several immune cells present in AT from different body sites, differing each one in the proportion of cell types. Lymph nodes present in AT are sorrounded by perinodal adipocytes, which have a higher proportion of polyunsaturated fatty acids (PUFA) than adipocytes far from the nodes.  Lymphoid clusters within AT include milky spots (MS) and fat-associated lymphoid clusters (FALCs), which have an active role determining whole-body immune responses. Adipocytes are also able to secrete adipocytokines (leptin, resistin, adiponectin, etc.) and classical cytokines (IL-6, TNF-a, etc.). 
3. Adipose tissue regulates energy intake: Cytokines secreted by AT regulate appetite and energy balance, acting through neural pathways involved in energy homeostasis.
4. Nutrition affects the gut flora: Microbial composition of the human gut flora is very responsive to diet. Small changes in either macronutrient distribution or food choices affect differently not only the relative proportion of certain species, but also their metabolism and gene expression patterns.
5. Gut flora regulates fat mass and metabolism: Digestion of plant cell walls, oligosaccharides and other food components by gut bacteria increases the energy yield of food, contributing to energy intake. Acetate and propionate, produced by the fermentation of soluble fiber, are metabolized (predominantly) in skeletal muscle and the liver, respectively. Gut microbiota also supress FIAF activity and promotes hepatic triglyceride synthesis. Metabolism of drugs and xenobiotics is dependent on the composition of the gut flora. 
6. Gut flora regulates immunity: The presence of specific bacteria shapes the immune system and regulates mucosal and peripheral immune responses. The gut flora also competes with enteropathogens directly and by the action of antimicrobial peptides. Evolutionary co-adaptation has given gut bacteria and other microorganisms essential roles for mammalian health. 
7. Nutrition regulates immunity: Energy availability and macronutrients regulate the function, maturation and differentiation of immune cells. 

Nutrition has the potential to act on all levels mentioned above. This is why a good diet is very important not only for prevention, but also for treatment of diseases of civilization.

Important components

The nutritional immunotherapy protocol integrates concepts from immunology, molecular and evolutionary biology. The first two help us answer the "how" question, while the latter helps us understand the "why". 

What differences this protocol from other diets is that it takes into account the fact that people who already have developed an inflammatory and/or autoimmune disorder respond differently to any diet. This means that the response to a diet is individual, and more importantly, in this case, the starting point is not a natural one. This point is important for understanding the recommendations given hereafter.  

The history of the patient, specially those aspects that would compromise the response to certain macronutrients and the normal development of a tolerant immune system, needs to be addressed before trying to make any nutritional adjustment. Nowadays, with genetic tools (like 23andMe), tailoring the diet according to the genotype is possible and helpful. 

Important factors for the success and application of the protocol are shown below.

Relevant factors for the protocol

Mode of birth

Hygiene practices during childhood

Family diet, diet history and maternal environment

Medical history

Genotype

Social/lifestyle experiences

Symptoms

Self-assessment

Bloodwork

 Mode of birth

This important but commonly overlooked factor is determinant for immune development and future health. Vaginal birth is the natural mode of birth because it stimulates not only hormonal responses in the mother and the child, but because it promotes an adequate colonization of the neonate, one that we have been adapted for. At birth, the newborn is sterile*, so it can be colonized virtually by any species. Babies born by cesarean section have an abnormal microbiota, as they harbor bacteria from the hospital's environment, medical practitioners and the mother's skin. Normally, during the passage through the birth canal, the infant is exposed to vaginal and cervical flora. Because of its proximity, newborns are also rapidly colonized by maternal gut microbiota, which seems to be the predominant source of bacteria. Pre-term infants also display a different pattern of microbial colonization.

Hygiene practices during childhood 

Gut development is a continuous process that has its last phase during late infancy/early childhood, as the child transitions from breastmilk to complementary foods. Exclusive breastfeeding (and ingestion of colostrum) is very important for preventing inadequate colonization, as it has bacteria, immune and growth factors which promote immune development. Breastmilk also has a perfect nutritional composition, with oligosaccharides (and other components) that promote the growth and establishment of commensal bacteria (predominately Bifidobacteria). Formula-fed infants display an aberrant gut microbiota and normal colonization is severly delayed (if not completely disrupted).

Microbial exposure favors diversification and exposure to pathogens, which is necessary for stimulation of immune memory and tolerance. Contact with animals, eating raw food and playing in the dirt are a necessary part of a healthy lifestyle in infancy. Excessive hygiene and antibiotic use promote dysbiosis. 

Family diet, diet history and maternal environment 

The diet eaten by your father, mother and grandparents influences the expression of genes involved in energy metabolism. These effects are transmitted intergenerationally and lasting during adulthood. Inadequate dietary patterns followed during childhood and adulthood worsen immune and metabolic function. Additionally, maternal status during pregnancy (stress, nutrition, etc.) has profound effects on many genes. 

Medical history 

Previous diseases, antibiotic abuse and utilization of other substances can influence both the normal functioning of the immune system as well as metabolism.

Genotype 

The presence of certain alleles are important for tolerance of specific food components (ie. lactose) and variability in immune responses (ie. MHC alleles, cytokine gene polymorphism).

Social/lifestyle experiences 

Having bad social relationships, lack of optimism, stress and other common lifestyle experiences affect the inflammatory status of the body. For example, losing a game in very competitive persons increases the levels of inflammatory cytokines higher than in non-competitive subjects. Mental stress also seem to affect the composition of the gut microbiota and gut permeability. 

Symptoms, self-assessment and bloodwork

Any symptom (either bad or good) is valuable for trying to identify potential problems. Self-assessment, including anthropometric measures, emotional status or the characteristics of feces can also help narrowing the spectrum of possible disorders. Bloodwork and biomarkers are helpful for confirming assumptions and health status.

Summary

Before beginning any nutritional therapy, it is important to check for past events and factors that affect the metabolic and immune status. This will aid in finding the right dietary composition that helps the most with a given problem and reducing the time of experimentation needed for finding the adequate nutrition and supplementation for an individual.  

*Although recent evidence suggests that bacteria colonize the gut in uterus.

Human microbiota and atherosclerosis

I've been wanting to post about this study for a while now. I think its a good update while I finish my first post on my nutritional immunotherapy protocol. This study was performed given the preliminary evidence linking infections and atherosclerosis, and the association of the human microbiota with the atherosclerotic plaque. For example, bacterial DNA has been observed in atherosclerotic plaques from young and old subjects (1, 2). This relationship has been investigated with more focus on oral bacteria, due to the association of periodontal disease and cardiovascular disease (CVD) (3, 4) and the presence of periodontal pathogens in  atherosclerotic plaques (5). 

The authors tried to answer the following questions:

Is there a core atherosclerotic plaque microbiota? 

Are bacteria present in the plaque also detectable in the oral cavities or guts of the same individuals?

Do the microbiotas of the oral cavity, gut, and atherosclerotic plaque relate to disease markers such as plasma levels of apolipoproteins and cholesterol? 

Is an altered oral or fecal microbiota associated with atherosclerosis?

Using 16S rRNA sequences (from patients with clinical atherosclerosis and controls) and the unweighted UniFrac distance metric (qualitative instead of quantitative), they found strong clustering of samples according to body site, suggesting that the oral, gut and atherosclerotic plaque (AP) sites have different microbial communities:

PC1 and PC2 refer to the first two principal coordinates from the principal coordinate analysis of unweighted UniFrac, plotted for each sample (See also Fig S1). Of these sites, bacterial diversity was higher for the gut microbiota. 

The analysis of the atherosclerotic plaque microbiota revealed that there was a positive correlation between the amount of bacterial 16S rRNA and the number of leukocytes present in the AP, and there was significantly higher levels of Proteobacteria and fewer Firmicutes compared with the oral and gut samples. Supporting the role for a "core" AP microbiota, several OTUs were present in all AP samples, which differentiated these samples from oral or fecal samples: Chryseomonas was detected at high levels in the AP samples, but not in gut or oral samples, being the most discriminative genus between sites and driving the differences between body sites. Other OTUs, three for the genus Staphylococcus, three classified as Propionibacterineae and one belonging to the genus Burkholderia, were specific for AP samples and were present in all AP samples analyzed.

There were no OTUs differentiating oral samples from healthy subjects and patients, but there were correlations between the abundances of OTUs in the oral cavity and CVD markers: the abundance of Fusobacterium was positively correlated with levels of cholesterol (P = 0.028) and LDL (P = 0.005), the abundance of Streptococcus was positively  correlated to HDL (P = 0.0001) and ApoAI (P = 0.01) levels and the abundance of Neisseria was negatively correlated to levels of these last two markers (P = 0.02 and 0.005, respectively). This is interesting, given that Fusobacterium has been associated with periodontal disease (6). As with oral samples, there were no differentiating OTUs between gut samples from controls and patients (in terms of OTU abundances). In gut samples, the abundance of two OTUs classified as uncharacterized members of Erysipelotrichaceae and Lachnospiraceae families were positively correlated with cholesterol (P = 0.009 and 0.001, respectively) and LDL (P = 0.012 and 0.007, respectively). 

Finally, inter-individual comparisons between sites showed that some OTUs were shared among sites. These included OTUs for Veillonella (in AP and oral samples in 11 of 13 patients, detected also in the gut sample of two patients) and Streptococcus (in AP and oral samples in 6 of 10 patients, detected also in the gut of four patients). Within patients, the AP samples contained OTUs shared with oral (Propionibacterium, Rothia, Burkholderia, Corynebacterium, Granulicatella, Staphylococcus) and gut (Bacteroides, Bryantella, Enterobacter, Ruminococcus) samples. 

Summary

• The study identified a "core" atherosclerotic plaque microbiota, comprising higher levels of Proteobacteria and fewer Firmicutes, compared with the gut and oral samples. 
• The AP microbiota contained specific OTUs not shared with the analyzed body sites.
• The abundance of some OTUs in the gut and oral cavity was correlated with CVD markers. 
• Shared OTUs among sites included Streptococcus and Veillonella, the correlation being stronger among the oral cavity and the AP, and these OTUs were also found in the gut samples from some patients. Across patients, the abundance of both were correlated in the oral cavity and the AP.

Commentary

I find this study very interesting because it supports the role of infection on the pathogenesis of atherosclerosis and CVD. The "infection hypothesis" of atherosclerosis has been proposed before (7). The fact that specific bacteria is present in AP and not in other body sites analyzed and that the amount of bacterial 16S rRNA was positively correlated with leukocyte counts, support the notion that these pathogens support directly atherosclerosis progression. However, the study only analyzed the oral cavity and the gut, so it is impossible to conclude that these pathogens couldnt have been derived from other body sites (for example, the skin). Moreover, primers commonly utilized to amplify 16S rRNA sequences are limited to some species, an inherent property of the method (8). Nevertheless, it seems more feasible to suppose that the origin of AP bacteria is the oral cavity because of the close proximity of the bacterial communities in the mouth to the highly vascularized gingival lining and because of the thickness of the subgingival epithelium, which differs from other protective layers such as the skin or the gut mucosa (9). Accordingly, any mechanical disruption of oral bacterial biofilms can trigger bacteremia, and these include oral procedures (periodontal probing, tooth extractions, etc), oral hygiene activities (such as brushing) and physiological phenomena (like chewing) (9). This, coupled with the findings that the abundance of OTUs in the AP were correlated with that of the oral cavity support this hypothesis. Gut bacterial origin is more complicated but feasible (as shown by the presence of gut bacteria in AP samples). The authors suggest that one possible way of this transfer is by phagocytosis of macrophages at epithelial linings. 

If indeed bacteria play a role in the formation and/or progression of atherosclerosis, the million dollar question is why do these specific pathogens adhere to the vascular endothelium? Moreover, is this colonization the initial trigger for the localized inflammatory response or just aggravates the condition? With the available evidence it is hard to answer these questions. It has been proposed that atheromas might act as mechanical sieves, collecting bacteria from the cirulation (10). This would have deleterious consequences, as bacterial accumulation in the AP would lead to an increased inflammatory response. It could also link the fact that endotoxemia increases the risk of CVD (11), for which periodontal pathogens seem to play an important role (12). Supporting the role of infection as secondary to atherosclerotic inflammation, fungal DNA has been observed in AP (13), with some species correlated with that found on human microbial communities. It is of worth noting that in this study, fungal richness was not associated with classical CVD risk factors. Because not all normal residing oral bacteria are found in AP samples, AP invasion might be related to the virulence properties of some species (9). This seems to be the case, as in the study reviewed here, there was a common abundance of Streptococcus and Veillonella in AP samples. Streptococcus is able to adhere to the endothelium, while Veillonella is able to change its adherence capacity in the presence of some factors from Streptococcus (14). In fact, there is a tight relationship between Streptococcus and Veillonella in the oral cavity as some strains co-aggregate, partially because Veillonella seems to be metabolically dependent on Streptococcus (15). This relationship is so important that Veillonella is unable to establish an infection without Streptococcus.

In conclusion, microbial accumulation in AP might contribute to the progression of atherosclerosis. Although the mechanism by which these microorganisms colonize this site is not defined, it is clear that several microbes found in other body sites are also found in AP, which suggests that the normal human microbial communities are an important source of pathogens contributing to atherosclerosis progression. Translocation from these sites, in turn, is controlled by the host inflammatory status. This seems to be relevant to translocation from the oral cavity: transient bacteremia is experienced by everyone because of mechanical disruption of microbial biofilms (for instance, when eating), but it is controlled quickly. However, there are always some persisters which resist by-standing immune mechanisms. Obviously, a higher microbial load facilitates dissemination into the bloodstream and could possible influence the degree of transient bacteremia. A higher bacterial load coupled with a compromised subgingival epithelium barrier increases the risk of bacteremia and secondary colonization. So, in order to reduce the risk of AP colonization by oral pathogens, it is wise to target these two factors. For reducing oral bacterial load (overgrowth), reducing sucrose intake might be of benefit, as bacterial glucosyltransferase (GTF) plays a crucial role in plaque formation (16) and S.mutans is only pathogenic in the presence of sucrose (17). Dietary sucrose has been shown to increase total viable microbial density and S.mutans population in human dental plaque (18). Sucrose alone seem to be more cariogenic than sucrose plus fructose (19, 20). Additionally, sucrose alters the ionic concentration in the biofilms' matrix, altering the normal de- and re-mineralization process of enamel and dentin  (21). The role of starches in dental plaque formation is controversial (22), although some authors are in agreement with the Cleave & Yudkin hypothesis, which states that an excess of fermentable carbohydrate intake (in the absence of dental interventions) promotes dental diseases and then systemic diseases (23). Nevertheless, starchy foods commonly ate might promote dental plaque formation and disease. Pollard (24) showed that cornflakes, branflakes and wholemeal bread produced the minimum dental plaque pH peak, while all foods tested promoted enamel demineralization*. This might be related to the fact that, although starches can reduce plaque pH and induce demineralization, sucrose accelerates this effects (25). This is probably mediated by the interaction between bacterial GTF and salivary amylase (26). In contrast to what some might expect, whole fruit and fruit juices induce enamel demineralization by the same magnitude (27). This has been also found in some observational studies, where high fruit consumption is associated with increased caries risk (28). On the contrary, cheese and nuts have shown a negative association (29). Finally, inflammation increases the risk of oral bacterial growth and translocation, which might induce and/or aggravate systemic diseases (30). Periodontal disease has been positively associated with obesity (31), metabolic syndrome (32), type 2 diabetes (33), Alzheimer's disease (34), among other. Thus, controlling inflammation is key to avoid secondary diseases caused by pathogenic oral bacteria. 

* "Test foods were oranges, apples, bananas, Cornflakes, Branflakes, Weetabix, Alpen (no added sugar), white bread, wholemeal bread, rice, and spaghetti, with positive and negative controls of sucrose and sorbitol."

Koren O, Spor A, Felin J, Fåk F, Stombaugh J, Tremaroli V, Behre CJ, Knight R, Fagerberg B, Ley RE, & Bäckhed F (2011). Human oral, gut, and plaque microbiota in patients with atherosclerosis. Proceedings of the National Academy of Sciences of the United States of America, 108 Suppl 1, 4592-8 PMID: 20937873