Category Archives: Scare Stories

Neu5gc, Red Meat, and Human Disease: Part I

A number of people asked for comments on the most recent red meat scare, including Nicole, Ryan, and Mishkin on the blog, JT Olds on Twitter, and others on Facebook. You probably saw some of the headlines:

The article Nicole linked is a bit more scientifically inclined: “Possible Link Between Red Meat Consumption And Increased Cancer Risk Identified” (IFL Science). Here’s the press release version from UCSD: “Sugar Molecule Links Red Meat Consumption and Elevated Cancer Risk in Mice”. In the blogosphere, Stephan has summarized the issue in the context of a post on red meat and cancer.

The headlines are based on a paper [01] that reported that, in mice genetically altered to lack a sugar (Neu5gc) that humans also lack, feeding Neu5gc and injecting anti-Neu5gc antibodies generates inflammation which can promote the growth of cancers.

Significance of Neu5gc

The paper itself is a rather artificial scenario whose significance will be determined by future work. So analyzing this single paper would not be interesting. But I think it’s worthwhile to look into the broader idea that eating Neu5gc-bearing meats might be inflammatory or a source of autoimmunity.

In terms of PHD recommendations, this could affect the relative emphasis we place on different meats. If Neu5Gc is a true health risk, then we would want to emphasize seafood more and red meat less.

Another benefit to thinking about Neu5gc is that it may give us some insight into what a PHD “autoimmune protocol” should look like.

Background: Evolutionary History of Neu5gc

All cells in multicellular organisms are coated in carbohydrates, and the carbohydrates terminate in one of 43 sialic acids. In mammals, two forms of predominate: Neu5Ac and Neu5Gc. Each mammalian cell has tens or hundreds of millions of molecules of Neu5Ac and Neu5Gc on its surface. [02]

Neu5Gc is made from Neu5Ac, but the gene for making Neu5Gc was inactivated in the human lineage shortly before the emergence of Homo. The mutation occurred 3.2 million years ago and reached fixation – that is, all ancestral hominids had come to have the mutated gene – 2.9 million years ago. [03] This very rapid fixation indicates there was strong selection in support of the mutation.

In fact, this mutation by itself may have led to a speciation event, after which our ancestors could no longer mate with other apes. From that point on, Neu5gc-less females had difficulty producing children with males who retained the Neu5gc gene, because they would form antibodies against Neu5gc-coated sperm, making fertilization unlikely. [04]

Why was losing Neu5gc selected in our ancestors? Two possibilities are likely:

  • Loss of Neu5gc improved brain function.
  • Loss of Neu5gc (temporarily) reduced vulnerability to (ancestral) pathogens.

It should be noted that Neu5Gc has been lost independently in some other mammals as well – ferrets and new world monkeys. New world monkeys such as capuchins and spider monkeys also experienced a brain expansion, and ferrets are notably smart, so either explanation might be relevant to these cases of “convergent evolution.”

Neu5gc and Brain Function

Carbohydrates are extremely important for intercellular interactions. Indeed, the incorporation of carbohydrates into cell membranes and extracellular matrix is what made possible the rise of multicellular life.

In no organ are intercellular interactions as complex or consequential as in the brain. Not surprisingly, then, carbohydrates including the sialic acids are important to brain function.

The human brain is extraordinarly rich in sialic acids: neural membranes have 20 times more sialic acids than membranes of other human cell types. Animal brains are also enriched in sialic acids relative to their other tissues, but not as much as in humans; the human brain has 2-4 times more sialic acids the brains of other mammals. [05]

Curiously, though, Neu5gc is rare in the brains of all animals. Neu5gc is strongly suppressed, by about 98%, in the brains of all vertebrates, suggesting that its presence inhibits brain function. [06] It appears that Neu5gc is somehow toxic to brain function.

Loss of the gene for Neu5gc completely eliminated Neu5gc from the hominid brain. If Neu5gc does impair brain function, mutational inactivation of Neu5gc would have improved brain function. If so, the mutational inactivation of Neu5gc could have been driven by the same evolutionary forces that, soon after, selected for the tremendous expansion of the hominid brain.

Incidentally, dietary sialic acids — except for Neu5Gc – appear to be nutritious for humans, and especially for the developing infant brain. Breast milk is exceptionally rich in sialic acids, almost all of it Neu5Ac. Formula, by contrast, has much lower levels of sialic acids (0.21 mmol/L compared to 3.72 mmol/L in breast milk). Breast fed infants have nearly twice as many sialic acids in saliva than formula fed infants, confirming that milk sialic acids are taken up by the body and utilized.

Animal studies show that sialic acids in breast milk nourish the brain. Sialic acids facilitate neurotransmission between neurons. When piglet milk is supplemented with sialic acids, brain sialic acid levels are increased, and the piglets learn faster and make fewer mistakes in maze tests. [05] Rodents also perform better on tests of learning and memory after sialic acid supplementation. [07]

Not only does formula have fewer sialic acids than breast milk, cow milk based formulas have some Neu5Gc. [05] It has been observed that formula-fed infants have lower IQs than breast-fed infants. Sialic acids might help explain that. The lack of nourishing Neu5Ac and the presence of toxic Neu5Gc in formula might lastingly impair brain function in formula-fed infants.

Neu5Gc and Infection Risk

As the outermost molecules in the carbohydrate coat surrounding cells, sialic acids are the first contact point for pathogens seeking entry to the cell, and for immune cells seeking to detect whether the cell is native or foreign.

There has been a “Red Queen” evolutionary arms race in which pathogens evolved ways to utilize sialic acids for cell entry, or to hide from the immune system; and animals evolved changes to their sialic acids to frustrate the pathogens. [08]

Many pathogens interact with sialic acids in order to adhere to and gain entry into the cell. Pathogens generally rely on a single specific endocytic route for cell entry. This often requires binding to a specific sialic acid as the initial point of attachment.

Pathogens that specifically utilize Neu5Gc to enter cells include canine and feline parvoviruses [09]; pathogens that specifically utilize Neu5Ac include adeno-associated viruses and the minute virus of mice (MVM) [10].

A human pathogen that uses sialic acids to enter cells is the malaria protozoan. Plasmodium falciparum causes severe disease in humans and enters cells via Neu5Ac; Plasmodium reichenowi causes milder disease in chimpanzees and gorillas and enters cells via Neu5Gc. Plasmodium falciparum appears to have evolved recently – possibly reaching its current form only 10,000 years ago when the rise of agriculture and animal husbandry brought humans and mosquitos into closer proximity – while Plasmodium reichenowi is thought to resemble the ancestral form that would have afflicted hominids and apes 3.5 million years ago.

Possibly, the gene for Neu5Gc was inactivated to protect ancestral hominids from malaria. With the loss of Neu5Gc, hominids would have become immune to P. reichenowi. [11] [12]

Unfortunately, after P. falciparum’s adaptation to Neu5Ac which is overabundant in humans, we now suffer from more severe malaria than chimpanzees or gorillas (the “malignant malaria” mystery). [13]

In addition to entry points for microbes, sialic acids can be entry points for microbial toxins. For example, Shiga toxin from shigatoxigenic E. coli binds to Neu5Gc. [14]

Sialic Acid Concealment and the Gut Microbiome

The immune system is sensitive to the composition of the carbohydrate coat on a cell. White blood cells have a number of sialic acid detectors on their surfaces (called Siglecs, for sialic acid Ig-superfamily lectins). Some, which bind to human sialic acids, inhibit immune responses. Others, which bind to non-human sialic acids, activate immune responses.

Thus, when white blood cells contact a cell bearing human sialic acids, the immune system interprets it as “self” and tamps down immunity. When it detects foreign sialic acids, the immune system treats the cell as “foreign” and is more likely to attack it.

Some microbes – including commensal gut microbes – have been living in humans long enough that they have learned to take up sialic acids, chiefly Neu5Ac, and incorporate them into lipopolysaccharides on their cell membranes. This suppresses immunity toward them. [15]

A number of human pathogens have learned the same trick. Pathogens that incorporate sialic Neu5Ac into their cell membranes for the purpose of mimicking human cells and evading human immune defenses include Escherichia coli K1, Haemophilus influenzae, Pasteurella multocida, Neisseria spp., Campylobacter jejuni and Streptococcus agalactiae. [16]

Due to this “molecular mimickry” of human molecules, it has been suggested that these bacteria – especially Haemophilus influenzae and Neisseria spp. – may be sources of autoimmunity. [17]

While some bacteria can synthesize sialic acids themselves, most obtain it from their environment. These bacteria release enzymes called sialidases to cleave the sialic acids from food in the digestive tract, from surrounding cells, or from mucus. [15] Bacteria can obtain Neu5Ac from human tissue and mucus as well as food, but Neu5Gc only from food, chiefly beef and pork.

Neu5Gc in Human Tissue

Although humans can no longer synthesize Neu5Gc, we still have all the cellular machinery for utilizing it. When dietary Neu5Gc is absorbed into the body and enters cells, it can be incorporated into glycoproteins bound for the cell surface glycocalyx, just as Neu5Ac is.

As a result, Neu5Gc of dietary origin appears at low levels on the surface of human cells.

Neu5gc is found at high levels in all mammals except humans, ferrets, and new world monkeys; birds and reptiles do not produce Neu5Gc at all, and fish and shellfish produce only low levels. So, of the four major meat groups – beef, pork, chicken, and fish – Neu5gc is obtained predominantly from the red meats, beef and pork.

Among human cells, Neu5Gc appears at highest levels on tumor cells, especially metastatic cells. [21] This makes Neu5Gc a potential target for cancer therapy.

Neu5Gc as an Immunogen

Neu5gc expressed on cell walls is a potential immunogen. When pig organs are transplanted to humans, Neu5Gc is the second most important cause of rejection, after the α1,3-galactose (αGal) epitope. [20]

Anti-Neu5gc antibodies have been found in 85% of humans. [18] It is thought that antibodies form in early childhood after dietary Neu5Gc is incorporated by certain gut bacteria into lipooligosaccharides that can generate antibodies. Some of these antibodies may cross-react with compounds human cells form from dietary Neu5Gc; these human molecules are then known as “xeno-autoantigens.” [21]


Neu5Gc from mammalian meats, such as beef and pork, is incorporated into the cell surface coats and walls of gut microbes and some human cells, mainly in the gut and in tumors. Neu5gc in bacterial walls is immunogenic and 85% of people have detectable antibodies to Neu5Gc. Eating beef and pork supplies antigens for these antibodies, potentially triggering inflammation. There are concerns that this inflammation may have negative health effects.

Next up: Neu5Gc and autoimmunity.

Note: May Perfect Health Retreat

Interested in a luxury vacation combined with an education in how to be healthy for the rest of your life? Science classes, cooking classes, movement and relaxation classes, personal health coaching, gourmet food, all on a magnificent beach, with hot tubs and salt water pools? Come to the Perfect Health Retreat! Spaces are currently 2/3 filled and the retreat is expected to sell out early. Visit here for more info or email me at



[01] Samraj AN et al. A red meat-derived glycan promotes inflammation and cancer progression. Proc Natl Acad Sci U S A. 2014 Dec 29. pii: 201417508. [Epub ahead of print].

[02] Kraemer PM. Sialic acid of mammalian cell lines. J Cell Physiol. 1966 Feb;67(1):23-34. Was 21

[03] Hayakawa T, Aki I, Varki A, Satta Y, Takahata N. Fixation of the human-specific CMP-N-acetylneuraminic acid hydroxylase pseudogene and implications of haplotype diversity for human evolution. Genetics. 2006 Feb;172(2):1139-46. Full text: Was 22

[04] Ghaderi D et al. Sexual selection by female immunity against paternal antigens can fix loss of function alleles. Proc Natl Acad Sci U S A. 2011 Oct 25;108(43):17743-8. was 2

[05] Wang B. Molecular mechanism underlying sialic acid as an essential nutrient for brain development and cognition. Adv Nutr. 2012 May 1;3(3):465S-72S. Full text: Was 31

[06] Davies LR, Varki A. Why Is N-Glycolylneuraminic Acid Rare in the Vertebrate Brain? Top Curr Chem. 2013 Mar 8. [Epub ahead of print] was 8

[07] Wang B. Sialic acid is an essential nutrient for brain development and cognition. Annu Rev Nutr. 2009;29:177-222. was 32

[08] Varki A. Colloquium paper: uniquely human evolution of sialic acid genetics and biology. Proc Natl Acad Sci U S A. 2010 May 11;107 Suppl 2:8939-46. was 51

[09] Löfling J et al. Canine and feline parvoviruses preferentially recognize the non-human cell surface sialic acid N-glycolylneuraminic acid. Virology. 2013 May 25;440(1):89-96. was 54

[10] Wu Z et al. Alpha2,3 and alpha2,6 N-linked sialic acids facilitate efficient binding and transduction by adeno-associated virus types 1 and 6. J Virol. 2006 Sep;80(18):9093-103. was 55

[11] Varki A, Gagneux P. Human-specific evolution of sialic acid targets: explaining the malignant malaria mystery? Proc Natl Acad Sci U S A. 2009 Sep 1;106(35):14739-40. was 57

[12] Martin MJ et al. Evolution of human-chimpanzee differences in malaria susceptibility: relationship to human genetic loss of N-glycolylneuraminic acid. Proc Natl Acad Sci U S A. 2005 Sep 6;102(36):12819-24. was 58

[13] Rich SM et al. The origin of malignant malaria. Proc Natl Acad Sci U S A. 2009 Sep 1;106(35):14902-7.

[14] Byres E et al. Incorporation of a non-human glycan mediates human susceptibility to a bacterial toxin. Nature. 2008 Dec 4;456(7222):648-52.

[15] Varki A, Gagneux P. Multifarious roles of sialic acids in immunity. Ann N Y Acad Sci. 2012 Apr;1253:16-36. Full text:

[16] Severi E, Hood DW, Thomas GH. Sialic acid utilization by bacterial pathogens. Microbiology. 2007 Sep;153(Pt 9):2817-22. Full text:

[17] Harvey HA, Swords WE, Apicella MA. The mimicry of human glycolipids and glycosphingolipids by the lipooligosaccharides of pathogenic neisseria and haemophilus. J Autoimmun. 2001 May;16(3):257-62.

[18] Zhu A, Hurst R. Anti-N-glycolylneuraminic acid antibodies identified in healthy human serum. Xenotransplantation. 2002 Nov;9(6):376-81.

[19] Takahashi T et al. N-glycolylneuraminic acid on human epithelial cells prevents entry of influenza A viruses that possess N-glycolylneuraminic acid binding ability. J Virol. 2014 Aug;88(15):8445-56.

[20] Park JY et al. α1,3-galactosyltransferase deficiency in germ-free miniature pigs increases N-glycolylneuraminic acids as the xenoantigenic determinant in pig-human xenotransplantation. Cell Reprogram. 2012 Aug;14(4):353-63.

[21] Samraj AN, Läubli H, Varki N, Varki A. Involvement of a non-human sialic Acid in human cancer. Front Oncol. 2014 Feb 19;4:33.

Lessons From The Latest Red Meat Scare

I’ve had about ten requests for thoughts on the new paper in Nature Medicine [1] that finds red meat can promote atherosclerosis by a roundabout route: carnitine in the meat is metabolized by gut bacteria into a compound called TMA, which the liver converts to TMAO, which in high doses promotes growth of atherosclerotic plaques.

The same group has done similar studies with other molecules; two years ago the culprit was not carnitine but phosphatidylcholine. [2]

The Scare

Some of the news stories:

It sounds like red meat is dangerous! The best line came from the New York Times:

Lora Hooper, an associate professor of immunology and microbiology at the University of Texas Southwestern Medical Center, who follows the Paleo diet, heavy on meat, exclaimed, “Yikes!”

The Big Picture

The issue here is closely related to one discussed in page 77 of our book:

Protein is not food for us alone; gut bacteria can ferment protein.

Although fermentation of carbohydrates by gut bacteria is usually beneficial, fermentation of protein is not: it generates toxic compounds, including amines, phenols, indoles, thiols, and hydrogen sulfide, which make a foul-smelling stool.

It seems likely, therefore, that high protein intakes are suboptimal for gut health.

When protein is fermented, nitrogen is released, and many nitrogenous compounds are toxic.

The group behind the new research, led by Stanley Hazen, has been looking at another pathway by which bacterial fermentation of meat might be dangerous – the pathway that runs through Trimethylamine. Trimethylamine (TMA) has a simple structure; three methyl groups bonded to a nitrogen atom: N(CH3)3.

Compounds such as choline and carnitine that contain both methyl groups and nitrogen are potential precursors to TMA.

TMA is responsible for the fishy smell of decaying fish. It is highly abundant in fish.

The liver converts TMA into its oxide, TMAO. The Hazen group in a series of papers has argued that higher TMAO levels in blood are associated with atherosclerosis. In a recent paper they assert, “TMAO levels explain 11% of the variation in atherosclerosis.” [3]

So, the equation they are putting together is: fermentation of meat in the gut produces TMA leading to TMAO production which may increase your chance of atherosclerosis by 11%.

Risk is Highly Dependent on the Nature of Your Gut Flora

Here is the key data from the new paper [1]. The “d3” prefix means that the carnitine was labeled with deuterium, an isotope of hydrogen, to help trace its molecular destinations.

In the left panel of part (e), subjects have been fed a steak (eaten in 10 minutes) plus a 250 mg deuterated carnitine supplement – in total, the carnitine equivalent of 1.5 pounds of meat. Deuterated TMAO levels in blood rise to about 1.8 parts per million after 24 hours.

Then subjects are given antibiotics for a week to suppress their gut flora, and fed steak and deuterated carnitine again. On antibiotics, their blood has no deuterated TMAO at all 24 hours after the meal.

In the right panel, 3 weeks after coming off antibiotics to allow gut flora to regrow, subjects are challenged again with steak and deuterated carnitine. Their blood level of deuterated TMAO now exceeds 12 parts per million – 7 times higher than before the antibiotics.

They go on to test the mix of flora in subjects, and show that flora composition is closely correlated with blood levels of deuterated TMAO after consumption of deuterated carnitine. Some types of gut bacteria produce a lot of TMA from food carnitine, others produce little.

So the amount of TMAO entering the blood from bacterial metabolism of food carnitine is highly dependent on the nature of the gut flora. If you kill off normal flora with antibiotics, then eat meat and carnitine, you will get an overgrowth of bacteria specialized to feed on meat and carnitine. That might not be good for you.

The Vegan vs Omnivore Comparison

Food carnitine is far from the only source of blood TMAO. In fact, TMA is a natural breakdown product of choline, one of the most abundant molecules in the body, and the body has evolved an enzyme for converting TMA into TMAO — the gene is FMO3. So we should ask, how much does metabolism of carnitine by gut bacteria affect blood TMAO levels? For that we need measurements of normal TMAO, not just deuterated TMAO.

We can see what that data looks like in a plot comparing blood TMAO levels between vegans and omnivores (panel c of their Figure 2):

These are the TMAO levels normally circulating in the fasting blood of SAD omnivores and vegetarians. As you can see, there’s considerable overlap between the two distributions. 75% of the omnivores had TMAO levels within the same range as 90% of the vegetarians.

So about 75% of omnivores and 90% of vegetarians have normal TMAO levels. What about the 25% of omnivores and 10% of vegetarians whose TMAO levels are elevated?

In panel e, you can see that the enterotype of the gut flora is a much better predictor of blood TMAO levels than whether someone eats meat. Those with high Prevotella, low Bacteroides averaged about triple the TMAO levels of those with low Prevotella, high Bacteroides flora.

So it really is the gut flora that determine blood TMAO levels.

What Drives the Gut Flora?

What determines whether you have the bad gut flora?

The general picture is this. The immune system regulates the number of microbes living in the gut. When levels become high, antimicrobial peptides are released into the gut to kill some off. When levels are low, antimicrobial peptide production is reduced to let microbes multiply.

This means that if the proportion of bacteria who feed on protein, carnitine, and choline is too high, it’s probably because there is insufficient food for the competing bacteria who feed on carbohydrate forms of fiber. If you have a lot of gut bacteria feeding on fiber, there’s no room in the gut for large amounts of bacteria who feed on meat.

So the 25% of omnivores and 10% of vegan/vegetarians with high TMAO levels are probably the people on low-fiber diets – the ones who get their carbs from flour and sugar. On such a diet, the good bacteria are starved and the bad bacteria that produce TMA multiply.

How Does TMAO Produce Atherosclerosis?

The explanation offered by the Hazen group is that TMAO suppresses “reverse cholesterol transport” conceived broadly as the process of migrating excess cholesterol out of macrophages for transport to the liver and excretion in feces via the bile.

Basically, the idea here is:

  1. Atherosclerosis begins with metabolic syndrome, a state characterized by high LDL levels and caused by endotoxemia (high levels of endotoxins entering the body from the gut).
  2. As we’ve discussed (“Blood Lipids and Infectious Disease, Part II,” July 12, 2011), LDL particles have an immune function. They are oxidized by microbial cell wall components. The resulting oxLDL particles are taken up by macrophages, which then present the microbial cell wall components to other immune cells for antibody formation.
  3. Endotoxemia initiates the process of atherosclerosis by (a) poisoning the liver to cause metabolic syndrome which raises LDL levels, and (b) oxidizing LDL – since endotoxins are bacterial cell wall components that can oxidize LDL – and driving the oxLDL into macrophages.
  4. After macrophages have separated the microbial cell wall components from their accompanying LDL particle, the cholesterol and fat have to be exported to keep them from building up in the cell.
  5. If cholesterol and fat cannot be exported quickly enough, the macrophage is injured and becomes a “foam cell.” Disabled foam cells accumulate in specific locations and form atherosclerotic plaques.
  6. TMAO suppresses bile acid creation, reducing the excretion of cholesterol from the body and leading to higher LDL levels and a greater likelihood that macrophages will become foam cells.

If this is true, then TMAO is not intrinsically atherosclerotic. TMAO in blood only becomes atherosclerotic in the context of metabolic syndrome brought on by endotoxemia.

What causes endotoxemia? A dysbiotic flora generated by a diet high in sugar, flour, and omega-6 fats (see our book, pp 220-222).

Conclusion: Lessons Learned

The lessons of this study are:

  1. Don’t eat a high-sugar, high-flour, low-fiber diet.
  2. Do eat natural whole foods that have the kind of fiber we and our probiotic gut flora co-evolved eating; mainly, resistant starch from in-ground starches like potatoes and soluble fiber from fruits and vegetables.
  3. Don’t eat excessive amounts of meat. As we noted in the book, excess protein is available to gut bacteria for fermentation and that produces a number of toxic byproducts.
  4. Do eat PHD levels of meat – one-half to one pound per day. This level of meat consumption will provide healthful and nourishing amounts of protein, choline, and carnitine, and will not cause any harm if accompanied by PHD levels of healthy plant foods.

None of these lessons is new. This study doesn’t overturn any established dietary wisdom. It is just one more piece of data reminding us to eat a balanced diet consisting of the foods we evolved eating – plant as well as animal.


[1] Koeth RA et al. Intestinal microbiota metabolism of l-carnitine, a nutrient in red meat, promotes atherosclerosis. Nat Med. 2013 Apr 7.

[2] Wang Z et al. Gut flora metabolism of phosphatidylcholine promotes cardiovascular disease. Nature. 2011 Apr 7;472(7341):57-63.

[3] Bennett BJ et al. Trimethylamine-N-oxide, a metabolite associated with atherosclerosis, exhibits complex genetic and dietary regulation. Cell Metab. 2013 Jan 8;17(1):49-60.