Summary of the 24th Marabou Symposium: Nutrition and the Human Microbiome


WPT James, London School of Hygiene and Tropical Medicine and the International Association for the Study of Obesity. Charles Darwin House, 12 Roger Street, London WC1N2JU UK. E-mail:, Phone: +44-207-685-2586, Fax: +44-207 685 2581.


This summary covers the articles and attributed discussion in the present supplement, which resulted from the 24th Marabou Symposium titled “Nutrition and the Human Microbiome”, which was held in Stockholm in 2011 with the participation of about 40 global experts in microbiology, physiology, biology, and medicine. The individual articles address a number of topics related to the human microbiome; the attributed discussion, however, offers much more on the nature of the current scientific debate and provides insight into new opportunities for research as well as possible effects of the gut microbes, ranging from possible prenatal epigenetic effects to brain function and behavior.


It has been known for over three centuries that distinctly different bacteria are found in the mouth and in feces, but only very recently has it been shown that the human gut, with its extraordinary array of trillions – probably 1011– of microbes, has a remarkably distinct population unlike any other in the biosphere.1 Bacteria make up the bulk of the biomass in the human gut and in other microbial sites, but there are also other organisms, such as archaea, eukaryotes, and viruses, present in much smaller numbers that nevertheless have potentially important functions. For decades, many unsuccessful attempts were made to culture and identify all of these organisms, but now there is increasing evidence that nearly all of them could be cultured if efforts to this end, for example, in the collection and culture of anaerobic organisms in particular, were more rigorous. Nevertheless, new approaches, which are not dependent on culturing the organisms, are not only rapidly advancing the entire field, they are also providing completely new insights. These derive from a multiplicity of molecular and other techniques, e.g., proteomic and metabolite analyses. However, by scanning the huge number of genes in the human microbiome, it is now possible to devise a phylogenetic approach, with the array of gene data allowing allocation of genes to particular classes of microorganisms. These new molecular and mathematical approaches have revolutionized the study of the microbiome; in fact, based on the latest assessment by the major international project the MetaHIT Consortium, there are about 3.3 million nonredundant genes in the human gut.2

The term “metagenomics” refers to the array of DNA genes whose identities were established by the initial genome scans; now, however, there is also an emphasis on the assessment of 16S rRNA as appropriate markers. The current debate focuses on how best to discriminate between microbial species and whether scans should be performed for particular genes or for metabolic functions, since there is increasing evidence that there may be common metabolic capacities associated with similar genes but derived from very different organisms. Combinations of techniques are also being utilized, with proteomics being used to assess gene expression. Metabolite analyses are beginning to highlight the interaction between the compounds secreted or contained within bacterial remnants and the host's multiple responses to microbial metabolism and to signaling from both unabsorbed organisms and absorbed fragments. Adding to this complexity is the multiplicity of interbacterial molecular transfers occurring within the gut, including transfers of molecular components from ingested plants, which may also contribute to bacterial gene plasticity and metabolism. Given this plasticity in the microbiome, the question is raised as to what is meant by the seeming “stability” of a person's microbiome.

This challenge of how to identify and discriminate factors controlling the diversity, resilience, and specificity of organisms in different sites in the body, in different people, and in different global settings has suddenly become much more interesting because new and cheaper molecular scanning methods have become available since 2010. So, instead of arduously studying the precise nature of a few organisms by a time-consuming assessment of their characteristics as ascertained by culture, it is now possible to monitor a thousand or more individuals in different settings and to begin to evaluate the variety of organisms and how they respond to the prevailing diet and other environmental conditions. Given the huge data set from genetic analyses of millions of different organisms, however, a number of issues must be addressed, including the following: 1) whether scans should be performed for particular genes, 2) how many different genes need to be scanned to be certain that particular species are identified, 3) how particular genes are clustered phylogenetically, and 4) the extent to which studies must be conducted with ever-more detailed analyses to show the resilience of particular species when there are perturbations in the microbial patterns. Bioinformatics is, therefore, extremely important in metagenomic analysis, as elsewhere in molecular biology; however, the identification of a particular range of apparently distinct clusters of organisms may turn out to reflect the detail of the molecular analyses and the bioinformatics used when, on more detailed study, no such clustering of species is found. Thus, it is necessary to question the validity of the approach chosen, as emphasized by Rob Knight in the 24th Marabou Symposium and reported in the present supplement.3 There is now major collaborative research under way with the aim of standardizing the techniques used and documenting the microbiome in children and adults under very different circumstances, and the presently reported symposium stimulated further collaborations.


Given all the caveats in interpreting the new techniques and analyses, it should be highlighted that the gut of the newborn is sterile at birth. However, it is now apparent that the route of birth matters, because babies delivered by Caesarean section promptly acquire the organisms associated with the mother's skin, whereas normally delivered babies have a completely different set of organisms derived from the mother's vagina. Then, as an infant is breastfed, additional organisms are acquired from the mother's skin, but many more organisms will build up, depending on the nature of the environmental microbial load. It is well known that mother's milk contains an array of antibodies and other molecules that selectively suppress a variety of species, and the plentiful lactose in milk also promotes the establishment of lactobacilli, which, as Mulder et al.4 show, limit colonization by pathogens and promote normal immune function. Nevertheless the child's mouth, throat, and upper intestine also begin to respond to the inflow of organisms.

The neonatal period is the critical time for gut colonization, which sets in motion a sequence of host responses along with the establishment of what are termed “nutrient niches” for clusters of specific organisms in very circumscribed microenvironmental sites. Depending on the microenvironment of pH, metabolites, nutrient availability, immune functions, interacting and responsive mucin and epithelial receptors, specific organisms come to occupy the surface of the mouth, throat, and upper and lower intestine and influence the cascade of other microbes that may either be swept on into the colon or become established as what seem to be commensal organisms in the upper intestine. So specific is the establishment of these early organisms that, as David Relman5 documents, the microbes associated with the left third molar tooth may be very different from those habitually on the right third molar. By virtue of these early interactive events, the organisms seem to establish their own priority. This not only influences the ability of other organisms to remain in situ but also seems to entrain the overall host response and immune system. However, the inflow of microbes during the first few months of life means that, as the competition from and the entraining by different and more dominant organisms proceeds, the metabolome of the infant changes rapidly and does not acquire a more stable adult array of microbes until about 3 years of age.

The early inflow of environmental organisms is known to be dramatically different in different parts of the world, but diet also has a profound effect. Thus, within a few days after birth, the gut microbial pattern of a bottle-fed baby, even one fed in a clean environment, is totally different from that of a breastfed baby. If a child is reared in a poor environment, as most of the world's children are, then the microbial content of the gut is dramatically different within a few days. Depending on the magnitude of the bacterial and pathogenic load, the gut's microbes will alter the entire intestinal response and, in some cases, the gut structure, as observed in some poor parts of the world, such as Asia. Yet individuals from these poor populations are far less likely to have allergies and other autoimmune problems or inflammatory disease, even if they migrate to a clean environment in later childhood or adult life. Therefore, this microbial entraining of the immune system early in life seems to be a crucial developmental process.


Ingested microbes, after traversing the stomach – where some will be killed by the stomach's acidity and others will bind to the stomach mucosa – then encounter the mucus of the epithelium of the small intestine. This single layer of mucus is secreted from the goblet cells in the crypts of the small intestine and consists of sulfated glycans, which differ depending on the blood group of the baby. In the newborn, there is little mucin breakdown in the first few months of life under normal environmental conditions. This mucus layer is unattached in the small intestine and fills up the space between most of the villi, so normally most of the small intestine is covered by mucus. The ileum has a thicker layer of mucus that extends about 50 µm above the tip of the villi and therefore protects this part of the intestine from ready bacterial penetration. To some extent, the nature of the bacteria's own surface structure determines whether bacteria bind to the glycans and take up residence; if this occurs, bacteria slowly hydrolyze the mucin as a food source while also having access to luminal nutrients. However, bacteria can then, by their production of exoglycosidase, hydrolyze and remove a single sugar at a time from the mucin molecule. Some bacteria also have proteolytic action, which allows them to permeate the loose mucin layer to interact with the underlying epithelial surfaces.

Breast milk contains a wide array of molecules that not only bind bacteria but also mimic in their structure the glycan molecules of the intestinal surface, so breast milk thereby prevents many bacteria from attaching to the surface of the intestine. The defense against mucosally attached bacteria also, however, involves secreted molecules, the defensins, as well as antibodies of the immunoglobulin A class derived from the crypts. These are often trapped within the mucin layer and offer a further barrier to bacterial penetration. However, there are also innate immune-like receptors within the epithelium itself, and this array of receptors, including the multiple Toll-like receptors (TLRs), bind a wide range of both commensal and pathogenic organisms, thereby preventing them from penetrating the epithelium. The specificity of these innate receptors varies; for example, TLR2 recognizes lipoproteins and zymosan, TLR3 recognizes double-stranded RNA, TLR4 recognizes lipopolysaccharide, and TLR5 recognizes bacterial flagellin, so there are multiple means whereby these receptors can lock on to different types of bacteria without any need for acquiring immunity by developing antibodies or specialized immune cells. Some microbes, however, have nonreactive flagellins, and these have been suggested as a mechanism by which the Helicobacter pylori bacterium so readily becomes a commensal in the stomach in early life.

By the time the bacteria reach the colon, the availability of oxygen is markedly reduced, so anaerobic bacteria proliferate. The glycans of the MUC 2 mucin of the large intestine of humans, unlike glycans in the mucin of the small intestine, are identical in everyone, so bacteria using these glycans as attachment sites constitute a relatively constant component of colonic flora. The human colonic mucus is organized in two layers: an inner, stratified mucus layer that is firmly adherent to the epithelial cells and is approximately 150–200 µm thick, as reported by Gunnar Hansson (see attributed discussion in the present supplement6), and an outer layer. The inner mucus layer is dense and does not allow bacteria to penetrate, thus keeping the colonic epithelial cell surface free from bacteria. The inner mucus layer is progressively converted into the outer layer, which is the habitat of the commensal flora. Experimental colitis can be induced when there is a deficiency in the production of the O-linked glycans in the gut, which emphasizes the importance of the mucinous protective layer.


Just beneath the epithelium, in the lamina propria of the villi and crypts, there is an extremely extensive immune system. Farther down the small intestine, there are special collections of lymphatic cells in Peyer's patches, where there is no overlying mucus, so the bacteria gain more ready access to the immune system. These Peyer's patches are the specialized sites of the immune system where immune responses are both induced and regulated. Special folding M cells in the epithelium of the villi and above the Peyer's patches allow the controlled transfer of bacteria to the dendritic cells, located below the epithelium in the lamina propria and Peyer's patches. These dendritic cells have special recognition receptors that bind bacteria and then orientate these antigens for presentation to the thymus-derived T cells in the intestinal immune network. Excessive bacterial contact seems to promote exaggerated proinflammatory immune responses, whereas limited bacterial or no bacterial exposure, as exists in germ-free conditions, can dramatically impair immune development and function. Thus, controlled and regulated bacterial contact and the mucosal process of sampling bacteria seem to be needed to allow the induction of optimum immune function.

There are also regulators of the innate TLRs, some of which also inhibit dendritic cell activity. Some of these regulators, when stimulated by commensal organisms, can limit the development of autoimmune reactions and inflammation. There is now intense interest in how the commensal bacteria affect T cell activation and how, after T cell activation, normal immune function is maintained but then becomes altered in inflammatory bowel disease. Sometimes, altered immune function is a prelude to autoimmune diseases, which are increasing in incidence.


One might expect to see marked differences among individuals in the types of bacteria they host, but a striking finding from the microbial patterns of both monozygotic and dizygotic twins and their mothers is that family, i.e., environmental factors rather than genetics, seems to determine the range of bacteria that emerge both in the mucosal layer of the small intestine and in the lumen. Dusko Ehrlich, the coordinator of the European MetaHIT project, however, reported that he and his group have developed a gene catalog that contains far more genes than were in their previous human genome library (see attributed discussion6), and they have been able to capture about 85% of the genes from the microbial analyses conducted on fecal samples from the 124 individuals in their European cohort.2 About 40% of an individual's genes were found to be shared by at least half the European cohort under study, and in over 90% of the individuals, there are about 57 shared species. However, the abundance of these species varies remarkably, with 2,000-fold differences. In a prior report, the MetaHIT Consortium identified three robust clusters, which they call “enterotypes,” that proved to be neither nation nor continent specific.7 When the analyses were applied to other studies, the same three enterotypes were revealed. This shows that the intestinal microbial genetic variation is generally stratified and is not just a continuous blend of many different ranges of bacterial genes. Furthermore, when assessed in detail, the characteristics of the enterotypes depended on specific abundant molecular functions derived from several different organisms and not necessarily provided by a few species present in abundance. Thus, the enterotype was based more on a molecular functional analysis rather than on a classification of different specific microbial organisms. The challenge now is to determine the significance of these three enterotypes and whether they apply to African and Asian populations as well as those in Europe and the Americas. Already, the MetaHIT Consortium has extended these studies and observed that the three enterotypes are robust, with each dominated by one main species: one by Bacteroides, one by Prevotella, and one by Ruminococcus.2 What determines the clustering of shared species is unknown, as is the physiopathological significance of the three enterotypes, although there is some relationship to obesity (see below).


One remarkable message that emerged from the discussion was the very substantial variation in the normal microbial flora found in different laboratories. There seemed to be very different microbiomes, which might be ascribed to the subtle differences in the genetic type of mice used by a particular laboratory. However, it soon became apparent that each laboratory has its own unique microbial environment that seems to become established in the colony of animals. So, care is now needed before one extrapolates data from one laboratory to another without taking the prevailing microbiome into account.

A plethora of studies were also highlighted which used rodent models to assess the digestive and metabolic roles of the gut bacteria by studying germ-free animals and then assessing the significance of colonizing these animals with different controlled microflora. Using these models, it has been known for decades that the germ-free animal is dramatically different physiologically and in its response to diets. However, one of the difficulties is that rodents routinely engage in coprophagy, i.e., they eat their own feces, and this, of course, introduces a host of different metabolic factors and signals, some or many of which may not readily pass into the body from the colon. With coprophagy, however, the colonic contents can have marked effects once they are ingested and absorbed from the small intestine. Furthermore, the metabolisms of rodents and humans are very different, with rodents normally expending a substantial amount of energy to maintain their body temperature via the marked use of brown adipose tissue, whereas under normal conditions, adult humans have very little need for much heat generation from their modest amounts of brown adipose tissue.


The breastfed infant's intestine is soon dominated by bifidobacteria, whereas, in most formula-fed infants, similar amounts of Bacteroides and bifidobacteria (approx. 40%) are found, with new evidence of a complex interaction between the factors in breast milk and the interaction between these factors and the baby's epithelial cell immune responses in affecting the intestinal microbial pattern. After weaning under normal circumstances, the “fiber” portion of the diet, i.e., the nonstarch polysaccharides and lignin in the cell wall of the plant, passes through the intestine into the colon. Inulin and oligosaccharides, which contain bonds that are resistant to mammalian hydrolytic enzymes, also pass into the colon. However, it is now recognized that a surprising amount of starch, if made resistant by the physical rearrangements of the carbohydrate structure while cooling after a food has been cooked, will also pass through into the colon as “resistant” starch. This unabsorbed starch fraction may be increased by rapid gut transit, and the nondigested portion of carbohydrate intake becomes the major source of energy for microbial growth in the colon and logically could be expected to have profound effects on the differential growth of different microbial species. It is impossible to conduct a coherent analysis with uncontrolled diets that change daily, but controlled dietary studies are now being used to assess these effects, as described by Flint in the present supplement.8 Clearly, if the intake of carbohydrates is reduced and the type of carbohydrate eaten is refined and is therefore low in cell wall constituents, there will be far smaller amounts of substrate for the colonic bacteria to use as an energy source. It is therefore not surprising that, given the generation of the short-chain fatty acids from the unabsorbed carbohydrates, a decrease in the proportion and total numbers of butyrate-producing Lachnospiraceae related to Roseburia and of bifidobacteria is found when the intake of nondigestible carbohydrate falls. Feeding a diet high in resistant starch to overweight/obese adults in the United Kingdom led within 3–4 days to a 10-fold increase in ruminococci related to Ruminococcus bromii. This increase, to about 25% of the total bacterial 16S rRNA, does not occur in everyone but is rapidly reversible when individuals revert to their original diet. New data also suggest that long-term, high-fat, low-fiber diets are associated with the development of the Bacteroides enterotype and long-term high-carbohydrate diets with the Prevotella enterotype. Short-term dietary manipulation does alter the microbial composition, but the enterotype of the individual is sustained, implying that the enterotype is related to early or long-term dietary experience.9


Traditionally, repeated analyses of fecal samples have suggested that, although the use of antibiotics results in marked changes in the density and patterns of microbes in the intestine, individuals eventually revert to their usual resilient microbial enterotype. However, David Relman,5 in studies of human volunteers, found that this may not hold true if the density of organisms is studied and the effects of not just a single antibiotic exposure but of repeated treatments with antibiotics are assessed. Furthermore, the timing of antibiotic use may be important. It seems possible that, when an infant or young child is exposed to antibiotics, the pathogenic infection may be overcome, but at the same time, the antibiotic may also modify the entire gut microbiome and immune responsiveness during a critical period in the establishment of the individual's normal microbiome. This is now clearly an area that requires further study.


When a pathogen is ingested, it may have specific effects that allow greater penetration through the mucosal layer. For example, amoebiasis affects the colon, where Entamoeba histolytica secretes proteases,10 one of which cleaves mucus so that the entire mucus barrier is disrupted, thereby allowing the parasite to penetrate down to the epithelium and invade. Other pathogens exert their effect by binding to the cell wall, proliferating, and then secreting toxins that induce marked damage to the epithelium and often elicit a secretory response that persists until the immune response is activated to inhibit and then usually eliminate the invading pathogen.


The currently accepted hypothesis is that the inflammatory bowel diseases Crohn's disease and ulcerative colitis result from an abnormally regulated immune response to the commensal microbiota. The early age of onset of Crohn's disease and the development of ulcerative colitis in the second or third decade of life suggest that the environmental influence is exerted prior to this and is probably greatest during the earliest stages of life. This is supported, as Bernstein and Shanahan11 highlight, by studies of migrants from low-income to high-income countries, which show that the earlier the individual migrates, the greater the risk of acquiring inflammatory bowel disease. When migration is delayed until adult life, these individuals retain the expected low risk found in their country of origin.11 Recent studies in twins have revealed that, when one twin develops ileal Crohn's disease, there is a loss of bacteria belonging to the core bacterial groups such as Faecalibacterium and Roseburia and increased amounts of Enterobacteriaceae and Ruminococcus gnavus.12 The common core of microbes found in ulcerative colitis is much smaller, with a different general core set of organisms. Extensive genetic studies of both Crohn's disease and ulcerative colitis have been conducted, with several dozen genes being found in the ill subjects. These genes may play some role that is yet to be determined. There are, however, marked differences between the two diseases, with Crohn's disease patients having reduced numbers of Faecalibacterium prausnitzii, which is considered an anaerobic butyrate-producer with anti-inflammatory properties.13 When the mucosal bacteria are analyzed, a marked reduction in an obligatory mucus-degrading and propionate-producing bacterium belonging to the Verrucomicrobia, i.e., Akkermansia muciniphila, has been found. This bacterium is known to stimulate the immune system and promote a better barrier function experimentally.

The potential role of individual susceptibility to bacterial involvement in ulcerative colitis is suggested by studies of knockout mice in which an intestinal epithelial cell-specific deficiency of core-1-derived O-glycans, the core constituent of the mucosal layer, was found. These animals developed an ulcerative colitis type of illness in their distal colons, very much like human ulcerative colitis. In an article in the present supplement, Shanahan14 links this to his studies of biopsies from patients with ulcerative colitis, in which a subset of patients had altered core-1 O-glycosylation with somatic mutations in C1GALT1C1, which is essential for core-1 O-glycosylation.


In the present supplement, Willem de Vos15 notes that there is only a single layer of epithelial cells between adhering bacteria and the complex enteric nervous system, which is found throughout the intestine and has multiple interactions with the brain. It is clear that the intestinal microbes communicate with the epithelial cells, as shown both experimentally16 and in humans.17 Indeed, new evidence shows that specific intestinal microbes interact with the enteric plexus, with very clear effects being evident in animal behavior, including anxiety-like responses.18 Currently, there are some remarkable but as yet unpublished scanning studies of human brain function and emotional responses that suggest human behavior can be modulated by changes in the gut microflora, so perhaps the sense of feeling generally unwell in association with a gut disorder is not simply the result of inflammatory effects within the intestine. A consistent reduction in anaerobic gram-positive bacteria belonging to Dorea spp. and Ruminococcus spp. is being found in patients with irritable bowel syndrome, whereas healthy controls have greater numbers of Bifidobacterium spp. and Faecalibacterium prausnitzii, which have previously been associated with the absence of abdominal pain and inflammation.


The bacterial changes in patients with Crohn's disease and ulcerative colitis have stimulated interest in the concept of trying to change the microbial flora by providing suitable microbial drinks or meals that could redress a microbial imbalance in the gut. These microbial products or mixtures are referred to as probiotics, whereas prebiotics involve the provision of suitable carbohydrate or other substrates to preferentially stimulate particular groups of beneficial microbes already present but presumably in lower numbers or with lower metabolic activity. Experimental treatments vary in their apparent effectiveness. Although sometimes a particular cocktail of bacteria has been shown to induce remission in patients with ulcerative colitis, a recent Cochrane review concluded that the current probiotics are only modestly beneficial in ulcerative colitis19; nevertheless, the choice of cocktail would appear to be critical. The usefulness of both probiotics and prebiotics in pediatric patients has been recently reviewed by the American Society of Pediatrics,20 who concluded that the use of probiotics is modestly effective in randomized clinical trials in treating acute viral gastroenteritis in healthy children and in preventing antibiotic-associated diarrhea in healthy children. There is also some evidence that probiotics prevent necrotizing enterocolitis in very-low-birth-weight infants (<1.5 kg), but studies of other intestinal infections and conditions, including ulcerative colitis and irritable bowel disease, while encouraging, lacked sufficient data to draw any coherent conclusions. No evidence of benefit was demonstrated in individuals with Crohn's disease.

Analysis of the use of prebiotics to stimulate favorable growth and/or activity of indigenous probiotic bacteria is still in its infancy. Human milk is recognized to contain substantial quantities of prebiotics, but the effects of all the molecular forms of prebiotics in human milk can only prevent intestinal damage in the newborn gut, prior to the development of the adult microbiome.


Colitis caused by Clostridium difficile infection typically develops during or just after a hospital stay, although some cases are unassociated with hospitalization. Often, the condition arises in patients who have taken antibiotics, and it may develop in symptomless carriers who are already harboring the organism. Up to 10% of elderly subjects may be carriers, but when routine antibiotics are given or another factor impacts the gut flora, a virulent form of the Clostridium organism may dominate the colonic mucosa, resulting in a marked reduction in the other more usual organisms as well as the development of abdominal pain, cramps, and diarrhea. Repeated efforts to treat these patients with, for example, vancomycin or metronidazole, often fail, so radical measures are now being tried. When a colonic washout is followed by the transfer of a relative's colonic contents into the patient, the benefit can be substantial: de Vos15 highlights his dramatic results, with claims of an over 95% “cure” rate, or at least a seemingly permanent remission. This infusion of a much more diverse flora was not always successful in eliminating the virulent organisms, but another person's colonic contents, upon transfer, produced favorable results, with analysis of fecal flora months later showing that the microbes present were those originally found in the donor's feces. At present, similar attempts are being made in patients with irritable bowel syndrome and inflammatory bowel disease to see if this approach might work.


Backhed et al.21 have reported dramatic results in experimental animals. Depending on the microbial flora in the gut of the experimental mice, there may be either an extreme or a minimal tendency to gain weight and become obese when the animals are provided with a highly appetizing fat- and sugar-rich diet of junk food and confectionary. Whereas the microbially colonized mice rapidly gained weight, the germ-free mice did not gain more weight than control mice fed a low-fat diet, and recent findings show that this effect is dependent on the presence of sucrose in the diet. The microbiome of obese mice also has an increased capacity to harvest energy from a polysaccharide-rich diet, but these animals are also recycling the short-chain fatty acids and other metabolites with potential appetitive or energy partition effects through their usual ingestion of feces. Some differences in the microbiome of obese versus lean individuals have also been suggested; for example, an inverse relationship was found between obesity and the Bacteroides/Firmicutes ratio, but with the ratio increasing further following weight loss. The specific effects of any relationship of this or any other microbial mix on the propensity for obesity or on the state of obesity itself was difficult to identify, and further analyses have not confirmed the original relationship. If increased harvesting of energy from short-chain fatty acids is considered a possible mechanism, rather than an effect on gut hormones affecting appetite, then doubt was expressed in this workshop about the magnitude of this potential harvesting effect in the progressive weight gain in humans that has resulted in the current obesity epidemic.

More dramatic results, however, are now being found in subjects with the metabolic syndrome and insulin resistance. Willem de Vos15 described in a workshop how the replacement of colonic contents in an individual affected by the metabolic syndrome with the colonic contents from a lean relative or colleague induces a rapid improvement in insulin sensitivity, which is sustained for many weeks. It is still not known how this occurs mechanistically, and whether it depends on colonic events and metabolites absorbed from the colon, on concomitant changes in the microbiome in the lower ileum with alterations in gut hormones, or on important interactions of the microbiome with the enteric plexus.


Although much of the focus on intestinal microbes has traditionally been related to pathological events, the beneficial role of commensal organisms and the need for them to be established as an important buffer against pathological events from early childhood was emphasized at the the 24th Marabou Symposium. In the symposium discussion, Alan Jackson recalled one of the great nutritional concerns of the post-World War II era, which was how to satisfy the protein needs of the world (see attributed discussion6). Using isotopic turnover techniques to identify the quantitative flow and demand for essential and conditionally essential amino acids, he showed how urea, normally considered simply as the end product of protein metabolism, passes into the gut. There, it is immediately used by bacteria to synthesize their own essential and nonessential amino acids, which, in turn, are reabsorbed and used as important contributors to protein synthesis when dietary supplies are short.22 Thus, in individuals on vegetarian diets, two-thirds of the nitrogen excreted as urea is salvaged, with colonic hormone-sensitive urea transporters helping to redirect the turnover of urea into the colon. New estimates of lysine requirements show that about 30 mg/kg/day are needed, but under vegetarian conditions 20–30 mg/kg/day of this need can be met by microbial synthesis in the gut.23 This remarkable finding is matched by the recognition that the newborn breastfed baby receives a substantial proportion of its nitrogen as urea and other non-amino-acid forms in breast milk. In the past, this intestinal recycling of urea by intestinal microbes was overlooked as a crucial part of an infant's early survival and growth.


The remarkable developments in new, recently available sequencing techniques should help to transform the understanding of the microbiome. It is already evident, however, that the normal establishment of microbes is of extraordinary importance in conditioning the human immune system and establishing a prior claim to intestinal surfaces and niches that serve as robust barriers to a host of potentially pathological organisms. Only now, however, are the effects of the microbiome on the immune system and the link between the enteric plexus of nerves in the intestine and the brain becoming apparent, along with the multiple related effects on metabolism. It is now becoming possible to study individual differences in enterotypes along with their significance, the response to tightly controlled diets, and the differences in microbiomes across the world and their role in conditioning, protecting, or causing a multitude of diseases. The importance of multidisciplinary interaction is evident and opens up a new nutritional field of potentially great public health significance.