SEARCH

SEARCH BY CITATION

Philip James: This discussion covers not only all the lectures with their papers but also any new topics that are considered important. The topic is very wide ranging so let us begin with some aspects of methodology. What tools should we be using to assess the nature and actions of the microbiome: sophisticated DNA sequencing, proteomics, and/or metabolomics?

METHODOLOGY: A GENE SCANNING APPROACH

  1. Top of page
  2. METHODOLOGY: A GENE SCANNING APPROACH
  3. DIFFERENT APPROACHES TO THE ANALYSIS OF THE MICROBIOME
  4. METABOLIC PROFILING
  5. DEVELOPMENT OF THE MICROBIOME AFTER BIRTH AND THE INFLUENCE OF HOST GENETICS
  6. THE IMPACT OF EARLY FEEDING
  7. GUT PERMEABILITY IN THE NEWBORN
  8. EARLY MACRONUTRIENT EFFECTS ON THE HUMAN MICROBIOME
  9. FAT ABSORPTION AND BILE SALTS IN THE NEWBORN
  10. POPULATION DIFFERENCES IN THE MICROBIOME ACROSS THE GLOBE
  11. GUT MICROBIOTA AND THEIR POTENTIAL ROLE IN DEVELOPMENTAL PROGRAMING OF THE BRAIN
  12. DIETARY STANDARDIZATION
  13. NUTRIENT NICHES AND MICROBES THROUGHOUT THE LENGTH OF THE INTESTINE
  14. MUCOSAL BACTERIA
  15. MUCIN'S STRUCTURE AND INTERACTION WITH MUCOSAL BACTERIA
  16. GUT IMMUNOLOGY IN RELATION TO MUCIN AND MICROBIAL COLONIZATION
  17. DIETARY INTERVENTIONS: EFFECTS ON COLONIC MICROFLORA
  18. AUTISM – A GUT ROLE?
  19. ROLE OF THE MICROFLORA IN SYNTHESIZING ESSENTIAL AMINO ACIDS AND ALLOWING THE MAINTENANCE OF PROTEIN BALANCE ON VERY LOW INTAKES
  20. CHILDREN'S STUNTING: THE ROLE OF GASTROINTESTINAL MICROBIOTA
  21. THE ETHICS OF STUDYING DEPRIVED CHILDREN
  22. ENERGY BALANCE AND THE MICROBIOME
  23. PHYSICAL ACTIVITY AND THE MICROBIOME
  24. INSULIN SENSITIVITY AND THE METABOLIC SYNDROME
  25. PROBIOTICS AND THEIR EFFECTS
  26. CROHN'S DISEASE AND ULCERATIVE COLITIS
  27. REFERENCES

Dusko Ehrlich: I have been coordinating a major European collaborative study – the MetaHIT [Metagenomics of the Human Intestinal Tract] Consortium1 involving multidisciplinary teams from 12 European institutions and one Chinese group. The Consortium is funded mostly by the European Union [EU] and focuses on assessing gut bacterial species by their genetic profiles. We have assessed the relationship of these genetic profiles to some diseases, particularly obesity, as set out in this presentation.* First, we constructed a whole reference gene catalogue of microbes from our digestive tract in order to develop profiling tools, which we call quantitative metagenomics. By screening fecal samples, we then mapped the communities of sequences to obtain information about gene abundance. Our gene catalogue contains 150 times more genes than exist in our own human genome and captures about 85% of genes from the microbial analyses conducted on fecal samples from 124 individuals in the European cohort that we examined.2 Then, we found in a small group of American and Japanese subjects that we could capture about 80% of their sequences in our catalogue. African and Asian populations have not been examined, so we do not know how well they are covered by our gene catalogue.

Each individual has about half a million microbial genes of the 3.3 million genes in our catalogue, and about 40% of an individual's genes are shared by at least half of our European cohort. When we use our quantitative metagenomics pipeline to examine other shared species, we find that in over 90% of the individuals there are about 57 shared species, but the abundance of these species varies a lot, so the variation among individuals is quite large with up to 2,000-fold differences. This means that if you do not have appropriate analytical systems it is easy to miss the presence of some species in individuals when the species are not really abundant. This then gives the erroneous impression that people are strikingly different when one of the key differences is in the relative abundance of particular microbial species.

The next important finding from the study is that some species co-vary in abundance with other species, and we found that we could look at the overall pool of bacteria in different individuals and see different patterns which we then extended and refined analytically, thereby defining particular bacterial communities, which we describe as “enterotypes.”3 Enterotypes are defined as bacterial communities which we find in different people.

These developments occurred by assessing the whole microbiome in 33 individuals, using genetic sequencing techniques by Sanger. By combining 22 newly sequenced fecal metagenomes of individuals from four countries with previously published data sets, we identified three robust clusters, i.e., the enterotypes, which proved not to be specific to any nation or continent. We also confirmed the presence of these enterotypes in two published, larger cohorts. This shows that the intestinal microbial genetic variation is generally stratified and not just a continuous blend of many different ranges of bacterial genes. This, in turn, means there may be only a limited number of well-balanced host-microbial symbiotic states that respond differently to diet and drug intake. The enterotypes are characterized mostly by the composition of species types. However, it became apparent that specific abundant molecular functions, derived from several different organisms, are not necessarily provided by a few species present in abundance. This means that by using specific probes for particular genes with known metabolic functions, one derives a functional view rather than simply classifying the gut in terms of different specific microbial species.

We observed that measures such as the body mass index, age, or gender did not explain the observed overall enterotypes, but certain marker genes or functional modules could be identified which were associated with each of these three individual characteristics. Thus, for example, twelve genes significantly correlated with age and three functional modules with the body mass index. This hints at a diagnostic potential for some microbial markers. The studies were then extended to include 85 Danes, 154 US adults, and 454 individuals from 60 other nationalities. In all of these individuals, the three previously identified enterotypes were evident with a main driving species for each enterotype: one Bacteroides, one Prevotella, and one Ruminococcus. So, as individuals, we are similar but we can also be subdivided by our bacterial communities into three groups. These groups can be likened a little to a blood group, but we have no inkling as to why there are these clusters and the reason for the dominance of one or another enterotype. This seems to be a new facet of human biology which may allow patient stratification and aid the development of personalized medicine and nutrition.

Our analyses of two European population groups do not, of course, tell us that this is a universal phenomenon. However, preliminary results from adults in China seem to indicate that this might be so. Nevertheless, we still do not know the reasons for the marked differences in species types and in gene abundance between individuals. It has been suggested that early exposure in childhood to antibiotics may be very important. We are also now finding in correlation studies that 33 meta-species correlate with obesity. There are six meta-species which tend to be present in lean and not in obese individuals, whereas other meta-species are present in the obese rather than in the lean.4

Rob Knight: I want to emphasize the importance of standardizing methods and analytical techniques. Dusko Ehrlich's International Human Microbiome Standards [IHMS] consultation is very important as there is an exchange of samples and it ensures that a large database, like the one Willem de Vos5 described, can be run through many different techniques and the results compared. Our data assessing the microbial diversity through the agrarian negro microbial mat shows it to have the most diverse microbial community that we know. The main factor is the oxygen gradient through the mat, so going from the top to the bottom one can plot the metadata of microbes in relation to the oxygen and sunlight gradients. Then, on the basis of our multiple statistical analyses, we can cluster the data. However, we have to remember that this clustering may be an artifact of our analytical approach and not reflecting the real continuous gradient effects. This cautious approach to the mathematical analyses of different microbial populations is critical, especially as we start moving into larger microbial populations such as you find in the gut. Then, when we start trying to work out what happens to the microbiota throughout the lifespan and especially during those periods when it is very rapidly changing, in infancy and in old age, we need to be very cautious.

You need to remember that the techniques we are using have only been available for about a year, so many of the detailed studies we need have still to be done. So part of the reason why my lab focuses so much on developing these technology platforms is that we are very aware that there are a lot of great cohorts and a lot of great ideas out there. By making the techniques broadly available we can have a far greater impact than by trying to maintain control of those studies ourselves. I would definitely encourage everyone who has access to such a cohort to start sampling much more densely now than you would have thought feasible a couple of years ago.

Over the last few years, we have moved from studies of say 3–5 individuals in whom the original analyses were very valuable to studying 300 subjects. We still have only a handful of study sites, and in some ways each study site is only one data point reflecting the combined effects of diet, genetics, and environmental exposure. So we need to have 1,000 sites rather than 1,000 people at one site. We do not know the effect size of any of these factors. We know that people differ to an extraordinary degree in their microflora, but we do not yet know what the main factors guiding those differences are. We have seen pretty consistently that infants have very different gut microbiota from adults. This can give us a scaling of differences with which to compare differences between populations – those on different diets, the effects of genetic differences, etc. This will also help in determining the size of sampling.

Ruth Ley: I agree. We also have a very smooth gradient of change with time in our baby series. We have also recreated cluster patterns and we see distinct clusters depending on the type of analysis we undertake.

Rob Knight: If one knows what one has to start with and then uses statistical techniques to simulate patterns, then one can rapidly discover what the statistical analysis is doing. This is based on Justin Kuczynski's Nature Methods paper.6

DIFFERENT APPROACHES TO THE ANALYSIS OF THE MICROBIOME

  1. Top of page
  2. METHODOLOGY: A GENE SCANNING APPROACH
  3. DIFFERENT APPROACHES TO THE ANALYSIS OF THE MICROBIOME
  4. METABOLIC PROFILING
  5. DEVELOPMENT OF THE MICROBIOME AFTER BIRTH AND THE INFLUENCE OF HOST GENETICS
  6. THE IMPACT OF EARLY FEEDING
  7. GUT PERMEABILITY IN THE NEWBORN
  8. EARLY MACRONUTRIENT EFFECTS ON THE HUMAN MICROBIOME
  9. FAT ABSORPTION AND BILE SALTS IN THE NEWBORN
  10. POPULATION DIFFERENCES IN THE MICROBIOME ACROSS THE GLOBE
  11. GUT MICROBIOTA AND THEIR POTENTIAL ROLE IN DEVELOPMENTAL PROGRAMING OF THE BRAIN
  12. DIETARY STANDARDIZATION
  13. NUTRIENT NICHES AND MICROBES THROUGHOUT THE LENGTH OF THE INTESTINE
  14. MUCOSAL BACTERIA
  15. MUCIN'S STRUCTURE AND INTERACTION WITH MUCOSAL BACTERIA
  16. GUT IMMUNOLOGY IN RELATION TO MUCIN AND MICROBIAL COLONIZATION
  17. DIETARY INTERVENTIONS: EFFECTS ON COLONIC MICROFLORA
  18. AUTISM – A GUT ROLE?
  19. ROLE OF THE MICROFLORA IN SYNTHESIZING ESSENTIAL AMINO ACIDS AND ALLOWING THE MAINTENANCE OF PROTEIN BALANCE ON VERY LOW INTAKES
  20. CHILDREN'S STUNTING: THE ROLE OF GASTROINTESTINAL MICROBIOTA
  21. THE ETHICS OF STUDYING DEPRIVED CHILDREN
  22. ENERGY BALANCE AND THE MICROBIOME
  23. PHYSICAL ACTIVITY AND THE MICROBIOME
  24. INSULIN SENSITIVITY AND THE METABOLIC SYNDROME
  25. PROBIOTICS AND THEIR EFFECTS
  26. CROHN'S DISEASE AND ULCERATIVE COLITIS
  27. REFERENCES

Jens Nielsen: If we wish to get to the mechanisms we will have to have studies where we actually analyze all the different types of analyses, i.e., DNA, proteomics, and metabolic profiling in the same study. Then, we can start to find some real mechanistic explanation for the observed correlations. This has been done with a conglomerate of different microbial species but not with in vivo systems in human or mice studies. I am pretty sure we can use the same approaches to study more complex microbial systems.

Gunnar Hansson: The key actors are the proteins, so we need to focus on proteomics. But we also need to have the exact sequences from the genomes, which we do not have at present, so we have to work through this and it will take some time.

Dusko Ehrlich: There are different issues to consider. First, there are the technologies which provide the array of transcriptomic, proteomic, and metabolomic data. But then we have to consider how to assess the mechanistic functions. Some would argue that we should go systematically through the technologies, whereas others opt for preliminary assessments on the basis of current knowledge and focus on the mechanisms by looking at particular genes, their alleles, particular mutants, and selective species. This might short circuit the need to go systematically through all the -omic techniques first. The different approaches may well depend on resources and individuals’ enthusiasm.

Ruth Ley: We have also seen how focusing on different specific techniques gives us very different views. We can find differences between individual subjects and between the healthy and unhealthy with some techniques but not others. Willem [de Vos] also showed differences in the outcomes of obesity studies depending on which method was used. So it is critically important to ensure we are using exactly the same methods across studies in order to corroborate or validate results.

Willem de Vos: One does need a good database. The genomic database is expanding remarkably, so I think we are getting somewhere and proteomics is now feasible. I agree, of course, that there are different methods, but the phenotyping as well as the genotyping of the individuals being monitored is important.

Jens Nielsen: We need to consider the use of systems biology and move beyond just traditional correlation analysis. Then, we also need to recognize that there is a tendency to do larger and larger studies when all this may do is emphasize the large variability between subjects. In learning some of the key factors controlling metabolomics function, we may get further doing in-depth studies with much smaller subject groups.

Cutberto Garza: Not only are there the various methods to think about, but there are issues to consider about kinetics and dynamics. If you are talking about a variable where your only interest is whether it is present or absent and, if present, its magnitude, then this is very different from considering the basic conditions leading to its abundance, e.g., are you looking at these things in the fed or fasted state? Is it more interesting to make sure that you understand the diurnal variations, especially of proteins and their expression? And how do those elements come into the interpretation if we are moving towards mechanistic explanations?

Harry Flint: I think there is a need for more quantitative approaches, in particular using modeling that will relate the new information we get from the omics technologies to previous information on gut function. As Philip knows very well, there is a lot of extremely good work being done on gut transport and the way that fiber influences it, etc. So we need to think about what is surprising and what would we predict will happen. Often, many of the changes may be explicable on the basis of existing knowledge; others may not be.

Philip James: So not only have we got different technologies with no clear understanding of the relationship between the analytical approaches but we also have not come to the question of diversity and magnitude of particular species. It sounds as though we are looking at the problem in a completely chaotic way because we have not defined anything properly.

Harry Flint: Yes – there are two important areas of theoretical effort: one is bioinformatics, which is largely based on the analysis of sequences. We have heard of some extremely good examples of how that has progressed and how sophisticated it has become, but the other whole area is systems modeling in terms of rates of events happening: biochemical reactions, absorption by gut surfaces, turnover, etc. That needs to be brought into the picture.

METABOLIC PROFILING

  1. Top of page
  2. METHODOLOGY: A GENE SCANNING APPROACH
  3. DIFFERENT APPROACHES TO THE ANALYSIS OF THE MICROBIOME
  4. METABOLIC PROFILING
  5. DEVELOPMENT OF THE MICROBIOME AFTER BIRTH AND THE INFLUENCE OF HOST GENETICS
  6. THE IMPACT OF EARLY FEEDING
  7. GUT PERMEABILITY IN THE NEWBORN
  8. EARLY MACRONUTRIENT EFFECTS ON THE HUMAN MICROBIOME
  9. FAT ABSORPTION AND BILE SALTS IN THE NEWBORN
  10. POPULATION DIFFERENCES IN THE MICROBIOME ACROSS THE GLOBE
  11. GUT MICROBIOTA AND THEIR POTENTIAL ROLE IN DEVELOPMENTAL PROGRAMING OF THE BRAIN
  12. DIETARY STANDARDIZATION
  13. NUTRIENT NICHES AND MICROBES THROUGHOUT THE LENGTH OF THE INTESTINE
  14. MUCOSAL BACTERIA
  15. MUCIN'S STRUCTURE AND INTERACTION WITH MUCOSAL BACTERIA
  16. GUT IMMUNOLOGY IN RELATION TO MUCIN AND MICROBIAL COLONIZATION
  17. DIETARY INTERVENTIONS: EFFECTS ON COLONIC MICROFLORA
  18. AUTISM – A GUT ROLE?
  19. ROLE OF THE MICROFLORA IN SYNTHESIZING ESSENTIAL AMINO ACIDS AND ALLOWING THE MAINTENANCE OF PROTEIN BALANCE ON VERY LOW INTAKES
  20. CHILDREN'S STUNTING: THE ROLE OF GASTROINTESTINAL MICROBIOTA
  21. THE ETHICS OF STUDYING DEPRIVED CHILDREN
  22. ENERGY BALANCE AND THE MICROBIOME
  23. PHYSICAL ACTIVITY AND THE MICROBIOME
  24. INSULIN SENSITIVITY AND THE METABOLIC SYNDROME
  25. PROBIOTICS AND THEIR EFFECTS
  26. CROHN'S DISEASE AND ULCERATIVE COLITIS
  27. REFERENCES

Elaine Holmes: I want to highlight the usefulness of metabolic profiling. We have used urinary metabolic profiling because this can highlight the role of microbial metabolism and reveal how different nutritional states and early experiences can affect long term the microbial population's metabolism. For example, in the INTERMAP Study we assessed 5,000 people in 4 countries – China, Japan, United Kingdom, and United States – and looked at urinary metabolites in relation to blood pressure. With the NMR spectra from normal-weight people, i.e., those with BMIs < 25, we compared them with the obese group, i.e., BMIs ≥ 30. About 30–40% of the 23 discriminatory metabolites were either from gut microbial metabolites or gut microbial and mammalian cometabolites. So, for example, para-cresol sulphate and phenyl-acetyl-glutamine were very different between the lean and obese. We know from animal studies also that the more hippurate you excrete, the greater the lean body mass.

Another example relates to the effects of bypass surgery for obesity. We find in animal studies a big change in bacterial metabolites as a result of the new interactions of the diet with gut bacteria. However, more intriguing is the fact that we can distinguish in 18–27 year olds those born either normally or preterm because there are substantial differences in their metabolic profiling in urine.7 Preterm subjects had a greater total fat mass, more internal fat, and this is seen especially in men with greater abdominal, hepatic, and intramuscular adipose tissue. Diastolic blood pressures were also higher as were the significant differences in the urinary metabolome, with higher methylamines and acetyl-glycoproteins and lower hippurate levels. So these types of analysis can provide biological markers which may indicate greater long-term risks and may, in due course, suggest particular neonatal remedies.

DEVELOPMENT OF THE MICROBIOME AFTER BIRTH AND THE INFLUENCE OF HOST GENETICS

  1. Top of page
  2. METHODOLOGY: A GENE SCANNING APPROACH
  3. DIFFERENT APPROACHES TO THE ANALYSIS OF THE MICROBIOME
  4. METABOLIC PROFILING
  5. DEVELOPMENT OF THE MICROBIOME AFTER BIRTH AND THE INFLUENCE OF HOST GENETICS
  6. THE IMPACT OF EARLY FEEDING
  7. GUT PERMEABILITY IN THE NEWBORN
  8. EARLY MACRONUTRIENT EFFECTS ON THE HUMAN MICROBIOME
  9. FAT ABSORPTION AND BILE SALTS IN THE NEWBORN
  10. POPULATION DIFFERENCES IN THE MICROBIOME ACROSS THE GLOBE
  11. GUT MICROBIOTA AND THEIR POTENTIAL ROLE IN DEVELOPMENTAL PROGRAMING OF THE BRAIN
  12. DIETARY STANDARDIZATION
  13. NUTRIENT NICHES AND MICROBES THROUGHOUT THE LENGTH OF THE INTESTINE
  14. MUCOSAL BACTERIA
  15. MUCIN'S STRUCTURE AND INTERACTION WITH MUCOSAL BACTERIA
  16. GUT IMMUNOLOGY IN RELATION TO MUCIN AND MICROBIAL COLONIZATION
  17. DIETARY INTERVENTIONS: EFFECTS ON COLONIC MICROFLORA
  18. AUTISM – A GUT ROLE?
  19. ROLE OF THE MICROFLORA IN SYNTHESIZING ESSENTIAL AMINO ACIDS AND ALLOWING THE MAINTENANCE OF PROTEIN BALANCE ON VERY LOW INTAKES
  20. CHILDREN'S STUNTING: THE ROLE OF GASTROINTESTINAL MICROBIOTA
  21. THE ETHICS OF STUDYING DEPRIVED CHILDREN
  22. ENERGY BALANCE AND THE MICROBIOME
  23. PHYSICAL ACTIVITY AND THE MICROBIOME
  24. INSULIN SENSITIVITY AND THE METABOLIC SYNDROME
  25. PROBIOTICS AND THEIR EFFECTS
  26. CROHN'S DISEASE AND ULCERATIVE COLITIS
  27. REFERENCES

David Relman: I am a great supporter of Rob Knight's approach and the value of expanding the numbers of individuals studied. On the other hand, a complementary role with intense longitudinal studies with a smaller number of individuals, where there are factors that can be manipulated, is also valuable. There, you have the advantage of controlling for individual genetics and, indeed, epigenetics. So I am advocating the introduction of a birth cohort study and taking the long-term approach we have seen in Framingham for cardiovascular disease.8

Rob Knight: Agreed. Tying in with larger birth cohorts that are already approved and involve genetic analysis makes it easy to add microbiome analysis. This is where twin studies are also of exceptional value. When we looked at mono- and dizygotic twins,9 we knew that it had been reported that identical twins shared many common features in their microflora, but when we added dizygotic twins and mothers to our analyses, we completely changed our views and found more of a family effect than primary genetic factors.

Philip James: What about the studies of identical twins separated at different intervals after birth? This should be a powerful way of assessing the establishment of microbial patterns, which I gleaned from David Relman's talk might occur very early in life.

Rob Knight: To my knowledge, these studies have not been done. My impression is that there can be huge changes as a baby grows, but whether those differences are chance variations or reflecting environmental influences and whether they matter long term, we do not yet know.

Philip James: Can we say whether there are differences between babies depending on their form of delivery? The Caesarian-delivered baby will not have been exposed to all the bacteria encountered during normal delivery but how long do these microbial differences last? We know that the gut of a breastfed baby is completely different in microbial terms from that of an artificially fed baby. So have studies gone on through the first couple of years of life for these different groups of children?

Rob Knight: We only now just have the techniques to do this. I would definitely encourage everyone who has access to such a cohort to start sampling much more densely now than you would have thought feasible a couple of years ago.

Dusko Ehrlich: I think David Relman's seminal work on babies is quite difficult to interpret because of the chaotic behavior of the microbes during the first couple of years. I agree fully with what Rob said about insufficient data, but our analysis, supported by the Japanese study,10 which, with a very small sample (13 individuals) who were studied in detail, found the gut microbiota from unweaned infants to be simple with a high inter-individual variation in taxonomic and gene composition, whereas the weaned children and adults were more complex but showed a high functional uniformity regardless of age or sex. This implies that the first few days’ events may not be that important.

Rob Knight: We have known for decades about the infant being programmed by the mother's smell with marked subsequent behavioral programming but we really do not know yet whether we are dealing with critical periods for establishing microbial specificity.

David Relman: There is enough circumstantial evidence to suggest that the early period is important, but we do not understand the behavior of taxa without looking at the underlying genetic potential or the metabolites and all the downstream products. I just think that we know enough to say that this is where we are to focus some special effort, but as to what it means yet, I don't think we can say.

THE IMPACT OF EARLY FEEDING

  1. Top of page
  2. METHODOLOGY: A GENE SCANNING APPROACH
  3. DIFFERENT APPROACHES TO THE ANALYSIS OF THE MICROBIOME
  4. METABOLIC PROFILING
  5. DEVELOPMENT OF THE MICROBIOME AFTER BIRTH AND THE INFLUENCE OF HOST GENETICS
  6. THE IMPACT OF EARLY FEEDING
  7. GUT PERMEABILITY IN THE NEWBORN
  8. EARLY MACRONUTRIENT EFFECTS ON THE HUMAN MICROBIOME
  9. FAT ABSORPTION AND BILE SALTS IN THE NEWBORN
  10. POPULATION DIFFERENCES IN THE MICROBIOME ACROSS THE GLOBE
  11. GUT MICROBIOTA AND THEIR POTENTIAL ROLE IN DEVELOPMENTAL PROGRAMING OF THE BRAIN
  12. DIETARY STANDARDIZATION
  13. NUTRIENT NICHES AND MICROBES THROUGHOUT THE LENGTH OF THE INTESTINE
  14. MUCOSAL BACTERIA
  15. MUCIN'S STRUCTURE AND INTERACTION WITH MUCOSAL BACTERIA
  16. GUT IMMUNOLOGY IN RELATION TO MUCIN AND MICROBIAL COLONIZATION
  17. DIETARY INTERVENTIONS: EFFECTS ON COLONIC MICROFLORA
  18. AUTISM – A GUT ROLE?
  19. ROLE OF THE MICROFLORA IN SYNTHESIZING ESSENTIAL AMINO ACIDS AND ALLOWING THE MAINTENANCE OF PROTEIN BALANCE ON VERY LOW INTAKES
  20. CHILDREN'S STUNTING: THE ROLE OF GASTROINTESTINAL MICROBIOTA
  21. THE ETHICS OF STUDYING DEPRIVED CHILDREN
  22. ENERGY BALANCE AND THE MICROBIOME
  23. PHYSICAL ACTIVITY AND THE MICROBIOME
  24. INSULIN SENSITIVITY AND THE METABOLIC SYNDROME
  25. PROBIOTICS AND THEIR EFFECTS
  26. CROHN'S DISEASE AND ULCERATIVE COLITIS
  27. REFERENCES

Philip James: We seem to be hearing that the reactivity of the intestine is highly dependent on the early exposure to the microbiome input. Do we know the real impact of the mode of delivery, caesarean versus normal delivery, and then of breastfeeding in relation to long-term programming of the immune system and how the microbiome might be involved?

Alan Jackson: We know a great deal about the difference between breastfed and non-breastfed infants in terms of their gut function, their gut microflora, and the associations with gastrointestinal and respiratory disease, which presumably all might be mediated through some of the mechanisms that have been discussed.

Ole Hernell: I think the microflora is not what you would expect: breastfed infants have a more diverse microbiota than formula-fed infants even in the flora of the mouth.

Karen Madsen: We see decreases in microbial diversity in almost every single disease state: inflammation, diabetes, but what induces the decrease in microbial diversity? Is it an immune-mediated action decreasing certain species, or is it a microbial-induced phenomenon not allowing certain species to grow?

Dusko Ehrlich: If you are interested in the impact of early feeding on the composition of microbiota in later life – then we think that perhaps by the age of two years there is a reasonably stable community and this remains reasonably stable afterwards. Once again, I am referring to the gut. So how can variables in early exposure affect these communities later on when there are so many changes in the first 1–2 years?

Denise Kelly: We have undertaken pig studies11 looking at the impact of early exposure to bacteria and we are now looking at animals of the same genotype in two different environments, either reared intensively indoors or simply outdoors. The ones that were outdoors and exposed to highly diverse soil microbes actually had a less diverse gut microflora, i.e., the results were counterintuitive. We followed this with an experiment where we took animals that were basically exposed to maternal and environmental microbes for 3 days, then placed in isolators and followed to maturity. They maintained their high diversity, which was greater than that of those exposed outdoors to the normal range of microflora and reared naturally. So my interpretation now is that part of the microbial–immune normal development involves a decreased diversity with a later greater stability in the microflora. That is what we have reported.

Ole Hernell: That is absolutely fascinating – but on the other hand, you did introduce a variable which is not so realistic for humans who don't spend their lives in isolators after 3 days.

Cutberto Garza: I just want to caution us, because if we compare the bottle-fed and the breastfed and breastfeeding in relation to their gut flora, we have to recognize that there is a variety of feeding patterns ranging from exclusive breastfeeding to mixed feeding at various ages. So if we are going to try to predict the long-term functional outcome, it becomes complicated. If you look at the breastfed and the bottle-fed infant, from early on their BMRs are different, their heart rate regulation is different, the blood pressure is different, the time they spend in REM sleep and non-REM sleep is different, and their patterns of growth in height and weight are completely different. We have no clues as to whether these effects are due to the different microbiota, or whether, in fact, it is due to all the active components that are in breast milk.

Alan Jackson: The formula industry has spent a lot of effort looking to see how different oligosaccharides added to milk may modulate the flora against the background of breast milk, which has a much more complex array of oligosaccharides. I just wonder whether there is specific information about the effect of either oligosaccharides generally, or particular oligosaccharides on the pattern of microflora that might evolve?

Gunnar Hansson: The human milk oligosaccharides are different from those in cow's milk. Furthermore, the oligosaccharides in the milk mimic the structures of the mucins in the small intestine but not in the large intestine, so the binding of bacteria by these oligosaccharides may be protecting them from small intestinal interactions so these wash down the lumen into the colon where the mucin structures are different from those in the milk.

Ruth Ley: I just wanted to mention a study we did on American breastfed infants.12 We had a time series over 2.5 years and if one defines diversity as taxonomic richness, that increases very steadily over time from birth to 2.5 years with biomass increases and a big shift when solid food comes in accompanied by a big expansion of bacteroides.

Seppo Salminen: I would actually support the observations of Dusko Ehrlich and Ruth Ley: we have looked at German and Finnish infants during breastfeeding and the diversity seems to be lower than in bottle-fed infants.13 I guess it is partly due to the fact that you have the microbes from the mother's milk, which are continuously being fed to the infant during the breastfeeding time. Willem de Vos has also found that during the time of breastfeeding the diversity is relatively stable, and if you increase the diversity, then you also increase the risk of allergic diseases later on.14

Per Brandtzaeg: We need to remember that breast milk contains secretory IgA antibodies with very broad activity. They will coat bacteria immediately, so in the breastfed baby's gut the bacteria are coated by IgA. This coating may reflect dependence on the antibody coating itself or there may be polysaccharide lectin binding. Blaise Corthésy in Switzerland has shown that on the M-cells in the Peyer's patches there are receptors for IgA and they actually take up the coated bacteria and induce a homeostatic immune response in mice.15 These receptors are also found in humans. The secretory IgA binds and is internalized by dendritic cells in the subepithelial dome region, with the IgA-bound antigens inducing mucosal and systemic responses associated with the production of anti-inflammatory cytokines and a limitation in the activation of dendritic cells. So in terms of humoral immunity at mucosal surfaces, the IgA appears to combine the properties of a neutralizing agent, i.e., immune exclusion and of a mucosal immunopotentiator, which induces effector immune responses in a noninflammatory context favorable to preserving the local homeostasis of the gastrointestinal tract. So this means that how the bacteria are handled and influence the immune system of the baby is very complex and cannot be reproduced in artificially fed babies.

Nathalie Delzenne: I would like to come back to an old story because when we talk about the interaction of oligosaccharides with the gut microbiota in babies, we talk about microbial diversity but, in practice, a lot of data have focused on the bifidobacteria, so perhaps we should be careful to emphasize that we have not studied extensively other bacteria and how all these influence immunity.

Ingemar Ernberg: Hans Boman is a pioneer of antibacterial peptides16 as the effector molecules of innate immunity. The dominating targets are bacterial membranes and the killing reaction is faster than the growth rate of the bacteria. Some antibacterial peptides are clearly multifunctional. Humans need two classes of these proteins or defensins and clinical cases show that deficiencies in these peptides give severe symptoms and atopic allergy. They may reduce diversity, but I don't think that has to be the case. It could also be that they suppress dominance and allow a more balanced flora. I do not know enough about antibacterial peptides in breast milk. However, in evolutionary terms, the antibacterial peptides seem to have been a major component to balance the normal flora. They may be, in evolutionary terms, a very old component to balance normal floras in the microbial niche specific to the individual's locality.

Cutberto Garza: There are a number of these peptides, including lysozymes, and then a number of antibodies such as IgG and IgM in milk. There are also the polysaccharides in human milk which mimic the receptors on the intestinal wall and, therefore, bacteria attach to the polysaccharides in the milk before they reach the intestinal receptors. These polysaccharides vary not only within the day, but within a feeding period. They also vary across ages, and one of the most remarkable things about human milk is its environmental specificity. If you look at the immunological composition of human milk made in Stockholm, it is very different from that made in Boston, or very different from that even made in Wales! So, one has to be very careful when describing milks’ characteristics. It is a very dynamic fluid, but there is excellent evidence of the antibacterial effects of breast milk17 with what appears to be a finely orchestrated series of responses throughout early development.

Ole Hernell: There are also differences between breastfed and bottle-fed children in many other features. Thus, the exocrine pancreas seems to be downregulated in the breastfed infant, but as soon as you start to give complementary feeding, it is turned on. You have a lower pH in the formula-fed infant's stomach and you have, probably, different bile salt concentrations. The whole physical chemistry of the small intestine is probably different in the breastfed compared with the formula-fed infant. The type as well as the timing of complementary feeding may affect the microbiota and immunity.

Seppo Salminen: Breast milk also provides a significant amount of bacteria like the Lactobacillus and Bifidobacteria plus the bacteria from the mother's skin. So those are certainly also modulating the gut microbiota in early life and may be reducing the diversity of the microflora.

Harry Flint: If we are talking about what controls diversity, we need to be careful to define what we mean. We also know that we have a diversity of substrates, oligosaccharides, antimicrobial factors, and we also have temporal changes in the diet. So, for the host, the more frequent the dietary change, the more you have a fluctuating community and selection for different bacterial types. But there is another factor that we haven't touched on, i.e., the bacteriophages, and these viruses can turn over bacterial strains at a massive rate. In aqueous communities, it is quite well-documented now that the most abundant bacterial species are one of the first to be eliminated by a virus. So viruses are themselves agents of diversity in bacterial communities.

Karen Madsen: I just wanted to add to the importance of breast milk. We did a series of experiments with interleukin-10 knockout mice18 which spontaneously develop colitis. However, if they were cross-fostered to a normal mother and were then fed on this mother's milk, they did not develop colitis. So the early breast milk has a clear modulatory role on the developing immune system which lasts, at least in an animal model, for the rest of its life.

David Relman: Do we know about the degree of variation in milk oligosaccharide structures in milk which relate to genetics?

Gunnar Hansson: There are blood group influences, especially relating to the Lewis system.

Tore Midtvedt: Remember, meconium is filled with defensins, so the early interactions of maternal colostrum with bacteria are very complex.

Cutberto Garza: Colostrum is rich in IgA and a variety of other components. I do not think we understand the dynamics of colostrum production as it varies from woman to woman. In some women, it lasts for more than 48 h, whereas in others it is over within the first 12 h. Yet, it seems to play a very important role in establishing those early protective effects of human milk feeding in young infants. Most other mammalian milks have high levels of IgG, but in the human, the IgG is transferred in the last few weeks before birth rather than through the milk.

Olle Hernell: I think colostrum is much richer in the antimicrobial factors IgA, lactoferrin, and lysozyme than mature milk.

Philip James: What do we know, Cutberto, about the differences between caesarean section and normally delivered babies? Do we have any sense of any long-term health impacts? We are hearing that, in fact, they have completely different microbial patterns when you look at them early on, but can we even crudely relate this to anything we know about differences between babies.

Cutberto Garza: There are some very classic studies now looking at asthma, for example, as an outcome between caesarean and vaginally delivered infants. Caesarean babies have much higher rates of asthma, and it is speculated that it is all due to the lack of impact of the microbiota and this implies that it could affect other immune-related disorders.

Philip James: If I have been delivered by caesarean section and then I am breastfed rather than bottle-fed, what does that do?

Cutberto Garza: They haven't done that at all, I don't think: I have not seen those outcomes being controlled for feeding mode in that way in the asthma studies.

Richard Black: If there is a planned caesarean, that is different from an emergency caesarean where the amniotic lining has ruptured and bacteria can track back and then provide bacterial input to the baby. That is a subtle distinction, but I think it is important.

GUT PERMEABILITY IN THE NEWBORN

  1. Top of page
  2. METHODOLOGY: A GENE SCANNING APPROACH
  3. DIFFERENT APPROACHES TO THE ANALYSIS OF THE MICROBIOME
  4. METABOLIC PROFILING
  5. DEVELOPMENT OF THE MICROBIOME AFTER BIRTH AND THE INFLUENCE OF HOST GENETICS
  6. THE IMPACT OF EARLY FEEDING
  7. GUT PERMEABILITY IN THE NEWBORN
  8. EARLY MACRONUTRIENT EFFECTS ON THE HUMAN MICROBIOME
  9. FAT ABSORPTION AND BILE SALTS IN THE NEWBORN
  10. POPULATION DIFFERENCES IN THE MICROBIOME ACROSS THE GLOBE
  11. GUT MICROBIOTA AND THEIR POTENTIAL ROLE IN DEVELOPMENTAL PROGRAMING OF THE BRAIN
  12. DIETARY STANDARDIZATION
  13. NUTRIENT NICHES AND MICROBES THROUGHOUT THE LENGTH OF THE INTESTINE
  14. MUCOSAL BACTERIA
  15. MUCIN'S STRUCTURE AND INTERACTION WITH MUCOSAL BACTERIA
  16. GUT IMMUNOLOGY IN RELATION TO MUCIN AND MICROBIAL COLONIZATION
  17. DIETARY INTERVENTIONS: EFFECTS ON COLONIC MICROFLORA
  18. AUTISM – A GUT ROLE?
  19. ROLE OF THE MICROFLORA IN SYNTHESIZING ESSENTIAL AMINO ACIDS AND ALLOWING THE MAINTENANCE OF PROTEIN BALANCE ON VERY LOW INTAKES
  20. CHILDREN'S STUNTING: THE ROLE OF GASTROINTESTINAL MICROBIOTA
  21. THE ETHICS OF STUDYING DEPRIVED CHILDREN
  22. ENERGY BALANCE AND THE MICROBIOME
  23. PHYSICAL ACTIVITY AND THE MICROBIOME
  24. INSULIN SENSITIVITY AND THE METABOLIC SYNDROME
  25. PROBIOTICS AND THEIR EFFECTS
  26. CROHN'S DISEASE AND ULCERATIVE COLITIS
  27. REFERENCES

Cutberto Garza: The duration of gut permeability in the infant is still uncertain. About 20 years ago, we infused mothers with doubly-labeled N15 and C13 amino acids and documented in a variety of ways that the infants were absorbing either complete or fragments of lactoferrin, for example in the third month of life, and secreting these fragments in urine.19 Usually, however, we do not think permeability lasts much beyond 2 months of age, at most.

Philip James: If you have lactoferrin and other macromolecules, including bacterial proteins, permeating the intestine in the first few weeks of life this may have a major role in the generation of tolerance or other immunological effects with long-term implications?

Per Brandtzaeg: There is experimental evidence of increased permeability in the newborn, which is actually promoting induction or tolerance and regulatory adhesins, so this process may be very beneficial.

Fergus Shanahan: We need to be careful about assuming that permeability is very short-lived. Seven percent of people, we know from capsule endoscopies, have ulceration in the small bowel as a normal event. We also have diseases, such as Crohn's disease, with extensive ulceration and ulcerative colitis but people are not dying from sepsis. Celiac disease also affects 1–2% of the Western population, but they do not have septicemia. In contrast, ulceration of the skin with a cut or abrasion often does lead to infection. The gastrointestinal tract, which is only one cell layer thick, has tremendous reserves, so in many ways it does not matter. The surface area of the small intestine is huge and any malabsorptive condition involves marked permeability. When you do permeability studies for macromolecules, even for particulates, they show markedly increased permeability, yet they cannot absorb fat and proteins, and the intestine is not then acting like a sieve. This permeability phenomenon is overstated.

Philip James: About 40 years ago while I was working on intestinal function in malnourished babies I was told in India that by 6 weeks of age Indian babies had a flat small intestine that looked like celiac disease despite being breastfed. However, they also received routinely herbal teas which contained not only toxins but viruses and bacteria, and this was a routine finding throughout Southeast Asia. This was not tropical sprue.

Irv Rosenberg: I agree with Fergus in relation to permeability in the mature intestine but in terms of the permeability characteristic of the infant intestine in the first 6 weeks of life, I think general permeability closes down before about 2 months. So in my own field, I know that folic acid bound to folate-binding protein in breast milk will pass across that barrier in the first 6 weeks with a high degree of efficiency; but once this permeability shuts down, then you have to use other forms of folate transport. So my suspicion is that there is a very broad range of molecular sizes capable of permeating the intestine and involving many different kinds of molecules in the first few weeks of life.

Tore Midtvedt: I gather from my veterinarian friends that the period of permeability in a newborn animal partly reflects the time when trypsinogen is activated. In a newborn animal, you have no activation of trypsinogen, so you haven't the protein-degrading enzymes there. When trypsinogen is activated, then the permeability starts to change.

Philip James: When is it activated?

Tore Midtvedt: The activation depends on the species: it is excreted from the pancreas as inactivated trypsinogen – it is activated usually in the upper part of the small intestine.

Cutberto Garza: I thought it was very highly dependent on the IgG composition of the milk, because, in fact, that is the mechanism that a species uses to transfer IgG to the infant, rather than using the transplacental route. Is that right?

Per Brandtzaeg: That again depends on the species. For instance, in rodents they have their Fc series of gut receptors in the neonate, which are expressed for a few weeks and take breast milk IgG back into the neonate.20 Cows and horses are less dependent on these receptors, which are also found in the maternal gland. Human breast milk contains significant amounts of secretory immunoglobulin A [sIgA] and many other factors which are only partially digested but the products inhibit pathogens. In the human, there is also uptake of proteins which regulate the G protein stimulated receptors and these proteins – the RGS proteins – can be taken up by a receptor in the Peyer's patches stem cells. There is some leakage also of IgG into the gut, but not through any receptor system. Human milk also contains cytokines, cytokine receptors, Toll-like receptor agonists and antagonists, hormones, anti-inflammatory agents, and nucleotides, which all modulate inflammation and immunity. The Toll-like receptors are pattern recognition receptors that recognize conserved structures in pathogens, trigger innate immune responses, and prime antigen-specific adaptive immunity.21 Human milk is also rich in complex carbohydrates, i.e., glycans, and these indigestible glycans allow or stimulate the colonization of what are, in effect, probiotic organisms which modulate mucosal immunity and protect against pathogens. The glycans have a structural homology to intestinal cell surface receptors so they inhibit pathogen binding to the mucosa.22

EARLY MACRONUTRIENT EFFECTS ON THE HUMAN MICROBIOME

  1. Top of page
  2. METHODOLOGY: A GENE SCANNING APPROACH
  3. DIFFERENT APPROACHES TO THE ANALYSIS OF THE MICROBIOME
  4. METABOLIC PROFILING
  5. DEVELOPMENT OF THE MICROBIOME AFTER BIRTH AND THE INFLUENCE OF HOST GENETICS
  6. THE IMPACT OF EARLY FEEDING
  7. GUT PERMEABILITY IN THE NEWBORN
  8. EARLY MACRONUTRIENT EFFECTS ON THE HUMAN MICROBIOME
  9. FAT ABSORPTION AND BILE SALTS IN THE NEWBORN
  10. POPULATION DIFFERENCES IN THE MICROBIOME ACROSS THE GLOBE
  11. GUT MICROBIOTA AND THEIR POTENTIAL ROLE IN DEVELOPMENTAL PROGRAMING OF THE BRAIN
  12. DIETARY STANDARDIZATION
  13. NUTRIENT NICHES AND MICROBES THROUGHOUT THE LENGTH OF THE INTESTINE
  14. MUCOSAL BACTERIA
  15. MUCIN'S STRUCTURE AND INTERACTION WITH MUCOSAL BACTERIA
  16. GUT IMMUNOLOGY IN RELATION TO MUCIN AND MICROBIAL COLONIZATION
  17. DIETARY INTERVENTIONS: EFFECTS ON COLONIC MICROFLORA
  18. AUTISM – A GUT ROLE?
  19. ROLE OF THE MICROFLORA IN SYNTHESIZING ESSENTIAL AMINO ACIDS AND ALLOWING THE MAINTENANCE OF PROTEIN BALANCE ON VERY LOW INTAKES
  20. CHILDREN'S STUNTING: THE ROLE OF GASTROINTESTINAL MICROBIOTA
  21. THE ETHICS OF STUDYING DEPRIVED CHILDREN
  22. ENERGY BALANCE AND THE MICROBIOME
  23. PHYSICAL ACTIVITY AND THE MICROBIOME
  24. INSULIN SENSITIVITY AND THE METABOLIC SYNDROME
  25. PROBIOTICS AND THEIR EFFECTS
  26. CROHN'S DISEASE AND ULCERATIVE COLITIS
  27. REFERENCES

Cutberto Garza: If we are comparing infant formula and human milk composition, the most variable component in human milk is fat, both in content and in quality. Protein levels are fairly stable, lactose levels are absolutely steady, and fat is generally the most variable. On the other hand, because the infant's intake is often driven by its energy needs the compositional differences may be compensated for by changes in volume which the baby demands of the mother. But fat is, by far, the most variable, and possibly the least well absorbed of all the components. Many of the proteins in human milk are relatively resistant to digestion.

Alan Jackson: If you have an infant that has sufficient energy to allow it to grow at a rapid rate, then that will set the metabolic behavior of the host and of the microbiome to behave in a particular way. So an infant who is growing faster will upregulate the urea salvage compared to one who is growing more slowly and, therefore, there is an interaction where the needs of the host are set by the overall energy and the proportion of macronutrients that are taken in the diet. So the load passing into the gut is probably the important factor.

FAT ABSORPTION AND BILE SALTS IN THE NEWBORN

  1. Top of page
  2. METHODOLOGY: A GENE SCANNING APPROACH
  3. DIFFERENT APPROACHES TO THE ANALYSIS OF THE MICROBIOME
  4. METABOLIC PROFILING
  5. DEVELOPMENT OF THE MICROBIOME AFTER BIRTH AND THE INFLUENCE OF HOST GENETICS
  6. THE IMPACT OF EARLY FEEDING
  7. GUT PERMEABILITY IN THE NEWBORN
  8. EARLY MACRONUTRIENT EFFECTS ON THE HUMAN MICROBIOME
  9. FAT ABSORPTION AND BILE SALTS IN THE NEWBORN
  10. POPULATION DIFFERENCES IN THE MICROBIOME ACROSS THE GLOBE
  11. GUT MICROBIOTA AND THEIR POTENTIAL ROLE IN DEVELOPMENTAL PROGRAMING OF THE BRAIN
  12. DIETARY STANDARDIZATION
  13. NUTRIENT NICHES AND MICROBES THROUGHOUT THE LENGTH OF THE INTESTINE
  14. MUCOSAL BACTERIA
  15. MUCIN'S STRUCTURE AND INTERACTION WITH MUCOSAL BACTERIA
  16. GUT IMMUNOLOGY IN RELATION TO MUCIN AND MICROBIAL COLONIZATION
  17. DIETARY INTERVENTIONS: EFFECTS ON COLONIC MICROFLORA
  18. AUTISM – A GUT ROLE?
  19. ROLE OF THE MICROFLORA IN SYNTHESIZING ESSENTIAL AMINO ACIDS AND ALLOWING THE MAINTENANCE OF PROTEIN BALANCE ON VERY LOW INTAKES
  20. CHILDREN'S STUNTING: THE ROLE OF GASTROINTESTINAL MICROBIOTA
  21. THE ETHICS OF STUDYING DEPRIVED CHILDREN
  22. ENERGY BALANCE AND THE MICROBIOME
  23. PHYSICAL ACTIVITY AND THE MICROBIOME
  24. INSULIN SENSITIVITY AND THE METABOLIC SYNDROME
  25. PROBIOTICS AND THEIR EFFECTS
  26. CROHN'S DISEASE AND ULCERATIVE COLITIS
  27. REFERENCES

Tore Midtvedt: In terms of fat absorption, the 7 alpha dehydroxylation of bile acid molecules only starts after 6 months of age in a breastfed baby, so this implies that the microbial flora are not interfering with the fat absorption mechanism until after several months. However, we may need to be concerned because the mother's milk contains any toxins she has stored in her fat tissue which is then mobilized during lactation. In some parts of Europe there are considerable amounts of antimicrobial xenobiotics associated with the fat content of the mothers’ breast milk and we need to recognize this.

Alan Jackson: In ileostomy patients, assessing the effects of dietary fat may be a secondary consequence of bile salts, malabsorption, or depletion in the bile salt pool.

Michiel Kleerebezem: I agree that ileostomy patients have a different bile acid cycling and quite a lot of bile acids are excreted in the ileostomy effluent, so this could affect the microflora.

Irv Rosenberg: The deconjugation of bile salts by bacteria will influence their effect on fat micelle formation and absorption as well as the interplay of bacteria and bile acids. The infant may well also have some delay in its ability to generate the full enterohepatic circulation of bile salts because the permeability of the infant's intestine will affect the nature of both fat and protein absorption with its implications for the generation of passive immunity.

Ole Hernell: Monoglycerides also have an antimicrobial effect, at least in vitro. We know that if you have a fat composition as identical as possible to breast milk in an infant formula, it is less well-absorbed than the fat from human milk. At this age, what probably happens is that fat absorption occurs well down the small intestine and even in the ileum and that, of course, could affect the microbiota.

The pancreatic lipase, which has the positional specificity for the one and three positions on the glycerol backbone of the triglyceride in milk is not expressed in the newborn infant. The intestinal concentration of pancreatic triglyceride lipase [PTL] and bile salts is lower in newborns compared to later in life. Instead, the PTL-related protein 2 and bile salt-stimulated lipase [BSSL] are the key enzymes secreted from the pancreas, which in concerted action with gastric lipase operate to achieve efficient fat absorption during infancy. BSSL is also present in human milk, which affects fat absorption and growth in breastfed preterm infants.23 We have shown that neither of these have a positional specificity so both can hydrolyze the fatty acid in their central position.

Philip James: So you then have a marked inflow of glycerol and independent free fatty acids?

Ole Hernell: The extent to which that occurs – that really depends on the bile salt concentration, because if you have a mycellar phase, you will probably absorb the monoglycerides. If you have low bile salt concentration, you will probably form lamellar vesicles and then fatty acids are preferred in terms of absorption – so you have then more hydrolysis in the glycerol 2 position, but that is my speculation.

POPULATION DIFFERENCES IN THE MICROBIOME ACROSS THE GLOBE

  1. Top of page
  2. METHODOLOGY: A GENE SCANNING APPROACH
  3. DIFFERENT APPROACHES TO THE ANALYSIS OF THE MICROBIOME
  4. METABOLIC PROFILING
  5. DEVELOPMENT OF THE MICROBIOME AFTER BIRTH AND THE INFLUENCE OF HOST GENETICS
  6. THE IMPACT OF EARLY FEEDING
  7. GUT PERMEABILITY IN THE NEWBORN
  8. EARLY MACRONUTRIENT EFFECTS ON THE HUMAN MICROBIOME
  9. FAT ABSORPTION AND BILE SALTS IN THE NEWBORN
  10. POPULATION DIFFERENCES IN THE MICROBIOME ACROSS THE GLOBE
  11. GUT MICROBIOTA AND THEIR POTENTIAL ROLE IN DEVELOPMENTAL PROGRAMING OF THE BRAIN
  12. DIETARY STANDARDIZATION
  13. NUTRIENT NICHES AND MICROBES THROUGHOUT THE LENGTH OF THE INTESTINE
  14. MUCOSAL BACTERIA
  15. MUCIN'S STRUCTURE AND INTERACTION WITH MUCOSAL BACTERIA
  16. GUT IMMUNOLOGY IN RELATION TO MUCIN AND MICROBIAL COLONIZATION
  17. DIETARY INTERVENTIONS: EFFECTS ON COLONIC MICROFLORA
  18. AUTISM – A GUT ROLE?
  19. ROLE OF THE MICROFLORA IN SYNTHESIZING ESSENTIAL AMINO ACIDS AND ALLOWING THE MAINTENANCE OF PROTEIN BALANCE ON VERY LOW INTAKES
  20. CHILDREN'S STUNTING: THE ROLE OF GASTROINTESTINAL MICROBIOTA
  21. THE ETHICS OF STUDYING DEPRIVED CHILDREN
  22. ENERGY BALANCE AND THE MICROBIOME
  23. PHYSICAL ACTIVITY AND THE MICROBIOME
  24. INSULIN SENSITIVITY AND THE METABOLIC SYNDROME
  25. PROBIOTICS AND THEIR EFFECTS
  26. CROHN'S DISEASE AND ULCERATIVE COLITIS
  27. REFERENCES

Elaine Holmes: Is there a developmental difference, either in the microflora or in the immunology between a baby born in sub-Saharan Africa, where you are going to need, lifelong, more immunity per se, and a babe in Europe?

Rob Knight: We have birth cohort data from Bangladesh, although not from Malawi, which is still in the process of being analyzed. When we look cross-sectionally, what we see is that the rate of development in terms of the rate of approach of the infant microbiota to the adult microbiota is essentially the same in the United States, Malawi, and Bangladesh, although the end states are very different in those populations. So it is interesting that the trajectory seems to be very similar, although these data are still very preliminary and based on cross-sectional rather than longitudinal studies at this point.

David Relman: Although the trajectory and rate of change may be similar, is the end point similar? I suspect not. We have looked at the microbiota of some Bangladeshi adults but I have no feel for the geographical differences across the world.

Rob Knight: The endpoints in the three populations we have studied are different, but we do not know whether this reflects their relatively restricted lifestyles, their genetics, environmental exposures, or diet.

Alan Jackson: We also have to remember the complexity of dietary data and its analysis, so if you want to do something like principal component analysis on dietary patterns across the world in relation to particular microbial patterns across the globe, this will be a challenge.

Rob Knight: So far, we are in a very rudimentary phase of characterizing microbial patterns across the word. We know that the Malawi diet is about 90% corn with 10% from leaves. This is very different from the South American diet, which has a lot more root vegetables, and both of those are very different from the Western diet. We have not, however, yet characterized extensively the microbial flora in the subjects that we have access to.

Alan Jackson: We haven't characterized the Malawi diet, but we have characterized the South African diet. When you do that, you get three or four broad patterns dependent on rural living, urbanization, socioeconomic status, and so on. So within a population you can get big differences.

Rob Knight: My impression is that the studies which undertake careful analyses of diet and of microbiota are not the same studies. We clearly need combined evaluations.

Philip James: My recollection is that the Malawi diet is extraordinarily low in fat, very modest in protein, with a very substantial intake of maize and starchy root crops, such as cassava, which are fiber-rich starches. In South America, one has some ancient South American diets, but these are being overwhelmed now by Western diets. We are being told that we have a resilient gut microbial flora, presumably established early on in life, but there are huge dietary changes taking place globally as populations transfer from a rural village environment into a slum setting, which applies to 30 to 50% of the world's population. In the slums, the bacterial load is enormous and the dietary changes very substantial, so are your studies dealing with the teeth microflora, the small intestine microflora, or the mass organ responsiveness of the colon?

Rob Knight: In our Malawi study site, which is run by Mike Manary of the Blantyre College of Medicine, there are four separate villages, but the diet is extremely homogenous within most villages. Our South American site includes the Platanillal in Venezuela.24 We did look explicitly at a gradient from people who are living in very traditional communities involved in gathering food and a light agricultural lifestyle through to those that are more integrated with cities and living in the slums. However, although we did see some systematic microflora changes with Westernization, these American Indian populations were more closely related in their microflora to their country counterparts than they were to the US population.

Richard Black: Do we know what the effects are of being on a long-standing maize diet versus a rice diet versus a meat diet as the main staple versus a tuber?

Rob Knight: I don't know of any microbiota studies at that level that have used reasonably current techniques. Old frozen samples are fine if stored at minus 80 degrees. I think these studies would be of immense value in setting out a framework so that we could find out if the microfloral exposure soon after birth is the key to the individual's spectrum of microbes and their persistence. Even 2 years ago it would have been prohibitive to contemplate sequencing DNA from that number of people.

GUT MICROBIOTA AND THEIR POTENTIAL ROLE IN DEVELOPMENTAL PROGRAMING OF THE BRAIN

  1. Top of page
  2. METHODOLOGY: A GENE SCANNING APPROACH
  3. DIFFERENT APPROACHES TO THE ANALYSIS OF THE MICROBIOME
  4. METABOLIC PROFILING
  5. DEVELOPMENT OF THE MICROBIOME AFTER BIRTH AND THE INFLUENCE OF HOST GENETICS
  6. THE IMPACT OF EARLY FEEDING
  7. GUT PERMEABILITY IN THE NEWBORN
  8. EARLY MACRONUTRIENT EFFECTS ON THE HUMAN MICROBIOME
  9. FAT ABSORPTION AND BILE SALTS IN THE NEWBORN
  10. POPULATION DIFFERENCES IN THE MICROBIOME ACROSS THE GLOBE
  11. GUT MICROBIOTA AND THEIR POTENTIAL ROLE IN DEVELOPMENTAL PROGRAMING OF THE BRAIN
  12. DIETARY STANDARDIZATION
  13. NUTRIENT NICHES AND MICROBES THROUGHOUT THE LENGTH OF THE INTESTINE
  14. MUCOSAL BACTERIA
  15. MUCIN'S STRUCTURE AND INTERACTION WITH MUCOSAL BACTERIA
  16. GUT IMMUNOLOGY IN RELATION TO MUCIN AND MICROBIAL COLONIZATION
  17. DIETARY INTERVENTIONS: EFFECTS ON COLONIC MICROFLORA
  18. AUTISM – A GUT ROLE?
  19. ROLE OF THE MICROFLORA IN SYNTHESIZING ESSENTIAL AMINO ACIDS AND ALLOWING THE MAINTENANCE OF PROTEIN BALANCE ON VERY LOW INTAKES
  20. CHILDREN'S STUNTING: THE ROLE OF GASTROINTESTINAL MICROBIOTA
  21. THE ETHICS OF STUDYING DEPRIVED CHILDREN
  22. ENERGY BALANCE AND THE MICROBIOME
  23. PHYSICAL ACTIVITY AND THE MICROBIOME
  24. INSULIN SENSITIVITY AND THE METABOLIC SYNDROME
  25. PROBIOTICS AND THEIR EFFECTS
  26. CROHN'S DISEASE AND ULCERATIVE COLITIS
  27. REFERENCES

Sven Pettersson: I, with Joseph Rafter, coordinate an EU Tornado [Molecular Targets Open for Regulation by the Gut Flora –New Avenues for improved Diet to Optimize European Health] Consortium where we attempt to evaluate whether the very early exposure to gut microbes programs the individual's future responses including the closure of the blood brain barrier and even the programming of behavior.25 Before birth, the human gut is only exposed to amniotic fluid, but immediately after birth, the newborn is rapidly exposed to an array of complex microbes. This process has been shown to contribute to the developmental programming of epithelial barrier function, gut homeostasis, and angiogenesis, as well as the innate and host adaptive immune function that Denise was talking about. Our recent data indicate that gut microbiota also have effects on liver function,26 so other organs might well be also involved. Using measures of motor activity and anxiety-like behavior, germ-free [GF] mice were found to have increased motor activity and reduced anxiety, compared with specific-pathogen-free [SPF] mice with a normal gut microbiota. This behavior was associated with altered expression of genes known to be involved in second messenger pathways and synaptic long-term potentiation in brain regions implicated in motor control and anxiety-like behavior. For example, GF mice have elevated noradrenaline, dopamine, and 5-hydroxy tryptophan levels turnover in the brain's striatum. The GF mice also show altered expression of synaptic plasticity-related genes, which relate to the development of anxiety behavioral patterns in test animals. Then, when GF mice were exposed to gut microbiota early in life, they showed similar characteristics to those of the SPF mice, including reduced expression of PSD-95 and synaptophysin in the striatum. So our results suggest that the microbial colonization process initiates signaling mechanisms that affect neuronal circuits involved in motor control and anxiety behavior. If the microbial exposure of GF animals occurs later in life, it does not affect behavior in adult life. This implies that there is a time window where you can have some impact on the brain's response to environmental cues.

In our analyses we have focused on the neurotransmitters. If we take the serotonin that is derived from tryptophan then in fetal life, the growing organism's needs are satisfied, but after birth there is a huge developmental switch and the newborn is suddenly in need of tryptophan. The metabolism of tryptophan is complex and during postnatal development, the levels of serotonin peak around weaning and then rise again to adult levels. Gut bacteria alter tryptophan metabolism. Jeffrey Gordon showed 10 years ago that these microflora also affect gut barrier closure.27 We have also shown that gut microbes affect the control of the blood brain barrier after birth, so this will affect all manner of functions relating to brain development. Diamond et al.28 have shown that the B cell antibodies responsive to environmental infections cross-react with brain antigens and, normally, the blood brain barrier ensures that brain reactivity is not induced. However, in some infective circumstances this blood brain barrier is opened, as in cases of rheumatic fever. This is when you see the abnormal motor movements of chorea induced by the abnormal basal ganglia because of cross reactivity when the blood brain barrier has opened. Furthermore, Serrats et al.29 at the Salk Institute have shown that the perivascular macrophages can signal the brain's hypothalamic axis and induce stress responses. So there are several mechanisms whereby microbes can induce substantial brain changes.

So perhaps the maintenance of the blood brain barrier is an active process. That is, it requires constantly a secretion of metabolites from bacteria in the environment in order to maintain its integrity. What then happens if you introduce antibiotics? There are also accumulating data to suggest that as you get older, you get more and more pro-inflammatory responses, and the issue is then whether the microbiome will gradually change the barrier integrity and particularly affect the brain of the elderly?

Fredrik Bäckhed: Sven, all of these gut metabolites are produced either in the gut or in the brain. Do you think it is the gut-derived serotonin that is the more important, or is the brain-derived serotonin the key in these behavioral phenotypes?

Sven Pettersson: We showed that you increased the steady state levels of noradrenaline, dopamine, and serotonin within the brain. That is intrinsic to the brain and it is very hard to see how this molecule would actually penetrate the blood brain barrier from the periphery.

Philip James: So you are implying that all sorts of molecular species and other components are going to get into the brain under these early life conditions and that in some way this programs the metabolic profiling of the brain and the next phase of brain development?

Sven Pettersson: Absolutely. And this also occurs during the phase in the uterus when the blood brain barrier is open.

Irv Rosenberg: The metabolism of tryptophan involving its degradation or to some extent even its conversion to serotonin has a number of steps, some of which are vitamin B6 dependent. It raises once again the question of how the infant manages these conversions and whether, after birth, we have special nutritional requirements which are affected by the microbes.

Cutberto Garza: We worry, for example, about the amount of glycine in human milk and the role that it can play. But Alan Jackson would be the best person to comment because he has looked at this very carefully30 and how the recycling of urea and the roles the methyl groups all play affect that very complex interaction.

Elaine Holmes: If you look at brain composition, then the hippocampus, in particular, has a different composition in germ-free and conventional animals, i.e., higher GABA and glutamate levels than conventional animals. There is a very distinct difference in all brain regions, brain stem, and hippocampus. I don't know why.

Alan Jackson: One of the more important issues about rodents is that they eat their own stools as a matter of course and, therefore, generate nutrients and organisms which are recycled and absorbed by the small intestine – that is important and substantially different from humans. I just wonder the extent to which one is looking at a consequence of having an increased bacterial or metabolite load in the small intestine. It suggests we need to be careful when comparing animals and humans.

I think it is probably worth just making the point about changing behavior in relation to events which affect the gut. Meaney's group has shown very clearly that grooming behavior actually influences the behavior of the offspring and the subsequent generations by inducing specific epigenetic changes to the promoter region of particular genes in certain loci in the brain.31 So here you have a potential link between a nutrient-related behavior, which may link to specific microbiome changes, and is then communicated or associated with an outcome in the offspring which has very hard biological markers. At the very least, it suggests that the relatively short period of exposure at early time points can have a long-term effect, and it offers a way in which the microbiome might potentially interact through behavior with the parent/offspring relationship to make it at least plausible.

Tore Midtvedt: Sudo in Japan brilliantly showed that the ability to counteract stress was completely different in germ-free and conventional animals32. When he then conventionalized newborn germ-free animals, they became conventional in their stress response but not if the mice were conventionalized later. So that taught me that you do have a window for establishing gut microbial-induced changes to long-term brain behavior.

DIETARY STANDARDIZATION

  1. Top of page
  2. METHODOLOGY: A GENE SCANNING APPROACH
  3. DIFFERENT APPROACHES TO THE ANALYSIS OF THE MICROBIOME
  4. METABOLIC PROFILING
  5. DEVELOPMENT OF THE MICROBIOME AFTER BIRTH AND THE INFLUENCE OF HOST GENETICS
  6. THE IMPACT OF EARLY FEEDING
  7. GUT PERMEABILITY IN THE NEWBORN
  8. EARLY MACRONUTRIENT EFFECTS ON THE HUMAN MICROBIOME
  9. FAT ABSORPTION AND BILE SALTS IN THE NEWBORN
  10. POPULATION DIFFERENCES IN THE MICROBIOME ACROSS THE GLOBE
  11. GUT MICROBIOTA AND THEIR POTENTIAL ROLE IN DEVELOPMENTAL PROGRAMING OF THE BRAIN
  12. DIETARY STANDARDIZATION
  13. NUTRIENT NICHES AND MICROBES THROUGHOUT THE LENGTH OF THE INTESTINE
  14. MUCOSAL BACTERIA
  15. MUCIN'S STRUCTURE AND INTERACTION WITH MUCOSAL BACTERIA
  16. GUT IMMUNOLOGY IN RELATION TO MUCIN AND MICROBIAL COLONIZATION
  17. DIETARY INTERVENTIONS: EFFECTS ON COLONIC MICROFLORA
  18. AUTISM – A GUT ROLE?
  19. ROLE OF THE MICROFLORA IN SYNTHESIZING ESSENTIAL AMINO ACIDS AND ALLOWING THE MAINTENANCE OF PROTEIN BALANCE ON VERY LOW INTAKES
  20. CHILDREN'S STUNTING: THE ROLE OF GASTROINTESTINAL MICROBIOTA
  21. THE ETHICS OF STUDYING DEPRIVED CHILDREN
  22. ENERGY BALANCE AND THE MICROBIOME
  23. PHYSICAL ACTIVITY AND THE MICROBIOME
  24. INSULIN SENSITIVITY AND THE METABOLIC SYNDROME
  25. PROBIOTICS AND THEIR EFFECTS
  26. CROHN'S DISEASE AND ULCERATIVE COLITIS
  27. REFERENCES

Sven Pettersson: In the EU TORNADO program, the need to specify in any study the exact nature of the diet became crucial. In mice studies, dietary differences are a major problem. Furthermore, the typing of the mice is crucial and a so-called Black-6 mouse may be very different from one lab to another because somebody has been sloppy. Then, after obtaining suitable omic profiles it is very easy to misinterpret these from a mechanistic point of view. So we need concerted action from the animal feed industry to help us make tailor-made diets so we can do reliable mechanistic studies in rodents.

Karen Madsen: This issue is just as important in humans when we consider physiological and functional analyses.

Tore Midtvedt: All the experimental results that have been presented have been obtained using specific animals with conventional flora. The conventional flora in Gothenburg differs strongly from conventional flora in the United States and strongly from the conventional flora at the Karolinska. So we need to have a far better definition of what the conventional flora is in our so-called “conventional” animals. Many of them, in practice, are somewhere between germ-free and conventional.

Philip James: How am I going to define the parameters for specifying “conventional”?

Tore Midtvedt: So far, it is more or less the absence of some specified microbes! Conventionality is defined as signifying the absence of something unusual or pathogenic, so the definition means almost nothing!

Fredrik Bäckhed: If you buy mice from a vendor, they are all mice with a specific but unknown flora and this is very different from mice that you have been breeding in your facility for the past 40 years. So I think at least you should have your control mice within your facility for several generations before you even consider them as controls. Also, bear in mind that these are not wild mice – these are mice that have been in captivity for many generations, though it would be interesting to start looking at the wild mice and their microflora.

NUTRIENT NICHES AND MICROBES THROUGHOUT THE LENGTH OF THE INTESTINE

  1. Top of page
  2. METHODOLOGY: A GENE SCANNING APPROACH
  3. DIFFERENT APPROACHES TO THE ANALYSIS OF THE MICROBIOME
  4. METABOLIC PROFILING
  5. DEVELOPMENT OF THE MICROBIOME AFTER BIRTH AND THE INFLUENCE OF HOST GENETICS
  6. THE IMPACT OF EARLY FEEDING
  7. GUT PERMEABILITY IN THE NEWBORN
  8. EARLY MACRONUTRIENT EFFECTS ON THE HUMAN MICROBIOME
  9. FAT ABSORPTION AND BILE SALTS IN THE NEWBORN
  10. POPULATION DIFFERENCES IN THE MICROBIOME ACROSS THE GLOBE
  11. GUT MICROBIOTA AND THEIR POTENTIAL ROLE IN DEVELOPMENTAL PROGRAMING OF THE BRAIN
  12. DIETARY STANDARDIZATION
  13. NUTRIENT NICHES AND MICROBES THROUGHOUT THE LENGTH OF THE INTESTINE
  14. MUCOSAL BACTERIA
  15. MUCIN'S STRUCTURE AND INTERACTION WITH MUCOSAL BACTERIA
  16. GUT IMMUNOLOGY IN RELATION TO MUCIN AND MICROBIAL COLONIZATION
  17. DIETARY INTERVENTIONS: EFFECTS ON COLONIC MICROFLORA
  18. AUTISM – A GUT ROLE?
  19. ROLE OF THE MICROFLORA IN SYNTHESIZING ESSENTIAL AMINO ACIDS AND ALLOWING THE MAINTENANCE OF PROTEIN BALANCE ON VERY LOW INTAKES
  20. CHILDREN'S STUNTING: THE ROLE OF GASTROINTESTINAL MICROBIOTA
  21. THE ETHICS OF STUDYING DEPRIVED CHILDREN
  22. ENERGY BALANCE AND THE MICROBIOME
  23. PHYSICAL ACTIVITY AND THE MICROBIOME
  24. INSULIN SENSITIVITY AND THE METABOLIC SYNDROME
  25. PROBIOTICS AND THEIR EFFECTS
  26. CROHN'S DISEASE AND ULCERATIVE COLITIS
  27. REFERENCES

Alan Jackson: Presumably the organisms occupy different ecological niches and one of the main pressures as to what lives in those niches is the availability of nutrients. Now, we have talked about nutrients in terms of the host diet, but clearly the microenvironments of the organisms is much more refined than the host diet and the organisms themselves operate amongst themselves to not only use that environment but also to condition that environment. So to what extent is what we heard about the resilience of microbial flora in specific sites a statement about the nutrient environment that is imposed, given the nature of the ecological niche? Thus, one might be able to predict the microbial pattern based on the observed nutrients or the limiting nutrient in the niche, but I am not sure that we are measuring these potentially crucial environmental nutrient niches.

Philip James: David Relman: can you answer this issue of nutrient niches in relation to microbial stability? Is the stability because that microbe has got into that particular location? What is the process whereby it gets the lock-in, and is it actually related to the micronutrient availability and the extent to which that conditions other organisms?

David Relman: There must be other environmental factors that are equally important in determining the nature of the niche and how the habitat shapes the niche. We are not very good at measuring these features.

Alan Jackson: I accept that, but just from the point of view of this meeting, which is about nutrition and the microbiome, do we not need to know the extent to which nutrition is a major determinant of the resilience and stability of the microflora in particular niches and how the nutritional environment selects the patterns of organisms?

Karen Madsen: This microenvironment, e.g., involving the redox potential and oxygen and nutrient availability, is a vital requirement which will influence the microflora of the epithelium. It is clear that the pH will change in this microenvironment and modulate the ionic effects at the mucosal barrier, so this may play a huge role in modulating which gut microbes are sustained.

Tore Midtvedt: These are important issues. In the duodenum, you have a different pH, you have some millivolt increases in the redox potential, and you have a little more oxygen tension. So you may have microbes that are capable of doing things, but they cannot do it under those prevailing conditions. Then, if you are considering bile acids a 7-alpha dehydroxylation step can never occur at the pH that is established up in the duodenum. So these key physiochemical factors need to be considered when we think about metabolism in the small intestine.

Olle Hernell: It is probably very relevant to start with the small intestine microbiota. In jejunal biopsies taken from patients with potential celiac disease, the biopsies have very different microflora from that seen lower down the gut. I assume that there is cross-talk with the immune system, and this is even more likely to occur in the small rather than the large intestine.

Philip James: In the old days we looked at the microflora, as best we could, in small intestinal biopsies of malnourished children with diarrhea in Jamaica33 and in adults returning from India with tropical sprue.34 The mucosal microflora of the sick children with gastroenteritis was distinctive but not pathognomonic, but we knew that Indian children were so microbially contaminated that by 6 weeks of age, despite purported breastfeeding, they had a flat mucosa with an intense jejunal microflora. Then, when we studied adults returning to the United Kingdom with tropical sprue, the extent of microfloral growth found within the mucosa was the index of whether they came back and recovered rapidly or not. It seems to me that as we have talked about mouth, feces, different parts of the skin, we could do with something about the new evidence on different parts of the small intestine and the upper colon.

Tore Midtvedt: Between the mouth and the anus, we know little about the microbiome. One challenge is that the presence of a gene doesn't mean the presence of a function. So from a metabolic point of view, we need far more data about the functional status of the flora between the mouth and the anus.

Fredrik Bäckhed: The microbial metabolites that we actually study that have an impact on human or mouse physiology are largely expressed in the ileum and not the colon. So it should be the ileal flora that we should study for a lot of these responses and they may then be reflected in the colonic flora. So the challenge will be to start collecting samples from the ileum or the jejunum. Then, the next challenge will be to differentiate between the luminal and mucus-associated microbiota.

Michiel Kleerebezem: We do tend to focus on the colonic flora when the small intestinal microbiota needs studying. In Wageningen University we have put a lot of effort into trying to get at least some information on this in human ileostomy patients. Of course, these patients are not normal, but we have evidence that their ileal flora reflects that in the more proximal small intestine, so more jejunal-like microbial communities are found in the ileostomy effluent of these subjects.

Willem de Vos: I agree with Michiel Kleerebezem that the ileum is very important and a major place for microbial interaction with metabolism. When we look at the ileal stoma effluent, it is quite rich in short-chain fatty acids – it has 10–20 mmol of acetate and butyrate, so this reflects substantial bacterial activity. It is therefore possible that the signaling, e.g., relating to obesity, also plays a role in the ileum.

Bo Angelin: Of course it is vital to stress the importance of studying the small intestine. Coming from the bile acid field, I have to point out that the small intestine is completely different from the colon. I have always assumed that the bacterial composition of the Peyer's patches in the small intestine is important in immune function and very different from the flora in the rest of the small intestine. The metabolic environment will mean that a certain bunch of microbes go together and form a community and then exert an important regulatory function. Studying the colon will tell us very little about the small intestine, but we also need to study the various parts of the small intestine in relation to what is going on metabolically.

Nathalie Delzenne: In terms of the relationship between the upper part of the gut and the colon, the host's physiology is really crucial. There are many response elements to the microbial flora in the colon and these responses may feed back to the small intestine by neuronal or hormonal routes, so colonic events may condition the host response to upper gut bacteria and we must not assume they are acting in the upper gut independently of what is happening further down.

Dennis Kasper: Immunologists have primarily focused on the small intestine in terms of mucosal immunology whereas microbiologists have focused on the colon. So there is this enormous discrepancy between these two supposedly related disciplines. There is also the issue of the type of microbiome. This is of extreme importance in terms of gut immunology since, for example, the segmented filamentous bacteria [SFB] induce enormous differences depending on whether the mice are derived from the Jackson Laboratory or Taconic Farms. First, there is a genetic model of rheumatoid arthritis where in the KXB mouse the colonization with SFB induces much more rapid and severe disease. This is mediated through the TH17 cells. On the other hand, Dan Littman35 showed in his Cell paper that a pathogen of the gut, Citrobacter rodentium given to a mouse already colonized by SFB, does not induce the expected gastrointestinal disease because the Th17 cells are protective in that model. Thus, colonization with SFB, which adhere tightly to intestinal epithelial cells in the terminal ileum, is associated with the presence of Th17 cells in the intestinal lamina propria, and these mice are then more resistant to the growth of C. rodentium. So this shows that specific species of bacteria can induce or protect the animal from different diseases depending on whether the mouse is colonized with this one species. However, there could be a hundred other organisms that have a significant effect with enormous differences in outcome. So when you look at the immunologic data over the years, you find similar labs, with similarly conducted experiments producing widely different results, but they have never considered that the microflora could be very different and be playing a major role.

Denise Kelly: I see numerous papers for example using the TLR5 knockout mouse where some labs report that they develop metabolic syndrome and obesity whereas other labs see nothing. I think there comes a point where we have to say we may have made great progress but we now need to be systematic and collaborative with multi-site studies using and testing under exactly the same conditions, including diet. The other issue is that when monitoring the role of commensal microbes, we culture them and sometimes do not realize that as you subculture bacteria, you are losing genetic information all the time. So its fitness for an ecosystem is changing. I think there isn't enough attention being paid to that aspect, so we really have to look at the whole picture: the mice, the genetics, the culturing, the diet, i.e., everything, if we are going to get results that stand the test of time.

Ruth Ley: Agreed, but before we go to more complex studies where we are swapping mice and so on, we first need to create a database of our results so that we can perform meta-analyses. So it is increasingly important for people generating these data to annotate them well so that they can be centrally deposited and allow the conduct of meta-analyses. Then we can look for host genotype–microbiota interactions that lead to different phenotypes by looking across our multiple studies. This is true also for dietary issues: we have been involved in several studies where we had diet information but this did not give us any correlates and so we never reported it. With such an emphasis on positive results, we are missing vital information, so we need a database of unpublished as well as published data because this will help with meta-analyses in the future.

Jens Nielsen: I completely agree. Many people see the variability in results and say that this is extremely complex so because of this complexity we need to bring in mathematical models that can capture key things. Rob Knight showed us how there may be a large variation in species, but when metabolic functions are analyzed one starts to see that there is more consistency. So I am pretty sure that we can generate a few representative metabolic models for different species that are representative of what is happening. We have started to do that – building models for many different bacteria that exist in the gut and trying to discern the common denominators of the specific function in these different models. We may well end up describing the metabolic conversions in the microbiome in the gut with relatively few descriptions of different bacteria and have a very good assessment of the overall metabolism. So I think that is the kind of way we need to go.

Harry Flint: I think this is a very interesting approach where you try to define a minimal number of functional groups. Of course that depends on being able to correlate activity and culture, preferably, with what is happening in vivo. We have been doing this quite successfully and using it together with functional modeling. You can then estimate how good the prediction is from the information you have on cultural groups. However, you can't hope to deal with all functions in one set of models

Dusko Ehrlich: In humans, when we look for associations of bacteria in disease, what we have observed is that the numbers of subjects matter. We need to examine sufficient numbers of humans in order to detect reasonable associations but we cannot, of course, routinely obtain small intestinal samples. The next point relates to diet. We seem to be convinced that diet influences the microbiota, but frankly I am not aware of deep and broad analyses in humans which really demonstrate that.

MUCOSAL BACTERIA

  1. Top of page
  2. METHODOLOGY: A GENE SCANNING APPROACH
  3. DIFFERENT APPROACHES TO THE ANALYSIS OF THE MICROBIOME
  4. METABOLIC PROFILING
  5. DEVELOPMENT OF THE MICROBIOME AFTER BIRTH AND THE INFLUENCE OF HOST GENETICS
  6. THE IMPACT OF EARLY FEEDING
  7. GUT PERMEABILITY IN THE NEWBORN
  8. EARLY MACRONUTRIENT EFFECTS ON THE HUMAN MICROBIOME
  9. FAT ABSORPTION AND BILE SALTS IN THE NEWBORN
  10. POPULATION DIFFERENCES IN THE MICROBIOME ACROSS THE GLOBE
  11. GUT MICROBIOTA AND THEIR POTENTIAL ROLE IN DEVELOPMENTAL PROGRAMING OF THE BRAIN
  12. DIETARY STANDARDIZATION
  13. NUTRIENT NICHES AND MICROBES THROUGHOUT THE LENGTH OF THE INTESTINE
  14. MUCOSAL BACTERIA
  15. MUCIN'S STRUCTURE AND INTERACTION WITH MUCOSAL BACTERIA
  16. GUT IMMUNOLOGY IN RELATION TO MUCIN AND MICROBIAL COLONIZATION
  17. DIETARY INTERVENTIONS: EFFECTS ON COLONIC MICROFLORA
  18. AUTISM – A GUT ROLE?
  19. ROLE OF THE MICROFLORA IN SYNTHESIZING ESSENTIAL AMINO ACIDS AND ALLOWING THE MAINTENANCE OF PROTEIN BALANCE ON VERY LOW INTAKES
  20. CHILDREN'S STUNTING: THE ROLE OF GASTROINTESTINAL MICROBIOTA
  21. THE ETHICS OF STUDYING DEPRIVED CHILDREN
  22. ENERGY BALANCE AND THE MICROBIOME
  23. PHYSICAL ACTIVITY AND THE MICROBIOME
  24. INSULIN SENSITIVITY AND THE METABOLIC SYNDROME
  25. PROBIOTICS AND THEIR EFFECTS
  26. CROHN'S DISEASE AND ULCERATIVE COLITIS
  27. REFERENCES

Olle Hernell: I think that it is very important that you really know where you are sampling in the small intestine. One must also distinguish between the luminal sample and the mucosal sample. If you take a biopsy and treat it rather rigorously you will probably have left only those microbes which are attached or are adhering to the mucosa, so perhaps these are the species that are really interesting; they are clearly different from what you will find in the lumen.

MUCIN'S STRUCTURE AND INTERACTION WITH MUCOSAL BACTERIA

  1. Top of page
  2. METHODOLOGY: A GENE SCANNING APPROACH
  3. DIFFERENT APPROACHES TO THE ANALYSIS OF THE MICROBIOME
  4. METABOLIC PROFILING
  5. DEVELOPMENT OF THE MICROBIOME AFTER BIRTH AND THE INFLUENCE OF HOST GENETICS
  6. THE IMPACT OF EARLY FEEDING
  7. GUT PERMEABILITY IN THE NEWBORN
  8. EARLY MACRONUTRIENT EFFECTS ON THE HUMAN MICROBIOME
  9. FAT ABSORPTION AND BILE SALTS IN THE NEWBORN
  10. POPULATION DIFFERENCES IN THE MICROBIOME ACROSS THE GLOBE
  11. GUT MICROBIOTA AND THEIR POTENTIAL ROLE IN DEVELOPMENTAL PROGRAMING OF THE BRAIN
  12. DIETARY STANDARDIZATION
  13. NUTRIENT NICHES AND MICROBES THROUGHOUT THE LENGTH OF THE INTESTINE
  14. MUCOSAL BACTERIA
  15. MUCIN'S STRUCTURE AND INTERACTION WITH MUCOSAL BACTERIA
  16. GUT IMMUNOLOGY IN RELATION TO MUCIN AND MICROBIAL COLONIZATION
  17. DIETARY INTERVENTIONS: EFFECTS ON COLONIC MICROFLORA
  18. AUTISM – A GUT ROLE?
  19. ROLE OF THE MICROFLORA IN SYNTHESIZING ESSENTIAL AMINO ACIDS AND ALLOWING THE MAINTENANCE OF PROTEIN BALANCE ON VERY LOW INTAKES
  20. CHILDREN'S STUNTING: THE ROLE OF GASTROINTESTINAL MICROBIOTA
  21. THE ETHICS OF STUDYING DEPRIVED CHILDREN
  22. ENERGY BALANCE AND THE MICROBIOME
  23. PHYSICAL ACTIVITY AND THE MICROBIOME
  24. INSULIN SENSITIVITY AND THE METABOLIC SYNDROME
  25. PROBIOTICS AND THEIR EFFECTS
  26. CROHN'S DISEASE AND ULCERATIVE COLITIS
  27. REFERENCES

Gunnar Hansson: We have talked a lot about the microbes in the lumen and then about host metabolism – but we have not considered the mucosal barrier. We made some progress in that area recently.36,37 The normal intestinal microbiota inhabit the colon mucus without triggering an inflammatory response. The colonic mucus is organized in two layers: an inner, stratified mucus layer that is firmly adherent to the epithelial cells and approximately 50 µm thick in the mouse. It is around 100 µm in rats and it is probably 150 to 200 µm in humans in the distal colon – it is a little bit thinner further up the colon and varies a little along the colon. These mucus layers are organized around the highly glycosylated MUC2 mucin, which has 5,200 amino acids and this glycoprotein is glycosylated with 80% glycans, so it carries a lot of good stuff for the bacteria to live on. The mucin forms a large, net-like polymer derived from monomers of mucin type O-linked oligosaccaharides [O-glycans]. These are then converted to dimers and trimmers before forming net-like structures as glycoproteins. O-glycans are synthesized post-translationally in the Golgi apparatus. All O-glycans are initiated with a primary structure referred to as Tn antigen [GalNAcα-O-Ser/Thr], which is normally masked by additional glycosylation to form the main type of O-glycans, core 1-derived structures. The biosynthesis of core 1 is controlled by core 1 β1,3-galactosyltransferase (also called T synthase). This mucin is secreted by the goblet cells, which are derived from stem cells in the bottom of the crypt. The mucin is a very packed material within the goblet cells, but on secretion it expands about 1,500 times in volume. The inner mucus layer is dense and does not allow bacteria to penetrate, thus keeping the epithelial cell surface free from bacteria. The inner mucus layer is converted into the outer layer, which is the habitat of the commensal flora. The outer mucus layer, which, in mice, is a non-attached layer approximately 100 µm thick, is expanded due to proteolytic activities provided by the host but probably also caused by commensal bacterial proteases and glycosidases. The numerous O-glycans on the MUC2 mucin not only serve as nutrients for the bacteria but also as attachment sites and, as such, probably contribute to the selection of the species-specific colon flora.

In the small intestine, you do not have two mucus layers but only one mucus layer and it fills up the space between most of the villi. The problem often in the proximal small intestine is that sometimes the villi are pointing out of the mucus, but normally most of the small intestine is also covered by mucus. In the distal ileum, it extends around 50 µm above the tip of the villi. The interesting thing is that in the small intestine the mucus is not attached and is permeable to bacteria, unlike in the colon. So there are different organizational structures and functions in the small and large intestines. The glycans that the bacteria encounter and use as attachment sites are different in the small intestine, unlike the large intestine, so blood group A mucus is different from blood group B mucus in the small intestine only.

The colonic mucosal wall has a stratified system of flat mucin nets derived by secretion of mucin from the mucosa with a generation time of about an hour as the inner mucus is changed to the outer mucus by a 4–5 times expansion due to cleavages inside the molecule. These cleavages are controlled by the host, but we do not yet understand exactly how that works at a distance from the cells. This outer mucus layer can be up to a millimeter thick and we now know that the loose structure allows the bacteria to penetrate but they only live in this layer – the inner mucus layer is totally devoid of bacteria because it is so dense.

The MUC 2 mucin in the large intestine of humans has identical glycans in all of us. In other parts of the body, such as the small intestine, our glycans vary according to our blood groups and other factors, but in the colon we are identical which is rather surprising. I think that tells us that there is a reason for this and the reason is most likely that bacteria are using these as attachment sites, because I think there are bacteria that live especially at the lower part of this outer layer where they attach to the glycans. And, of course, I think that is one of the main reasons why we have this relatively constant colonic flora. These glycans are the main source, or one of the main sources, of food for the bacteria. The bacteria produce enzymes, predominantly exoglycosidases, which remove one sugar at a time but there are also proteases that can disrupt the whole system and completely disrupt the mucus, and this relates to disease (see later).

GUT IMMUNOLOGY IN RELATION TO MUCIN AND MICROBIAL COLONIZATION

  1. Top of page
  2. METHODOLOGY: A GENE SCANNING APPROACH
  3. DIFFERENT APPROACHES TO THE ANALYSIS OF THE MICROBIOME
  4. METABOLIC PROFILING
  5. DEVELOPMENT OF THE MICROBIOME AFTER BIRTH AND THE INFLUENCE OF HOST GENETICS
  6. THE IMPACT OF EARLY FEEDING
  7. GUT PERMEABILITY IN THE NEWBORN
  8. EARLY MACRONUTRIENT EFFECTS ON THE HUMAN MICROBIOME
  9. FAT ABSORPTION AND BILE SALTS IN THE NEWBORN
  10. POPULATION DIFFERENCES IN THE MICROBIOME ACROSS THE GLOBE
  11. GUT MICROBIOTA AND THEIR POTENTIAL ROLE IN DEVELOPMENTAL PROGRAMING OF THE BRAIN
  12. DIETARY STANDARDIZATION
  13. NUTRIENT NICHES AND MICROBES THROUGHOUT THE LENGTH OF THE INTESTINE
  14. MUCOSAL BACTERIA
  15. MUCIN'S STRUCTURE AND INTERACTION WITH MUCOSAL BACTERIA
  16. GUT IMMUNOLOGY IN RELATION TO MUCIN AND MICROBIAL COLONIZATION
  17. DIETARY INTERVENTIONS: EFFECTS ON COLONIC MICROFLORA
  18. AUTISM – A GUT ROLE?
  19. ROLE OF THE MICROFLORA IN SYNTHESIZING ESSENTIAL AMINO ACIDS AND ALLOWING THE MAINTENANCE OF PROTEIN BALANCE ON VERY LOW INTAKES
  20. CHILDREN'S STUNTING: THE ROLE OF GASTROINTESTINAL MICROBIOTA
  21. THE ETHICS OF STUDYING DEPRIVED CHILDREN
  22. ENERGY BALANCE AND THE MICROBIOME
  23. PHYSICAL ACTIVITY AND THE MICROBIOME
  24. INSULIN SENSITIVITY AND THE METABOLIC SYNDROME
  25. PROBIOTICS AND THEIR EFFECTS
  26. CROHN'S DISEASE AND ULCERATIVE COLITIS
  27. REFERENCES

Per Brandtzaeg: Denise was referring to the IGA system in her talk and that is a most important immune system for the body. Eighty percent of all the plasma antibodies are generated in the gut mucosa and here is where the proximal or the distal small intestine's immunological response may be more important. I would like to emphasize that the immune system – the IGA system – per millimeter of gut length is almost the same in the proximal and distal gut. So what is stimulating this immunity? You talk, Gunnar, about the mucus layer being impermeable to bacteria. There must be gaps, though, because there are almost 30,000 small intestinal lymphoid follicles in the human gut and you must have access to those M cells in the lymphoid follicles to allow the bacterial stimulation of the IGA system.

Gunnar Hansson: When it comes to these lymphoid follicles in the mouse colon, we know that the bacterial flow is not a continuous system. I know, for example, that the mice have these fecal pellets – the mucus passes along with the fecal pellets, after which the colon is empty. But in humans with a colonic content of bacteria permanently there, the control of bacterial penetration must be a very different system.

Philip James: But you, Per, are still implying that you think that bacteria themselves have to get through as distinct from their metabolites to induce an immunological response.

Per Brandtzaeg: I think the bugs themselves must get access to the M-cells. . . . 

Gunnar Hansson: But that is in the small intestine.

Per Brandtzaeg: I think the follicle-associated epithelium in the distal small intestine and the small follicles in the large intestine are devoid of goblet cells. There is actually no mucus layer produced locally, topically, on the follicles.

Gunnar Hansson: On the follicles, I don't know because they are so small, but when it comes to Peyer's patches, we just recently looked and there is mucus on top of the Peyer's patches.

Per Brandtzaeg: But there are very few goblet cells compared to elsewhere in the intestine.

Gunnar Hansson: Yes. The mucus is thinner and probably very permeable, but there is at least some mucus on top of it.

Ingemar Ernberg: I am still interested in the design of the studies and what you said, of course, brings up the question, can you base any microbial characterization of human gut on biopsies after bowel cleansing? What has happened to the mucus after this procedure? In ulcerative colitis, you have a different quality of the mucus compared to a healthy colon, so would that affect bacterial sticking and bias what you find? In other words, should we not be using another technology to sample inside the gut?

Gunnar Hansson: I fully agree with you, because the experiment that I showed you was based on biopsies from cleaned patients, where we then put beads on top of the mucus to assess its permeability. However, we do the same procedures with control patients and then find their colonic mucosa impermeable. The biopsies easily lose their mucus, though, so we have to be very careful.

Dennis Kasper: We know so little about specific organisms and their impact, other than that they are remarkably different at the individual species level. The balance is between organisms involved primarily in pro-inflammatory-type-inducing immune responses and those inducing regulatory or anti-inflammatory immune responses. It is this balance which determines what the individual host will look like in their susceptibility to disease as they age. There is undoubtedly great redundancy in the immune system response to different organisms, but we know some of the effects of only four groups of bacteria: Bacteroides fragilis, Faecalibacterium prausnitzii, segmented filamentous bacteria, and the clostridia. A clostridia study involved 46 strains and induced colonic regulatory T-cells, but which organism was responsible, we do not know.38 So we have a long way to go before we have a reasonable understanding of the bacterial–immunological interplay.

DIETARY INTERVENTIONS: EFFECTS ON COLONIC MICROFLORA

  1. Top of page
  2. METHODOLOGY: A GENE SCANNING APPROACH
  3. DIFFERENT APPROACHES TO THE ANALYSIS OF THE MICROBIOME
  4. METABOLIC PROFILING
  5. DEVELOPMENT OF THE MICROBIOME AFTER BIRTH AND THE INFLUENCE OF HOST GENETICS
  6. THE IMPACT OF EARLY FEEDING
  7. GUT PERMEABILITY IN THE NEWBORN
  8. EARLY MACRONUTRIENT EFFECTS ON THE HUMAN MICROBIOME
  9. FAT ABSORPTION AND BILE SALTS IN THE NEWBORN
  10. POPULATION DIFFERENCES IN THE MICROBIOME ACROSS THE GLOBE
  11. GUT MICROBIOTA AND THEIR POTENTIAL ROLE IN DEVELOPMENTAL PROGRAMING OF THE BRAIN
  12. DIETARY STANDARDIZATION
  13. NUTRIENT NICHES AND MICROBES THROUGHOUT THE LENGTH OF THE INTESTINE
  14. MUCOSAL BACTERIA
  15. MUCIN'S STRUCTURE AND INTERACTION WITH MUCOSAL BACTERIA
  16. GUT IMMUNOLOGY IN RELATION TO MUCIN AND MICROBIAL COLONIZATION
  17. DIETARY INTERVENTIONS: EFFECTS ON COLONIC MICROFLORA
  18. AUTISM – A GUT ROLE?
  19. ROLE OF THE MICROFLORA IN SYNTHESIZING ESSENTIAL AMINO ACIDS AND ALLOWING THE MAINTENANCE OF PROTEIN BALANCE ON VERY LOW INTAKES
  20. CHILDREN'S STUNTING: THE ROLE OF GASTROINTESTINAL MICROBIOTA
  21. THE ETHICS OF STUDYING DEPRIVED CHILDREN
  22. ENERGY BALANCE AND THE MICROBIOME
  23. PHYSICAL ACTIVITY AND THE MICROBIOME
  24. INSULIN SENSITIVITY AND THE METABOLIC SYNDROME
  25. PROBIOTICS AND THEIR EFFECTS
  26. CROHN'S DISEASE AND ULCERATIVE COLITIS
  27. REFERENCES

Harry Flint: Dusko Ehrlich's challenge must be considered. We have fed a group of 14 obese human subjects with the metabolic syndrome – so not a normal group. Diets were fed in a crossover design testing the effects of either a wheat bran fiber-enriched diet, or a similar well-balanced diet but one enriched with resistant starch. The design is complicated by the need to change multiple dietary ingredients of the diet when one changes a sustained high-protein diet from a moderate- to a low-carbohydrate intake. We have found evidence of non-digestible protein inflow into the large intestine, because of the fecal branched fatty acids. Selective bacterial changes also occur with a low-carbohydrate, high-protein intake. Most of the changes seem to reflect the colonic inflow of carbohydrate, but a high-protein intake also seems to have selective effects. Dietary fat may be toxic to some microbial groups, but this is still not established. In this study, we added a specific commercially produced and defined resistant starch as an additive which went into lots of different food products – biscuits, cereals, and so forth. Resistant starch is that fraction of the starch generated by retrogradation when you heat and then cool a starch. This resistant starch does not get digested fully in the small intestine and it passes through to the colon where it then becomes a substrate for fermentation by colonic bacteria.

We have a lot of fecal samples from this study, and we looked at a subset of six out of the 14 individuals, with one fecal sample during each diet phase being analyzed by Sanger sequencing of the ribosomal RNA genes. This was done with Alan Walker at the Wellcome Trust Sanger Institute in Cambridge. We did a classic analysis, but out of interest, we plotted the frequency of the bacterial phylotypes or species that we picked up. The point I want to make here is that many people have emphasized the unculturability of human colonic bacteria. In agreement with that, we found that some two-thirds of the phylotypes that we picked up didn't correspond to anything that was cultured or in a collection. But if you look at the more abundant ones, those that occurred above, say, 1% of the total, the culturability is much, much higher. And these organisms have a particularly important role in the system because they are numerically dominant, so the top eight species account for about a third of all the bacteria in these people's feces. So it is lucky that we can culture these. Other organisms are probably not intrinsically unculturable, it is just that they are rather rarer than the others. If we had time and money to do a lot more cultivation, we would probably get most of them, and they may indeed have been isolated in the past by people like Sydney Finegold and the Moores – and then, regrettably, lost.

So the point of the study was to look at the response to diet. We got a significant response, on average, for two of these species: Eubacterium rectale doubled and Ruminococcus bromii went up some tenfold. Then, we used quantitative PCR to ascertain ribosomal gene sequences. As the diet changed to that containing resistant starch, we tracked the changes in each individual. Subjects showed an increase in the numbers of the clostridium leptum group, ruminococci, varying from a few percent to something like a quarter of the total bacterial RNA. Then, this reversed when they were on a low-resistant-starch intake. The kinetics were roughly what you would expect: it takes 2 or 3 days for the transition to occur, so this corresponds to an average gut transit time of about 60 h. The reversal has about the same time span. So we certainly modify our gut microbiota by what we eat, and these are fairly dramatic changes. Furthermore, not everybody responds in the same way. Two of our subjects did not have this group of bacteria and did not, therefore, respond with an increase. When we then analyzed the amount of residual resistant starch in the feces of the other 12 subjects, virtually all of the resistant starch had disappeared and had been completely fermented. For the two subjects with no ruminococci, around 50% of their resistant starch remained intact in the feces. So these subjects were not receiving so much short-chain fatty acids from that part of the diet. So on this basis, we suggest that the ruminococci could be a key species playing a crucial role in resistant starch fermentation.

If one fractionates stool samples into particulate fiber and liquid and then does a 16S library analysis, the ruminococci group stands out as being fiber-associated, so we conclude that they probably play a key role in the breakdown of particulate substrates, including particulate starch.

The final point which I think is important is the role of the gut environment in the selection of the microbial community. In an old, simple, single-stage fermenter study that we did years ago, we were feeding the fermenter with a mixture of carbohydrates as polysaccharides and started it off at pH 5.5 before inoculating the fermenter with a mixed fecal bacteria. We chose a pH of 5.5 as that measured in the proximal colon under conditions of active fermentation. A pH of 6.5 is more typical of the distal regions of the gut. We found that the lower the pH, the more dominant the production of butyrates with declining acetate production and rather little propionate. Change the pH and the products switch rapidly, and we now know this reflects a huge shift in the microbial community. The lower the pH, the greater proportion of butyrate-producing bacteria, which can represent half of the bacterial community, whereas at pH 6.5 they have virtually disappeared because they can no longer compete with the bacteroides. We then have more propionate production as the main product of these bacteria.

So whenever you change gut turnover by, for example, changing the fiber content of the diet, you will be impacting on the pH in the proximal colon and thereby affecting the competition between these major bacterial groups.

So, in conclusion, gut microbiota from the human may be under-cultured rather than unculturable. Diet composition does have an enormous impact on the colonic microbiota. You do not see this relationship easily in studies where you are recording dietary habits and plotting the response on a day-to-day basis because there are normal time delays in response and the diet is usually changing daily. Why, also, do we see this extraordinary inter-individual variation in microbes? David Relman has suggested some possible explanations and that is another interesting challenge.

Michiel Kleerebezem: One would expect a greater impact of fat in the small intestine, and this we see in ileostomy patients. Sampling only 5–6 h after a meal can show the microbial effect. With simple sugar-rich diets, the streptococci increase enormously from very low levels. The subjects that we look at have had their ileostomy already for at least 5 years. After the surgical procedure, one observes a loss of strict anaerobes with the emergence of the facultative aerobic microbes, but after a period of 5 years, the anaerobes are back.

Willem de Vos: Thanks, Harry. I also had the impression at first that diet did not do much in our early experiments and that the microbiota were stable, but with a few samples, we undertook unsupervised clustering with chip analysis and found in all our subjects, twice as many as yours, that there is a strong effect only from the resistant starch. I agree with you, it is especially Clostridium clostridioforme which is responding more than tenfold. So there is an effect. It depends, I think, a little bit on how you do the experiments, but you describe a very good trial and we should be doing studies like this and use the better technologies that Rob Knight and others have developed to learn more about the selective effects of diet.

Dusko Ehrlich: Thanks, Harry, for rising to the challenge and giving the data. Maybe I can try to offer a couple of thoughts. You clearly showed that you can induce perturbation of microbiota by probably fairly major interventions with a reversal in colonic numbers when you reverse the treatment. But the issue is whether this is important, because both you and Wilhelm both see only limited effects on a small proportion of the microbes. So if you consider that we may have 1,000 species in our colon, but you can influence the numbers of only ten of them, then is this a major change? We have observed similar changes on nutritional intervention, with significant increases in a limited number of meta-species followed by a reversal once the intervention stops. I am still undecided how important this is.

Harry Flint: I think an increase in a microbial group to become 25–30% of the total microflora is a major change. I would accept that these are fairly major dietary interventions because in our initial studies we were wanting to see an effect. So given that we do not all eat the same thing every day, it is inevitable that we have constant changes in our gut microbiota composition.

Arnold Berstad: Whether these changes are important or not depends, in part, on whether one observed symptoms, e.g., of flatulence and pain, in those subjects who did not ferment all their starch.

Harry Flint: I am not sure about gut symptoms, but remember these subjects also had the metabolic syndrome, so I would expect the resistant starch to improve their insulin resistance – we are still assessing the metabolic effects of these experiments.

Fredrik Bäckhed: I think Harry's study is marvelous in looking at the effect of some forms of fiber or similar molecules. If you use fat, for example, I would not expect any changes in the colonic microbiota, because the fats are all absorbed in the small intestine. However, we may well see effects on the small intestinal microbiota, but these are not easy to detect. My second point is about the unculturable issue – a paper from Andrew Goodman39 in Jeff Gordon's laboratory set out how they did this in quite a nice way: I think they cultured 70% of the microbiota – it is all about sample handling, how fast you get it into anaerobic chambers and so forth. So I think that Harry Flint is correct that we can probably culture everything if we do it correctly and have optimal conditions.

Ingemar Ernberg: I think you introduced another interesting problem with your elegant, albeit small, human study and that relates to the precise control of the diet in these subjects. How long do we need to study people on a constant diet to obtain baseline or intervention effects? Then, how many subjects should we use to assess these issues and is it important to do a crossover design? Should one start with the individual's natural diet, or should one first standardize the diet and then change it as part of the test? The other intriguing result is that you seemed to find changes within 3 days, so does this mean there is no need for a 3-week-long intervention with all the trouble and costs which that signifies?

Harry Flint: Any human study is a compromise – if you want very tight dietary control you sometimes have to sacrifice numbers.

Irv Rosenberg: You have induced a significant substrate change to the bacterial flora and then you show some change in their composition and substrate metabolism. What do we know about the effects of other nutritional factors in terms of essential nutritional factors for selective growth of particular bacteria? We know, for instance, in the malaria or tuberculosis story, that one can feed certain pathologic bacteria or parasites with iron in such a way as to change the dynamics of their interaction with the host organism. So I think we need some other studies that challenge the flora with varying amounts of their own nutritional needs for micronutrients, for certain kinds of protein and so forth. If one knew this from in vitro studies, would not this then affect some of the decisions about what should be a standard diet on which to study the flora.

Harry Flint: Iron is an excellent example of what you are talking about because it is a growth requirement for particular pathogens, so undoubtedly you have to think about iron levels. There will also be natural antimicrobials. For example, certain plant phenolics are antimicrobials but we know very little about their selective effects, so there are all sorts of dietary factors that we should be thinking about.

Richard Black: Are we getting lost in the forest of detail about specific microbes when we now can measure them superbly with several of our new techniques? Perhaps we are looking too closely at the microbes and should concentrate on the metabolic effects first and then work backwards?

AUTISM – A GUT ROLE?

  1. Top of page
  2. METHODOLOGY: A GENE SCANNING APPROACH
  3. DIFFERENT APPROACHES TO THE ANALYSIS OF THE MICROBIOME
  4. METABOLIC PROFILING
  5. DEVELOPMENT OF THE MICROBIOME AFTER BIRTH AND THE INFLUENCE OF HOST GENETICS
  6. THE IMPACT OF EARLY FEEDING
  7. GUT PERMEABILITY IN THE NEWBORN
  8. EARLY MACRONUTRIENT EFFECTS ON THE HUMAN MICROBIOME
  9. FAT ABSORPTION AND BILE SALTS IN THE NEWBORN
  10. POPULATION DIFFERENCES IN THE MICROBIOME ACROSS THE GLOBE
  11. GUT MICROBIOTA AND THEIR POTENTIAL ROLE IN DEVELOPMENTAL PROGRAMING OF THE BRAIN
  12. DIETARY STANDARDIZATION
  13. NUTRIENT NICHES AND MICROBES THROUGHOUT THE LENGTH OF THE INTESTINE
  14. MUCOSAL BACTERIA
  15. MUCIN'S STRUCTURE AND INTERACTION WITH MUCOSAL BACTERIA
  16. GUT IMMUNOLOGY IN RELATION TO MUCIN AND MICROBIAL COLONIZATION
  17. DIETARY INTERVENTIONS: EFFECTS ON COLONIC MICROFLORA
  18. AUTISM – A GUT ROLE?
  19. ROLE OF THE MICROFLORA IN SYNTHESIZING ESSENTIAL AMINO ACIDS AND ALLOWING THE MAINTENANCE OF PROTEIN BALANCE ON VERY LOW INTAKES
  20. CHILDREN'S STUNTING: THE ROLE OF GASTROINTESTINAL MICROBIOTA
  21. THE ETHICS OF STUDYING DEPRIVED CHILDREN
  22. ENERGY BALANCE AND THE MICROBIOME
  23. PHYSICAL ACTIVITY AND THE MICROBIOME
  24. INSULIN SENSITIVITY AND THE METABOLIC SYNDROME
  25. PROBIOTICS AND THEIR EFFECTS
  26. CROHN'S DISEASE AND ULCERATIVE COLITIS
  27. REFERENCES

Karen Madsen: Could we have comments on the potential role of early bacterial exposure in the development of autism, where early events as well as hereditary factors are becoming clearer?40,41

Tore Midtvedt: The relationship between the gut and autism is fascinating, with very nice Canadian work by MacFabe suggesting that the short-chain fatty acids, e.g., propionate, induce changes in the brain membranes and autistic-like behavior in animals.42,43

Natalie Delzenne: I would like to link potential bacterial/inflammatory factors with mood changes. Our clinicians had alcoholic patients subjected to alcohol withdrawal. The physicians analyzed the patients’ depression. On alcohol withdrawal, they showed an improvement in biofunction with a decrease in lipopolysaccharide [LPS], which correlated with their improvement in their emotional state. So, could the LPS play a key role in this behavioral change?

Sven Pettersson: If you get sunburned after a rich meal, which will put up the LPS, the sunburn has triggered a lot of pro-inflammatory release of, for example, high-mobility group box-1 (HMGB1), that can be sufficient to transiently open the brain barrier in a particular time when you don't want this to happen. That is not uncommon for patients with systemic lupus. So LPS would be one molecule, but there are bound to be a myriad of similar components.

ROLE OF THE MICROFLORA IN SYNTHESIZING ESSENTIAL AMINO ACIDS AND ALLOWING THE MAINTENANCE OF PROTEIN BALANCE ON VERY LOW INTAKES

  1. Top of page
  2. METHODOLOGY: A GENE SCANNING APPROACH
  3. DIFFERENT APPROACHES TO THE ANALYSIS OF THE MICROBIOME
  4. METABOLIC PROFILING
  5. DEVELOPMENT OF THE MICROBIOME AFTER BIRTH AND THE INFLUENCE OF HOST GENETICS
  6. THE IMPACT OF EARLY FEEDING
  7. GUT PERMEABILITY IN THE NEWBORN
  8. EARLY MACRONUTRIENT EFFECTS ON THE HUMAN MICROBIOME
  9. FAT ABSORPTION AND BILE SALTS IN THE NEWBORN
  10. POPULATION DIFFERENCES IN THE MICROBIOME ACROSS THE GLOBE
  11. GUT MICROBIOTA AND THEIR POTENTIAL ROLE IN DEVELOPMENTAL PROGRAMING OF THE BRAIN
  12. DIETARY STANDARDIZATION
  13. NUTRIENT NICHES AND MICROBES THROUGHOUT THE LENGTH OF THE INTESTINE
  14. MUCOSAL BACTERIA
  15. MUCIN'S STRUCTURE AND INTERACTION WITH MUCOSAL BACTERIA
  16. GUT IMMUNOLOGY IN RELATION TO MUCIN AND MICROBIAL COLONIZATION
  17. DIETARY INTERVENTIONS: EFFECTS ON COLONIC MICROFLORA
  18. AUTISM – A GUT ROLE?
  19. ROLE OF THE MICROFLORA IN SYNTHESIZING ESSENTIAL AMINO ACIDS AND ALLOWING THE MAINTENANCE OF PROTEIN BALANCE ON VERY LOW INTAKES
  20. CHILDREN'S STUNTING: THE ROLE OF GASTROINTESTINAL MICROBIOTA
  21. THE ETHICS OF STUDYING DEPRIVED CHILDREN
  22. ENERGY BALANCE AND THE MICROBIOME
  23. PHYSICAL ACTIVITY AND THE MICROBIOME
  24. INSULIN SENSITIVITY AND THE METABOLIC SYNDROME
  25. PROBIOTICS AND THEIR EFFECTS
  26. CROHN'S DISEASE AND ULCERATIVE COLITIS
  27. REFERENCES

Alan Jackson: Much has been made about the role of maldigested, malabsorbed carbohydrates passing to the microbiome through the ileo-cecal valve. I want to look at nitrogen and protein metabolism, where there is metabolic integration between the host and the microbiome. In this context, the small intestine and the large intestine function essentially as two different organs with very substantial fluxes of nitrogen through each, but with the two fluxes as separate entities. There is a flow down the small intestine of dietary nitrogen, e.g., in the form of protein, and its amino acids together with the secretion of digestive enzymes and the exfoliation of intestinal cells. However, by the time you get down to the ileo-cecal valve, this dietary and secreted/exfoliated protein has been essentially reabsorbed. When one gets to the colon, the overall net balance, i.e., nitrogen entering from the small intestine and that excreted in feces, appears to be very small, but if you calculate the flux of nitrogen-related compounds through the colon, there is a large flux of nitrogen from urea through the large bowel, with this flux being greater than the nitrogen flux found in the small gut. We have known for many years that only a proportion of the urea that is produced in the liver is excreted through the kidney under normal circumstances and 25–35% passes into the colon, where the urea nitrogen is potentially available as a source for microbial synthesis and metabolism. The question then arises as to its nutritional significance.

The rate of urea nitrogen excretion is often assumed to be the rate at which it is formed in the liver. If you take individuals on low-protein diets, the sort of diets that are habitually consumed by people on vegetarian diets, then the relative loss of urea through the kidney or into the colon is reversed from what we consider normal, because two-thirds of the liver's production of urea is salvaged through the colon, with only one-third being excreted through the kidney. The whole relationship between what is happening in the colon and the kidney is mediated through a series of specific urea transporters which are expressed in both the collecting duct of the kidney and in the colon.44 These transporters are under the influence of arginine vasopressin45 and, therefore, they are differentially responsive to dietary protein and water deprivation46 and enable a coordinated reabsorption of urea in the kidney and secretion of that urea into the colon. So, the rate of urea production is much less variable in humans across the usual range of nitrogen intakes than has been presumed with the relative routes of transfer through the kidney and into the colon explaining this ability to conserve nitrogen when on low-nitrogen intakes. Thus, the nitrogen that is being presented to the microflora in the colon through this mechanism changes based upon the individual's dietary intake of protein and, to a certain extent, their water intake.

Breastfed infants consume a diet which is very low in protein and which contains a significant amount of extra non-protein nitrogen, mostly in the form of urea. In the newborn there is a very active secretion of urea into the colon so that the colonic cycle dominates the fate and the whole handling of the end products of amino acid oxidation. This is important in understanding the requirements for dietary essential amino acids, particularly those limited essential amino acids in the diet, such as lysine. In 1995, international advice based upon balance, i.e., experiments where you feed diets of varying amino acid composition and look at the level that enables balance to be achieved, set the requirement for lysine in the region of 12 mg/kg/day.47 By 2000 to 2007, we had the ability to measure the rate at which lysine is oxidized in the body – this work being very much dependent upon the work of Vernon Young.48,49 It was found that the rate of utilization of lysine is substantially greater, something like 30 mg/kg/day. There is a substantial debate as to whether or not people need to eat much more high-quality protein in order to meet this requirement.

Now there have been some studies, albeit difficult to do, which have looked at the extent to which the urea nitrogen provides the substrate for microbes to make amino acids available for use by the host. When one assesses the amount of lysine that is potentially made available from the colonic microflora through this salvage process, then you end up with numbers in the region of 20–25–30 mg/kg/day.50,51 So, if you look at the amount that is derived from salvage plus the amount that needs to be taken in the diet, then you more than match the demands set by the oxidation of lysine in the body.

So I consider the small and large intestine as functionally very different – the small intestine is very much a supply-driven organ in terms of what is coming from the diet, whereas the large intestine is really a demand-responsive organ, both in terms of the demands of the host and of the bacteria. We have results that suggest that in obesity these relationships are changed, and certainly in pregnancy52 what happens to the pregnant woman between 18 and 28 weeks, in terms of this system, has an important effect upon the size of the baby at birth. Studies on women on low protein intakes in pregnancy highlight the importance of intestinal recycling of urea.

Irv Rosenberg: Alan, you used the words “demand-driven” for the colonic reaction as if to say that where lysine was limiting, the colonic microflora would respond to that and produce more lysine?

Alan Jackson: A change in the level of protein-nitrogen in the diet induces an adaptive regulation of urea transporters in the kidney and in the colon. That enables more nitrogen to be available to the colonic microflora. The question then is what do they do with the extra nitrogen? The presumption is that they use it for their own resources including the requirements for amino acids such as lysine. I used the specific example of lysine to demonstrate that lysine recovery could only have been derived from the bacteria, because the host, e.g., humans or rodents, cannot make lysine. We also see that bacterially derived lysine is used by the host and, therefore, has become available. In that situation, there is a demand by the microflora to use the available nitrogen for their own purposes. How bacteria make available lysine to the host, rather than simply retaining it for their own synthesis, I do not know.

Ruth Ley: Is there any evidence for nitrogen fixation in any populations with low-protein diets?

Alan Jackson: This was an issue 30 to 40 years ago. There have been studies where people have been locked into stabilized, topically enriched, nitrogen environments to see whether there is any evidence of nitrogen fixation without any functionally significant amounts being evident.

Tore Midtvedt: We have been considering urea metabolism, too, and the species involved is important. In man, 20 to 25% of urea passes into routes other than urine, e.g., tears, saliva, and into the stomach where quite a lot of urea is secreted. The first bacterium to meet the urea in the intestinal tract is helicobacter, which is present in 30 to 80% of the world's population, so I think one has bacterial synthesis of amino acids starting as high up as the stomach.

Alan Jackson: Our analyses suggest it would be difficult to achieve sufficient lysine by this method compared with the quantitative value of colonic events but I am not excluding other factors. When we give nitrogen-labeled urea orally, and it is hydrolyzed in the upper gastrointestinal tract, that label does not appear to go to amino acids in significant or substantial quantities. In terms of mammalian metabolism, we know there are only two ways in which substantial amounts of ammonia can be fixed: one is through glutamate dehydrogenase in the formation of glutamate, and the other is in the fixation of nitrogen in the reaction between serine and glycine. There are no other identified mammalian pathways.

Philip James: So you have got urea, for example, coming in through saliva and the stomach and that can, of course, provide the nitrogen substrate for the helicobacter or other bacteria in the small intestine where these bacteria can also be digested with subsequent bacterially derived lysine, but can lysine be absorbed from the colon?

Alan Jackson: Yes. This has been shown to occur.53 How the whole reabsorption of these essential amino acids is regulated, I am unsure, but we know that colonic urea nitrogen can end up in lysine and be used by the host in significant amounts.

Harry Flint: Your interesting analyses suggest that the output of urea was directly translating into bacterial protein, but I think the actual production of bacterial protein is limited largely by the carbon source, the energy source, so it is overall dietary balance that is going to dictate that. In some of the human studies I was talking about earlier, where we did weight loss studies that increased the protein but reduced the carbohydrate, we found that the ammonia level was relatively insensitive to those diets but the level of bacterial protein, as measured by total bacterial counts, related to the fermentable carbohydrate in the diet. So I am just suggesting that maybe the carbohydrate content is also relevant.

Alan Jackson: There is no doubt that the energy available can be limiting, and then there is the question relating to the energy saving by volatile fatty acid generation in relation to obesity, where we tried to calculate the energy balance over the large bowel in different clinical states and were unimpressed by the magnitude of that saving. If you have enough energy to drive bacterial growth, then the bacteria will need a source of nitrogen.

Cutberto Garza: In mammalian human milk, if one looks at the relative abundance of various amino acids, the one that is in least abundance is glycine, relative to need. If we look at the glutamate transferase, the nitrogen-salvaging enzyme, then interestingly enough, glycine is a negative feedback inhibitor for that transaminase. So if you wanted to conserve nitrogen, at least in the infant with such a very low supply in human milk, then one of the most effective mechanisms for doing that is to keep glycine levels low, as the urea is then being metabolized by bacteria and the ammonia is then absorbed, and on entering the infant's liver can attach to either glutamate or, possibly, to one of the shorter-chain carbon units. You do not in these circumstances have to involve the bacteria in synthesizing amino acids.

Alan Jackson: Except that the stoichiometry of what is happening is very difficult to achieve for total N. Now, if you feed babies on cows' milk, you give them a very large protein load which then presents the kidney with a large osmotic load and this requires a large amount of water to excrete. If the child had to excrete that amount of urea, it would dehydrate itself immediately. It is only by taking the urea and passing it into the colon that you can actually achieve reasonable water homeostasis.

Willem de Vos: In the camel, 97% of all urea is excreted into the upper gut and then utilized by microbial ureases in their rumen; very little urea has to be excreted in the kidneys, so they have next to no urine. That is the key point in explaining how a camel can live under those conditions.

Alan Jackson: Foregut fermenters obviously can use urea in a completely different way. The camel is a desert animal and has to conserve water as much as it has to conserve nitrogen as body protein. These desert animals have kidneys designed to generate substantial concentration gradients and a key element in their ability to retain water is the ability to retain urea. That is why the arginine vasopressin, the antidiuretic which regulates the urea metabolism, is also the same hormone that regulates the water metabolism. The water and the urea move together in response to similar signals and that is why the relationship between water homeostasis and urea metabolism and overall protein and nitrogen metabolism is critical.

CHILDREN'S STUNTING: THE ROLE OF GASTROINTESTINAL MICROBIOTA

  1. Top of page
  2. METHODOLOGY: A GENE SCANNING APPROACH
  3. DIFFERENT APPROACHES TO THE ANALYSIS OF THE MICROBIOME
  4. METABOLIC PROFILING
  5. DEVELOPMENT OF THE MICROBIOME AFTER BIRTH AND THE INFLUENCE OF HOST GENETICS
  6. THE IMPACT OF EARLY FEEDING
  7. GUT PERMEABILITY IN THE NEWBORN
  8. EARLY MACRONUTRIENT EFFECTS ON THE HUMAN MICROBIOME
  9. FAT ABSORPTION AND BILE SALTS IN THE NEWBORN
  10. POPULATION DIFFERENCES IN THE MICROBIOME ACROSS THE GLOBE
  11. GUT MICROBIOTA AND THEIR POTENTIAL ROLE IN DEVELOPMENTAL PROGRAMING OF THE BRAIN
  12. DIETARY STANDARDIZATION
  13. NUTRIENT NICHES AND MICROBES THROUGHOUT THE LENGTH OF THE INTESTINE
  14. MUCOSAL BACTERIA
  15. MUCIN'S STRUCTURE AND INTERACTION WITH MUCOSAL BACTERIA
  16. GUT IMMUNOLOGY IN RELATION TO MUCIN AND MICROBIAL COLONIZATION
  17. DIETARY INTERVENTIONS: EFFECTS ON COLONIC MICROFLORA
  18. AUTISM – A GUT ROLE?
  19. ROLE OF THE MICROFLORA IN SYNTHESIZING ESSENTIAL AMINO ACIDS AND ALLOWING THE MAINTENANCE OF PROTEIN BALANCE ON VERY LOW INTAKES
  20. CHILDREN'S STUNTING: THE ROLE OF GASTROINTESTINAL MICROBIOTA
  21. THE ETHICS OF STUDYING DEPRIVED CHILDREN
  22. ENERGY BALANCE AND THE MICROBIOME
  23. PHYSICAL ACTIVITY AND THE MICROBIOME
  24. INSULIN SENSITIVITY AND THE METABOLIC SYNDROME
  25. PROBIOTICS AND THEIR EFFECTS
  26. CROHN'S DISEASE AND ULCERATIVE COLITIS
  27. REFERENCES

Philip James: One of the great challenges for governments globally is the need to halve the number of children with malnutrition as part of the Millennium Development Goals [MDGs] by 2015. These children, although classified as malnourished, are actually predominantly short for their age and traditionally we have been taught that this has something to do with dietary deficiencies but also to recurrent infections. So the question is: do we have any data on the microbiome and the nature of the gut in these stunted children? This is a huge public health issue of enormous consequence. Southeast Asia has a high prevalence of stunting, but so do some African countries.

Rob Knight: We are currently studying children with either kwashiorkor or marasmus in the Malawi and Bangladeshi populations. We are looking at children's weight-for-age [WA] Z-scores and weight-for-height [WH] Z-scores. We also have height for age as well; that is not currently one of the things that we have been trying to correlate with the microbiome, but perhaps we should be.

Cutberto Garza: Height for age is the critical issue in relation to this problem of stunting. It is unlikely that any microbiome relating to stunting is dependent exclusively on geography because we found in a major international study of children's growth that children in Oman, Ghana, the United States, Norway, Brazil, and India had no differences in stature for the first 5 years across those six populations.54 However, what we did was to control very carefully for environmental influences, and so it would seem that if any population has a disproportionate number of short people, it is primarily environmental and not genetic. Obviously, an individual's growth is affected by genetics, but stature is a complex quantitative trait affected by very many genes, and population differences are environmentally related, for the most part, and could reflect the nature of the microbiota. But the stunting is not because they are Indian, Bangladeshi, or African.

Philip James: These populations of yours were, however, carefully selected in terms of having a normal birth weight with fully breastfeeding mothers and careful weaning practices, is that not right? What other environmental parameters did you control for?

Cutberto Garza: In India, for example, we undertook surveys to document the socioeconomic characteristics of populations where growth was not constrained. We had to canvas something like 40,000 households in South Delhi because we wanted to make sure that the study would be community based and not hospital based. We found in India that if both parents had 17 years of education, then we could not see any constraint on growth in the height of the children. The same thing was true in Oman for a different set of variables, and in Ghana for yet a third set of variables. So we did some preliminary surveys in each of these populations, determined what those socioeconomic characteristics were, and then, when we undertook our study on a population basis, we selected individuals that fit these profiles. This study was to develop new international standards and it was very clear that we could not simply reproduce the circumstances of the countries in general, where children, on average, grew at very different rates.

Alan Jackson: What that work also made clear is that although previously we had attributed a substantial amount of stunting by the age of 2 years to postnatal factors, maybe as much as half of it is prenatal, rather than postnatal. This is because, when you look at children who end up being short, a significant proportion are small and short at birth and, therefore, cluster at the lower end of the height distribution. So, there is a factor related to postnatal growth, but there is an interaction with prenatal growth and that question about prenatal growth is important in terms of the problems of preterm infants. Babies that are born short, whether they are preterm or term, tend to have disproportions in other parts of their body composition, and they tend to be relatively adipose and relatively low in lean tissues. Therefore, their ability to partition nutrients and metabolic behavior is different and is affected by prenatal events. Presumably, in real life, a significant proportion of the infants’ shortness is due to the effects of inflammation at the growth plate and the inability to utilize nutrients effectively as much as it is due to a limitation of nutrients of itself. Therefore, the interaction of systemic inflammation and nutrient availability is a combination that appears to be of particular importance in terms of failure of linear growth.

Ingemar Ernberg: This pre-delivery effect may also reflect the impact of epigenetic changes.

Tore Midtvedt: There is a huge project underway backed by the Bill & Melinda Gates Foundation covering eight places in the world – two in South America, two in Africa, and four in Asia; a cohort of at least 200 children in each place will be followed weekly for at least 2 to 3 years, fecal samples will be saved, and all the pathogens will be isolated. When these data are available, then we can start to answer some of these questions.

Rob Knight: Tore, you are referring to the MAL-ED Network, where we hope to use the new molecular techniques we have been discussing to monitor the microbiome.

Cutberto Garza: There is another study being done by Oxford University in conjunction with the Gates Foundation. They are following something like 200 children in eight or ten different sites around the world, but beginning to follow them at about 3 months of gestation, characterizing the mother's pregnancy in very, very careful terms, documenting the growth of the infant in utero using sonograms, and following them into postnatal life looking at both diet and growth. Not only are they documenting aspects of the infant's well-being, they are also collecting a number of biological samples in order to assess epigenetic effects.

David Relman: The March of Dimes Foundation has set up a new Prematurity Research Center at Stanford and we are collaborating with a number of sites around the United States to follow the human microbiome in mothers, beginning at the earliest points in pregnancy. In fact, we are trying to recruit women who are planning to become pregnant so we can start preconceptually to monitor multiple body habitats – fecal, vaginal, oral, and skin – with the idea of trying to understand how the trajectory of microbiome development during pregnancy might differ between those who go on to a premature outcome versus those that go to term. Clearly, on the basis of what we have discussed, we should also be looking at postnatal influences.

Alan Jackson: The epigenetic investigators have increasingly recognized the importance of the placenta and placentation – the whole question of the inflammatory state around placentation and the risk that bacterial microbiome-associated inflammation plays in that relationship. So the shift of concern as to what the microbiome looks like in these different sites is not of trivial interest; it might be actually of quite important fundamental interest to the establishment of pregnancy, the capability of pregnancy, the competence of the placenta, and how that then translates into the delivery of nutrients to enable fetal growth and establish the epigenetic profile.

Philip James: The early phase of establishing pregnancy is so critical in determining the whole trajectory of development.

THE ETHICS OF STUDYING DEPRIVED CHILDREN

  1. Top of page
  2. METHODOLOGY: A GENE SCANNING APPROACH
  3. DIFFERENT APPROACHES TO THE ANALYSIS OF THE MICROBIOME
  4. METABOLIC PROFILING
  5. DEVELOPMENT OF THE MICROBIOME AFTER BIRTH AND THE INFLUENCE OF HOST GENETICS
  6. THE IMPACT OF EARLY FEEDING
  7. GUT PERMEABILITY IN THE NEWBORN
  8. EARLY MACRONUTRIENT EFFECTS ON THE HUMAN MICROBIOME
  9. FAT ABSORPTION AND BILE SALTS IN THE NEWBORN
  10. POPULATION DIFFERENCES IN THE MICROBIOME ACROSS THE GLOBE
  11. GUT MICROBIOTA AND THEIR POTENTIAL ROLE IN DEVELOPMENTAL PROGRAMING OF THE BRAIN
  12. DIETARY STANDARDIZATION
  13. NUTRIENT NICHES AND MICROBES THROUGHOUT THE LENGTH OF THE INTESTINE
  14. MUCOSAL BACTERIA
  15. MUCIN'S STRUCTURE AND INTERACTION WITH MUCOSAL BACTERIA
  16. GUT IMMUNOLOGY IN RELATION TO MUCIN AND MICROBIAL COLONIZATION
  17. DIETARY INTERVENTIONS: EFFECTS ON COLONIC MICROFLORA
  18. AUTISM – A GUT ROLE?
  19. ROLE OF THE MICROFLORA IN SYNTHESIZING ESSENTIAL AMINO ACIDS AND ALLOWING THE MAINTENANCE OF PROTEIN BALANCE ON VERY LOW INTAKES
  20. CHILDREN'S STUNTING: THE ROLE OF GASTROINTESTINAL MICROBIOTA
  21. THE ETHICS OF STUDYING DEPRIVED CHILDREN
  22. ENERGY BALANCE AND THE MICROBIOME
  23. PHYSICAL ACTIVITY AND THE MICROBIOME
  24. INSULIN SENSITIVITY AND THE METABOLIC SYNDROME
  25. PROBIOTICS AND THEIR EFFECTS
  26. CROHN'S DISEASE AND ULCERATIVE COLITIS
  27. REFERENCES

Fergus Shanahan: I am just wondering whether there is an ethical dilemma in studying, or taking the opportunity to indulge our intellectual curiosity about, the microbiome in a group of people whose only misfortune is actually their socioeconomic or political status relating to being underprivileged, lacking in education, or short of food. We are studying for good reasons, but we are studying something without even trying an intervention of potential benefit. I am not entirely sure that there isn't an ethical dilemma there.

Cutberto Garza: This ethical issue was one of the reasons why we conducted our growth studies the way we did. We worked with very privileged groups to document what the potential was for society, thereby raising the ethical public health bar and being now able to highlight just how many children in a society are stunted when governments have considered their children just naturally short. We, therefore, decided to study a wide range of countries with exactly the same methodology so there was no question of bias. We wanted, in terms of Fergus’ challenge, to be able to say: “Look! All children have the potential to grow the same.” By doing this, we were able to highlight the need for political action to remedy a huge global problem.

Rob Knight: The interventions do not work equally well in all populations and there are some children who do not respond to interventions: even monozygotic twins can be discordant for their response to a given supplement. The question is: Why? It is possible that there is a pathogen that the non-responding twin has picked up or their microbiota are deficient in some respects or that it is just a mismatch between that specific supplement and what that person can metabolize. So with the much cheaper DNA sequencing, I think we are heading towards finding future interventions for some of the populations that are currently not responding to the interventions that are available now.

Alan Jackson: I think Fergus’ ethical question is important. We know we can repair a child's wasted condition, but we cannot yet make them taller in utero. Therefore, there is an issue about what specifically needs to be done in that context, in both a treatment and a preventive mode. It is absolutely true that there is a social dimension that is not being adequately addressed. But there is also a biological dimension where we are woefully inadequate because we simply do not have a clear enough understanding of how to manage the biology of that problem. And that is a problem that leads to children being short and fat, which when translated into adulthood is the whole question of the double-burden of malnutrition, diabetes, obesity, premature death from heart disease, and so on.

Cutberto Garza: I would agree with Alan. We do not fully understand why so many of the world's children are stunted, with this being evident in the vast majority by the end of the third year of life. There is some argument as to whether or not you can get even substantial catch-up after 3 years. There may be an early biological window of opportunity only. I must emphasize the importance of speaking about height for age being deficient and not misreading the finding of a low weight for age. So, if you enforce weight-for-age standards on a stunted population, you are going to get an overweight kid. I do not think we have any data to show that catch-up in stature is related to any bad metabolic outcome. We know that a disproportionate weight gain for a child's height may result in long-term adverse consequences.

Dusko Ehrlich: I am hearing in the last few minutes that there will be three studies which are relatively similar: the Bill & Melinda Gates' Foundation study, the Oxford Study, and the Stanford study. So we need to ensure that the results are comparable – I think we should get together perhaps under the umbrella of the International Human Microbiome Standards [IHMS] program, which is geared to getting people together to make sure that the protocols are used in a similar way and that the papers generated can be compared. We have a website for IHMS, so we can post these studies, make links, and try to diffuse the information in the best way we can.

Sven Pettersson: I would strongly support a program of standardization. We have this EU program, the TORNADO, which has been ongoing for the last 18 months, with the first reports on huge cohorts of microbiome analyses in children, adults, and the elderly, so we should be linking up with the coordinating people who run the adult studies to see that we make them all transparent.

Alan Jackson: Not only is there a need to standardize your approach to the microbiome, I would also encourage you to think about standardizing the outcomes in terms of these other areas relating to growth and body composition. I know that the International Atomic Energy Agency has over the past 5 years been trying to bring in and develop standardization of measurements of body composition.

We are also involved in two of the studies looking at prepregnancy effects on the growth and development of the fetus and in childhood, and exploring the effect of interventions prepregnancy on postnatal outcomes. In a current ongoing study led by Caroline Fall and others in the slums of Mumbai, there is a randomized controlled trial of different nutrient-rich foods to look at the progress through pregnancy and in the children subsequently. I just now wonder whether we should be looking at the microbiome in these children, too?

Philip James: There are published studies from the Indian slums with some beautiful mathematical analyses of the impact of different environmental effects on the development of the metabolic syndrome. One of the important components that emerged was the degree of environmental-induced inflammation.55

ENERGY BALANCE AND THE MICROBIOME

  1. Top of page
  2. METHODOLOGY: A GENE SCANNING APPROACH
  3. DIFFERENT APPROACHES TO THE ANALYSIS OF THE MICROBIOME
  4. METABOLIC PROFILING
  5. DEVELOPMENT OF THE MICROBIOME AFTER BIRTH AND THE INFLUENCE OF HOST GENETICS
  6. THE IMPACT OF EARLY FEEDING
  7. GUT PERMEABILITY IN THE NEWBORN
  8. EARLY MACRONUTRIENT EFFECTS ON THE HUMAN MICROBIOME
  9. FAT ABSORPTION AND BILE SALTS IN THE NEWBORN
  10. POPULATION DIFFERENCES IN THE MICROBIOME ACROSS THE GLOBE
  11. GUT MICROBIOTA AND THEIR POTENTIAL ROLE IN DEVELOPMENTAL PROGRAMING OF THE BRAIN
  12. DIETARY STANDARDIZATION
  13. NUTRIENT NICHES AND MICROBES THROUGHOUT THE LENGTH OF THE INTESTINE
  14. MUCOSAL BACTERIA
  15. MUCIN'S STRUCTURE AND INTERACTION WITH MUCOSAL BACTERIA
  16. GUT IMMUNOLOGY IN RELATION TO MUCIN AND MICROBIAL COLONIZATION
  17. DIETARY INTERVENTIONS: EFFECTS ON COLONIC MICROFLORA
  18. AUTISM – A GUT ROLE?
  19. ROLE OF THE MICROFLORA IN SYNTHESIZING ESSENTIAL AMINO ACIDS AND ALLOWING THE MAINTENANCE OF PROTEIN BALANCE ON VERY LOW INTAKES
  20. CHILDREN'S STUNTING: THE ROLE OF GASTROINTESTINAL MICROBIOTA
  21. THE ETHICS OF STUDYING DEPRIVED CHILDREN
  22. ENERGY BALANCE AND THE MICROBIOME
  23. PHYSICAL ACTIVITY AND THE MICROBIOME
  24. INSULIN SENSITIVITY AND THE METABOLIC SYNDROME
  25. PROBIOTICS AND THEIR EFFECTS
  26. CROHN'S DISEASE AND ULCERATIVE COLITIS
  27. REFERENCES

Dusko Ehrlich: We have conducted studies on obesity in a Danish study population which had been established as a cohort in 1999 and therefore called Inter99. This population has already been studied for their obesity characteristics and comorbidities three times.56 So there are historical data, but ecal samples were collected only during the last review, starting in 2009. In addition, my colleagues Karine Clement and Joël Doré have conducted a nutritional intervention in France with 49 overweight and obese individuals who received a calorie-restricted diet for 6 weeks and then a maintenance diet for 6 weeks with quantitative metagenomic analyses on their fecal samples.57 In the Danish obese cohort of 177 individuals, we looked at their gene sequences and found a bimodal distribution of individuals with respect to the number of fecal microbial genes they carry. We confirmed this in the French cohort – there are low-gene-count individuals and high-gene-count individuals. The difference amounts to a 35 to 40% difference in biomass. There are some genes that were only found in low-gene individuals and they had the Bacteroides-dominant enterotype, whereas the high-gene individuals seem to fall mostly in the Ruminococcus-driven enterotype. The two groups of individuals had about two-thirds of the genes from our reference gene catalogue.

Oluf Pedersen, who is the clinical PI for obesity in the MetaHIT program, tells us that low-gene obese Danish women with mostly a Bacteroides enterotype have increased indices of inflammation and Karine Clement finds the same feature in her French group. This raises the question as to whether low-grade inflammation correlates with lower bacterial diversity, because when we detect fewer bacterial genes that means that the diversity is lower. It has also been found that these Danish women with a lower microbial diversity had put on more weight over the previous 10 years, and in the French weight management study, those a with low gene diversity lost less weight in the intervention period when they were on an energy-restricted diet. In both the Danish and French study groups, their baseline studies showed that the low-gene individuals also tended to have higher BMIs than the high-gene individuals.

We then proceeded to assess whether we could detect distinct bacterial species associated with this high and low microbial diversity. For this purpose, we wanted to group genes into species – called meta-species and based simply on trying to reconstitute species from the genetic information without isolating the species, as such, from the microbial community. We used 50 genes as markers for a meta-species to construct the MetaHIT barcode, where each column related to an individual's bacterial genes and each line in the column indicated the specific gene present. We then found that 15 meta-species were strongly associated with the total abundance of genes among the sample of 99 MetaHIT obese individuals. So the hypothesis is that genes from the same species should have the same abundance in a given individual. Then, if gene abundance varies so much between individuals, we should be able to identify the meta-species responsible. We found, for example, that one of the species corresponded to Methanobrevibacter smithii, one of the main methanogens in humans, and was present in individuals with a low gene abundance. We also found that a number of other species were highly associated with gene number. However, for most gene constructs, we do not have reference genome sequences, so we cannot assign them taxonomically yet. Yet, if we take these constructs, we can relate different meta-species to gene abundance with a good specificity, as shown by our receiver operator characteristics [ROC] analyses of data from the cohort of 99 obese Danes. So these marker meta-species may prove, in due course, to have prognostic value in identifying people who are at high risk of obesity complications.

Harry Flint: As far as human energy balance is concerned, there are classic papers that suggest that we may derive, very roughly, about 10% of the dietary energy via the activities of the large intestinal microbes through the absorption of short-chain fatty acids. I think, interpreted naively, that says this is extra energy that may contribute to making us fat. The other way of looking at it is that we derive that energy from dietary components which include an appreciable amount of what we term “resistant starch.” However, if this starch were hydrolyzed and absorbed in the small intestine, then one would derive more than twice as much energy as from a starch molecule going into the large intestine and then getting fermented. So we derive energy less efficiently from fiber than we do from other readily digestible carbohydrate components of the diet. This energetic analysis does not take account of any satiety differences in the effects of these different routes of digestion.

I know that there has been a lot of discussion in the literature about the “energy harvest”; in other words, whether changing microbiota composition may determine how much energy we extract via fermentation. I think that the example I showed indicates that there may be some variations in microbiota composition that may affect net energy extraction, but I am also influenced by the old literature suggesting that gut transit responses, for example, may be far more influential. If you speed up transit through the gut, you alter the gut environment, the pH, and short-chain fatty acid production. You may also alter the fraction of carbohydrate which is digestible in the upper gut, and more of that may arrive in the colon. So I think we have not considered older studies and their analyses enough.

In terms of the animal experiments, we have heard how important both genetics and diets are, but I also think we have failed to consider different degrees of physical activity in these animals. This may also involve complex microbiota effects on activity and energy expenditure that may involve microbial metabolites and other mechanisms that we know very little about.

Philip James: We need to distinguish between metabolite or hormonal signaling and what we might term energy salvage in the colon – that is a completely different story.

Fredrik Bäckhed: In terms of the microbiota in the germ-free animal, one can get a fix on the problem; so if you do not see a difference in a certain parameter when comparing a germ-free with a conventionally raised mouse, then I would say it is less likely that the process will be microbially regulated in humans, whereas if you see a big difference between germ-free and colonized animals, I would say that that process may have a chance to be regulated by microbes in humans as well. Of course, germ-free animals are very different. Several processes in terms of energy partitioning are different – germ-free mice have denser bone mass, they have less fatness, but they seem to have the same amount of lean mass. So there is definitely something about energy partitioning that is very different. Some of this can be explained by energy harvesting, but other effects are clearly hormonally influenced, e.g., by glucagon-like peptide 1 [GLP1] and neuropeptide Y [PYY]. However, there are some problems measuring energy metabolism in germ-free compared with conventional mice; because the bacteria themselves can produce a lot of CO2, this will occur in a conventional but not in a germ-free animal. So I think that there are several processes that are microbially regulated in mice.

Philip James: What about leptin, which not only affects energy balance and the partitioning of energy but also bone mass?

Fredrik Bäckhed: We have done experiments where we assess leptin in germ-free and colonized mice and find the germ-free mice have lower leptin levels, but we have not sorted out all the factors involved.

Nathalie Delzenne: When germ-free and conventional mice are fed a diet supplemented with non-digestible and fermentable carbohydrate, we found differences in the gut hormonal responses but also in adipocytes which are, of course, exposed to absorbed short-chain fatty acids. So the story is complicated when we assess the role of fermentable carbohydrates.

Alan Jackson: I just wanted to add one other component to the energy balance equation relating to fermentable carbohydrate, and that is the induction of a bigger fecal mass of microbes, which constitute an appreciable energy loss as well as energy use.

Philip James: We showed a long time ago that the human colon can absorb volatile fatty acids [VFAs] avidly.58 Furthermore, Cummings and Stevens showed that at least 55% of the mass of colonic contents comprised bacteria59 and that the mechanism whereby the non-starch polysaccharides [NSPs], i.e., dietary fiber, increased stool weight was through fermentation of fiber by the anaerobic bacteria, thus stimulating microbial growth.60 So this microbial mass and the greater function and turnover of the colonic mucosa in response to higher fiber intakes will make greater energy demands. When comparing mice and humans, we also need to recognize the very different proportions of energy assigned to different tissues – as shown, for example, by the importance of the brown adipose tissue in rodents, whereas in man, despite the latest enthusiasm, it has a much smaller role in energy turnover. Furthermore, in rodents one is seeing the coprophagic effects, i.e., the impact of VFAs absorbed in the upper intestine and other microbial effects once the feces are eaten

Fergus Shanahan: My estimate is that there are remarkable differences in colonic size and length – one needs at least 20 more centimeters of the colonoscope if you are going to do colonoscopies in those eating consistently very-high-fiber diets. But in view of the remarkable reductions in physical activity over the last few decades, I wonder whether there are any data on the influence of physical activity on the microbiome?

PHYSICAL ACTIVITY AND THE MICROBIOME

  1. Top of page
  2. METHODOLOGY: A GENE SCANNING APPROACH
  3. DIFFERENT APPROACHES TO THE ANALYSIS OF THE MICROBIOME
  4. METABOLIC PROFILING
  5. DEVELOPMENT OF THE MICROBIOME AFTER BIRTH AND THE INFLUENCE OF HOST GENETICS
  6. THE IMPACT OF EARLY FEEDING
  7. GUT PERMEABILITY IN THE NEWBORN
  8. EARLY MACRONUTRIENT EFFECTS ON THE HUMAN MICROBIOME
  9. FAT ABSORPTION AND BILE SALTS IN THE NEWBORN
  10. POPULATION DIFFERENCES IN THE MICROBIOME ACROSS THE GLOBE
  11. GUT MICROBIOTA AND THEIR POTENTIAL ROLE IN DEVELOPMENTAL PROGRAMING OF THE BRAIN
  12. DIETARY STANDARDIZATION
  13. NUTRIENT NICHES AND MICROBES THROUGHOUT THE LENGTH OF THE INTESTINE
  14. MUCOSAL BACTERIA
  15. MUCIN'S STRUCTURE AND INTERACTION WITH MUCOSAL BACTERIA
  16. GUT IMMUNOLOGY IN RELATION TO MUCIN AND MICROBIAL COLONIZATION
  17. DIETARY INTERVENTIONS: EFFECTS ON COLONIC MICROFLORA
  18. AUTISM – A GUT ROLE?
  19. ROLE OF THE MICROFLORA IN SYNTHESIZING ESSENTIAL AMINO ACIDS AND ALLOWING THE MAINTENANCE OF PROTEIN BALANCE ON VERY LOW INTAKES
  20. CHILDREN'S STUNTING: THE ROLE OF GASTROINTESTINAL MICROBIOTA
  21. THE ETHICS OF STUDYING DEPRIVED CHILDREN
  22. ENERGY BALANCE AND THE MICROBIOME
  23. PHYSICAL ACTIVITY AND THE MICROBIOME
  24. INSULIN SENSITIVITY AND THE METABOLIC SYNDROME
  25. PROBIOTICS AND THEIR EFFECTS
  26. CROHN'S DISEASE AND ULCERATIVE COLITIS
  27. REFERENCES

Fredrik Bäckhed: We know in mice, both from our studies as well as in Sven's Pettersson's studies,25 that the germ-free mice move more, i.e., have a higher locomotive activity. They also have a lower body temperature. We don't know if the physical activity is a response to the need to raise the body's temperature, and the fecal output is much larger in a germ-free animal.

Philip James: In fact, there was a study by Sheila Bingham and John Cummings61 in which a group of healthy subjects underwent a quite vigorous training program whilst diet was maintained constant. Physical fitness improved, but there was no change in stool weight or transit time.

David Relman: I note that conventional and germ-free mice are usually maintained at a constant environmental temperature and this is not a mouse's normal environment. Chris Carr at the University of Cincinnati has been looking at the effects of ambient temperature, versus the mice's preferred temperature, on immune function and other metabolic features and finds that almost everything is affected by these temperature differences.

Elaine Holmes: I think the temperature question in our rodent studies is a huge one because we had an accident with a genetic group of db/db mice with a major propensity to develop diabetes, and when the environmental temperature was not controlled, the pregnant mice dropped their temperatures markedly and, to our surprise, we found the offspring did develop either insulin resistance or become fat. We were not allowed to repeat the experiment, but it implied profound long-term effects could be induced in utero.

Bo Angelin: When discussing whether exercise can change the microbiome, we ought to also consider whether the microbiome changes physical exercise. Because spontaneous non-exercise physical activity is, in humans, very responsive to changes in intake and affects energy balance,62 it seems pretty clear that the microbiome could change your behavior and induce more spontaneous activity.

Karen Madsen: We carried out a series of experiments where we put mice on different diets with different probiotics and then put them into a kind of a maze where they could either run around or escape into a hole. We find that the mice on the Western diet are very sedentary – they do not move around very much or explore; whereas if we provide them with certain probiotics, they scamper around for the whole of the time allowed, so we can definitely change physical activity and behavior by changing the diet and altering the microbiome.

INSULIN SENSITIVITY AND THE METABOLIC SYNDROME

  1. Top of page
  2. METHODOLOGY: A GENE SCANNING APPROACH
  3. DIFFERENT APPROACHES TO THE ANALYSIS OF THE MICROBIOME
  4. METABOLIC PROFILING
  5. DEVELOPMENT OF THE MICROBIOME AFTER BIRTH AND THE INFLUENCE OF HOST GENETICS
  6. THE IMPACT OF EARLY FEEDING
  7. GUT PERMEABILITY IN THE NEWBORN
  8. EARLY MACRONUTRIENT EFFECTS ON THE HUMAN MICROBIOME
  9. FAT ABSORPTION AND BILE SALTS IN THE NEWBORN
  10. POPULATION DIFFERENCES IN THE MICROBIOME ACROSS THE GLOBE
  11. GUT MICROBIOTA AND THEIR POTENTIAL ROLE IN DEVELOPMENTAL PROGRAMING OF THE BRAIN
  12. DIETARY STANDARDIZATION
  13. NUTRIENT NICHES AND MICROBES THROUGHOUT THE LENGTH OF THE INTESTINE
  14. MUCOSAL BACTERIA
  15. MUCIN'S STRUCTURE AND INTERACTION WITH MUCOSAL BACTERIA
  16. GUT IMMUNOLOGY IN RELATION TO MUCIN AND MICROBIAL COLONIZATION
  17. DIETARY INTERVENTIONS: EFFECTS ON COLONIC MICROFLORA
  18. AUTISM – A GUT ROLE?
  19. ROLE OF THE MICROFLORA IN SYNTHESIZING ESSENTIAL AMINO ACIDS AND ALLOWING THE MAINTENANCE OF PROTEIN BALANCE ON VERY LOW INTAKES
  20. CHILDREN'S STUNTING: THE ROLE OF GASTROINTESTINAL MICROBIOTA
  21. THE ETHICS OF STUDYING DEPRIVED CHILDREN
  22. ENERGY BALANCE AND THE MICROBIOME
  23. PHYSICAL ACTIVITY AND THE MICROBIOME
  24. INSULIN SENSITIVITY AND THE METABOLIC SYNDROME
  25. PROBIOTICS AND THEIR EFFECTS
  26. CROHN'S DISEASE AND ULCERATIVE COLITIS
  27. REFERENCES

Philip James: We have just had some astonishing data from Wilhelm de Vos, where he showed that when individuals had the colonic contents replaced by a healthy person's, then within 6 weeks there is a substantial improvement in insulin sensitivity, which I understand is maintained.

Fredrik Bäckhed: So far, they have only been allowed to do one time-point for some ethical reason, but the studies are now being extended. Also, when they look at the Clostridium difficile-infected patients, replacing the colonic contents of these patients leads to a normal microbiota which is then sustained. So the real issue is what the determinants of the sustainability of microbial replacements really are.

Bo Angelin: The most interesting experiment of all would, of course, be to make a colonic content transfer from a fat individual to an anorexia patient, because that would really tell you whether energy balance as well as some indices of metabolic sensitivity were changed.

Ruth Ley: In the last experiments we have done with Andrew Gewirtz,63 we have shown that mice genetically deficient in Toll-like receptor 5 [TLR5], a component of the innate immune system that is expressed in the gut mucosa and that helps defend against infection, exhibit hyperphagia and develop hallmark features of metabolic syndrome, including hyperlipidemia, hypertension, insulin resistance, and increased adiposity. These metabolic changes correlated with changes in the composition of the gut microbiota. By restricting the food intake we can limit the obesity, but the insulin resistance persists. Then, when we transfer the gut microbiota from the TLR5-deficient mice to wild-type germ-free mice, we induce the metabolic syndrome in the recipients. So the gut microbiota seems to have important contributors to metabolic disease and this is linked to the innate immune system. However, after transplanting a microbiota to induce insulin resistance and obesity, the microbiota converged with that of the lean donor after 7 weeks, but the phenotype of the mouse is now different; so, you have got an insulin-resistant, fat mouse, but its microbiota is no longer looking like the microbiota of an insulin-resistant fat mouse. So, whatever happened, you don't require the microbiota to be different at that stage to have the phenotype, so maybe with the humans, it is possible that insulin sensitivity is altered by a temporary change in the microbiota.

Philip James: I thought we had been finding that the microbiota is relatively stable unless there is a large colonic input of new substrate from non-starch polysaccharides or resistant starch, but the human findings are really surprising. The mice studies, I suppose, could reflect the impact of recycled metabolites from colonic contents which were then, as usual in rodents, eaten, with these metabolites inducing changes in small intestinal hormonal output, thereby altering insulin resistance? Coprophagy confuses the potential mechanisms in rodents I would guess.

Tore Midtvedt: With Arnold Berstad, we did fecal transplantation in a cohort of patients with recurrent antibiotic diarrhea and found that they recovered normal bowel function within 4 days.64,65 The production of short-chain fatty acids, the breakdown of mucin, the inactivation of tryptic activity, the conversion of cholesterol to coprostanol, and the bile acid metabolism were normalized. Having done this, the patients were then subsequently stable.

Harry Flint: I think fecal bacteria therapy seems to have become popular, particularly in the United States, for treatment of C. difficile, and the 94% success rate found by Wilhelm is typical from what I have heard, so it does seem to be a long-term solution.

Fredrik Bäckhed: I think we are talking about two different conditions where, first, we deal with patients who have a very abnormal microbiota, whereas other relatively healthy people have much more stable microflora which are difficult to change.

David Relman: I agree. It may be that some of these patients with recurrent or persistent C. difficile overgrowth and diarrhea are stuck in one of these alternative microbial states.

Dusko Ehrlich: My feeling is that there is quite a consensus in the community that, in healthy adults, the microbiota is stable.

PROBIOTICS AND THEIR EFFECTS

  1. Top of page
  2. METHODOLOGY: A GENE SCANNING APPROACH
  3. DIFFERENT APPROACHES TO THE ANALYSIS OF THE MICROBIOME
  4. METABOLIC PROFILING
  5. DEVELOPMENT OF THE MICROBIOME AFTER BIRTH AND THE INFLUENCE OF HOST GENETICS
  6. THE IMPACT OF EARLY FEEDING
  7. GUT PERMEABILITY IN THE NEWBORN
  8. EARLY MACRONUTRIENT EFFECTS ON THE HUMAN MICROBIOME
  9. FAT ABSORPTION AND BILE SALTS IN THE NEWBORN
  10. POPULATION DIFFERENCES IN THE MICROBIOME ACROSS THE GLOBE
  11. GUT MICROBIOTA AND THEIR POTENTIAL ROLE IN DEVELOPMENTAL PROGRAMING OF THE BRAIN
  12. DIETARY STANDARDIZATION
  13. NUTRIENT NICHES AND MICROBES THROUGHOUT THE LENGTH OF THE INTESTINE
  14. MUCOSAL BACTERIA
  15. MUCIN'S STRUCTURE AND INTERACTION WITH MUCOSAL BACTERIA
  16. GUT IMMUNOLOGY IN RELATION TO MUCIN AND MICROBIAL COLONIZATION
  17. DIETARY INTERVENTIONS: EFFECTS ON COLONIC MICROFLORA
  18. AUTISM – A GUT ROLE?
  19. ROLE OF THE MICROFLORA IN SYNTHESIZING ESSENTIAL AMINO ACIDS AND ALLOWING THE MAINTENANCE OF PROTEIN BALANCE ON VERY LOW INTAKES
  20. CHILDREN'S STUNTING: THE ROLE OF GASTROINTESTINAL MICROBIOTA
  21. THE ETHICS OF STUDYING DEPRIVED CHILDREN
  22. ENERGY BALANCE AND THE MICROBIOME
  23. PHYSICAL ACTIVITY AND THE MICROBIOME
  24. INSULIN SENSITIVITY AND THE METABOLIC SYNDROME
  25. PROBIOTICS AND THEIR EFFECTS
  26. CROHN'S DISEASE AND ULCERATIVE COLITIS
  27. REFERENCES

Philip James: If the microflora is stable, how do probiotics work?

Fergus Shanahan: It is a reasonable question. I think the term “fecal transplantation” is just another way of saying a fecal enema. Tore Midtvedt is right: many doctors tried the impact of fecal enemas even before they had identified the C. difficile organism as the cause of the recurrent diarrhea. In the United States, physicians are able to charge for these enemas or fecal transplantations, but what is actually happening is that one is giving a probiotic. I need to declare that I am patent holder and founder member of a company with two probiotics on the market in the United States. In my opinion, there are two broad indications where you might consider a probiotic, and I do regard the fecal transplantation as, in essence, flooding the system with commensals that you have harvested. In desperately ill patients, you are trying to buy time and reduce the C. difficile numbers back into a reasonable range. These organisms are, however, detectable in 10% of institutionalized elderly who have no symptoms, but they are a time-bomb waiting to get carrier's sick if an antibiotic is administered. So C. difficile is probably best considered as a normal resident in the gut of a lot of people. It just becomes a problem in the wrong context.

So one broad indication for the use of probiotics is where you have got a major disruption of the microbiota. In this situation, your choice of organism is probably not critical. The second situation is where you are trying to get a very defined effect, where the choice of organism is absolutely vital. An example is in the very early premature baby, below 1.5 kg, who is exposed to organisms in an intensive pediatric facility when their digestive tract, their immune system, their blood brain barrier, and indeed their brain are not fully developed for birth. They are then easily colonized with fatal results. In these circumstances, I think we should be offering them an organism that has less pathogenic potential, e.g., avoid the colonization with a Clostridium or a Bacteroides species, which are associated with necrotizing enterocolitis, and substitute with something like a Lactobacillus or Bifidobacterium. Even though the trials are quite variable, I think the evidence for probiotic therapy in these very desperately sick, low-birth-weight babies is very compelling. Not only are they being protected from necrotizing enterocolitis, but their overall death rate is also lowered. So I actually think we will look back some day and wonder why we did not control the organisms that are colonized at early birth.

The third use of probiotics is actually in the developing world, where there is an overload of potential pathogens in both food and water. Most trials of probiotics have actually been done in the developed world, but there is one very fine community-based, double-blind, placebo-controlled trial in the slums of Calcutta. Nearly 4,000 children aged 1–5 years were involved and given either a probiotic drink containing Lactobacillus casei strain Shirota or a nutrient drink daily for 12 weeks with a follow-up period of another 3 months. The prevention of a first episode of diarrhea was modest (14%) but statistically significant and was not shown to relate to any specific organism.66 Nevertheless, this modest effect, when translated onto a population basis in developing countries, can make a substantial difference.

A fourth area for considering specific probiotics is in those with oxalate kidney stones – a third of us produce quite a lot of oxalic acid. It has been shown that almost all 6- to 8-year-old children have Oxalobacter formigenes in their feces, but this organism is retained by only 60 to 80% of adults, perhaps because of the effects of previous antibiotic use. When volunteers without detectable O. formigenes ingested an oxalate-degrading organism, e.g., O. formigenes, and then received a standard oxalate load, the urinary excretion of oxalate was reduced. Oxalate degradation is then apparent in the feces of these individuals, and this oxalate degradation persists after the specific oxalate-degrading organism has been given.67 Patients with inflammatory bowel disease have a 10- to 100-fold increased risk of nephrolithiasis, with enteric hyperoxaluria being the major risk factor, and clinical studies now show the efficacy of a suitable probiotic for reducing urinary oxalate excretion.68,69

Elaine Holmes: Another area where probiotics are being tried is in pancreatitis, but we need to be cautious because the impact seems very uncertain at present.70,71 Another example from our own laboratory is in rats that were undergoing ischemia reperfusion: we applied a probiotic, expecting it to aid recovery, but the probiotic proved lethal, so we need to be cautious.

Denise Kelly: As the initiator of a company and a patent holder in probiotics I also have specific interests in this field, but we have to remember that the European Food Safety Authority has not offered a single positive opinion on any of the probiotics that are currently marketed, despite the fact that it is a $31 billion global market. So I think the problem I have is that when you look at the work being conducted, these probiotics are often not selected and it is not surprising that a single organism is unable to do everything being asked. I just think now that the whole approach to screening probiotics needs to be revised and we need to look for organisms that are fit for specific purposes.

Seppo Salminen: The real problem is that EFSA's assessment has nothing to do with the efficacy of probiotics in the way we have been discussing the issue. This is because the European Parliament ruled that food health claims have to apply to healthy European populations, whereas we have been talking about their use in patients. This patient use, does not qualify for a health claim.

Fergus Shanahan: I agree completely. If, however, you just want to substitute an organism that has less pathogenic potential in the baby's susceptibility to necrotizing enterocolitis, the choice may not be critical. But if you want to do something very specific, such as immunomodulation, we have a different challenge. Probably the best example where a mechanism has been defined is in the experimental work with mice of Colin Hill's where the experimental evidence suggests that a single gene in a potential probiotic organism is responsible for the production of bacteriocin, which inhibits the growth and deleterious effects of listeria.72 Humans are very susceptible to listeria.

CROHN'S DISEASE AND ULCERATIVE COLITIS

  1. Top of page
  2. METHODOLOGY: A GENE SCANNING APPROACH
  3. DIFFERENT APPROACHES TO THE ANALYSIS OF THE MICROBIOME
  4. METABOLIC PROFILING
  5. DEVELOPMENT OF THE MICROBIOME AFTER BIRTH AND THE INFLUENCE OF HOST GENETICS
  6. THE IMPACT OF EARLY FEEDING
  7. GUT PERMEABILITY IN THE NEWBORN
  8. EARLY MACRONUTRIENT EFFECTS ON THE HUMAN MICROBIOME
  9. FAT ABSORPTION AND BILE SALTS IN THE NEWBORN
  10. POPULATION DIFFERENCES IN THE MICROBIOME ACROSS THE GLOBE
  11. GUT MICROBIOTA AND THEIR POTENTIAL ROLE IN DEVELOPMENTAL PROGRAMING OF THE BRAIN
  12. DIETARY STANDARDIZATION
  13. NUTRIENT NICHES AND MICROBES THROUGHOUT THE LENGTH OF THE INTESTINE
  14. MUCOSAL BACTERIA
  15. MUCIN'S STRUCTURE AND INTERACTION WITH MUCOSAL BACTERIA
  16. GUT IMMUNOLOGY IN RELATION TO MUCIN AND MICROBIAL COLONIZATION
  17. DIETARY INTERVENTIONS: EFFECTS ON COLONIC MICROFLORA
  18. AUTISM – A GUT ROLE?
  19. ROLE OF THE MICROFLORA IN SYNTHESIZING ESSENTIAL AMINO ACIDS AND ALLOWING THE MAINTENANCE OF PROTEIN BALANCE ON VERY LOW INTAKES
  20. CHILDREN'S STUNTING: THE ROLE OF GASTROINTESTINAL MICROBIOTA
  21. THE ETHICS OF STUDYING DEPRIVED CHILDREN
  22. ENERGY BALANCE AND THE MICROBIOME
  23. PHYSICAL ACTIVITY AND THE MICROBIOME
  24. INSULIN SENSITIVITY AND THE METABOLIC SYNDROME
  25. PROBIOTICS AND THEIR EFFECTS
  26. CROHN'S DISEASE AND ULCERATIVE COLITIS
  27. REFERENCES

Gunnar Hansson: We have recently published a paper together with Fu and Lijun Xia73 reporting how we induced knock-out mice with intestinal epithelial cell-specific deficiency of core 1-derived O-glycans as the predominant form of O-glycans. The animals developed an ulcerative colitis type of illness in their distal colons – like human ulcerative colitis. So the inner mucus layer is probably extremely important in protecting the colon from bacterial penetration and inflammation. If we treat these mice with antibiotics, then we can see that the mucus can come back, which seems to point to the primary role of the bacteria in inducing the inflammation once the protective mucin is not there. There are some other data from mouse models of colitis showing that, probably, the immune system influences the quality of these mucus layers.

Then, when we looked at patients with ulcerative colitis, we found a subset in whom the Tn antigen was expressed in their colon biopsy samples. Tn-positive and -negative crypts often occurred within a single colon section, suggesting a clonal abnormality of O-glycosylation, and some of the Tn-expressing epithelia contained C1GALT1C1 missense mutations. These findings support a molecular mechanism for colitis development in a subset of UC patients, whereby core 1 O-glycosylation is altered through somatic mutations in C1GALT1C1, which is essential for core 1 O-glycosylation.

Olle Hernell: Thank you Gunnar for very interesting data. The defense against mucosal bacteria seems also to involve the defensins trapped in the mucin layer. In humans, the mucin breakdown only occurs in the newborn after several months.

Gunnar Hansson: The defensins and other antimicrobials, as well as IGA, also come up from the crypts.

Denise Kelly: I wanted to ask you – as I understand it, you are saying that the intrinsic composition of the mucin in the inner and outer layer is identical and it is really its thickness or the gel-ness of it which determines access to the bacteria? The bacteria with mucolytic activity can degrade the mucin, but you do not see any of this degradation in the inner layer?

Gunnar Hansson: There are several bacteria which have very specific proteases that are able to cleave these polymers at very specific sites. We have already published74 on Entamoeba histolytica, which secretes one of its proteases that cleaves at the site, so the whole mucus system falls apart and allows the parasite to penetrate down to the epithelium and invade. I am predicting that there will be a number of bacteria that can do similar things.

Karen Madsen: The luminal environment affects directly the activity of the probiotic and it differs in normal people and in patients with Crohn's disease and ulcerative colitis.75 We know that in our clinical trials, 20% of people taking probiotics show no evidence of the microorganisms’ survival. We also know that if we take biopsies when we have used Bifidobacterium infantis, which in animals and cell culture have an immunosuppressive effect, then, when we take biopsies from controls, we see the expected immunosuppressive effects. But when we take biopsies from Crohn's patients, we find the exact opposite – we actually turn on NF-κB and IL-8 secretion. This response is, again, different from that seen in patients with ulcerative colitis. So a diseased person does not respond to probiotics in the same way a normal person does: the microenvironment in the lumen alters the probiotic activity.

Sven Pettersson: I agree that it is dangerous to extrapolate from the normal to the diseased state. We also know that germ-free mice live longer than mice that are exposed to bacteria. Both Nicholson and our group have published data showing that when you start to colonize germ-free mice, the first thing that happens is they downregulate the whole xenobiotic metabolism – with high levels of xenobiotic metabolism, you live longer.

Fredrik Bäckhed: It may not relate simply to xenobiotic metabolism. It is well known in all living organisms that if you calorie restrict an animal, they live longer, and a germ-free animal is calorie restricted. The current problem with probiotics is in part the food industry's doing, because they really want to market their yogurt as appropriate for everybody and this is not true. What we need are criteria for specifying the benefit for individuals with specific problems. So we need to be really good at identifying the conditions when probiotics will work and which we should use.

Dusko Ehrlich: The major stumbling block for me is that we don't understand why our responses to probiotics differ, so it is difficult to make logical progress. Fergus identified specific conditions where there is benefit from specific probiotic types, and I agree. When we do have these conditions with good data, we are clear. The second relates to the transfer of a whole mass of different species in colonic content transfers, where we change everything. Now it seems that it does work, although we don't know enough about the stability of the microflora. The third situation is where we are trying to shift the microflora to a modest degree, but where we still are very uncertain about the fundamental controls affecting the microflora and their effects.

Seppo Salminen: We have used fecal transplants in Finland for chronic C. difficile diarrhea and the donors have mostly been family members, but if we are now going to test this technique with any donor one has to ask what else are we transferring with the microbiota? What do we know about the viruses, what do we know about other things that we are transplanting?

Philip James: Let us now go on to such conditions as the irritable bowel syndrome or the frequency of abdominal pain, which Willem was also trying to tackle and where he was able to show with four different microbial probes that he could predict 81% of cases with irritable bowel syndrome. If so, how common is this problem in the community?

Fergus Shanahan: Very frequent – in Europe in a survey of 40,000 people, the estimated prevalence was around 14%. About half of them would seek medical attention but half do not. The condition is only diagnosed on history and there is no official biomarker. The European Food Safety Authority does not accept biomarkers, but if Willem's analyses are correct, then this is a real step forward.

Philip James: So now, let's come to ulcerative colitis. We have not talked about the potential role of sulphite in the diet, which I know Cummings and his colleagues suggested might be involved after a major analysis of the diets of different patients where their dietary patterns were ranked with the degree of colonic disease, as judged by the colonic mucosa's appearance on sigmoidoscopy.76 They had previously suggested that hydrogen sulphide in the colon was toxic77 and influenced not only by protein intake78 but also by dietary sulphite, used as a preservative in the diet, as well as other forms of dietary sulphate. We also know that some of the drugs used in the treatment of ulcerative colitis reduce the generation of sulphide in the colon of patients being treated for ulcerative colitis.79 There was also a small intervention study in Australia, with a substantial response in the majority of patients to a diet which limited protein and sulphide intake80– nearly all the patients with ulcerative colitis remitted one sulphide and protein intake were no longer reduced. Daily intake of inorganic sulfate varies enormously,81 with inorganic sulphate in the diet taking many forms – sulfite, sulfur dioxide, bisulfate, or metabisulfite, as used routinely in the preservation of processed foods and beverages. In addition, there are the sulphur-containing amino acids. Where have we got to with this story and the potential issue of individual differences in the microbial metabolism of sulfur compounds in the colon and their relationship to nitric oxide metabolism in determining ulcerative colitis?82

Fergus Shanahan: I have not seen any duplicated trials on sulphite/sulphate intake in relation to ulcerative colitis, but this is worth looking at.

Dusko Ehrlich: In unpublished data from a cohort of 100 patients, my colleagues are finding that the variety of microbial species is lower in patients with ulcerative colitis than in healthy people and there seem to be differences in microbial composition in those who suffer frequent relapses.

Harry Flint: We have recently shown, in vitro, that lactate is a very good cosubstrate for sulphate reduction by the major bacterium Desulfovibrio piger83 that is the dominant sulphide producer in the colon. Intriguingly, lactate accumulation is a feature of severe colitis, and this has been known for a long, long time, so I think sulphide mechanism needs more investigation.

Gunnar Hansson: I agree it is very important to separate Crohn's disease from ulcerative colitis, but Crohn's disease is most likely a mishandling of intercellular bacteria which are then not processed by the immune system.84 We still don't know exactly what is the basis for colitis, but I am quite sure that it has to do with this mucus layer. We know, for example, that there are spontaneous mouse models which get colitis because they have mutations in the cytokine production locus,85 but we don't know about any humans yet. We also know that there are certain bacteria that can produce enzymes that can degrade the mucus layer. So I think there will be a number of different reasons for the disease, but you have to find the common theme as the underlying factor.

David Relman: Just a note of caution about the linkage between microbiota and disease states, for example in the gut or in the mouth. The majority of the data show associations between a changed microbiota and a disease at the time the disease has already been diagnosed. Then, of course, the real challenge will be to show that the change in the microbiota precedes, or somehow leads to, the pathology. It is obviously difficult to deal with this challenge, but we need to build not just a temporal relationship, but a dosage relationship, i.e., the association, specificity, sensitivity, biologic plausibility, and all the other criteria that Sir Austin Bradford Hill and others set out a long time ago.

Philip James: That is a good point to end this discussion. What has become apparent is the extraordinary value of interactions across several disciplines. We are dealing not only with complex issues relating to the identity and interactions of different microbes, but how these different species have established themselves in the gut early in life, the importance of establishing a niche for themselves, and the complex interaction of dietary factors and the host's immune responses and the interplay with such an array of metabolically highly active microbes. Some time ago, this was seen as simply a fascinating challenge where our understanding of the geographical differences in the prevalence of different microflora was interesting. However, from our discussions it has become apparent that these microbes have an astonishing influence on our normal metabolic processing and, indeed, there is a major challenge for us all, as it becomes ever more apparent that the subject of this Marabou Trust “think tank” has many potentially major public health implications. Thank you for engaging so robustly in the discussion and for starting what promises to be very exciting interdisciplinary collaborations across the globe.

REFERENCES

  1. Top of page
  2. METHODOLOGY: A GENE SCANNING APPROACH
  3. DIFFERENT APPROACHES TO THE ANALYSIS OF THE MICROBIOME
  4. METABOLIC PROFILING
  5. DEVELOPMENT OF THE MICROBIOME AFTER BIRTH AND THE INFLUENCE OF HOST GENETICS
  6. THE IMPACT OF EARLY FEEDING
  7. GUT PERMEABILITY IN THE NEWBORN
  8. EARLY MACRONUTRIENT EFFECTS ON THE HUMAN MICROBIOME
  9. FAT ABSORPTION AND BILE SALTS IN THE NEWBORN
  10. POPULATION DIFFERENCES IN THE MICROBIOME ACROSS THE GLOBE
  11. GUT MICROBIOTA AND THEIR POTENTIAL ROLE IN DEVELOPMENTAL PROGRAMING OF THE BRAIN
  12. DIETARY STANDARDIZATION
  13. NUTRIENT NICHES AND MICROBES THROUGHOUT THE LENGTH OF THE INTESTINE
  14. MUCOSAL BACTERIA
  15. MUCIN'S STRUCTURE AND INTERACTION WITH MUCOSAL BACTERIA
  16. GUT IMMUNOLOGY IN RELATION TO MUCIN AND MICROBIAL COLONIZATION
  17. DIETARY INTERVENTIONS: EFFECTS ON COLONIC MICROFLORA
  18. AUTISM – A GUT ROLE?
  19. ROLE OF THE MICROFLORA IN SYNTHESIZING ESSENTIAL AMINO ACIDS AND ALLOWING THE MAINTENANCE OF PROTEIN BALANCE ON VERY LOW INTAKES
  20. CHILDREN'S STUNTING: THE ROLE OF GASTROINTESTINAL MICROBIOTA
  21. THE ETHICS OF STUDYING DEPRIVED CHILDREN
  22. ENERGY BALANCE AND THE MICROBIOME
  23. PHYSICAL ACTIVITY AND THE MICROBIOME
  24. INSULIN SENSITIVITY AND THE METABOLIC SYNDROME
  25. PROBIOTICS AND THEIR EFFECTS
  26. CROHN'S DISEASE AND ULCERATIVE COLITIS
  27. REFERENCES
Footnotes
  • *

    Dusko Ehrlich presented a lecture which formed the basis for considerable discussion but did not produce a paper. This is a brief summary of his lecture as it helps explain subsequent commentaries.

  • Since the meeting, one of the participants, Rob Knight, and his colleagues have published evidence that long-term, high-fat, low-fiber diets are associated with the Bacteroides enterotype and a long-term, high-carbohydrate diet with a Prevotella enterotype. Short-term dietary manipulation did alter the microbiome, but the enterotype of the individual was sustained, implying that the eneterotype relates to early or long-term dietary experience.4