Gastrointestinal Microbiota and Advances in Prebiotic and Probiotic Research Conference

Session I: Function of Gut Microbiota
Session Chair: Cathryn Nagler, Ph.D.
Massachusetts General Hospital
Charlestown, MA

Dr. Cathryn Nagler introduced the speakers in the session and noted that the immune system is powerfully shaped by the presence of commensal intestinal bacteria. Mice that have been raised germ-free (GF), without commensal microbiota, have deficits in T cells and antibody producing B cells in both the gut associated lymphoid tissue and the spleen. Recent work from Dennis Kasper's lab has demonstrated that a bacterial polysaccharide from the ubiquitous gut microorganism Bacteroides fragilis can partially correct these deficits and direct lymphoid organogenesis.

During immune system development, tolerance to the presence of gut microorganisms and foreign antigens, such as those from food, must be maintained. Dysregulation of mucosal homeostasis leads to disease at both ends of the Th1/Th2 spectrum. An immune response to food results in food allergies, which is a growing problem for unknown reasons. Immune responses to commensal microorganisms characterize inflammatory bowel diseases (IBD). Emerging evidence suggests that the commensal microflora induces populations of regulatory cells that help maintain host tolerance.

Modifying Immune Function
Bengt Björkstén, M.D., Ph.D.
Karolinska Institute
Stockholm, Sweden

Dr. Björkstén demonstrated that pregnancy can cause deviation of immune system to occur in mothers due to changing progesterone and other hormone levels. It was found that levels of interleukin (IL) 4 and IL10 increase, and interferon gamma (INF ) levels decrease during pregnancy. In utero, the fetal immune system is skewed toward Th2 cells. After birth, Th2 and Th0 cells are elevated; sometime during childhood, the immune system is modified but approximately one-third of children retain a Th2 response to allergens and develop respiratory allergies. Differences in the development of allergies have been observed between Estonian and Swedish children. Estonian children have levels of pollution and smoking similar to those in Sweden 40 years ago and have lower rates of allergies and type 1 diabetes. Stronger Th1 and Th2 responses to allergens are observed in Swedish compared to Estonian infants. Both Th1 and Th2 responses were exaggerated in Swedish infants who developed allergies, indicating that these children failed to dampen the responses as they aged. Estonian infants had higher levels of secretory IgA than Swedish infants. Secretory IgA is the last part of the immune system to mature, indicating that Estonian infants had an initially higher rate of postnatal immune system maturation than Swedish children, although these differences are no longer evident by 5 years of age. Swedish and Estonian infants expressed both Th1- and Th2-like immune responses to allergens regardless of atopic status; early Th2-like responses were down regulated, except in children who developed clinical allergies.

Environmental factors believed to be protective against allergies include pets, living on a farm, or having older siblings. Air pollution, traffic, an indoor environment, lack of breast feeding, and tobacco smoke were believed to be risks for allergy development, but currently there is no evidence that they increase risk. Living on a farm, going to day care, or having three or more older siblings is protective against atopic dermatitis, but respiratory infections occurring in children younger than 6 months of age increased risk. Farming also appears to be protective against asthma and continues to be protective despite changes in farming practices that may reduce exposure to potential allergens.

The gut microbiota influences the development of oral tolerance. Differences in the acquisition of gut microbes have been observed in Estonian compared to Swedish infants, with Estonian infants having a more pronounced colonization, theorized to be related to differences between the two countries in hospital stays for childbirth. Healthy infants and those without allergies had more extensive gut colonization and a higher prevalence of Bifidiobacteria. Differences in gut ecology between allergic versus non-allergic children still were evident at 5 years of age. The vaginal flora of mothers at 22 to 24 weeks gestation also appeared to affect the rate of asthma medication use in their children, with asthma being more common in children of mothers with vaginal Staphylococci and mothers who had used antibiotics during pregnancy. Atopic infants also showed consistently less gut microbial diversity, even before the appearance of symptoms.

Exposure to bacterial species is essential for proper immune system development. The bacteria essential to this process likely were present during the evolution of the immune system in mammals and were distributed around the world throughout the history of man. Based on differences in immune function and the gut microbiota in Estonian versus Swedish infants, the presence of these bacteria may have declined during the second half of the 20th century. Probiotics may have similar effects to those of "poor hygiene" and may help moderate the increase in "immunologically mediated diseases of affluence" in the future.

Animal Models of Immune System Development and Modulation Throughout Lifespan
Gary B. Huffnagle, Ph.D.
University of Michigan
Ann Arbor, MI

Dr. Huffnagle acknowledged that a number of inflammatory mucosal diseases have been associated with disturbances in the microbiota. Gut responses could affect immunity in the lungs because fluids, particles, and microbes introduced into the nasal cavity are found largely in the gastrointestinal (GI) tract shortly thereafter; what is inhaled also is swallowed. Ingestion of allergens can induce systemic tolerance to that allergen ("oral tolerance"), and the GI tract can act as a "sensor" for the development of tolerance to inhaled allergens. Gnotobiotic mice (mice lacking a microbiota) do not develop oral tolerance but introducing a microbiota into them (conventionalization) can restore this ability. This leads to the question of whether the microbiota regulates tolerance to allergens. To test whether alteration of the microbiota can affect pulmonary allergic responses, C57BL/6 and Balb/c mice were given a short dose of antibiotic (cefoperazone) and then Candida by oral administration and exposed to the mold Aspergillus by inhalation. IgE levels were highest in C57BL/6 mice that were exposed to the antibiotic, Candida, and Aspergillus. These mice also had higher levels of mucus production. Microbiota disruption was associated with higher levels of several allergy-associated cytokines (IL5, IL13, lung eosinophils, and mast cells) after exposure to Aspergillus. Hypersensitivity induced by Aspergillus in microbiota-disrupted mice was observed in nasal septal tissue as well as in the lungs. Exposure to Candida alone had little effect on the microbiota but Candida plus cefoperzone changed the microbiota composition to include fewer Lactobacilli.

Mouse models also can be used to study the effects of microbiota disruption on chronic versus resolving infections. Cryptococcus neoformans can cause both chronic and resolving pulmonary infections in mice. Chronic infections are characterized by a stronger Th2 response, whereas resolving infections are characterized by stronger Th1 activity. CBA/J mice are more likely to develop a resolving infection, and C57BL/6 mice develop chronic infections. These two mouse strains initially were housed separately and treated with cefoperzone. Untreated and antibiotic-treated mice then were co-housed and exposed to C. neoformans. C57BL/6 mice treated with the antibiotic developed higher levels of eosinophils and IL4 in response to C. neoformans infection. Neutrophils and IL12 levels were unaffected. Eosinophil and IL4 levels in CBA/J mice were unaffected by antibiotic treatment, but neutrophils and IL12 levels were increased. Thus, the microbiota of the mice appears to signal to the immune system to down-regulate inflammation.

Immune responses in lungs are linked to the GI tract through interactions of the innate and adaptive immune responses. In addition, swallowed allergens travel to the gut, where they are detected by the innate immune system; the immune response that occurs is influenced by the microbiota community. Interactions among the fungal and bacterial gut microbiota affect the development of the regulatory immune response. Although genetics have a role in innate immunity, the microbiota is the driving force in hypersensitivity or inflammatory diseases.

Session II: Normal Gut Ecology
Session Chair: Martin H. Floch, M.D., M.A.C.G.
Yale University
New Haven, CT

Before introducing the presenters in session II, Martin H. Floch of Yale University first pointed out that gut ecology is a dynamic process. The four major components of gut ecology are the gut wall, gut secretions, microflora (i.e., microbiota), and food. The gut wall and its secretions vary from mouth to esophagus to stomach to intestinal tract. The gut secretes approximately 6 to 10 liters of fluids per day, which contain electrolytes, enzymes, proteins, and water. The microbiota includes bacteria, but also fungi, parasites, and viruses. All of these components affect activities that occur in the gut lumen, such as digestion, absorption, hormone production, and detoxification.

Development and Maintenance of Gut Microbiota
David S. Newburg, Ph.D.
Massachusetts General Hospital
Boston, MA

Dr. Newburg mentioned that the intestinal mucosa is the most heavily glycosylated organ, and glycans are important for interactions between epithelial cells and microorganisms. Prior to weaning, Lactobacilli and Bifidobacteria predominate. After weaning, the species predominant in the microbiota and the type of glycans present on gut epithelial cells both change. In infants, the gut mucosa is heavily sialylated, whereas in adults, the gut mucosa is heavily fucosylated. These two types of glycans function in the anchoring of different types of enteric bacteria to the gut. At weaning, fucosyltransferase is induced and levels of sialyl transferase decrease, concurrent with large changes in gut flora. The mechanisms that drive this change could include genetic control of intestinal mucosal gene transcription, changes in diet, or stimulation by the microbiota.

Studies in germ-free (GF) mice showed that both nursing and weaned mice retain sialylated glycans, implying that the mechanism for the developmental change to fucosylated glycans occurs through colonization of the gut. Colonization of adult GF mice with normal mouse microbiota induced fucosylation of glycans, demonstrating that colonization participates in controlling the expression of cell surface glycans in the mammalian gut epithelium. In a bacterial depletion model, mice were treated with antibiotics for 2 weeks to deplete gut bacteria, and then repleted with normal mouse microorganisms. The repleted mice had microbiotas that were the same as in conventionally raised mice. UEA1 lectin, which binds to 1, 2-linked fucosyl groups, bound to the colonic mucosa of conventionally raised and repleted mice, but not depleted mice, confirm the link between colonization and expression of fucosylated cell surface glycans.

In mice and humans, fucosyltransferase 1 (fut1) and fucosyltansferase 2 (fut2) form fucosyl 1-2 linkages. Expression of fut1 is constitutive, whereas fut2 expression is sensitive to the composition of the gut microbiota. Mediation of fut2 expression in enterocytes by extracellular microorganisms involves the TLR signal transduction system. Colonization or bacterial repletion transiently activates ERK and JNK signaling, which are not activated in depleted animals. Binding of bacteria activates ERK and JNK, which in turn activates AP1; AP1 binding sites are found in the fut2 promoter region. Interestingly, NF B is not activated during this process, which therefore is non-inflammatory. Inhibition of ERK and JNK eliminates the recolonization-dependent induction of fut2 mRNA, confirming their critical role in transduction of this signal. TLR4 appears to be the crucial receptor in this process because colonization fails to induce fut2 in TLR4 knockout mice. The TLR4 ligand lipopolysaccharide (LPS) can induce fut2 activity in bacterially depleted mice but also activates the inflammatory response.

TLR4 is expressed on enterocytes. TLR4 sialylation is observed, but fucosylation is induced only in the absence of colonization. Fucosylated TLR4 is bound by fucosyl-binding microbes; this leads to activation of the JNK and ERK signaling pathways. In the DSS-induced colitis model, mice that lack functional TLR4 do not recover from colitis because fut2 expression is not activated. Similarly, bacterially depleted mice also do not survive DSS-induced colitis.

The gut microbiota communicates with intestinal epithelial cells through activation of ERK and JNK (but not NF B) pathways, with TLR4 signaling mediating the initial communication. TLR4 is fucosylated when the gut is not adequately colonized; during colonization, microorganisms could bind fucosyl residues on TLR4, thus inducing fut2 activity and resulting in fucosylation of the mucosal cell surface. Fut2 expression induces a fucosylated niche that favors colonization and maintenance of the intestinal microbiota, which helps protect the intestinal mucosa from injury.

Esophageal Bacterial Biota: Beyond and Below
Zhiheng Pei, M.D., Ph.D.
New York University School of Medicine
New York, NY

Zhiheng Pei of New York University presented research on the foregut microbiota. He suggested the increase in esophageal cancer could be due to changes in the distal esophageal microbiota, which differ between healthy people and patients with histopathlogical phenotypes of gastroesophageal reflux diseases (GERD). Phenotyping and metagenotyping, which refers to determining the genotypes of microbiota classified by phylogenetic distance-based normal reference range (NRR), can be used to classify distal esophageal microbiota and develop risk assessments related to GERD.

Calculating the mean phylogenetic distances among bacterial communities showed differences in microbiota between normal esophagus versus pathologic esophagus, which could indicate an association between host phenotype and bacterial metagenotype. The presence of a particular metagenotype (metagenotype II) showed a strong association with esophagitis and Barrett's esophagus. This metagenotype is characterized by an abundance of Gram-negative anaerobic/microaerophilic bacteria; whereas Metagenotype I is dominated by the genus Streptococcus and is associated with a normal esophagus. Overall, the two metagenotypes differ in population but are closely related to each other by lineage.

In the normal human foregut microbiome, the oral microbiota differs from the gastric biota in population as well as phylogeny but both are indistinguishable from the metagenotype I biota in phylogeny. In regards to population the oral biota is similar to metagenotype I biotas, but both differ from the gastric biota. In a pathologic state, the normal gastric biota is similar to the metagenotype II biota in both phylogeny and population. The foregut is dominated by Firmicutes, but the gastric and metagenotype II biotas harbor a large proportion of Gram-negative anaerobes/microaerophils from Bacteroidetes and Proteobacteria.

Metagenotype II may represent an altered metagenotype I biota or an ectopic gastric biota. Since metagenotype II is strongly associated with pathological changes associated with GERD, GERD could represent a microbial ecological disease; thus, a new type of treatment for reflux could include converting the metagenotype II biota to metagenotype I through the use of antibiotics, probiotics, or prebiotics.

Analysis of the Distal Gut in Adults
Ruth E. Ley, Ph.D.
Center for Genome Sciences
Washington University School of Medicine
St. Louis, MO

Dr. Ley presented research on the distal gut microbes found in primates and adults. She found that the dominant species in both the human and primate gut are Firmicutes and Bacteroidetes. These species predominate in stool samples, which are representative of mucosal communities present in the distal gut.

She discussed a study in which the UniFrac Metric was used to compare gut microbial communities based on the fraction of unique branch length using 16S rRNA sequences. A dendogram was developed from this analysis and showed microbial community relatedness across a number of species. Based on microbiota, humans cluster together despite differences in age, sex, nationality, or diet. Human microbiotas are included in the large omnivore group that contains most of the primates. Hindgut fermentors cluster together, as do foregut fermentors. Bears cluster with carnivores, regardless of diet (for example, pandas cluster with the carnivores).

Network analysis of the mammalian gut microbiota involved binning of sequences into Operational Taxonomic Units (OTUs). Each genus-level OTU then is assigned to an OTU-node, and each animal host is assigned to an animal node. Connections between genus and animal modes then are made, and the structure of the network is analyzed. Dietary information also can be incorporated into the network; for example, the ways in which diet shapes mammalian gut communities can be shown by network analyses that cluster omnivores, carnivores and herbivores. Mammalian phylogeny also shapes gut microbial communities, based on clustering of the individual orders (Primate, Carnivora, Artiodactyla, etc.). Whether the animals were zoo-raised or born in the wild did not affect clustering. Diet also did not affect clustering based on order; more links were observed within a taxonomic order than across orders, regardless of diet.

Co-evolution implies reciprocal adaptation of symbionts. Showing genetic changes in a host in response to changes in microbes is difficult; for example, humans may colonize with suitable microbes rather than co-evolve. An alternative approach is to look for co-diversification of communities and hosts and compare bacterial similarities to host phylogenies. Codivergence and adaptation to a novel host, host-swapping, or diversification within a lineage all can produce splits on the microbial phylogenetic tree. Some evidence of co-evolution between mammals and their microbial communities has been found; for example, the microbial community found in bears mirrors phylogeny and is not affected by diet.

Diet and host phylogeny both impact the composition of the gut microbial community. A large degree of host-specificity was found in this analysis, along with evidence of co-diversification in specific clades. Humans are typical omnivores, and modern human lifestyle has had little effect on the human gut microbiota. The ability of microbes to jump between hosts may have helped to ensure the evolutionary success of mammals.

Elderly and Gut Microbiota
Adil E. Bharucha, M.D.
Mayo Clinic
Rochester, MN

Adil Bharucha of the Mayo Clinic evaluated the data on the effects of aging on the intestinal microflora. Although the dominant microbial composition in healthy individuals is constant over time, levels of Bifidobacteria begin to decline after 55 to 60 years of age. This decline is not due to changes in diet or hormone levels, but could be associated with immunological, physiological, or lifestyle factors. These changes could result in increased susceptibility to infections, gastrointestinal problems, or diseases associated with bacteria in the gut. Under these circumstances, functional foods could help protect against entero- and urogenital pathogens.

To determine whether intestinal microbiota are specific to individuals and stable over time, four studies (with a total of 23 subjects) assessed variation in intestinal bacteria over time periods that ranged from 3 months to 20 months, using 16S RNA, 16S rDNA, and fluorescent in situ hybridization (FISH) analysis for major anaerobes. Although these studies suggested that intestinal flora is specific to individuals, there was some variation in intestinal flora over time. A study of 10 individuals (between ages 20 and 50 years) in which samples were taken at 3 and 9 months showed that the total populations of Bifidobacteria and Lactobacilli generally were stable over time. The number of Lactobacilli varied greatly between subjects, however, and also varied between fecal samples in the case of one subject. This study also identified 29 strains of Lactobacilli and found that each subject had a unique collection of Lactobacilli strains and that the composition of Lactobacilli strains varied over time in most subjects. Each subject also had at least one Bifidobacteria strain that was found in both fecal samples, suggesting that some strains persisted in the intestinal tract for at least several months.

The effects of aging on intestinal flora also have been analyzed. Six studies with a total of 227 subjects examined the effects of age on intestinal flora in feces using cultures and molecular techniques. Some studies included only healthy subjects; others included subjects who were hospitalized or taking antibiotics. Older subjects were found to have decreased numbers of Bifidobacteria and less diversity in their intestinal microflora. Increases in Enterobacteria and some changes in Lactobacilli strains also were observed. These studies were small and not statistically significant; the variations observed could be due to interindividual variability. Additionally, bacterial flora mostly were assessed by culture, and only a minority (20%) of the intestinal flora can be cultured; 16S rRNA is a preferable phylogenetic marker for diversity analysis. Interindividual variability in bacterial flora could be due in part to lifestyle factors; for example, in one study, older individuals with high Bifidobacteria counts tended to lead healthier lifestyles.

Interindividual differences in intestinal flora during infancy also could explain differences observed in older people. Intestinal colonization that occurs a few hours after birth is influenced by factors such as birth by Caesarean section, antibiotic use, formula versus breast feeding, and hospitalization. One study of fecal samples from 1,032 infants found that Bifidobacteria counts were lower in infants born by Caesarean section or exposed to antibiotics or miconazole. It currently is unknown whether differences in intestinal flora in the elderly can be attributed to the persistence of differences observed in infancy because longitudinal studies have been performed only between birth and 6 months of age. Such a longitudinal study suggests that differences in Bifidobacteria resolve, but differences in B. fragilis persist at 6 months of age. A study of 90 children (5-13 years of age) from three European countries found no relationships between microbial composition and age, gender, geography, or breast feeding.

The effects of aging on the GI tract itself also may affect the intestinal flora. It has been proposed that reduced gastric secretion, intestinal dysmotility, and duodenal diverticulosis predispose to bacterial overgrowth and malabsorption in the elderly. The relationship between bacterial overgrowth and malabsorption is not, however, definitive. Bacterial overgrowth is diagnosed by breath tests, which are insensitive for detecting overgrowth in elderly people with hypochlorhydria. Increasing use of PPIs is replacing H. pylori infection as a cause of hypochlorhydria. Antibiotic use also is higher in the elderly and affects intestinal flora. Although there are many potential consequences of age-associated alterations in intestinal flora, such as impaired colonization resistance and systemic infections, antibiotic-associated and Clostridium difficile diarrhea, allergies, cancer, and immunosenescence, there is little direct evidence of these relationships.

Probiotics appear to be effective in preventing antibiotic-associated and C. difficile diarrhea and may improve immunity in the elderly. One study found that nutritional supplements decreased the incidence of respiratory, GI, skin, and genitourinary infections in the elderly, although there was no effect on several immune system markers. Treatment with B. lactis HN019 also enhanced phagocytosis and cell lysis in the elderly; these effects were more pronounced in people with poor baseline immunity.

In summary, there is some evidence that intestinal flora are specific to the individual and stable over time in adults. The elderly have fewer good (Bifidobacteria) and more harmful (Enterobacter) microorganisms in their intestines but it is unclear whether these observations reflect a cohort effect, association with aging, or association with conditions more prevalent among the elderly. Hospitalization and antibiotic use have been observed to modify the intestinal flora in the elderly, but the effects of these changes on GI physiology are not well understood. Although a few studies have suggested benefits, little is known about the effects of probiotics in the elderly.

Animal Models of Gastrointestinal Function
Simon P. Hogan, Ph.D.
Cincinnati Children's Hospital Medical Center
University of Cincinnati College of Medicine

Food anaphylaxis is a life threatening acute inflammatory reaction that leads to approximately 150 to 200 deaths per year in the United States. The clinical affects of food allergies include elevated levels of antigen specific IgE, increase in intestinal mast cell and eosinophil levels, increased levels of mast cell derived mediators such as vasoactive amines and proteases. These mast cell derived mediators are thought to induce the intestinal symptoms including abdominal pain, diarrhea, and intestinal dysmotility and extra intestinal manifestations including hypertension, urticaria, and intravascular leakage.

Exposure to food antigens in non-allergic people generally results in oral tolerance. Tolerance arises when food antigens promote a T regulatory response which suppresses immunological activity and the induction of allergic immune reaction. In contrast, in atopic allergic individuals, exposure to food antigens leads to the arise of a CD4⁺ Th2 response. CD4 T cells generate Th2 cytokines including IL-4, IL-5, IL-13, and IL-9. These cytokines lead to the subsequent activation of effector cells including mast cells, eosinophils, and IgE-secreting cells.

One important Th2 cytokine in the induction of food allergic reaction is the cytokine IL-9. IL-9 is derived from CD4 Th2 cells and regulates a number of functions closely associated with allergy including the regulation of mast cells and eosinophils. IL-9 enhances stem cell factor (SCF) dependent growth of mast cells progenitor cells and is also a potent mast cell survival factor. IL-9 can also enhance IL-4 mediated production of IgE and regulates the expression of FcεRI α-chain on mast cells. Furthermore, it enhances the MCP-1 and MCP-2 expression and promotes histamine release from mast cells.

To begin to delineate the role of IL-9 in the intestinal anaphylaxis, we employed an experimental model of oral antigen induced anaphylaxis in mice deficient in the cytokine IL-9. Treatment of WT mice with oral antigen leads to the induction of oral antigen induced anaphylaxis which is associated with diarrhea formation, mast cell activation, and elevated levels of IgE. Notable, mice deficient in IL-9 have attenuated anaphylaxis. These studies suggest that IL-9 plays an important role in experimental anaphylaxis.

To further assess the role of IL-9 in food allergy, we generated a mouse that over expresses IL9 in the intestine. These mice presented with intestinal mastocytosis, increased mast cell activation, increased intestinal permeability, and increased susceptibility to oral antigen-induced anaphylaxis. Collectively, these studies demonstrate an important role for IL-9 and mast cells in intestinal anaphylaxis.

Session III: Impact of Disease on Gut Microbiota
Session Chair: Martin J. Blaser, M.D.
Department of Medicine
New York University School of Medicine
New York, NY

Before introducing the speakers in the session, Martin Blaser stated that microbes in the human intestine are ancient, and signals provided by both the microbes and the human host likely led to co-evolution of the colonizing microbiota and the host. Two potential models may have led to this co-evolution. In the first, signals between the host and microbiota are coordinated. In the second, signals are not coordinated, leading to an evolutionary "arms race." It is likely that the persistence of the microbiota in the host required some degree of coordination and cross-signaling; perturbation in this signaling may lead to disease.

H. pylori was the dominant microorganism present in the gastric microbiome, but it is disappearing, which represents a major microbial ecological shift with important implications. For example, H. pylori is intimately associated with epithelial cells and affects cyclin D1 expression, which mediates cell cycle progression. H. pylori binds to fucosylated Lewis antigens on epithelial cells and produces a signaling molecule (CagA) that is injected into the epithelial cells. CagA is tyrosine phosphorylated by host cell kinases, and then upregulates MAP kinases, and then transcription factors such as c-Jun and c-Fos, leading to transcription of cyclin D1 in epithelial cells. H. pylori also affects secretion of MMP-1 in gastric epithelial cells, depending on its cagA genotype. Early ERK activation is CagA-independent; late activation is CagA-dependent, and in this way, H. pylori affects levels of matrix metalloproteinase-1 (MMP-1) secretion. These are two examples about how the nature of the microbiota affects important signals to host cells, with resulting changes in the cells' phenotypes.

Necrotizing Enterocolitis
Cathy Hammerman, M.D.
Hebrew University
Shaare Zedek Medical Center
Jerusalem, Israel

Dr. Hammerman said that ninety percent of Necrotizing Enterocolitis (NEC) cases occur in preterm infants and between 4 and 20 percent of low birth-weight infants will be diagnosed with NEC. The mean age at diagnosis is approximately 20 days after birth in infants born at less than 30 weeks' gestation. Mortality from NEC is between 20 and 50 percent and increases as gestational age decreases. She also mentioned NEC survivors are at risk for future neurological and developmental impairments.

Introduction of enteral feeds is delayed for premature infants, which reduces gut motility and leads to intestinal stasis, which may permit pathogens to multiply. Many preterm infants also are treated with broad-spectrum antibiotics that may disrupt normal intestinal colonization. These factors contribute to delayed and abnormal gut colonization that may favor pathogenic microorganisms. Preterm infants also have weakened intestinal barrier function, characterized by impairment at tight junctions, which can adversely affect peristalsis and mucin production. Mucin has beneficial functions, including lubrication, mechanical protection, and protection against the acidic environment provided by gastric and duodenal secretions. Mucin also aids in the fixation of pathogenic bacteria, viruses, and parasites. The degree of protection conferred on the GI tract by mucins relates in part to the maturity of the mucins. Normal colonization is necessary for normal gut development. Colonization with pathogenic microorganisms that reach an impaired intestinal barrier can activate an inflammatory cascade that is upregulated and excessive in premature infants. This leads to further tissue necrosis, perforation, and eventual death.

Probiotics have been considered as a potential treatment or preventive for NEC. Probiotics are well tolerated by premature infants and will colonize the premature gut. An early study found that enteral supplementation with Bifidobacteria decreases formation of NEC-like lesions in a neonatal rat model. Probiotics may help normalize the gut microbiota and prevent NEC. Potential benefits of probiotics include increased colonization with favorable microflora, competition with pathogenic bacteria for binding sites and nutrients, improvement in mucosal barrier function, decreased gut permeability, promotion of fermentation of lactose, and production of antimicrobial substances.

Several small studies have tested the effects of probiotics on NEC incidence. The combinations of species used include Lactobacillus acidophilus and Bifidobacterium infantis (two studies); Lactobacillus GG alone; and B. infantis, Streptococcus thermophilus, and Bifidobacterium bifidus. Three of the four studies showed a significant decrease in NEC incidence. The study that did not show a significant decrease was a multicenter study, using only a single organism (Lactobacillus GG), and had a low baseline incidence of NEC. A meta-analysis of these studies showed a significant decrease in NEC incidence associated with the use of probiotics. Probiotics use also was associated with a decrease in NEC severity. A meta-analysis of eight studies of probiotics and NEC found a significant decrease in NEC incidence and no impact on mortality or sepsis. Long-term follow up studies to clarify the benefits of probiotics for NEC include comparison studies of different strains of probiotics, determining whether there is an additive or synergistic effect of multistrain probiotic cocktails, determining the best time to begin treatment and optimal duration, dosing and safety studies, and determining which infants to treat.

Effects of Probiotics in Mouse Models and IBD
Fabio Cominelli, M.D., Ph.D.
University of Virginia Health System
Charlottesville, VA

Dr. Cominelli presented information on pouchitis, a condition that develops after surgery for ulcerative colitis (UC). After reconstruction of the anus using the small intestine, 5 to 10 percent of patients develop inflammation that resembles UC. Probiotics have some efficacy in treating this condition. There is good evidence that treatment with VSL#3 after antibiotic-induced remission can maintain remission and prevent pouchitis. There also is evidence that probiotics can help treat active UC. Although probiotics may be as effective as mesalamine for maintenance of remission, relapse rates in probiotic/control groups are near placebo rates. VSL#3 has been tested for treatment of UC only in small, open-label studies. Probiotics are less effective in treatment of active Crohn's Disease (CD) and maintenance of remission. CD patients have surgery approximately every 4 years, and it is difficult to maintain remission after surgery. Probiotics may be most effective in treating CD that is in remission. Probiotics may function in IBD by competitive exclusion of bacterial adhesion or translocation, stimulation of a Th2/T regulatory immune response, secretion of products with antimicrobial activity, enhancement of epithelial barrier integrity, or by increasing T cell apoptosis.

Dr. Cominelli also explained that probiotics may help maintain gut homeostasis and induce gut health by stimulating epithelial innate immune responses. The SAMP1 mouse serves as a model of CD. This mutation arose spontaneously without genetic or immunological manipulation. The phenotype of the SAMP1 mouse closely resembles human CD in terms of disease location and histological features. Penetrance is nearly 100 percent by 10 weeks of age, and the phenotype persists for up to 80 weeks. Certain genetic and environmental factors are required for the pathology to occur, and the mice also must have functional T cells and cytokines/adhesion molecules. The SAMP1 phenotype also is responsive to CD therapies such as steroids or anti-tumor necrosis factor (TNF) agents. The SAMP1/Yit/Fc mouse spontaneously develops CD-like ileitis. Histologically, infiltration of lymphocytes is observed and mice develop perianal disease with fissures, rectal prolapse, and fistulas. These mice also develop fibrostenotic strictures in their intestines. No disease is observed until 5 weeks of age, and then an inductive phase develops, followed by a chronic phase.

There is evidence of a defect of the innate immune system in IBD. Mutations in NOD2 have been observed in humans with CD. Additionally, mice deficient for NFκB, IL1/TNF, and IL17/IL18 are more susceptible to DSS-induced colitis. Experimental ileitis (perhaps resembling CD) is characterized by two distinct immunological phases, with epithelial barrier dysfunction perhaps constituting the primary defect in both experimental ileitis and IBD. Treatments that have the potential to stimulate the innate immune system, such as probiotics, also are effective in IBD treatment.

VSL#3 contains eight species: four are Lactobacilli (casei, plantarum, acidophilus, and delbrueckii), three are Bifidobacteria (longum, breve, and infantis), and one is a Streptococcus species (salivarius thermophilus). These microorganisms are present at a concentration of 300 billion per gram. VSL#3 shows synergism and reduced toxicity and has been shown to be beneficial in human and animal studies. If given before the onset of symptoms, VSL#3 can prevent ileitis in SAMP1 mice. VSL#3 suppressed inflammation in 5 out of 11 treated mice and provided complete protection against ileitis in the responsive mice. This compound also restored normal permeability in the mice and increased and gene expression of TNF IkBα. VSL#3 did not, however, treat established disease in SAMP1 mice. Probiotic treatment with high-dose probiotics (such as VSL#3) may be effective in maintaining "permanent" remission or may prevent disease in individuals predisposed to CD by stimulating the innate epithelial immune system.

Metagenomic Studies of the Gut Microbiome and Role in Energy Balance
Jeffrey Gordon, M.D.
Center for Genome Sciences
Washington University School of Medicine
St. Louis, MO

Jeff Gordon explained that variation in the gut microbial ecology in humans may affect the efficiency of harvesting nutrients and energy from food, thus implying a role for the microbiome in obesity and malnutrition. For example, mice that are genetically obese because they are homozygous for a null allele of the leptin gene (ob/ob mice) have a shift in their gut microbial ecology. Using 16S rRNA sequencing, ob/ob mice were found to have an increase in the ratio of Firmicutes to Bacteroidetes species compared to their lean +/+ or ob/+ littermates. This shift in proportional representation was observed across virtually lineages in both the Firmicutes and Bacteroidetes. A study of 12 obese humans found that Bacteroidetes species increased as the participants lost weight when placed on either fat- or carbohydrate-restricted low calorie diets. These findings demonstrate a possible relationship between host adiposity and gut microbial ecology. This microbial shift may be a biomarker of host adiposity or may be a contributory mechanism that influences energy balance. Metagenomic studies of the cecal microbiomes of ob/ob and +/+ showed an enrichment of genes involved in the degradation of otherwise indigestible polysaccharides. Bomb calorimetry of cecal matter showed a small but statistically significant decrease in the amount of energy in the cecal matter of ob/ob mice. Gut microbial community transplant experiments were performed in the cecal contents from lean +/+ or ob/ob mice and were placed into GF wild type recipients. GF mice that received cecal contents from ob/ob mice showed a larger increase in adiposity; there was no difference in food consumption between the two groups of transplanted mice. Thus, changes in the microbiota observed in genetically obese mice are associated with an increased capacity for energy harvest. Other studies in gnotobiotic mice have shown that the microbiota also affects host genes that determine how absorbed energy is metabolized and stored.

To investigate the relationship between diet, the gut microbiome, and energy balance, the Gordon lab studied two groups of C57Bl/6J +/+ mice: those fed a standard polysaccharide-rich chow diet, and those fed a high-fat/simple-sugar prototypic Western diet. The latter group develops obesity. 16S rRNA sequencing of cecal contents showed that mice on the Western diet had an increased relative abundance of Firmicutes and a division-wide decrease in the Bacteroidetes. The increase in Firmicutes was primarily attributed to bloom in the Mollicutes class of Firmicutes. This bloom did not depend on either a functional innate or adaptive immune system; the increase in the Mollicutes class also was observed in conventionally-raised RAG1 knockout mice (who lack mature T and B lymphocytes) and MyD88 KO animals. Transplantation of the cecal contents of mice on a Western diet into GF wild-type recipients resulted in a greater increase in adipose tissue than did transplantation of a microbiota from lean mice fed a standard chow diet. Switching mice from a Western to a carbohydrate- or fat-restricted diet halted the increase in adiposity and reduced the Mollicutes bloom.

The Mollicutes lineage of the Firmicutes normally are found in low abundance in the mouse (and human) gut microbiota. However, members represented up to 60 percent of bacterial species in animals fed the Western diet. The Mollicute lineage that bloomed in mice with diet induced obesity (DIO) could not be cultured. Metagenomic sequencing of the cecal microbiomes of mice with DIO and controls fed a standard chow diet, plus sequence analysis of the genome of Eubacterium dolicum, a cultured member of the human gut microbiota that is a close relative of the lineage that bloomed in mice with DIO revealed features in the Mollicute lineage that accounts for its success in the Western diet nutrient milieu. Together, these findings indicate that DIO-associated microbiota has an increased ability to promote host adiposity - further illustrating the interrelationship between diet, gut microbial ecology, and host energy balance.

Additional studies related to the role of the microbiome in energy balance are planned or underway which includes metagenomic analysis of the fecal microbiotas of young adult female monozygotic and dizygotic twin pairs (concordant and discordant for obesity or leanness). One hundred cultured representatives of the major phylotypes in the human gut microbiota also are being sequenced and the data will be used to predict their niches (professions). The role of Archaea and energy balance also is under investigation. Colonization of germ-free mice with Bacteroides thetaiotaomicron with or without Methanobrevibacter smithii (the principle archaeon in the human gut microbiota) was performed. GeneChip analysis showed that the pattern of B. thetaiotaomicron foraging for polysaccharides in the distal gut is altered in the presence of M. smithii. Expression of B. thetaiotaomicron genes involved in the degradation of fructans (commonly used as food sweeteners) was increased. Co-colonization also resulted in increases in SCFA production, increased de novo hepatic lipid production, and increased host adiposity compared to mice colonized with a single microbial species. Sequencing the M. smithii genome revealed ways in which it has adapted to life in the gut and includes the capacity to consume a variety of bacterial metabolites. M. smithii also have extensive machinery related to surface variation and genomic evolution.

The talk ended with a discussion of whether the gut microbiota and microbiome may also have a role in malnutrition. Studies by others have shown that siblings in a family may differ in their tendency to develop states of severe malnutrition (e.g., kwashiorkor), despite consumption of the same diet. Analysis of the gut microbiota of children with kwashiorkor and their siblings and mothers, prior to and after dietary intervention, may help to determine whether the gut microbial ecology can predispose individuals to this disease. These types of studies are currently being pursued by members of the Gordon lab and their collaborators.

Session IV: Factors Influencing Gut Microbiota
Session Chair: Mary Ellen Sanders, Ph.D.
Dairy and Food Culture Technologies
International Scientific Association for Probiotics and Prebiotics
Centennial, CO

Mary Ellen Sanders commenced session four by introducing the presenters in the session. She also noted that session five continued consideration of the normal intestinal microbiota and how it can be affected by probiotics, prebiotics, nutritional status, antibiotic use, inflammatory disease and the special situation with the premature infant [Previous talks in Session II considered the case of gut microbiota and adult mammals (Ley) and the unique situation in the elderly (Bharucha)].

Factors Influencing Gut Microbiota: Probiotics
Mary Ellen Sanders, Ph.D.
Dairy and Food Culture Technologies
International Scientific Association for Probiotics and Prebiotics
Centennial, CO

Dr. Sanders defined probiotics as live microorganisms that when administered in adequate amounts provide benefits to the host. She went on to explain that a strict interpretation of the definition for probiotics does not require that its impact on health be exerted by direct influence on the gut microbiota. However, it is likely that in many cases, health effects due to probiotics are so mediated. It is notable, however, how little is known about the impact of probiotics on the dominant groups of colonizing microbes. This is largely because only recently have techniques become available that will enable such analysis.

In general, the fed probiotic strain can be isolated from fecal samples, but is not retained as part of the colonizing microbiota. The fed strain is usually not recovered until 1-4 weeks after feeding has stopped. This is true for studies conducted with either fecal or mucosal samples, although longer term recovery from mucosal samples has been documented. Studies on the impact of probiotics on certain subgroups of colonizing bacteria have in general shown a transient increase in the genus of the fed bacterium. Although different studies have shown statistically significant changes in certain other subgroups of bacteria, these changes appear to not be consistent between studies and likely are dependent on the probiotic strain fed, dose, and the methods used to characterize the groups (culture-based techniques, FISH, PCR).

Whether probiotic strains can change fecal biochemical parameters is another issue related to efficacy studies. Bacterial metabolites that form in the large intestine depend on the characteristics of the bacterial populations, transit time through the colon, and substrate availability. Some changes have been observed, including increases in SCFAs, which are beneficial to colonocytes, along with decreases in amines, ammonia, phenolic compounds, and thiols, which have been implicated in the pathogenesis of diseases such as cancer.

Thus, the effects of probiotics depend on strain, dose, and methods (culture, FISH, PCR) used to characterize the microbiota. During probiotic feeding, increases in the fed strain in feces are observed, but this increase is lost by 1 to 4 weeks after feeding ends. Definitive studies on probiotic impact on dominant bacterial communities are lacking. Changes in biochemical parameters, such as SCFAs, ammonia, amines, phenols, and some enzymatic activities have been observed. Probiotics appear to impact pathogens and decrease infectivity and toxicity and thus could be useful in decreasing morbidity and health care costs if used preventively.

Factors Affecting the Microbiota of the Preterm Infant
Frank R. Greer, M.D.
University of Wisconsin School of Medicine
Madison, WI

Dr. Greer from University of Wisconsin School of Medicine highlighted that little is known about bacterial colonization of the intestine of premature infants. The piglet intestine is similar to that of a human infant, and piglets have been used as model organisms to study this issue. Studies in piglets found that delayed feeding, as commonly occurs for preterm infants, resulted in reduced jejunal mass, villus height, and crypt depth. Breast milk was believed to result in differences in intestinal colonization compared to formula, but in at least one study of preterm infants, no differences in proportional representation of microflora species were found by 30 days of age. Delay in initiating enteral feeding and antibiotic use appeared to have larger effects, primarily a paucity of bacterial species in the gut, especially Lactobacillus and Bifidobacteria.

Potential influences on the infant microbiota prior to delivery include maternal diet, maternal use of probiotics or prebiotics, duration of rupture of membranes, and maternal antibiotic treatment. After delivery, the microbiota is affected by the mode of delivery, local environment (home versus hospital birth), gestational age and birth weight, acuity of illness, type of feeding (breast milk or formula), delay of oral feedings, and antibiotic treatment. Steroids, PPIs, probiotics, and prebiotics also may influence the infant microbiota. A study of 1,032 infants at 1 month of age (11 were born at less than 32 weeks' gestation) found that infants fed formula had a somewhat higher prevalence of E. coli, C. difficile, B. fragilis, as well as Lactobacilli compared to infants fed breast milk. Maternal diet, duration of rupture of membranes, and maternal antibiotic use had no effect on the biota. Infants delivered by Cesarean section had a slightly lower prevalence of Bifidobacteria, but the presence of older siblings slightly increased the prevalence of Bifidobacteria. Premature infants and infants born by Caesarean section also had a slightly elevated prevalence of Clostridia, which may be related to time spent in the hospital. Antibiotic use increased the prevalence of B. fragilis.

Probiotics could affect the infant microbiota by reducing the presence of pathogenic species in the bowel and increasing colonization with desirable microflora. Probiotics also could improve the effects of enteral nutrition, resulting in reduced dependence on parenteral nutrition. Probiotics also increase gut mucosal barrier function to prevent translocation of bacteria and bacterial products, upregulate or modulate immune responses, and increase anti-inflammatory cytokine production. Specific probiotics studied for their effects on the infant GI tract microbiota include Lactobacillus GG, Bifidobacterium breve, Bifidobacterium subtilis, Bifidobacterium lactis, and Saccharomyces boulardii. A study of 91 infants born at 28 weeks' gestation in which half were given B. breve ,once oral feeding began, found that colonization with B. breve improved the growth of the infants. Only four infants who received treatment were not colonized, and all four had received antibiotics for an extended period.

Prebiotics such as breast milk, inulin, galacto-oligosaccharides, and fructo-oligosaccharides have also been tested for their ability to influence the infant microbiota. Breast milk contains lactose, the fermentation of which acidifies the gut and stimulates growth of some bacteria while inhibiting growth of others. Breast milk also contains fatty acids, which can destroy some viruses and pathogens such as Giardia; lactoferrin, which is active against pathogenic bacteria; and glycol-conjugates, including oligosaccharides that contain lactose. Prebiotics can enhance colonization by Bifidobacteria and Lactobacillus, inhibit specific pathogens, and modulate the immune system through production of intestinal fermentation products. There are, however, few RCTs that have tested prebiotics on preterm infants. One small study found that infants born at 31 weeks' gestation who were fed an experimental formula containing a mixture of 90 percent galacto-oligosaccharide and 10 percent fructo-oligosaccharide, had significantly increased levels of Bifidobacteria. Based on the results of available studies, the most productive strategy for altering the microbiota of preterm infants may be to curtail antibiotic use and begin early feedings with human breast milk, which is the ultimate prebiotic.

Dietary Inputs and the GI Microbiota
Randy Buddington, Ph.D.
University of Memphis
Memphis, TN

Dr. Buddington emphasized that a diverse gut microbiota is a more stable microbiota. In the developing gut, appropriate colonization maintains a balance of diverse microorganisms that prevents overgrowth by potential pathogens. He acknowledged that the gut microbiota is influenced by dietary inputs, which affect bacterial strains and also intestinal function. Competition and cooperation between the bacteria and the host also are involved in maintaining the microbiota. Digestive problems such as pancreatitis and gall bladder disease can disturb the balance of the microbiota. Birth (when nutrients begin to be received orally versus through the placenta), weaning, starvation (including parenteral nutrition and hibernation), diminished function (short bowel syndrome and gastric bypass procedures), and extreme diets all are examples in which interactions among diet, GI function, and GI bacteria affect the microbiota.

Studies using piglets born at 92 percent of gestation found that providing colostrum during total parenteral feeds and when an infant was transitioned to full enteral feeds decreased the incidence of NEC. The rapidity of onset of NEC and lack of NEC during total parenteral nutrition (TPN) implies that it may not be due solely to infection with a pathogen. Instead, SCFAs or other compounds produced by bacteria may damage the premature epithelium and render it vulnerable to pathogens. The composition of formula fed to preterm infants also affects NEC incidence. Nearly 100 percent of infants fed a formula containing polycose as its sole carbohydrate developed NEC. Polycose may be digested poorly by these infants and may be used by bacteria in the gut to produce damaging products.

Parenteral nutrition and hibernation provide examples of the effects of diminished input. TPN effectively starves the gut. Patients receiving TPN have fewer aerobes and more anaerobes (primarily due to an increase in Clostridia) in their gut microbiota; differences in prevalence of Gram-negative bacteria also have been observed. Diminished input appears to decrease the abundance and diversity of species present in the gut. Hibernating frogs have fewer bacteria in the mucosa and gut and also a higher percentage of aerobic organisms.

Different sources of carbohydrates also affect the gut microbiota. In rats, feeding with starch as the primary source of carbohydrates results in a higher proportion of aerobes in the gut than feeding with sucrose as the primary carbohydrate. Sucrose is digested more quickly in the proximal bowel, resulting in fewer carbohydrates in the cecum, compared to the starch diet. This deprives downstream bacteria of a carbohydrate source; these bacteria thus digest other available compounds, and the metabolic characteristics of the gut are changed. If obese humans are placed on a carbohydrate-restricted diet, production of SCFAs and butyrate by fermentation of carbohydrates in the gut is decreased, resulting in changes in the densities of fecal bacteria. Thus in humans, changing the diet can result in changes in the numbers and proportions of gut microbial species, which in turn may affect host physiology. Dietary input is needed for proper function of the human GI tract and proper composition of the gut microbiota. Bacteria present in the gut affect both digestive and other physiological functions. Probiotics and prebiotics may be useful tools for managing GI health.

Factors Affecting Gut Microbiota: Antibiotic Use
Josef Neu, M.D.
University of Florida
Gainesville, FL

Dr. Neu maintained that manipulating the microbiota of the newborn can have a significant impact throughout life. Antibiotic use by pregnant women, which is common in cases of premature labor, prolonged rupture of membranes, and in pregnant women who test positive for Group B streptococcus, can affect the newborn microbiota. Ante- and peripartum antibiotic treatments are useful, but can increase the infants' susceptibility to coliform infections and NEC, particularly in premature infants. Breast milk is a source of beneficial microorganisms. Analysis of breast milk, the mother's blood and feces, and the infant's feces found the same species of microorganisms, suggesting that microbes may migrate from the mother's intestine to the breast milk and may represent microbes transferred to the infant through the milk. Transfer of bacteria through milk may be a means by which maternal microbes colonize the neonatal gut; thus, maternal antibiotic use could affect the colonization process.

Of the 10 most commonly used medications in neonatal intensive care units (NICUs), four are antibiotics. Dr. Neu affirmed that Ampicillin and aminoglycosides are given to more than 90 percent of infants born at less than 30 weeks' gestation. A Cochrane review of five trials of oral aminoglycosides to reduce NEC incidence performed in the late 1970s to early 1980s found that antibiotic used reduced the incidence of NEC significantly, but also resulted in increased incidence of colonization with resistant bacteria. There also was concern that this treatment could lead to delayed or altered colonization by serving as an "unnatural" selection pressure. Correct initial colonization of the intestine is crucial because it is one of the most important immunological exposures faced by the newborn infant. Crosstalk between bacteria and the cells of the epithelium induces gene expression in both the epithelium and the immune system, and niches are formed as part of a potentially long-lasting biofilm within the luminal glycocalyx. A comparison of infants born by Caesarean section versus vaginal delivery 7 years after birth found that the microbiota of those born by Caesarean section, which deprives the infant of contact with the maternal/vaginal microbiota, were characterized by a lack of strict anaerobes and the presence of facultative anaerobes such as Clostridia.

To analyze the effects of neonatal antibiotic treatment on GI tract development, rats were treated with Clamoxyl, which significantly reduced bacteria levels - particularly Lactobacillus - in the colon. Affymetrix gene array analysis showed that 10 to 30 percent of genes undergoing maturational changes showed modulation by the antibiotic such that their normal pattern of maturation either was accelerated or slowed. Major histocompatibility complex genes, which are required for tolerization to luminal antigens, were significantly affected. Expression of mast cell-associated genes was increased. To determine the effect of duration of initial empirical antibiotic use on risk of NEC, data about first antibiotic course from a cohort of extremely low-birth weight infants admitted to Neonatal Research Network centers between 1999 and 2001 were analyzed. Factors associated with prolonged empirical therapy were identified, and associations between therapy duration and NEC or death were evaluated. For every day of intravenous antibiotic use, the risk of NEC increased by 7 percent. Prolonged initial antibiotic use also significantly increased NEC incidence.

The long-term effects of biofilm formation by the initial colonizing bacteria may be a concern. Nonculture-based metagenomic techniques are needed to broadly analyze the microbes present in the biofilm. An analysis of four clindamycin-treated patients and four controls found that there were persistent long-term impacts (2 years after treatment). The increasing use of antibiotics over the last 50 years has decreased the incidence of several infectious diseases, but incidences of conditions such as CD, multiple sclerosis, and type 1 diabetes have increased. The increased use of antibiotics, as well as changes in agricultural practices and the environment, could be affecting the incidence of these conditions. A study of BBDP rats (which spontaneously develop diabetes) and BBDR rats (which require a stimulus such as viral infection to induce diabetes) found differences in the intestinal microbiota of these animals. Antibiotic treatment also decreased the incidence of diabetes. This appears to be a species-specific effect, because treatment of non-obese diabetic mice with doxycycline increased the incidence of diabetes. Together with evidence from human studies, this work suggests that manipulating the intestinal microbiota early in life with antibiotics may have significant short- and long-term consequences. It also implies that the nonspecific use of antibiotics in pregnant or lactating women and in infants should be reconsidered.

Session V: New Developments in Prebiotic and Probiotic Research
Session Chair: James Versalovic, M.D., Ph.D.
Baylor College of Medicine
Houston, TX

The final session summarized recent advances in prebiotic and probiotic research. Dr. Versalovic from Baylor College of Medicine initiated session five with the introduction of the speakers and he mentioned that the GI tract displays differential distribution of bacteria, and prebiotics and probiotics can affect this distribution. Dr. Versalovic pointed out that probiotic-fed piglets retain supplemental Lactobacillus reuteri only in the proximal gut even in the presence of endogenous L. reuteri, although fecal counts are the same, which demonstrates how analytical methods can impact data interpretation. Bacteria present in the gut modify the immunological and physiological characteristics of this organ by secreting mediators. Prebiotics affect bacterial growth and gene transcription, thus also impacting immunological and physiological parameters.

Probiotics-Mediated Regulation of Immune Signaling Pathways
James Versalovic, M.D., Ph.D.
Baylor College of Medicine
Houston, TX

The session chair, James Versalovic, described classes of probiotics that have different immunomodulatory capacities. In a study to determine the effects of probiotics on immune signaling pathways, human-derived probiotic L. reuteri strains were tested for the ability to suppress TNF and block human immune signaling pathways. L. reuteri can both enhance and suppress cytokine production. Some strains of L. reuteri inhibit pro-inflammatory cytokines such as TNF, thus ameliorating chronic inflammation in the gut. Other strains stimulate the innate immune system and signaling pathways that lead to upregulation of TNF and other proinflammatory cytokines. Immunostimulatory L. reuteri have been effective in preventing acute inflammation in Salmonella-induced enteritis in pigs, and rotavirus infection in mice. Strains can be distinguished by chromosomal DNA profiling using repetitive element PCR; and strain genotypes can be linked to immunomodulatory phenotypes . Strain-specific activation of NFκB also is observed among strains that modulate TNF production. In the presence of the TLR2-agonist Pam3Cys, NFκB may not be activated by L. reuteri; in contrast, AP-1 activity may be suppressed, thus leading to downregulation of TNF independent of NFκB signaling .

Immune-Enhancing Effects of Prebiotics
Bernhard Watzl, Ph.D.
Institute of Nutritional Physiology
Federal Research Centre for Nutrition and Food
Karlsruhe, Germany

Dr. Watzl described immunomodulatory effects of prebiotics. Intestinal bacteria may directly modulate local and/or systemic immunity or may have indirect immunomodulatory effects, such as generation of SCFAs or immunogenic carbohydrate structures through bacterial metabolic activity. Prebiotics are selectively fermented and allow specific changes in the composition or activity of the GI microflora that are beneficial for the health of the host. Examples of prebiotics include inulin, fructo-oligosaccharides (FOS), short-chain FOS, lactulose, raffinose, galacto-OS, and nigero-OS. These compounds occur naturally in foods such as bananas, leeks, garlic, onions, and chicory, or are enzymatically produced from simple carbohydrates. Differences in diet thus may result in significant differences in the microflora. Adding prebiotics to the diet promotes the production of immunomodulatory mediators by the microflora.

The fermentation of prebiotics increases the production of the SCFA butyrate. Butyrate colon and serves as an energy source for colonocytes. It promotes colonocyte proliferation and differentiation (promotion of a normal phenotype), cell cycle arrest and apoptosis of colorectal cancer cells; alters expression of signaling genes; and has immunomodulatory effects. Butyrivibrio fibrisolvens MTD1 is a strong producer of butyrate. Feeding B. fibrisolvens to mice resulted in a 3.5-fold increase in splenic natural killer (NK) cells and a 3.9-fold increase in NK T cell numbers, along with suppression of aberrant crypt foci in the colon and rectum. Supplementation of rats with SCFAs by TPN increased NK cell activity of splenocytes in vivo, but did not affect NK cell activity in vitro; this could be due to GI-related effects such as cytokine production in response to SCFAs. Butyrate has been shown to decrease TNF- ß, IL6, and IL1ß mRNA levels in the presence of LPS and also to decrease secretion of TNF-α protein and increase secretion of IL10. Studies in humans have shown that inulin and FOS may modulate activity of gut associated lymphoid tissue (GALT) and reduce tumor incidence and disease activity in UC and CD.

Inulin, FOS, and synbiotics also increased NK cell-like activity and IL10 production by Peyer's patch cells in rats. Modulation of T cell mediated immunity in jejunal intraepithelial lymphocytes of pigs by prebiotics and synbiotics also has been observed. Synbiotic treatment resulted in antitumorigenic activity in a rat model of colon carcinogensis, and it increased NK cell activity significantly in these tumor-bearing rats. A recent study in humans found that consuming the same synbiotic (inulin enriched with oligofructose in combination with L. rhamnosus and B. lactis) had minor effects on the lytic activity of NK cells and other immune parameters measured in immune cells isolated from peripheral blood of colon cancer patients and of patients previously treated for polyps. Overall, immunomodulatory effects of prebiotics seem to be restricted to the immune system associated with the gut.

Interrelationship Between Soy Isoflavone-Metabolizing Phenotypes and Human Genetics
Johanna Lampe
Fred Hutchinson Cancer Research Center
Seattle, WA

Phytochemicals modulate human health, and the gut microbiota is involved in metabolism of phytochemicals found in food; metabolism of phytochemicals varies among individuals. Bacteria in the gut participate in the metabolism of the soy isoflavone daidzein. Daidzein is similar to estrogen, can bind to both estrogen receptor (ER)-α and ER-ß and has estrogenic effects. The major daidzein metabolites produced by human intestinal bacteria are equol (produced in 30 to 50% of individuals) and O-desmethylangolensin (O-DMA), which is produced in 80 to 90 percent of people. Equol is more strongly estrogenic than both daidzein and O-DMA and has been associated with effects such as greater lengthening of the menstrual cycle follicular phase and improved maintenance of bone mineral density in postmenopausal women. Equol producers also have lower mammographic density, and equol production was inversely associated with prostate cancer risk in a study of Japanese men.

The ability to produce equol is affected by the bacteria present in the gut. Factors that may allow a person to harbor equol-producing bacteria include diet, gut function, antibiotic exposure, and genetic factors. Supplementation with isoflavones did not promote equol production in nonequol producers. A higher prevalence of equol producers has been observed among vegetarians and in Asian populations that consume soy. Higher intakes of dietary fiber, carbohydrate, protein, polyunsaturated fatty acids, and alcohol have been observed among equol producers.

A system to phenotype equol and ODMA-producers has been developed that poses little burden to participants and can be used in large, population-based epidemiologic studies. Participants are challenged with soy, and urine is collected after 3 days. Equol and ODMA can be measured in urine and are highly stable. This approach was used to examine familial correlations to determine whether there was a genetic component to equol production. Neither equol- nor ODMA-producer phenotypes were associated with gender, smoking, exposure to animals, or a majority of dietary factors. Some association was observed with age; equol production decreases in older individuals, but it is unclear whether this is a cohort effect or related to changes in the microbiota that occur with age. Equol production was associated with increasing years of education, use of oral contraceptives by women, and consuming a low fat diet. ODMA production was associated with high consumption of fried foods and low consumption of caffeinated beverages and inversely associated with age, height, and body mass index. Intraclass correlations and model fitting from segregation analysis of equol- or ODMA-producer phenotype suggests that a Mendelian autosomal dominant form of inheritance provides the most parsimonious fit with the data. The equol-producing phenotype appears to be fairly stable over time.

Studies have sought to determine if the prevalence of equol-producing phenotypes differs between populations that consume high or low amounts of soy. A higher prevalence of equol-producers is observed among the Japanese compared to Westerners, and higher rates of equol production are observed among vegetarians versus non-vegetarians. A study in which Korean Americans living in Seattle were compared to Caucasians found that the Korean Americans had a higher prevalence of equol-producers and lower prevalence of ODMA-producers. Observational studies thus have found that racial or ethnic differences that may reflect lifestyle differences, level of education, diet (although this is not consistent across populations), and constipation and slower gut transit times may be associated with the equol-producer phenotype.

Efforts have been made to distinguish between daidzein-metabolizing phenotypes using fecal microbial community profiles. Multivariate analysis of terminal restriction fragment length polymorphism (tRFLP) patterns derived from 16S rRNA gene sequences generated using fecal aliquots incubated with daidzein in vitro allowed equol-producing individuals to be distinguished from ODMA-producers. Preliminary data from tRFLP analysis thus suggest that microbial community structure differs by equol-producer phenotype.

Functional Genomic and Metabolic Studies of Host-Microbial Symbiosis in the Mammalian Gut
Justin L. Sonnenburg, Ph.D.
Center for Genome Sciences
Washington University School of Medicine
St. Louis, MO

Dr. Sonnenburg mentioned that at least 400 bacterial species are present in the human gut microbiota with the two dominant bacterial divisions being Bacteroidetes and Firmicutes, which represent 90 percent of gut bacteria. The microbiota has important modulatory effects on the host immune system and metabolism, but little is known about its operating principles. The microbiota likely is affected by perturbations or differences arising in the gut environment, including diet, ingestion of microbes, and host genotype. These variables and their relationship with the microbiota and the host need further investigation. Questions to be answered to facilitate understanding of microbiota function include: how the microbiota is affected by changes in host diet, how changes in community membership affect resident microbes and the host, and how a specific change in host genotype could influence the microbiota.

Justin Sonnenburg discussed the abilities of a model intestinal commensal bacterium Bacteroides thetaiotamicron to affect carbohydrate utilization patterns in gnotobiotic mouse models. B. thetaiotamicron shifts its gene expression profiles towards different carbohydrate utilization patterns depending on changes in diet and local nutrient availability. Fundamental changes in diet alter the functional capacities of commensal bacteria, and these alterations in bacterial metabolism may determine whether host-derived carbohydrates in the mucus layer, for example, are effectively scavenged by intestinal bacteria. The administration of probiotics may fundamentally alter functional capacities of indigenous bacteria. Each of three probiotic strains had a distinct impact on the transcriptome of B. thetaiotamicron and suggests that functional effects on the microbiome may differ dramatically depending on which probiotic strain is administered in vivo. These studies are being extended to examine effects of mixed bacterial species interactions on gene expression profiles of probiotics.

DNRC Director: Dr. Van S. Hubbard || DNRC Home || Contact DNRC

The Division of Nutrition Research Coordination is part of the National Institutes of Health, Bethesda, MD, USA.

Privacy || Disclaimer || Copyright || Credits || Accessibility

Department of Health and Human Services