D Kelly, Rowett Institute of Nutrition & Health, University of Aberdeen, Greenburn Road, Foresterhill, Aberdeen AB25 2ZD, Scotland, UK. E-mail: firstname.lastname@example.org. Phone: +44-1224-437475. Fax: +44-1224-437465.
The healthy human gut supports a complex and diverse microbiota, dominated by bacterial phylotypes belonging to Bacteroidetes and Firmicutes. In the inflamed gut, overall diversity decreases, coincident with a greater representation of Proteobacteria. There is growing evidence supporting an important role for human gut bacteria in mucosal immunity; interactions at the level of both intestinal and colonic epithelial cells, dendritic cells, and T and B immune cells have been documented. These interactions influence gut barrier and defense mechanisms that include antimicrobial peptide and secretory IgA synthesis. The functional effects of commensal bacteria on T helper cell differentiation have led to the emerging concept that microbiota composition determines T effector- and T regulatory-cell balance, immune responsiveness, and homeostasis. The importance of this biology in relation to immune homeostasis, inflammatory bowel disease, and the rising incidence of autoimmune diseases will be discussed. The detailed description of the human gut microbiota, integrated with evidence-based mechanisms of immune modulation, provides an exciting platform for the identification of next-generation probiotics and related pharmaceutical products.
The gut is colonized by a very diverse population of microorganisms that include the three domains Bacteria, Archaea, and Eukarya as well as viruses. Increasingly, more is known about the bacterial communities that colonize the mammalian gut, but relatively less is known about viruses and other less abundant groups, including methanogenic archaea. The increased awareness of the contributions of these less prevalent groups of microorganisms in modulating human physiology is becoming more apparent with the growing number of publications in this area.1–3 This review, however, focuses on the description of the bacterial communities that colonize the human gut and highlights some of the exciting developments that have shed light on how these gut microbes enhance the functionality of the gut and its associated immune system.
Initiation of the gut colonization process starts at birth, when microorganisms from maternal body surfaces and the environment are acquired by the infant.4,5 These early colonization events have profound effects on the overall phylogenetic composition of the gut microbiota and the immune system of the adult.5 During the neonatal period, the microbial community structure is variable and unstable and has been termed “chaotic” in relation to composition and structure.6 Interestingly, this unstable phenotype lasts for a considerable period of time before a stable, adult-type microbial composition emerges.7 This transition has been reported to occur at around 3 years of age.8
Immense progress has been made in the last 5 years in describing the human gut microbiota, mainly as a result of culture-independent molecular investigations.The bacteria that colonize the human gut have now been well described. These belong predominantly to the bacterial phyla Firmicutes, Bacteroidetes, and Actinobacteria.9,10 Under certain disease conditions, however, an increasing presence of proteobacteria has been reported.7,11,12 Microbial diversity, particularly at the phylum level, seems to be consistent across many Westernized populations. Much greater complexity, however, is observed when the microbiota is evaluated at the genus, species, and subspecies levels. The microbiota of an individual comprises around 1,000 different bacterial species, which represent a unique signature for that individual. This variation is explained by the diverse number of factors that influence the type of bacteria able to successfully colonize the human gut, including birth mode, genotype, age, biogeographical location, living environment, diet, and the nutrition and health status of the individual. These factors include both deterministic niche-related inputs and historical environmental events and collectively influence the microbial species present in the gut, particularly during the time of early inoculation and colonization.13 Recent work has also provided some understanding of the importance of less abundant taxa, such as methanogenic archaea and sulfate-reducing bacteria, which colonize the human colonic mucosa and are thought to be metabolically significant as well as relevant to gut health.2,3
Interestingly, although certain bacterial phyla dominate the human gut, the relative percentages of Firmicutes and Bacteroides vary within the human population, and this seems to reflect dietary effects and differences in macronutrient consumption.14,15 For example, compositional differences in the microbiota can be induced by diets rich in carbohydrates and fats; these diets contribute to microbiota shifts that result in increased energy harvest from the diet. Furthermore, transplantation of the microbiota from obese mice results in increased adiposity in lean mouse recipients.16,17 The responsiveness of the microbiota to dietary energy load has also been described in humans.18 Furthermore, the introduction of prebiotic substances can significantly influence key bacterial groups in the gut, such as members of the Firmicutes and Verrucomicrobia, including Akkermansia.19
Diet and nutrition are clearly important determinants of microbial diversity and functionality and have important consequences for health. A recent study evaluating the microbiota of two culturally distinct populations from the United States and Korea has revealed significant differences in microbial diversity at the class, order, and family levels of the fecal microbiota.20 This difference most likely reflects dietary differences between these populations. Alterations of the microbiota related to both overnutrition (obesity) and malnutrition have profound consequences for immune function and human health. The effects of malnutrition still handicap up to a billion people in the world today.21 Nutrient statuses of fatty acids, glucose, and amino acids as well as micronutrients such as vitamins A, D, and E and iron have all been shown to affect immune functions, mainly due to nutrient deficiencies.
Antibiotic usage also influences microbial diversity.22 Ciprofloxacin rapidly and profoundly induces a loss in bacterial diversity within a few days of administration. Following removal of this antibiotic, the microbiota recovers, but to a compositional state distinct from the initial community structure. This effect is exaggerated following several consecutive courses of antibiotic treatment.
Host genetics and environment are also important determining factors of microbiota composition. The microbiota of healthy monozygotic twins is very similar to that of dizygotic twins, suggesting that factors such as environment play a particularly important role in determining microbial composition.6,23–25
GUT MICROBIOTA AND THE EPITHELIAL CELL LAYER
Epithelial cells provide the first line of contact for gut bacteria and physically separate the large diversity of microorganisms that colonize the gut from the internal gut milieu. These cells also maintain a barrier that limits the translocation of moieties between epithelial cells and is composed of tight junctions made up of claudins, occludins, and F-actin.26 (Figure 1a,b). Interestingly, it was recently observed in IL-10KO mice with active colitis that tight junction integrity was compromised, and this appeared to correlate with altered expression of syndecan-1 and syndecan-4 in gut epithelial cells.27 The functional importance of these heparin sulfate proteoglycans in gut barrier function and their role as sensitive markers of barrier integrity is considerable. Their detection, in soluble forms, in the blood following dysregulated expression may increase their potential value as clinical biomarkers of gut function.
Epithelial cells also make contact with cells of the gut immune system and line the lamina propria of the small and large intestines and the Peyer's patch regions; these Peyer's patch sites of organized lymphoid tissues play a critical role in direct antigen sampling from the gut. These regions are considered the inductive sites of the immune system and are specialized areas where immune responses are both induced and regulated. Exposure of immune cells to bacteria has important consequences for gut health: excessive contact promotes exaggerated proinflammatory immune responses, whereas limited bacterial or no bacterial exposure, such as in germ-free conditions, can dramatically impair immune development and function. Hence, controlled and regulated bacterial contact and sampling are required to maintain optimum gut immune functions.
To protect the gut from excessive bacterial exposure, the normal epithelium, particularly in the colon, is protected by a thick and continuous mucin layer that limits and restricts the exposure of gut surfaces to luminal gut microbes. In the ileum, the mucin layer is much more diffuse than in the colon and therefore facilitates more opportunity for microbe-epithelial contact (Figure 1c,d). This arrangement allows only modest microbial antigen access to the epithelium in the ileum and thus prevents immune hypersensitivity responses caused by excessive antigen load (bacterial and dietary). Interestingly, in addition to its function as a physical barrier that limits penetration of antigens and noxious agents, mucin also promotes bacterial binding.28 This is in part explained by the structural composition of mucin, which represents a complex supply of carbohydrate-rich glycolipids and glycoproteins. The glycosylation patterns of the gut epithelium and associated secreted mucins have stimulated great interest for years. This is due to the importance of post-translational modification in modifying the functional properties of glycosylated enzymes, receptors, and transporters in the gut.29–31 The epithelium's N-linked and O-linked glycosylation patterns, as well as its mucin content, are altered during development and with early feeding methods.32,33 The carbohydrate structures of mucin provide important binding sites for gut bacteria30 and represent rich sources of nutrients.34,35 Recently, deficiency of gut O-linked glycans has been found to trigger colitis.36
In addition to mucin, IgA and IgM, which are derived by both T-cell-dependent and T-cell-independent activation of B cells and their subsequent differentiation to immunoglobulin-secreting plasma cells, play an important role in controlling antigen penetration across the gut.37 Bacterial colonization of the gut is an important trigger for IgA and IgM production and drives the functional activation of IgA plasma cells from either B2 or B1 cells. B2 cells represent the most efficient generators of IgA in the organized lymphoid follicles or Peyer's patches in the gut.38 Recent work has found that the gut microbiota enhances the capacity of mucosal B2 cells to produce antibody.39 In situations of IgA deficiency, expansion of anaerobic bacteria, including segmented filamentous bacteria, occurs.40
The mucin layer of the gut also captures secreted antimicrobial peptides, including alpha and beta defensins, which are generated from Paneth cells and epithelial cells in the gut.41 Alpha defensins have broad-spectrum antimicrobial actions and control microbial numbers as well as microbial composition within the gut.42 For example, mice deficient in matrix metalloproteinase 7 show a decrease in Bacteroides and an increase in Firmicutes, whereas mice overexpressing HD5 have increased Bacteroides and decreased Firmicutes.43 Interestingly, the presence of HD5 also triggers the loss of segmented filamentous bacteria.44 One particular epithelial-derived beta defensin, HBD-1, which is constitutively expressed in the gut, becomes functionally active following the reduction of disulfide bonds at gut surfaces and exerts potent antimicrobial activity against gut microbes.45
EPITHELIAL CELLS AND INNATE IMMUNE RECEPTORS
A number of innate immune receptor classes involved in microbial recognition have been described thus far, including Toll-like receptors (TLRs), NOD-like receptors, Rig-like receptors, and C-type lectin receptors.46–48 These innate recognition receptors represent a universally conserved defense system that plays an important role in immune activation and defense against pathogens. Critically, however, the activation of such receptors can also induce important cytoprotective functions designed to limit tissue damage during inflammation. These innate immune receptors recognize conserved microbial structures found on microorganisms, both pathogens and commensals alike. These structures are frequently referred to as microbial-associated molecular patterns (MAMPS). TLRs are one important class of recognition receptors.49 The TLRs characterized to date include TLR1 to TLR11; TLR2, for example, recognizes bacterial lipoproteins and zymosan, TLR3 recognizes double-stranded RNA, TLR4 recognizes bacterial lipopolysaccharide (LPS), and TLR5 recognizes bacterial flagellin.49 TLRs are expressed both extracellularly and intracellularly on numerous cells in the gut, including intestinal epithelial cells and immune cells.50 The presence of these receptors, combined with their recognition of conserved structures on both commensals and pathogens and their activation of the innate immune response, raises the intriguing question of how the mucosal immune system distinguishes between harmful pathogens and beneficial commensal bacteria, particularly when immune activation and tolerance are the defining, but completely opposing, downstream responses to these distinct bacterial groups.
In the healthy gut, TLR3 and TLR5 appear to be constitutively expressed, while TLR2 and TLR4 are displayed at very low-level expression, inferring that the expression of TLR2 and TLR4 is carefully regulated to avoid unnecessary immune activation in response to the commensal microbiota. TLR4 was, in fact, the first TLR described. The normal downregulation of TLR4 and MD-2, a critical co-receptor for LPS-mediated signaling, is thought to be an important mechanism by which the gut maintains immune homeostasis in the presence of large numbers of commensal gram-negative microbes and bacterial LPS.51,52 Other mechanisms have been proposed to explain the lack of inflammatory responses to commensal gut microbes residing at gut epithelial surfaces; these include the compartmentalization and restricted expression of TLRs and the activation of negative regulators of TLR signaling cascades.53 For example, the minimal immune response induced by luminal bacteria has been explained by the restricted expression of TLRs to basolateral membranes of gut cells. This expression pattern has been described for TLR554 and is thought to explain the limited response of the gut epithelium to luminal flagellate pathogens and commensals in contact with the apical surface of gut cells. Hence, only when flagellate bacteria breach the gut barrier, as is the case with pathogenic salmonella, is TLR5 signaling induced. Other data, however, suggest that this explanation is potentially an oversimplification and may not be totally correct. Recently, TLRs have been demonstrated at the apical membranes of epithelial cells and shown to traffic internally in response to ligand engagement.55
A further important fact that may, in part, explain differences in immune outcome in response to commensal and pathogenic bacteria is that MAMP structures found on commensals and pathogens are not identical and are not functionally equivalent in terms of their ability to engage TLR activation and signaling. For example, some molecules of LPS are potent inducers of TLR4 signaling, while others are antagonistic.56,57 Similar biology explains the ability of certain bacterial flagellins to activate the TLR5,58 while other structures are nonactivating.59 Interestingly, nonactivating flagellins seem to assist pathogens such as Helicobacter pylori to evade the immune system, and this seems to represent an important mechanism by which this bacterium persists in the host gut.60
COMMENSAL REGULATION OF TOLL-LIKE RECEPTOR SIGNALING
TLRs are important for immune activation and defense against pathogens. TLR signaling is also important for gut homeostasis, cytoprotection, and health.61 The important functionality of these receptors is becoming more evident with the increasing number of reports indicating that TLR expression is altered in inflammatory bowel disease (IBD).62,63 Furthermore, deletion of TLRs such as TLR5 triggers spontaneous colitis,64 whereas flagellin treatment can protect against the damaging effects of chemicals and radiation.65 Clearly, TLRs and other microbial recognition receptors have an important role in diverse physiological responses such as immune defense and homeostasis, but their activation and signaling responses must be carefully controlled. Part of this control is explained by the functions of important negative regulators of TLR signaling. These negative regulators help put the brakes on inflammatory signaling. Example molecules include TOLLIP, SIGIRR, peroxisome proliferator-activated receptor gamma (PPARγ), and A20, some of which have been recently reviewed by Abreu.66,67
A20, also known as TNFAIP3, is a zinc finger protein that is a negative regulator of both TLR and NF-κB signaling (Figure 2a). It is also an NF-κB target gene. A20 is an intracellular protein that has been reported to limit TLR signaling, in particular TLR4- and TLR5-mediated responses.68,69 Conditional knockout of A20 in intestinal epithelial cells has been shown to increase susceptibility to experimental colitis.70 A20 also plays an important role in preventing autoimmune-driven inflammation by regulating the activation and antigen-presenting functions of dendritic cells (DCs).71,72 The mechanisms by which such proteins are regulated in response to gut bacteria are currently unknown. Insight into this regulation and, in particular, how it impacts TLR signaling may provide greater appreciation of immune homeostasis in the context of the gut microbiota, particularly since A20 has been identified as a susceptibility gene in human Crohn's disease.73
COMMENSAL REGULATION OF NUCLEAR FACTOR-KAPPA B PATHWAYS
The transcription factor nuclear factor kappa B (NF-κB) is activated downstream from a number of microbial recognition signaling receptors. NF-κB is involved in activating gene transcription of proinflammatory and survival pathways. In resting, unstimulated cells, this transcription factor, mainly the subunits p50 and p65 (RelA), resides in the cytosol as a complex associated with the inhibitor protein IκBa. Following stimulation of cells with proinflammatory agonists, including bacterial MAMP structures, the inhibitory protein is phosphorylated, ubiquitinated, and degraded, allowing the transcriptionally active subunits of NF-κB, mainly p50/RelA dimer, to enter the nucleus and activate transcription of NF-κB response genes (Figure 2b). The proinflammatory immune responses triggered by pathogenic bacteria are mediated, in part, by the transcription factor NF-κB, in particular the transactivating subunit RelA. In addition, pathogenic bacteria have also evolved mechanisms to suppress NF-κB activation and evade the host inflammatory response. One such example is an effector protein produced by pathogenic Yersinia, referred to as YopJ.74 This protein is injected into the cytosol of host cells to prevent the phosphorylation of NF-κB. Similarly, the Salmonella AvrA effector inhibits the NF-κB pathway.75 The first evidence that commensal microbes may also influence NF-κB signaling established that gram-negative Bacteroides vulgatus could induce the activation of NF-κB and the phosphorylation of the transcriptionally active RelA subunit.76 Following this report, it was then demonstrated that, like certain pathogens, commensal bacteria can inhibit and modulate the NF-κB pathway.77–79 The role of nuclear receptors, such as PPARγ, in mediating the anti-inflammatory effects of commensals was also demonstrated.77 Since then, the importance of PPARγ in regulating inflammatory and anti-inflammatory responses to bacteria, including commensals, and its role in human IBD has been established.80,81 The role of other nuclear receptors, including the vitamin D receptor, has also been highlighted recently.82–84 The molecular mechanisms by which commensal bacteria, and indeed probiotic bacteria, regulate NF-κB are not fully elucidated, but recent evidence suggests the action of bacterial-derived secreted factors may be involved.85–87
COMMENSAL INTERACTIONS WITH GUT DENDRITIC CELLS
Gut DCs, situated below the epithelium in the lamina propria and Peyer's patches, sample bacteria and present bacterial antigens to T cells. They also recognize microbes through microbial recognition receptors that include TLRs, NOD-like receptors, and C-type leptin receptors. The major route of bacterial uptake in the gut is via epithelial microfold (M) cells located over the follicular-associated epithelium of Peyer's patches (Figure 3a). Using the M cell route, bacteria can gain access to underlying immune cell populations, including DCs (Figure 3b), which then induce effector T cell functions within the Peyer's patch and mesenteric lymph nodes (MLNs), the inductive immune sites of the gut.88 Bacterial antigen can also be taken up across the lamina propria by a number of proposed mechanisms, including via special DCs that have dendrites that can traverse the gut epithelium89 or via villus M cells.90 Within the lamina propria and Peyer's patch, distinct populations of DCs are found that respond rapidly to antigen and migrate to the MLNs to activate specific T cell responses. For example, lamina-propria-derived DCs expressing the integrin CD103 appear to have the unique capacity of inducing gut-homing T cells in the MLNs. CD103- DCs, on the other hand, can also prime T cells but fail to induce gut-homing molecules such as CCR9.91 More recently, it has been shown that CD103+ cells migrate to MLNs and are classical antigen-presenting cells.92
Antigen-presenting cells, including DCs and macrophages, have been characterized using a number of cell markers, including CD11c, CD11b, CD8, CD103, F4/80, and CX3CR1.93 These markers define unique cell subsets that have differing responses to microbial antigens and can induce specific T helper (Th) cell responses such as Th1 (producing IFNγ), Th2 (producing IL-4, IL-5, and IL-13), Th17 (producing IL-17), and regulatory T cells (Tregs).94 The type of immune response is influenced by a number of factors, including the type of microorganism encountered, the local cytokine/chemokine milieu, and the type of pattern recognition receptors expressed on DCs, such as TLRs, C-type leptin receptors, and NOD-like receptors.95 Commensal bacteria seem to favor the induction of FoxP3+ Tregs via lamina propria and MLN DCs expressing CD103.92,96,97 Important work highlights the function of CX3CR1+ intestinal macrophages in IL-10 production, expansion of Tregs, and maintenance of tolerance.91,98 Epithelial-derived cytokines such as transforming growth factor β and retinoic acid also condition CD103+ DCs to induce Tregs (Figure 4).99 A number of DC subsets can exhibit this function, depending on the ratio of T cells to antigen-presenting cells and the regional localization.93
COMMENSAL MICROBIOTA AND T CELL DIFFERENTIATION
Immune regulation supports the presence of a complex microbiota in the gut but also ensures that effector immune responses are activated in response to invading pathogens. The importance of Tregs in maintaining tolerance to the gut microbiota has been known for some years.100 The commensal microbiota has also been shown to drive the local expansion of CD4+ Th cell populations, including Th17 cells.101 Segmented filamentous bacteria are key components of the microbiota responsible for driving the expansion of T cells, including Th17 cells.102,103 Furthermore, the commensal bacterium Bacteroides fragilis has been shown to induce CD4+ CD25+ Foxp3+ Treg cells.104 This biological function has been extended to include members of the Clostridium clusters IV and XIVa, which are also represented in the Altered Schaedler Flora.105,106 Thus, microbial composition is an important factor determining the intricate balance between different T cell subsets in the gut. Supporting this notion is evidence that, in disease states such as IBD, microbial composition is altered, with a loss of diversity associated with reduced Firmicutes and Bacteroidetes phylotypes.9,107 This association between microbial composition and systemic inflammatory conditions agrees with reports that segmented filamentous bacteria promote autoimmune arthritis108 and that aggressive proinflammatory T cell responses driven by the microbiota promote experimental autoimmune encephalomyelitis.109,110 Interestingly, B. fragilis, with its propensity to promote Treg responses, was found to protect against experimental autoimmune encephalomyelitis.111
INFLAMMATORY BOWEL DISEASE
Microbial pathogens have been long implicated in human IBD. However, no associated pathogen has fulfilled Koch's postulate, and the current accepted hypothesis is that Crohn's disease and ulcerative colitis, both forms of IBD, are caused by dysregulated immune responses directed towards the commensal microbiota in genetically susceptible individuals. Hence, both genetic and environmental factors contribute to this disease. Studies in twins have helped unravel the relative contributions of genetics and the microbiota. For example, a large twin-cohort study has evaluated the contribution of the microbiota to human IBD.25 This study revealed that healthy twins, living in different environments, have similar microbiota at the genus level but different microbiota at the finer levels of resolution, such as the bacterial species and strains. These data illustrate that environment plays an important role in dictating bacterial colonization. However, microbiota differences were identified in twins with IBD, namely the potentially important loss of bacteria belonging to Roseburia, Alistipes, and Ruminococcaceae.25
The loss of certain bacteria in human IBD and the evidence of general dysbiosis, which creates an imbalance between protective beneficial and more “proinflammatory” commensals, have generated significant interest in the use of probiotics to redress microbial imbalances in the gut of IBD patients.112 A probiotic cocktail of bacteria has been shown to induce remission in patients with ulcerative colitis,113 although a recent Cochrane review concluded that the probiotics evaluated thus far provide only modest benefits in ulcerative colitis, indicating that larger randomized control trials are needed to establish efficacy.114 In addition, greater consideration must be given to the selection of probiotics for evaluation.115 The scientific progress in recent years will help to identify natural bacteria of the human gut and to clarify which properties of gut bacteria make them ideal candidates for probiotic evaluation. It will then be possible to make a more informed selection of bacterial species and strains that can be used to confer important health benefits.
In addition to IBD, many autoimmune diseases are also increasing in incidence, particularly in Westernized societies. These include type 1 diabetes, multiple sclerosis, and rheumatoid arthritis. The increase in these diseases, as well as the increased frequency of IBD, has been associated with an increase in environmental hygiene.116 More recently, however, the presence of specific bacterial populations in the gut has been linked to the development of autoimmune diseases.111
UNRAVELING THE BENEFITS OF COMMENSALS FOR THE GUT
Gut bacteria have evolved to colonize and persist within the mammalian gut. Those that have successfully achieved residence have evolved in their fitness and competitiveness to survive in this ecosystem over millions of years. The factors that enable a bacterium to colonize the gut are not fully elucidated, but there is a general acceptance that the mutual benefits for both the host and bacteria provide the key to their successful partnership. Hence, it is important to investigate both sides of the host-microbe partnership.
UNDERSTANDING THE HOST RESPONSE
The last couple of decades have seen huge advances in the field of genome analysis. The completion of the Human Genome Project in 2000, in particular, has meant a giant leap forward. It is now estimated that humans have between 20,000 and 25,000 genes, and the roles of these genes in health and disease are only beginning to be identified.
High-density DNA microarrays enable the study of genome-wide transcriptome responses due to the high number of genes present on these arrays. Affymetrix GeneChips® are the most frequently used platform for this expression profiling. These GeneChips consist of high-density oligonucleotide arrays, and each transcript is detected by a probe set of between 11 and 20 probes. More than a million individual probes can be synthesized on a glass chip about 1.5 cm2 in size. The latest Genome arrays analyze around 40,000 transcripts, while the new-generation exon arrays measure both gene expression and alternative splicing, and tiling arrays enable transcript mapping.
ANALYSIS AND PROCESSING OF HOST GENE DATA
Various computational tools facilitate the analysis of the large data volumes produced in microarray experiments. Affymetrix GeneChip Operating Software® uses the fluorescence scanning image (DAT file) obtained after sample hybridization to calculate the intensities for each probe on the chip. Normalization removes nonbiological variation and is a critical step for obtaining reliable data. In addition, normalization converts the 11–20 probe intensities to a single intensity for the targeted transcript. Multiple normalization strategies are available, with some of the most popular ones being Affymetrix® MAS 5.0, Robust Multichip Average (RMA),117 and gcRMA. Quality control usually starts with inspection of the scanned images, grid alignment, and Affymetrix quality control parameters. Subsequently, different diagnostic plots can be used to monitor array quality, including RNA digest plots, density distribution plots, MvA plots, normalized unscaled standard error plots, and relative log expression plots.118 Further data analysis can take the form of numerous methods and includes both supervised detection of differentially expressed genes and gene sets (t-test, F-test) and unsupervised cluster analysis (hierarchical clustering, self-organizing maps, k-mean clustering).
BIOLOGICAL INSIGHTS OBTAINED FROM PATHWAY ANALYSIS
Microarray data analysis can result in thousands of differentially expressed genes. The typical procedure is to sort these genes on the basis of fold-change and then focus on unraveling the biology of only a small number of genes for practical reasons. This means that most of the data on the array are never used. Pathway analysis tools can help uncover functional patterns in large groups of differentially expressed genes by listing processes or pathways that are overrepresented between treatments. Maps show complex molecular interactions and regulations in an intuitive, graphical way.119 There are two major types of biological pathways: metabolic pathways, which incorporate protein-based interactions and modifications, and signal transduction and transcriptional regulatory pathways, which provide information on mechanisms of transcription.120
A number of different commercial (MetaCore®, Ingenuity® Pathway Analysis, GeneSpring® Pathways) and publicly available (BioCarta, KEGG, WikiPathways) pathway software programs are now available. As an example, MetaCore® is a proprietary, manually curated database containing human protein-protein, protein-DNA, and protein compound interactions, metabolic and signaling pathways, and the effects of bioactive molecules. MetaCore® software contains approximately 450 canonical signaling and metabolic pathways and also enables the creation of interactively built biological networks. Multiple experiments can be visualized simultaneously in individual maps.
FUNCTIONS IDENTIFIED USING GENE ONTOLOGY
Functional interpretation of differential gene expression can be further aided by the use of the Gene Ontology database (http://www.geneontology.org). The Gene Ontology project is a bioinformatics initiative that has enabled the standardization of genes across species and databases. Genes are described and annotated using a controlled vocabulary of terms. One of the most widely used sites based on Gene Ontology, DAVID, or the Database for Annotation, Visualization and Integrated Discovery (http://david.abcc.ncifcrf.gov), is a free online bioinformatics resource developed by the Laboratory of Immunopathogenesis and Bioinformatics at the National Cancer Institute. An expanded version of this original web-accessible program is described by Dennis et al.121 DAVID provides a comprehensive set of functional annotation tools to understand the biological meaning behind large lists of genes. Apart from identifying enriched biological themes based on Gene Ontology, DAVID also visualizes genes on BioCarta and KEGG pathway maps.
Most journals now require microarray data to be submitted to a public repository prior to publication. The purpose of public repositories is to store microarray data in such a way that it is fully available to the research community for analysis and interpretation. The largest databases are the National Center for Biotechnology Information Gene Expression Omnibus122 and the European Bioinformatics Institute ArrayExpress (http://www.ebi.ac.uk/arrayexpress). These databases require all data to comply with the Minimum Information about a Microarray Experiment (MIAME) guidelines (http://www.mged.org/Workgroups/MIAME/miame.html).
IDENTIFYING IMPORTANT IMMUNE-REGULATING COMMENSAL BACTERIA
As discussed, the human microbiome comprises the genes and genomes of the microbiota and contains at least 100 times as many genes as the human genome.123 Early work on analysis of the human microbiome used culture-based techniques to measure the diversity of gut commensal bacteria. These methods, however, were limited in that only the culturable fraction of the microbiota was assessed. The advent of culture-independent techniques with greater depth of analysis in the last decade has led to huge progress in human microbial diversity profiling.124 These techniques use the fact that the relative abundance of individual members of the gut microbiota determines the final depth of sequence coverage for any organism. Therefore, the genome sequences of abundant species will be highly represented in a set of random reads, whereas less abundant species are represented by a smaller number of sequences.125 Molecular techniques using 16S ribosomal RNA (16S rRNA) gene sequences as culture-independent biomarkers of microbial taxa,124 either by clone libraries or pyrosequencing, have shown how the microbiota is unique to each individual, fluctuating over time and in response to environmental perturbations.4,125 Eckburg et al.,10 in particular, utilized 16S rRNA sequence-based methods to show that two bacterial divisions, the Bacteroidetes and the Firmicutes, make up over 90% of the known gut phylotypes and dominate the distal gut microbiota. There is now a general consensus about the phylum level composition in the human gut, although the variation in species composition and gene pools is less defined.126 16S rRNA sequencing techniques are limited, however, by biases introduced by preferential PCR amplification of 16S rRNA genes and by the lack of functional interpretation of the uncovered gene sequences.125
Until recently, deep-sequencing metagenomic analysis of the microbiome was limited by the huge expense of both sequencing efforts and subsequent bioinformatics.127 The rapidly decreasing costs of both gene sequencing and data analysis, however, mean that an increased number of time points and body sites can be compared in a single study in order to further understand how the human microbiota changes over time and during health and disease.
Two major next-generation sequencing projects have been started over the last few years: the Human Microbiome Project and the MetaHIT (Metagenomics of the Human Intestinal Tract) project.9 The Human Microbiome Project, funded by the National Institutes of Health, has characterized over 600 microbial reference genomes, 70 million 16S sequences, 700 metagenomes, and 60 million predicted genes from at least 250 healthy adults.123 The MetaHIT project investigated fecal samples of 124 European individuals to describe a core, minimal gut metagenome.9 For this, Illumina®-based metagenomic sequencing was used to assemble and characterize 3.3 million unique open reading frames, derived from 576.7 GB of sequence, which was almost 200 times more than all previous studies of this kind.
Collectively, the progress in understanding host-microbe interactions, along with the availability of “omics tools,” means it will be possible to assign biological functions and activities both to isolated gut commensals and to communities of commensals. It is now possible to determine how commensal microbes influence immune health both in the gut and in systemic tissues.
Studies describing the human gut microbiota and the mucosal immune system continue to contribute to one of the most dynamic and progressive areas of scientific research. There have been tremendous recent advances in the understanding of microorganisms that occupy the human gut. The microbiota differences observed between healthy and diseased individuals offer even more clarity regarding the importance of gut bacteria to human health. Given the technical and scientific advances in gut microbiology and immunology, the opportunities for additional insight into the molecular mechanisms of bacterial-host symbiosis that deliver human health benefits have never been greater.
Improved culture methods for novel bacterial isolates have facilitated more detailed physiological investigations to elucidate how gut bacteria contribute to immune development and homeostasis. Molecular tools have enhanced the possibilities for the discovery of new biomarkers and have expanded knowledge of how the host systems benefit from microbial colonization. Analysis of the microbial metagenome, particularly in the context of the host response, provides tangible avenues for new drug discovery.
Collectively, this information should provide answers and solutions for how environment and diet can promote beneficial bacterial colonization that drives healthy immune function. The screening and isolation of health-promoting bacteria can also advance the development of both nutritional and pharmaceutical interventions that reverse the current trends in immune-related diseases, particularly in Westernized societies.
Funding. The authors are supported by the Rural and Environment Science and Analytical Services (RESAS) of the Scottish Government through funding received by The Rowett Institute of Nutrition and Health.
Declaration of interest. D. Kelly has a stake in patent WO/03 046580, which describes the novel anti-inflammatory activity of the commensal bacterium Bacteroides thetaiotaomicron.