Microbiota of the gastro-intestinal tract

The main authors of this section are Professor Gibson, Dr Hoyles and Dr McCartney and specifically Professor Rastall for the in vitro subsection.

The microbiota of the human gastro-intestinal (GI) tract inhabits a complex ecosystem⁽Reference Wilson and Blitchington²²⁾. Factors such as pH, peristalsis, nutrient availability, oxidation–reduction potential within the tissue, age of host, host health, bacterial adhesion, bacterial co-operation, mucin secretions containing Igs, bacterial antagonism and transit time influence the numbers and diversity of bacteria present in the different regions of the GI tract⁽Reference Kerckhoffs, Samson, van Berge Henegouwen, Ouwehand and Vaughan²³⁾. Until 20 years ago, our knowledge of the GI microbiota relied upon cultivation-based methods and recovery of bacteria from faecal samples. However, with the advent of molecular techniques and their application to biopsy and faecal samples, our knowledge of the GI microbiota has increased dramatically⁽Reference Suau, Bonnet and Sutren^5–Reference Green, Brostoff and Hudspith¹⁶⁾. An understanding of the bacteria making up the GI microbiota is important due to its involvement in the development of the GI mucosal immune system, maintenance of a normal physiological environment and for providing essential nutrients⁽Reference O'Connor, Barrett, Fitzgerald and Tamine²⁴⁾.

The small intestine (duodenum, jejunum and ileum)

The environment of the duodenum is acidic (pH 4–5) with lactobacilli and streptococci predominating, and numbers of bacteria are higher than those found in the stomach (10²–10⁴ CFU (per ml contents);⁽Reference O'May, Reynolds and Macfarlane²⁷⁾).

Cultivation studies have shown that lactobacilli, streptococci, veillonellae, staphylococci, actinobacilli and yeasts to be the most prominent in the duodenum and jejunum⁽Reference Kerckhoffs, Samson, van Berge Henegouwen, Ouwehand and Vaughan²³⁾. However, due to limitations in cultivation techniques and the ethical issues surrounding the obtention of biopsy samples from human subjects, our knowledge of the microbiota of the small intestine was poor until recently. Table 4 gives details of the results of recent molecular studies that have provided additional understanding of the microbiota of the small intestine. But these studies are only informative, because only one or a few donors have been used in each study, and their ages have not been representative of the general population. However, the results of the molecular studies appear to confirm those of cultivation-based work.

Table 4 Microbial diversity of the mucosa of the human small intestine as determined by 16S ribosomal ribonucleic acid gene sequence analysis

No., number.

* Numbers in parentheses represent proportion of clones ascribed to a particular phylum/genus/cluster where known. Names of nearest phylogenetic relatives are given.

The microbiota changes markedly from the duodenum to the ileum, as the velocity of the intraluminal content decreases, pH increases and oxidation–reduction potentials lower, with bacterial loads increasing to 10⁶–10⁸ CFU (per ml contents)⁽Reference Kerckhoffs, Samson, van Berge Henegouwen, Ouwehand and Vaughan²³⁾. As transit time in the small intestine is rather rapid (2–4 h) and the bacterial density relatively low, its impact in terms of overall fermentation is low compared with the large intestine (see later). The small intestine is also the site of many bacterial infections, such as salmonella and some Escherichia coli. For this reason, the small intestine is also a target for probiotics known to compete with pathogens. Similarly, sialylated acidic oligosaccharides from human milk can block the adhesion of pathogens on the epithelial surface.

The large intestine

The combination of increased transit time of the large intestine, increased nutrient availability (i.e. undigested food material from the upper GI tract, sloughed-off bacterial cells, microbial cell debris and by-products of microbial metabolism) and a more neutral pH ensures that the large intestine is a highly favourable environment for microbial colonisation. As the environment is strictly anaerobic (>100 mV), in particular obligate anaerobes prevail. Table 5 gives details of some bacteria that have been isolated from the GI microbiota. Table 6 gives details of molecular studies on biopsies from different regions of the large intestine. In addition to characterising the mucosa-associated microbiota, Zoetendal et al. ⁽Reference Zoetendal, von Wright and Vilpponen-Salmela¹¹⁾ demonstrated that the faecal microbiota differs from that inhabiting the GI mucosa.

Table 5 Bacteria, their substrates and products in the human large intestine Taken from Salminen et al. ⁽Reference Salminen, Bouley and Boutron-Ruault³⁷⁷⁾

aa, amino acid; Ac, acetate; Pr, propionate; Su, succinate; Bu, butyrate; La, lactate; f, formate; e, ethanol.

Table 6 Microbial diversity of the mucosa of the human large intestine as determined by 16S ribosomal ribonucleic acid gene sequence analysis

No., number.

* Numbers in parentheses represent proportion of clones ascribed to a particular phylum/genus/cluster where known. Names of nearest phylogenetic relatives are given.

Even today, due to the difficulty of obtaining samples from the different regions of the intestine, much of the work done in relation to the ecology and activity of bacteria within the GI tract is carried out using faecal samples. However, the faecal microbiota is not representative of that of the GI tract as a whole⁽Reference Zoetendal, von Wright and Vilpponen-Salmela^11, Reference Eckburg, Bik and Bernstein¹⁴⁾, and inferences made from in vitro studies in relation to specific GI diseases, particularly those involving the more-proximal regions of the intestine, should always be made with this in mind. However, a study examining the GI microbiota of sudden-death victims has shown that the faecal microbiota reflects that of the luminal contents of the descending colon in terms of the culturable component⁽Reference Macfarlane, Macfarlane and Gibson²⁸⁾. Molecular-based methods have been used to examine the faecal microbiota in recent years. Identification of specific strains isolated from faecal samples has become more accurate due to the use of 16S ribosomal ribonucleic acid gene sequence analysis and has improved taxonomic schemes and our understanding of the bacteria involved in specific metabolic processes (e.g. the role of Roseburia spp. in butyrate production⁽Reference Duncan, Aminov and Scott²⁹⁾, and the identification of the mucin-degrading bacterium Akkermansia muciniphila ⁽Reference Derrien, Vaughan and Plugge³⁰⁾). This improved characterisation of viable bacteria has also aided in the design of probes for use in fluorescence in situ hybridisation analysis (e.g. Rrec584 for Roseburia spp.⁽Reference Walker, Duncan and William Leitch³¹⁾).

Early cloning studies examined relatively small numbers of clones to generate a phylogenetic inventory of the faecal microbiota of healthy adults. Wilson & Blitchington⁽Reference Wilson and Blitchington²²⁾ generated two clone libraries (one from a 9-cycle PCR (fifty clones, twenty-seven operational taxonomic units) and the other from a 35-cycle PCR (thirty-nine clones, thirteen operational taxonomic units)) from a faecal sample from a healthy 40-year-old male. Of the clones they analysed, 35 % were related to the Bacteroides group, 10 % to the Clostridium coccoides group (Clostridium cluster XIVa) and 50 % to the Clostridium leptum group (Clostridium cluster IV). Less than a quarter of the sequences analysed were derived from a known bacteria. Suau et al. ⁽Reference Suau, Bonnet and Sutren⁵⁾ found that of the 284 clones they generated from a faecal sample from a 40-year-old male, the majority of the sequences fell into three phylogenetic groups: Bacteroides (31 %), C. coccoides (44 %) and C. leptum (20 %). The remaining clones were derived from Streptococcus salivarius and Streptococcus parasanguinis and bacteria related to Mycoplasma spp., clostridia, the Atopobium group, Verrucomicrobium spinosum and the Phascolarctobacterium faecium subgroup. Seventy-six per cent of the clones analysed were derived from previously unknown bacteria. Blaut et al. ⁽Reference Blaut, Collins and Welling³²⁾ used a cloning approach to demonstrate that microbial diversity in faeces increases with age. It was found that the number of operational taxonomic units corresponding to known molecular species was highest in infants and lowest in the elderly subjects, with 92 % of sequences from the elderly subjects corresponding to previously unknown bacteria.

As molecular methods have become more widely available and less time consuming and their relative costs have decreased, more ambitious cloning studies in which thousands of sequences have been examined have been carried out⁽Reference Eckburg, Bik and Bernstein^14, Reference Manichanh, Rigottier-Gois and Bonnaud³³⁾. The results of these studies in terms of the groups of bacteria represented by the largest number of clones and the identification of previously unknown bacteria are in accordance with those of Wilson & Blitchington⁽Reference Wilson and Blitchington²²⁾ and Suau et al. ⁽Reference Suau, Bonnet and Sutren⁵⁾, but are notable for the characterisation of several actinobacterial and proteobacterial sequences from human faecal samples.

Techniques such as temperature gradient gel electrophoresis and denaturing gradient gel electrophoresis (DGGE) allow higher numbers of samples from more donors to be examined than traditional cloning studies. Temperature gradient gel electrophoresis was used by Zoetendal et al. ⁽Reference Zoetendal, Akkermans and de Vos⁹⁾ to examine the total bacterial communities of faecal samples from sixteen adults. Host-specific fingerprints were generated, demonstrating interindividual variation in the composition of the faecal microbiota and confirming the results of cultivation studies. Some bands were seen in fingerprints from multiple donors, suggesting that species of the predominant microbiota were common across individuals. In addition, by obtaining samples from two donors over a 6-month period, the authors showed that the profiles of these donors did not differ significantly over time, demonstrating that predominant microbial species were relatively stable without dietary intervention. Excision and sequencing of bands of interest allowed the authors to perform a phylogenetic analysis on their samples, the results of which demonstrated that the majority of bacteria represented in their fingerprints did not correspond to known bacterial species. Of the prominent bands identified in almost all samples, most belonged to different Clostridium clusters, with the remainder identified as Ruminococcus obeum, Eubacterium hallii and Faecalibacterium prausnitzii. Zoetendal et al. ⁽Reference Zoetendal, Akkermans and Akkermans-van Vliet¹⁰⁾, using DGGE, demonstrated that host genotype affects the composition of the faecal microbiota. In that study, the authors examined faecal samples from fifty donors of varying relatedness. A higher similarity was seen between fingerprints from monozygotic twins living apart than between those of married couples or pairs of twins. There was a significant difference between the fingerprints of unrelated people grouped by either gender or living arrangements, and no relationship between the fingerprints generated and the age difference of siblings. Temporal temperature gradient gel electrophoresis and DGGE studies examining the faecal microbiota of children and infants have confirmed the impact of host genotype on the composition of the faecal microbiota⁽Reference Stewart, Chadwick and Murray³⁴⁾. Other studies employing DGGE have used primer sets that allow examination of the composition and dynamics of specific groups of bacteria (Table 7). The detection limit seems to be the main barrier to overcome in these studies, particularly when examining populations such as bifidobacteria and lactobacilli – the commonest prebiotic targets.

Table 7 Details of some TGGE and denaturing gradient gel electrophoresis studies of the faecal microbiota

With respect to the prebiotic concept, it is important to understand that apart from knowledge on the complexity of the gut microflora, it is also known that certain bacteria are associated with toxin formation and even pathogenicity when they become dominant. Others are associated with carcinogen generation and the metabolism of other xenobiotics. These potentially harmful bacteria belong to species within groups such as clostridia and bacteroides. Whereas knowledge on overt or latent pathogens has advanced markedly, due to the symptoms they can cause, there is less consensus on what characterises potentially harmful bacteria (without direct pathogenicity) and potentially healthy bacteria. Still potentially healthy bacterial groups are characterised by a beneficial metabolism to the host through their SCFA formation, absence of toxin production, formation of defensins or even vitamin synthesis. They may also inhibit pathogens through a multiplicity of mechanisms. Their cell wall is devoid of lipoplysaccharides or other inflammatory mediators (i.e. mainly Gram positive). Some may also compete with receptor sites on the gut wall and inhibit pathogen persistence and thus reduce the potential risk of infection. They may also compete effectively for nutrients with pathogens. One subject of intensive research is their stimulation of immunological defence systems, as discussed in section ‘Prebiotic effects and immune system’ of the present paper. Acknowledged examples are bifidobacteria and lactobacilli – known as useful probiotics. Intermediate genera like streptococci, enterococci, eubacteria and bacteroides can be classified as potentially beneficial to health or potentially harmful, depending on the species. With regard to some of the most recently identified genera in the major phylla (Firmicutes, Actinobacteria and Bacteroidetes), classification as potentially beneficial to health or potentially harmful still remains to be made. A scheme describing the hypothesis of a balanced microbiota has been proposed by Gibson and Roberfroid⁽Reference Gibson and Roberfroid³⁾ and recently endorsed by ISAPP (2008) even though it is stillsubject of ongoing discussion. A revised version of that scheme including the most recent knowledge on gut microbiota composition is presented in Fig. 1.

Fig. 1 Schematic representation of an adult gut microbiota. Major phylla and genera are located on a logarithmic scale as no. of CFU/g of faeces. Genera on the left site are likely to be potentially harmful whereas those on the right site are potentially beneficial to health. Those that sit both on the left site and the right site either contain species that are potentially harmful and species that are potentially beneficial to health or contain genera/species that still need to be classified. Indeed many of these have only recently been identified in the gut microbiota and their activity(ies) is/are still largely unknown.

The prebiotic concept is based on the selective stimulation of the host's own beneficial microflora by providing specific substrate for their growth and metabolism. Today, the effect is measured by using bifidobacteria or lactobacilli as markers, but may include others in the future, if their positive nature can be confirmed.

It has been shown by several studies (see section ‘Human studies showing prebiotic effects in healthy persons’ of the present paper) that dietary intervention can selectively modulate the indigenous composition of the gut microbiota. This is the basis of a prebiotic effect and this has been assessed through reliable molecular-based analyses.

Prebiotic effects and fermentation and physiology

In vitro tests for prebiotic effect

In vitro models aim at studying prebiotic effects independently from their passage through the upper parts of the GI tract even if digestion is sometimes partly simulated. These models are thus only indicative of a potential prebiotic effect, however, they do not prove the prebiotic attribute of a particular product as in vivo studies need to be performed to definitively demonstrate that the compound under investigation selectively stimulates the growth and/or activity(ies) of one or a limited number of microbial genus(era)/species in the gut microbiota that confers health benefits to the host. Since, as discussed earlier, the aim of the present paper is not to provide a list of food ingredients/supplements that classify as prebiotics, the following sections will only refer to a few examples to illustrate the potentials and the limits of in vitro tests as well as the advantages and disadvantages of the different experimental models.

Batch culture (pH or non-pH controlled) studies where different substrates are incubated with either pure culture of selected bacteria or faecal slurries subsequently analysed for microbial composition can be used:

1. (1) to study the selectivity of fermentation (including possible mechanism of selectivity) by, for example, bifidobacteria, lactobacilli of different substrates (e.g. main oligosaccharides contained in soyabeans are raffinose and stachyose which have been found to be good growth promoters of Bifidobacterium infantis but not E. coli, Streptococcus faecalis or Lactobacillus Lactobacillus acidophilus ⁽Reference Tamura⁶¹⁾) or similar substrates differing in molecular weights (e.g. wheat arabinoxylans) showing, e.g. that molecular weight can be an important factor in selectivity⁽Reference Hughes, Shewry and Li⁶²⁾.

2. (2) to show changes in faecal microbiota (e.g. increase in bifidobacteria) but also to compare the efficacy of different substrates (e.g. ITF, starch, polydextrose, fructose and pectin, galactans, xylo-oligosaccharides, soyabean oligosaccharides⁽Reference Wang and Gibson^63–Reference Hayakawa, Mizutani and Wada⁶⁵⁾).

3. (3) to measure and to compare the evolution of gas and SCFA production as a result of the fermentation of different substrates⁽Reference Rycroft, Jones and Gibson⁶⁴⁾.

Single-stage chemostat studies with ITF were used to compare differing techniques to analyse microbiota composition, demonstrating that discrepancies might exist between classical microbiological techniques and molecular approaches. Agar plate counts showed an increase in the combined populations of bifidobacteria and lactobacilli reaching 98·7 % of the total bacterial flora by steady state. However, 16S ribosomal ribonucleic acid genus-specific probes indicated an initial increase in the bifidobacteria population which decreased after 6 d, while lactobacilli thrived in the low pH fermenter (pH 5·2–5·4) maintaining a high population at steady state. Changes observed in the SCFA profile corresponded well with the population data obtained through probe methods⁽Reference Sghir, Chow and Mackie⁶⁶⁾.

Continuous culture systems inoculated with faecal slurries can be used to investigate fermentation profiles showing, for example that in accordance with earlier studies, bifidobacteria, and to a lesser extent lactobacilli preferred ITF to glucose, whereas bacteroides could not grow on these substrates⁽Reference Gibson and Wang^67, Reference Gibson and Wang⁶⁸⁾. By varying parameters in the chemostat, the conditions for growth of bifidobacteria and inhibition of bacteroides, clostridia and coliforms can be further analysed showing that low pH (pH 5·5), high culture dilution rate (0·3 h^− 1) and 1 % (w/v) concentration of carbohydrate (i.e. similar to the physico-chemical environment of the proximal colon) are optimum.

The three-stage gut model reproduces the three segments of the colon (proximal/ascending, transverse and distal/descending). It is used to confirm the effects observed in the previous models. Studies using this model show enhanced proliferation of bifidobacteria and/or lactobacilli by ITF and galactans in conditions resembling the proximal/ascending colon⁽Reference Gibson and Wang^67, Reference McBain and Macfarlane^69, Reference McBain and Macfarlane⁷⁰⁾. Whereas studies using models of vessels two and three (modeling transverse and descending colon respectively) displayed very little change in microbiota when fermenting galactans⁽Reference McBain and Macfarlane⁷⁰⁾. In the same model, changes in enzyme activities (β-glycosidase, β-glucuronidase, azoreductase and arylsulphatase) can also be monitored showing their suppression after fermentation of galactans⁽Reference McBain and Macfarlane⁷⁰⁾ or soyabean–oligosaccharides⁽Reference Wada, Watabe and Mizutani⁷¹⁾. Investigating the effect of pH and substrate concentration on the fermentation selectivity of galactans alongside other products, Palframan et al. ⁽Reference Palframan, Gibson and Rastall⁷²⁾ reported a strong bifidogenic effect at pH 6 and at 2 % (w/v) and suggested that they may be well fermented in the distal colon. In another study, galactans of rather low molecular weight (1 % w/v) had a strong bifidogenic effect which showed good persistence through the first two vessels, with a weaker response in the third⁽Reference Tzortzis, Goulas and Gee⁷³⁾.

The simulator of the human intestinal microbial ecosystem model consists of a series of five temperature and pH-controlled vessels that simulate the stomach, small intestine, ascending, transverse and descending colons, respectively. It can be fed with a complex growth medium containing selected substrates (e.g. ITF) to study their fermentation including the monitoring of metabolites and to analyse their effect on enzyme activities and composition of the microbiota by using a multiphase approach consisting of plate counting, quantitative PCR and DGGE⁽Reference van de Wiele, Boon and Possemiers⁷⁴⁾. Results have shown a significant increase in lactobacilli in the transverse and descending colon vessels. Low levels of bifidobacteria were recorded in the colon vessels. DGGE analysis revealed that bacteria in the ascending colon vessel grouped together as the bacteria in the other colon vessels did. Bifidobacteria clustered according to the time point rather than the vessel. Quantitative PCR, however, revealed a significant increase in bifidobacteria population in all three-colon vessels. ITF feeding also resulted in an increase in the production of SCFA, particularly propionate and butyrate, indicating a shift towards a more saccharolytic fermentation. The same model system and metabolic analysis can also be used to investigate the effect of different composition of the same substrates (e.g. of ITF with different molecular weight) on fermentation properties⁽Reference van de Wiele, Boon and Possemiers⁷⁵⁾.

A more sophisticated in vitro model of fermentation in the proximal large intestine is the TNO-intestinal model-2 model⁽Reference Minekus, Smeets-Peeters and Bernalier^76, Reference Venema, van Nuenen and van den Heuvel⁷⁷⁾. This consists of a series of linked glass vessels containing flexible walls. This arrangement allows simulation of peristalsis together with temperature regulation by means of pumping water through the space between the glass and flexible walls. The flow is controlled by computer to more accurately simulate peristalitc mixing. The vessels are further equipped with a hollow fibre membrane in the lumen to simulate absorption of water and SCFA. TNO-intestinal model-2 has been used to investigate the population changes on the fermentation of lactulose using culture-based methods coupled with DGGE⁽Reference Venema, van Nuenen and van den Heuvel⁷⁷⁾. Increases in lactobacilli and enterococci were seen.

Human studies showing prebiotics effect in healthy persons

By reference to the prebiotic concept as defined earlier, criteria for classification as a prebiotic are⁽Reference Gibson, Probert and Van Loo⁴⁾

1. (1) resistance to gastric acidity, hydrolysis by mammalian digestive enzymes and GI absorption;

2. (2) fermentation by intestinal microflora;

3. (3) selective stimulation of the growth and/or activity(ies) of one or a limited number of intestinal bacteria beneficially associated with health and well-being.

Any dietary component that reaches the colon intact (or partly so) is a potential candidate for prebiotic attribute, however, it is the latter of the three above criteria which is crucial but still the most difficult to fulfil (and which is often ignored when citing ingredients as ‘prebiotics’). Even if in addition to ITF and GOS, several dietary carbohydrates (e.g polydextrose, soyabean oligosaccharides, lactosucrose, isomalto-oligosaccharides, gluco-oligosaccharides, xylylo-oligosaccharides, gentio-oligosaccharides, mannan-oligosaccharides, lactose, hemicellulose, resistant starch, resistant dextrins, oat bran, oligosaccharides from melibiose, β-glucans, N-acetylchito-oligosaccharides, sugar alcohols such as lactitol, sorbitol and maltitol) show some fermentation selectivity when tested in laboratory systems (see section ‘In vitro tests for prebiotic effect’ in the present paper). However, the ultimate test for prebiotic activity (i.e. human volunteer trials) is lacking for the majority of these compounds. As for today, ITF and GOS are the compounds the most extensively tested in human trials that have confirmed their prebiotic effects as evidence by their ability to change the gut flora composition after a short feeding period at reasonably low doses⁽Reference Gibson and Roberfroid²⁰⁾ (Table 8). ITF, the most extensively tested forms in the literature, occur naturally in several foods such as leek, asparagus, chicory, Jerusalem artichoke, garlic, artichoke, onion, wheat, banana and oats, as well as soyabean. However, these foods contain only trace levels of ITF, so developments have taken the approach of removing the active ingredient from such sources (especially chicory roots) and adding them to more frequently consumed products in order to attain levels whereby a prebiotic effect may occur, e.g. cereals, confectionery, biscuits, infant foods, yoghurts, table spreads, bread, sauces, drinks⁽Reference Gibson, Probert and Van Loo⁴⁾. Other food ingredients/additives with potential prebiotic effects are already under investigations and will certainly be further developed in the future from dietary fibres and other non-digestible food ingredients. Very preliminary data already exist for some but many more replicate human studies including the quantitative analysis of a wide variety of bacterial genera in faecal microbiota using the more recent methodologies (as described in section ‘Microbiota of the gastro-intestinal tract – the large intestine’ of the present paper) are needed before this can be the case. Human trials may be carried out on volunteers who are on controlled diets, or are free living. To ensure consistency and exclude incidental findings, more than one human trial is needed and the totality of several human studies for a candidate prebiotic should be considered.

Table 8 Example of human studies (healthy persons) designed to determine the prebiotic effect of short-chain fructo-oligosaccharides (scFOS), fructo-oligosaccharides (FOS), galacto-oligosaccharides (GOS) and inulin

FISH, fluorescent in situ hybridization.

When evaluating a potential prebiotic effect it must be kept in mind that a dose–effect relationship and consequently a minimum effective dose are difficult to establish. Indeed, the major determinant that quantitatively controls the prebiotic effect is the number of targeted bacteria genus/species per gram of faeces the volunteers have before the supplementation with the compound presumed to show a prebiotic effect. This issue has been extensively discussed previously⁽Reference Roberfroid⁷⁸⁾.

Outline of benefit area

The main authors of this section are Professor Watzl and Dr Wolvers. To provide optimal resistance against a large variety of pathogenic encounters, the immune system has evolved to comprise multiple, functionally differing cell types enabling the development of an immune response that is specifically tailored to clear the pathogen involved. Consequently, a large spectrum of immune parameters involved in various types of responses exist, of which comprehensive descriptions can be found in many textbooks (e.g. Janeway's Immunobiology by Murphy et al. ⁽Reference Murphy, Travers and Walport⁷⁹⁾). Some of these may be measurable in human subjects and can be divided into innate v. adaptive, mucosal v. systemic, pro-inflammatory v. anti-inflammatory, etc. Modulating aspects of the immune system may, in theory, serve several clinical purposes. First, boosting or restoring the very purpose of immune function, i.e. the resistance against infections, may serve as a clinical tool to prevent or treat infectious diseases. Secondly, preventing or treating consequences of an aberrant or undesired immune response, such as those occurring with an allergic response or during chronic inflammatory diseases, are other targets with high clinical relevance.

Although there is no single immune marker that accurately reflects or predicts an individual's resistance to infection, parameters can be identified that play a more prominent role in certain types of infections or conditions than others. For instance, if resistance against the common cold, i.e. a viral upper respiratory tract infection, is the topic of interest, it seems appropriate to investigate natural killer (NK) cell and CD8+lymphocyte activity, whereas in case of IBD the balance between pro-inflammatory and immuno-regulatory cytokines will be of interest (see section ‘Prebiotic effects and IBD’ of the present paper). Moreover, in a previous ILSI Europe activity, the suitability of immune markers to measure immuno-modulation by dietary intervention in human subjects was assessed, leading to the identification of four high-suitability markers that are the result of an integrated immune reaction (vaccine-specific serum antibody production, delayed-type hypersensitivity response, vaccine-specific or total secretory IgA in saliva, the response to attenuated pathogens). In addition, a range of medium and low-suitability markers, such as functional activity of cells of the innate immune system (NK cell activity, phagocytosis, T-cell proliferation and various cytokines), were identified⁽Reference Albers, Antoine and Bourdet-Sicard⁸⁰⁾. Although the combined measurement of high- and medium-suitability markers may be a way to address aspects of immune status, the ultimate proof of accurate or even improved immune function in practice is a change in the incidence, severity or duration of infectious episodes or conditions with a prominent immune component such as allergies and chronic inflammation.

That modulation of certain aspects of the immune system may result from prebiotic effects and is based on the pivotal interaction between the intestinal microbiota and the host immune system. From several studies in germ-free and gnotobiotic animals, it is clear that the microbiota is essential for an optimal structural and functional development of the immune system⁽Reference Wagner^81–Reference Gaboriau-Routhiau, Rakotobe and Lecuyer⁸⁴⁾. The interactive co-existence of the immune system and the microbiota is especially apparent in the intestinal tract where the gut-associated lymphoid tissue (GALT) has evolved to provide optimal defense against intestinal pathogens, while at the same time tolerating dietary and self-antigens, as well as large populations of commensal non-pathogenic microbes.

Although specialised cells such as the M-cells and, as discovered more recently, also dendritic cells sample material directly from the intestinal lumen⁽Reference Rescigno, Urbano and Valzasina⁸⁵⁾, enterocytes are key intermediates that convey signals from the intestinal lumen to the mucosal immune system⁽Reference Sanderson^86, Reference Artis⁸⁷⁾ and are thus a target for a prebiotic effect on the immune system.

Prebiotic effects may influence the immune system directly or indirectly as a result of intestinal fermentation and promotion of growth of certain members of the gut microbiota. First, the mere presence of increased numbers of a particular microbial genus or species, or a related decrease of other microbes, may change the collective immuno-interactive profile of the microbiota. Through pattern-recognition receptors, such as the toll-like receptors (TLR), both immune cells and enterocytes interact with the so-called pathogen-associated molecular patterns, such as lipopolysaccharides (LPS, a membrane component of Gram-negative bacteria), lipoteichoic acids and unmethylated C-phosphate-G (CpG) DNA that are in fact present on all the micro-organisms surface regardless of pathogenicity. These interactions, possibly in combination with contextual cues of pathogenicity, result in a variety of downstream events eventually leading to cytokine production steering towards an appropriate immune response for the microbial event⁽Reference Medzhitov^88–Reference Rakoff-Nahoum, Paglino and Eslami-Varzaneh⁹⁰⁾.

Secondly, microbial products such as SCFA may interact with immune cells and enterocytes and modify their activity. G-protein-coupled receptors (GPR) 41 and GPR 43 are identified as receptors for SCFA and are expressed on leukocytes, especially polymorphonuclear cells⁽Reference Nilsson, Kotarsky and Owman^91, Reference Le Poul, Loison and Struyf⁹²⁾, as well as on enterocytes and enteroendocrine cells in the human colon⁽Reference Karaki, Tazoe and Hayashi^93, Reference Tazoe, Otomo and Karaki⁹⁴⁾. SCFA modulate chemokine expression in intestinal epithelial cells⁽Reference Sanderson⁸⁶⁾, differentially affect pro-inflammatory IL-2 and interferon (IFN)-γ and immuno-regulatory IL-10 production by rat lymphocytes in vitro ⁽Reference Cavaglieri, Nishiyama and Fernandes⁹⁵⁾, and a recent publication shows the importance of ligation to GPR43 in mice to maintain intestinal homeostasis⁽Reference Maslowski, Vieira and Ng⁹⁶⁾.

Thirdly, the potential direct ligation of pattern recognition receptors on immune cells by prebiotic carbohydrate structures may result in immunomodulation, although there is currently very little evidence for the presence of, for example, a fructose-receptor on immune cells.

In summary, there are plausible mechanisms by which prebiotic effects can modulate immune function parameters. The inaccessibility of the human GI immune system complicates the investigation in this area and most human studies rely on the measurement of ex vivo systemic immune markers, of which the predictive value for overall resistance to infections or outcome of immune-related disorders is limited.

Summary of key studies

Several comprehensive reviews have summarised the present knowledge of the immunomodulatory potential of prebiotic effects (especially ITF)⁽Reference Schley and Field^97–Reference Seifert, Watzl, Gibson and Roberfroid¹⁰¹⁾. A limited number of human studies have been performed but most have limitations as they investigated prebiotic effects in combination with the administration of other ingredients or did not include an appropriate control group.

The prebiotic effects on immune markers that represent a more or less integrated immune response, such as response to vaccination, were investigated in only a few studies (Table 9). Bunout et al. ⁽Reference Bunout, Hirsch and Pia¹⁰²⁾ supplemented healthy elderly with an oligofructose/inulin mix (6 g/d) in combination with a nutrient supplement, while the control group received maltodextrin with the nutrient supplement. No significant differences were observed in antibody titers after vaccination or on secretory IgA levels⁽Reference Bunout, Hirsch and Pia¹⁰²⁾. In a second study, the same authors investigated the effect of a supplement with oligofructose on various immune markers including delayed type hypersensitivity and vaccination. Elderly subjects attending a clinic received oligofructose as part of a complex nutritional supplement including Lactobacillus paracasei. Elderly subjects attending another clinic not receiving this supplement served as controls. Delayed type hypersensitivity response and antibody titres after vaccination did not differ between groups⁽Reference Bunout, Barrera and Hirsch¹⁰³⁾.

Table 9 The prebiotic effect on immune markers

R, PC, randomised, placebo-control; IFNγ, interferon γ PBMC, peripheral blood mononuclear cell; R, DB, PC, randomised, double-blind, placebo-control; GOS, galacto-oligosaccharides; FOS, fructo-oligosaccharides; NK, natural killer; CD, Crohn's disease; CO, crossover; OF, oligofructose; DPRPC, double-parallel, randomised, placebo-control; DTP, diphteria, tetanus, polio.

Human studies are detailed that allow interpretation of the effect of prebiotics alone, not of the combination of prebitocs with other ingredients.

Studies describe the effect on immune markers; studies that focus on clinical endpoints are summarized elsewhere in this paper (pediatrics, inflammatory bowel disease).

In infants aged 6–12 months (87 % breast-fed), the intake of oligofructose as part of an infant cereal had no effect on diarrhoea prevalence (see section ‘Use of prebiotic effects for paediatric disorders – diarrhoeal diseases’ of the present paper) and on vaccination-induced antibody titres to Haemophilus influenza when compared with the infant cereal alone⁽Reference Duggan, Penny and Hibberd¹⁰⁴⁾. Besides, the fact that a rather low dose of oligofructose was supplemented, breast-feeding may already have provided adequate amounts of human milk oligosaccharides in the present study. Also in infants at high risk for allergies, supplementation with a GOS/fructo-oligosaccharides (FOS) mixtures did not change antibody levels after a standard vaccination⁽Reference van Hoffen, Ruiter and Faber¹⁰⁵⁾. In contrast, early-life exposure of non-breast-fed infants to oligosaccharides had an effect on natural Ig production, as a mixture of GOS/FOS was shown to result in significantly higher faecal secreteory Immunoglobulin A (sIgA) concentrations as a consequence of the prebiotic effect⁽Reference Bakker-Zierikzee, Tol and Kroes^106, Reference Scholtens, Alliet and Raes¹⁰⁷⁾. Overall, there are currently no data that support beneficial prebiotic effects on the response to vaccination, but data on faecal secretory IgA in infants are promising when supplemented with a specific combination of compounds showing prebiotic effects.

In addition to the effects on integrated immune responses, the prebiotic effect on specific immune markers has been tested in a few studies of varying quality with differential outcomes (Table 9). In healthy elderly people receiving ITF-_DPav3-4 (6 g/d), a decrease in phagocytosis and IL-6 mRNA expression in peripheral blood mononuclear cell was found⁽Reference Guigoz, Rochat and Perruisseau-Carrier¹⁰⁸⁾. The present study was a one-arm study using baseline for comparison. Whether the tested ingredient induced the observed immunological changes cannot be answered from the present study. Increased NK cell activity and IL-2 production by peripheral blood mononuclear cell (Lymphokine production by mononuclear cells) were found in a synbiotic study in elderly⁽Reference Bunout, Barrera and Hirsch¹⁰³⁾. As this was a synbiotic intervention, a causal conclusion about an immunomodulation of the prebiotic intervention cannot be drawn. No effect was observed on secretion of IL-4, IFNγ and lymphocyte proliferation in cultured peripheral blood mononuclear cell⁽Reference Bunout, Hirsch and Pia¹⁰²⁾.

A study investigating the application of ingredients showing a prebiotic effect in pregnant women showed no effect on the composition of lymphocyte subsets or cytokine secretion patterns in circulating lymphocytes of the off-spring as assessed in cord-blood⁽Reference Shadid, Haarman and Knol¹⁰⁹⁾. For safety reasons, the dosage was relatively low in the present study.

A well-designed and controlled human intervention study investigated the effect of a mixture of galactans on the immune system of healthy elderly volunteers. The present study reported that intake of such GOS (galactans) (5·5 g/d) for 10 weeks significantly increased phagocytosis, NK cell activity and the production of the anti-inflammatory cytokine IL-10, while the production of pro-inflammatory cytokines IL-1β, IL-6, TNFα was reduced⁽Reference Vulevic, Drakoularakou and Yaqoob¹¹⁰⁾. A clear positive correlation between numbers of bifidobacteria in faecal samples and both, NK cell activity and phagocytosis, was observed. The present study suggests that a mixture of galactans beneficially affects the immune system and that the achieved effects may be indirect and mediated via a prebiotic effect, i.e. a change in microbiota composition. A few of the trials described earlier also show changes in immune markers alongside changes in the faecal microbiota, mainly increase in bifidobacteria. These studies thus provide data for the suggested link between a change in the flora and immunomodulation, but more studies showing correlative findings are required for convincing evidence.

Only a few studies that investigated the prebiotic effect on immune-related clinical end points such as resistance to infections, allergies and IBD have also included measurements on immune markers. Combining clinical end points with such functional markers may provide a possible mechanistic explanation for the observed effects. In a small number of patients with moderately active Crohn's disease, consumption of 15 g ITF/d reported positive clinical outcomes (see section ‘Prebiotic effects in Crohn's disease’ of the present paper), while IL-10 production by mucosal dendritic cells isolated from biopsies was increased as did expression of TLR-2 and TLR-4⁽Reference Lindsay, Whelan and Stagg¹¹¹⁾. Although some of the findings correlate with those found in animals studies⁽Reference Hoentjen, Welling and Harmsen¹¹²⁾, the open label character of the study needs to be considered.

In infants at high risk of allergies, a mixture of GOS/FOS supplemented for 6 months reduced plasma level of total IgE, IgG1, IgG2 and IgG3, whereas no effect on IgG4 was observed. In addition, cow's milk protein-specific IgG1 was significantly decreased⁽Reference van Hoffen, Ruiter and Faber¹⁰⁵⁾. This may be beneficial change in infants at risk of allergies, and although no direct correlations were reported, the same study found a significant reduction in the incidence of atopic dermatitis in a subpopulation of the GOS/FOS group⁽Reference Moro, Arslanoglu and Stahl¹¹³⁾.

Experimental data from animal studies indicate that, besides the systemic immune system, the GALT may be the primary target of immunomodulatory prebiotic effects. Biomarkers to assess functional changes in the GALT include sIgA, cytokine production and lymphocyte numbers. Prebiotic effects have been shown to increase sIgA concentration in the intestinal lumen, to increase B cell numbers in Peyer's patches, and, in intestinal tissues, to enhance IL-10 protein secretion and to decrease mRNA expression and protein concentrations of pro-inflammatory cytokines⁽Reference Watzl, Girrbach and Roller^98–Reference Seifert, Watzl, Gibson and Roberfroid¹⁰¹⁾. Genes related to intestinal immune responses seem to be a primary target of the prebiotic effects⁽Reference Fukasawa, Murashima and Matsumoto¹¹⁴⁾. Further, functional activities of NK cells and phagocytes isolated from various immune tissues were significantly increased but depending on the source of immune cells (Peyer's patches, mesenteric lymph nodes, intraepithelial lymphocytes) the prebiotic effects may differ⁽Reference Roller, Pietro and Caderni^115–Reference Girrbach, Schroeder and Breves¹¹⁷⁾. This illustrates the need to differentially study the prebiotic effects of on various immune compartments. The lack of sufficient tools to investigate prebiotic effects in the human GALT hampers insights into the possible differential impact on the mucosal v. the systemic immune system.

Key points

1. (1) Plausible hypotheses exist that ingredients showing a prebiotic effect may potentially affect the immune system as a direct or indirect result of the change in the composition and/or fermentation profile of the microbiota.

2. (2) There is currently limited, yet promising evidence that such ingredients modulate immune markers in human subjects. Well-designed human intervention studies are few.

3. (3) Data that show increased faecal sIgA levels in infants are promising and need to be confirmed.

4. (4) While several studies report changes in the faecal microbial composition alongside with changes in immune markers, only one study so far has correlated these findings. More studies addressing such correlation are needed to establish a firm link between changes in the microbiota and immune markers.

5. (5) Despite the wealth of evidence that compounds with prebiotic effects affect the intestinal microbiota, and modulate immune parameters, it is of importance to know whether these immunomodulatory effects result in a clinically relevant outcome, i.e. improved resistance against infections, or impairment of allergies and inflammation. Preliminary yet promising clinical end point studies exist which integrate the measurement of immune markers as possible explanation of prebiotic efficacy.

6. (6) Animal studies indicate that immunological effects may vary depending upon the anatomical site of origin of the immune cell (e.g. Peyer's patches v. intraepithelial lymphocytes). However, as the human GALT as primary target of the prebiotic effects cannot be easily addressed in human intervention studies, insights are difficult to obtain and thus still limited.

Recommendations

Data from well-designed, controlled human intervention studies with healthy subjects do not allow a final conclusion about the effects of ingredients showing a prebiotic effect on the immune system. Data so far are available for ITF and GOS, but few studies have been published so far. Therefore, further studies with adequate methodology, investigating immune parameters such as laid out by the ILSI Task Force on Nutrition and Immunity in Man⁽Reference Albers, Antoine and Bourdet-Sicard⁸⁰⁾, are warranted to obtain further insights on how prebiotic effects may modify immune function markers. Furthermore, tools should be developed to measure the impact of prebiotic effects on the GALT in human subjects, so an understanding of the tissue-specific effects can be achieved. Findings of such immunomodulation should lead to hypotheses on the potential use of compounds with prebiotic effects in relevant health-related conditions, which could then be tested in well-designed clinical end point studies. In addition, effects of different prebiotic chemical structures of prebiotics, dosing and timing of supplementation have to be studied.

Oligosaccharides and prebiotic effects in infant formulae

The main authors for this section are Professor Szajewska and Dr Stahl. The use of non-digestible carbohydrates in infant formulae and follow-on formulae has been commented on by the Committee on Nutrition of the European Society for Paediatric Gastroenterology, Hepatology and Nutrition⁽Reference Agostoni, Axelsson and Goulet¹¹⁸⁾. Based on the evidence obtained in a search up to January 2004, the Committee concluded that only a limited number of studies have evaluated the effects of the addition of substances with prebiotic effects to dietetic products for infants. Only one type of oligosaccharide mixture of galactans and ITF consisting of GOS and a high molecular weight fraction of inulin in a ratio of 9:1 (GOS/FOS) was evaluated. The Committee stated that although the administration of oligosaccharides with prebiotic effects has the potential to increase the total number of bifidobacteria in faeces and may also soften stools, there is no published evidence of any clinical benefits after addition of oligosaccharides with prebiotic effects to dietetic products for infants. No general recommendation on the use of oligosaccharide supplementation in infancy for preventive or therapeutic purposes can be made. The available data on the oligosaccharide mixtures in infant formulae do not demonstrate adverse effects. Validated clinical outcome measures of prebiotic effects in infants should be characterised in further well designed and carefully conducted randomised controlled trials (RCT), with relevant inclusion/exclusion criteria and adequate sample sizes. Such trials should also define the optimal quantities, types and intake durations.

A number of studies have been published thereafter on the addition of ingredients showing a prebiotic effect to dietetic products for infants and recently reviewed⁽Reference Boehm and Moro¹¹⁹⁾. These ingredients have been used either as one compound⁽Reference Yap, Mohamed and Husni^120–Reference Fanaro, Marten and Bagna¹²³⁾ or as a mixture of different neutral and acidic oligosaccharides⁽Reference Magne, Hachelaf and Suau¹²⁴⁾. Collectively, these studies confirm that the administration of oligosaccharides with prebiotic effects in dietetic products have the potential to increase dose dependently the total number of bifidobacteria in faeces, although at present, it is not possible to define the number of bifidobacteria that would constitute normal/optimal microbiota and to soften stools. Furthermore, prebiotic effects modulate stool pH, SCFA pattern similar to those of breast-fed infants. Whether any of these effects per se is of benefit is currently not well established. Clinical outcomes related to the use of dietetic products for infants supplemented with prebiotic effects are discussed in the forthcoming sections (e.g. effect on allergic diseases, infections).

Currently, the Directive 2006/141/EC on infant formulae and follow-on formulae specifically allows the addition of GOS–FOS in a ratio of 9/1 and in a quantity of 0·8 g/100 ml prepared product⁽¹²⁵⁾. The GOS and FOS were specified as ‘a combination of 90 % oligogalactosyl-lactose and 10 % high molecular weight oligofructosyl-saccharose’. This Directive also states that other combinations and maximum levels of FOS and GOS may be used if they satisfy the nutritional requirements of infants in good health as established by generally accepted scientific data.

Use of prebiotic effects in complementary foods for children

One controlled trial (RCT)⁽Reference Moore, Chao and Yang¹²⁶⁾ conducted in fifty-six healthy, term infants aged 4–12 months evaluated the tolerance and GI effects of an infant cereal supplemented with either ITF or placebo for 28 d. Compared with the control group, stool consistency was less often described as ‘hard’ and more likely to be described as ‘soft’ or ‘loose’ in the ITF-supplemented group. There was no difference between the groups in crying, spitting-up or colic. No difference in stool pH between the groups was found. There was also no significant difference in growth between the two groups. Clinical outcomes were not reported. The limitations of the present study include the use of non-validated tool for parental assessment of stool consistency, a small sample size and a short follow-up period.

Another double-blind RCT⁽Reference Scholtens, Alles and Bindels¹²⁷⁾ involving thirty-five infants aged 4–6 months studied the effect of adding GOS/FOS to solid foods results in an increase in the faecal proportion of bifidobacteria in the intestinal microbiota. Intention-to-treat analysis revealed no significant difference between the two study groups. Only per-protocol analysis involving twenty children who complied with the protocol showed that the faecal percentage of bifidobacteria increased from 43 to 57 % (P = 0·03) from weeks 0 to 6 but did non-significantly change in the control group (36 and 32 %, respectively, P = 0·4). There were no statistically significant differences in stool frequency and consistency.

More recently an indication for a prebiotic effect with ITF-fortified milk in children aged 7–8 years has also been reported⁽Reference Lien do, Nhung and Khan¹²⁸⁾.

Use of prebiotic effects for paediatric disorders

The effect of prebiotics in paediatric diseases has to be seen under the different aspect either of treatment or of prophylaxis. Theoretically, – and also clearly demonstrated in this part of the manuscript – prebiotics are more effective in prophylaxis more than in treatment. That seems logically because the prebiotic effect can only be seen after a certain period of time which is needed for the development of the microbiota (and which is significantly longer than the duration of an acute diarrhoea). In consequence, prebiotics are ideal candidates for prophylaxis but not for treatment.

Prebiotic effects and atopy

Atopic eczema is an itchy inflammatory skin condition with associated epidermal barrier dysfunction. Therapeutic options (emollients and topical steroids for mild-to-moderate eczema; topical or systemic calcineurin inhibitors, UV phototherapy, or systemic azathioprine for moderate-to-severe eczema) are relatively limited and often unsatisfactory, prompting interest in alternative treatment methods.

The rationale for using prebiotic effects in preventing atopic disorders is based on the concept that prebiotic effects modify the intestinal flora of formula-fed infants towards that of breast-fed infants. The intestinal flora of atopic children has been found to differ from that of controls with atopic subjects having more clostridia and tending to have fewer bifidobacteria than non-atopic subjects⁽Reference Kalliomaki, Kirjavainen and Eerola¹³⁹⁾. Thus, there is indirect evidence that differences in the neonatal gut microbiota may precede or coincide with the early development of atopy. This further suggests a crucial role for a balanced commensal gut microbiota in the maturation of the early immune system.

The Cochrane Review published in 2007⁽Reference Osborn and Sinn¹⁴⁰⁾ aimed at determining the effect of different ingredients showing a prebiotic effect (GOS/FOS, only FOS, GOS together with polydextrose and lactulose) on the prevention of allergic disease or food hypersensitivity in infants. Only two RCT of reasonable methodological quality according to the reviewers and involving 432 infants reported outcomes related to allergic disease. The reviewers concluded that there is insufficient evidence to determine the role of prebiotic supplementation of infant formula for the prevention of allergic disease and food hypersensitivity.

One of the included RCT⁽Reference Osborn and Sinn¹⁴⁰⁾ investigated the effect of the prebiotic mixture (GOS/FOS; dosage: 8 g/l) on the intestinal flora and the cumulative incidence of atopic dermatitis during the first 6 months of life in infants at risk for allergy (with at least one parent with documented allergic disease confirmed by physician). Of the 259 infants, 206 (79·5 %) infants who were randomly assigned to receive extensively hydrolysed whey formula supplemented either with 0·8 g GOS/FOS (experimental group, n 102) or with maltodextrin as placebo (control group, n 104) were included in the per-protocol analysis. The frequency of atopic eczema in the experimental group was significantly reduced compared with the placebo group (9·8 % v. 23·1 %, RR 0·42 (95 % CI 0·2, 0·8)), number needed to treat eight (95 % CI 5, 31). In a subgroup of ninety-eight infants, the parents provided fresh stool samples for microbiological analysis using plating techniques; the faecal counts of bifidobacteria were significantly higher in the group fed the GOS/FOS formula compared to the placebo group. No significant difference was found for the lactobacilli count between groups. Follow-up of the present study showed that at 2 years the cumulative incidences of atopic dermatitis, recurrent wheezing and allergic urticaria were higher in the placebo group (27·9, 20·6 and 10·3 %, respectively) than in the intervention group (13·6, 7·6 and 1·5 %) (P < 0·05). This is the first observation that prebiotic effects are able to reduce the incidence of atopic diseases and that this effect persists beyond the intervention period. This assessment is based on a per protocol evaluation which aims at testing efficacy; due to the high drop-out rate (20 % at 6 months and 48 % at 2 years of age) and lacking intention-to-treat analysis, effectiveness for field practice needs to be confirmed⁽Reference Arslanoglu, Moro and Schmitt¹⁴¹⁾ (see section ‘Prebiotic effects and mineral absorption’ of the present paper).

Prebiotic effects and gastro-intestinal infections

The main authors for this section are Professor Guarner and Dr Respondek (IBS), Dr Whelan (IBD) and Professor Rowland (colon cancer and bacterial activities). In adults, the use of ingredients showing a prebiotic effect in the fight against infections has hardly been studied. A few studies, dealing with different infectious problems, have been reported.

One study dealing with traveller's diarrhoea reports that consumption of 10 g ITF/d for a 2-week pre-travel period continued during a 2-week travel period to high and medium risk destinations had no effect on the prevention of traveller's diarrhoea, although the sense of ‘well-being’ was improved⁽Reference Cummings, Christie and Cole¹⁴²⁾. Furthermore, a study of patients consuming 12 g ITF/d while taking broad-spectrum antibiotics for 7 d, followed by another 7 d of the same treatment reported no difference from the placebo group regarding diarrhoea incidence, Clostridium difficile infection and hospital stay, while the number of faecal bifidobacteria increased significantly⁽Reference Lewis, Burmeister and Cohen¹⁴³⁾. In contrast, continued consumption of 12 g ITF/d for 30 d after the cessation of C. difficile-associated diarrhoea reduced the relapse rate, while increasing bifidobacteria levels⁽Reference Lewis, Burmeister and Brazier¹⁴⁴⁾.

Overall, the number of studies on the efficacy of ingredients showing a prebiotic effect in the prevention of infectious diseases is limited. Some positive outcomes exist alongside studies reporting no-effects. Clearly, a rationale is present for the use of such ingredients. However, any direct effect of the studied ingredients on the immune system cannot be excluded and the measurement of the putative associated effect on the microbiota is not always included in these studies, hindering the formation of any conclusions on possible underlying mechanisms.

Prebiotic effects and irritable bowel syndrome

The irritable bowel syndrome (IBS) is a functional bowel disorder manifested by chronic, recurring abdominal pain or discomfort associated with disturbed bowel habit, in the absence of structural abnormalities likely to account for these symptoms⁽Reference Spiller, Aziz and Creed¹⁴⁵⁾. The symptomatic array may include abdominal pain, discomfort, distension, cramping, distress, bloating, excess flatulence and variable changes in frequency and form of stools. Such symptomatic episodes may be experienced by almost every individual, and in order to separate IBS from transient gut symptoms, experts have emphasised the chronic and relapsing nature of IBS and have proposed diagnostic criteria based on the recurrence rate of such symptoms⁽Reference Longstreth, Thompson and Chey¹⁴⁶⁾. IBS is one of the most common intestinal disorders both in industrialised and in developing countries, and it is known to generate significant health care costs⁽Reference Spiller, Aziz and Creed¹⁴⁵⁾.

A precise aetiology for IBS is not recognised. However, epidemiological studies have identified a series of pathogenetic factors, including genetic and early environmental conditioning, cognitive/emotional adaptation, altered response to stress and inflammatory post-infectious processes of the gut mucosa, etc.⁽Reference Spiller, Aziz and Creed¹⁴⁵⁾. It has been shown that IBS patients have abnormal reflexes and perception in response to gut stimuli⁽Reference Serra, Salvioli and Azpiroz¹⁴⁷⁾. In subsets of patients, the underlying defects appear to be altered GI motility, visceral hypersensitivity, small bowel bacterial overgrowth, excess gas production, abnormalities in the composition of the gut microbiota (Table 10) or combinations of them⁽Reference Spiller¹⁴⁸⁾.

Table 10 Comparison of faecal microbiota between irritable bowel syndrome (IBS) and healthy control subjects

Among the modifications of the gut microbiota, a decrease of bifidobacteria and more specifically Bifidobacterium catenulatum has been observed in IBS patients in comparison to healthy subjects⁽Reference Balsari, Ceccarelli and Dubini^149–Reference Maukonen, Satokari and Mattö¹⁵⁵⁾.

Hypothetically, some of these disturbances may be corrected or counteracted by prebiotic effects. Indeed, compounds showing such effects are known to modulate the digestive microbiota and particularly to stimulate the growth of bifidobacteria especially when the initial level is low⁽Reference Rycroft, Jones and Gibson⁶⁴⁾. Furthermore, human studies with ITF or lactulose have shown that such prebiotics modulate gut transit⁽Reference Spiller^148, Reference Nyman¹⁵⁶⁾, decrease putrefactive activity within the gut lumen⁽Reference de Preter, Vanhoutte and Huys¹⁵⁷⁾, prevent GI infections⁽Reference Cummings, Christie and Cole^142, Reference Lewis, Burmeister and Brazier¹⁴⁴⁾ and mitigate inflammatory responses⁽Reference Lindsay, Whelan and Stagg^111, Reference Furrie, Macfarlane and Kennedy^158, Reference Casellas, Borruel and Torrejon¹⁵⁹⁾.

Indirect evidence for beneficial effects of ingredients showing a prebiotic effect on abdominal well-being was initially obtained in human trials addressing other primary end points. For instance, Cummings et al. ⁽Reference Cummings, Christie and Cole¹⁴²⁾ tested the effectiveness of ITF in preventing diarrhoea in 244 healthy subjects, travelling to high and medium risk destinations for travellers' diarrhoea (see section ‘Prebiotic effects and gastro-intestinal infections’ of the present paper for discussion of the effects on risk of intestinal infections). This randomised, double-blind, placebo-controlled study showed that consumption of 10 g ITF daily gave a significantly better sense of ‘well-being’ during the holiday, as recorded in post-study questionnaires. Likewise, Casellas et al. ⁽Reference Casellas, Borruel and Torrejon¹⁵⁹⁾ performed a prospective, randomised, double-blind, placebo-controlled trial to test the effect of ITF (12 g/d) in patients with active ulcerative colitis (UC). Interestingly, the study observed a significant decrease in abdominal symptoms with treatment but not with placebo, as assessed with the validated questionnaire of dyspepsia-related health scale⁽Reference Cook, Rabeneck and Campbell¹⁶⁰⁾.

Few studies have investigated the effect of ingredients showing a prebiotic effect in patients with IBS. The study by Olesen & Gudmand-Hoyer⁽Reference Olesen and Gudmand-Hoyer¹⁶¹⁾ tested a high dose of finally 20 g ITF during 12 weeks. The authors hypothesised that IBS symptoms may be provoked by large quantities of fermentable carbohydrates in the colon. After 4–6 weeks on treatment, IBS symptoms worsened, as expected, in patients on 20 g ITF/d and improved in patients on placebo. However, continuous treatment for 12 weeks resulted in adaptation and there were no differences between groups: symptoms improved in 58 % of the ITF group and in 65 % of the placebo group, and symptoms worsened in 8 % of the ITF group and in 13 % of the placebo group. Large doses of any fermentable carbohydrates should not be recommended to IBS patients.

Hunter et al. ⁽Reference Hunter, Tuffnell and Lee¹⁶²⁾ found no effect of 2 g ITF (three times daily) against placebo in a reduced group of IBS patients studied in a double-blind crossover trial. The Rome team of experts on functional bowel disorders do not recommend the use of a crossover design for IBS treatment trials as they have the potential disadvantages of carryover effects and unmasking the study product by differences in taste and palatability⁽Reference Irvine, Whitehead and Chey¹⁶³⁾. Dughera et al. ⁽Reference Dughera, Elia and Navino¹⁶⁴⁾ reported a positive effect of a synbiotic (including short-chain ITF at 2·5 g/d) on clinical manifestations and intestinal function in patients with IBS. However, this was an open-label and uncontrolled study, and IBS studies with subjective outcomes are prone to study bias⁽Reference Spiller¹⁴⁸⁾.

To date, there are two published studies of adequate study design reporting the effects of an ingredient showing a prebiotic effect in IBS. The first study screened 2235 subjects and recruited and randomised 105 patients with IBS fulfilling Rome II criteria with minor intensity of symptoms as assessed by an initial questionnaire. Treatment with short-chain ITF at 5 g/d for 6 weeks reduced incidence and intensity of symptoms as compared to the placebo product. Prebiotic treatment also improved functional digestive disorders related quality of life⁽Reference Paineau, Payen and Panserieu¹⁶⁵⁾.

The second study randomised forty-four subjects according to Rome II criteria into three groups either receiving 7 g/d placebo, 3·5 g/d of ingredient showing a prebiotic effect and 3·5 g/placebo and 7 g/d of the tested ingredient for 6 weeks. The prebiotic treatment significantly improved flatulence, bloating and composite score of symptoms as well-subjective global assessment. It also increased the proportion of bifidobacteria in faecal samples⁽Reference Silk, Davis and Vulevic¹⁶⁶⁾.

In summary, the two available studies with up to date standard provided positive outcomes for both the ITF and GOS tested up to 7 g. Results with less positive outcomes used either higher or lower doses.

Recommendations

Ingredients showing a prebiotic effect are likely to play a role in the symptomatic control of IBS. Evidence accumulated so far in well-designed clinical studies is limited, but suggests possible benefits at moderate doses. Further studies with adequate methodology are warranted.

Key points

1. (1) The IBS is a functional bowel disorder manifested by chronic, recurring abdominal pain or discomfort in the absence of structural abnormalities.

2. (2) The symptomatic array includes abdominal distension, cramping, distress, bloating, excess flatulence and variable changes in frequency and form of stools. Such symptomatic episodes may be experienced by almost every individual.

3. (3) The underlying defects appear to be altered GI motility, visceral hypersensitivity, small bowel bacterial overgrowth, excess gas production and abnormalities in the composition of the gut microbiota or combinations of these.

4. (4) Ingredient showing a prebiotic effect may counteract these disturbances as they were shown to modulate gut transit, decrease putrefactive activity within the gut lumen, prevent GI infections and mitigate inflammatory responses.

5. (5) To date, there are only two published studies of adequate study design testing such ingredient in IBS. Both studies improved the subjects' symptoms.

Prebiotic effects and inflammatory bowel disease

IBD is a chronic relapsing and remitting disorder characterised by inflammation, ulceration and stricturing of the GI tract. UC and Crohn's disease (CD) are the two main types of IBD. In Europe, the incidence ranges from 1·5 to 20·3 cases per 100,000 person-years for UC and from 0·7 to 9·8 cases per 100,000 person-years for CD, meaning that up to 2·2 million people in Europe currently live with IBD⁽Reference Loftus¹⁶⁷⁾.

UC causes continuous mucosal inflammation that is restricted to the colon, whereas CD causes discontinuous transmural inflammation anywhere throughout the GI tract, although it most frequently affects the terminal ileum⁽Reference Travis, Stange and Lemann¹⁶⁸⁾. Symptoms common to both UC and CD include diarrhoea, faecal urgency and incontinence. Severe abdominal pain and rectal bleeding are common and complications such as fissuring and abscesses may occur. These symptoms can have a profound impact on patients, with evidence of impaired nutritional status⁽Reference Lucendo and De Rezende¹⁶⁹⁾ and quality of life⁽Reference Irvine¹⁷⁰⁾.

The primary treatment approach in IBD is usually drug therapy. Patients can be treated with a variety of drugs, including 5-ASAs (e.g. mesalazine), steroids (e.g. prednisolone) and immunosuppressants (e.g. azathioprine). In addition, patients with CD may also receive new biological drugs such as monoclonal antibodies (e.g. the anti-TNF-α antibody infliximab) when standard drug treatment fails⁽Reference Schwartz and Cohen¹⁷¹⁾. Despite their general efficacy, such drugs can carry a significant burden. They are not only expensive but also side effects are common, with an incidence of 28 % for immunosuppressants, rising to 50 % for steroids⁽Reference Carter, Lobo and Travis¹⁷²⁾. In addition, approximately 30 % of patients with UC and 50 % of patients with CD will require surgery at some point in their life⁽Reference Carter, Lobo and Travis¹⁷²⁾. In the case of UC, a colectomy and formation of an ileo-anal pouch may be curative. However, following this procedure, a minority of patients will experience relapsing, remitting pouch inflammation, described as pouchitis.

Nutritional approaches to treating IBD have been investigated. In clinical trials, enteral nutrition has been shown to induce remission in 60–85 % of patients with CD, however, it remains less effective than steroids⁽Reference Zachos, Tondeur and Griffiths¹⁷³⁾ and patients report problems with palatability and abstinence from food⁽Reference Teahon, Pearson and Levi¹⁷⁴⁾. In view of these findings, safe and effective interventions that induce and maintain remission in IBD with a low incidence of side effects are urgently needed.

In order to identify potential therapeutic targets for IBD, examination of its pathogenesis is required. Although the precise mechanisms are not yet known, it appears that IBD results from a heightened mucosal immune response to the GI microbiota in genetically susceptible individuals.

The immunological processes underlying IBD involve alterations in the balance of proinflammatory and immuno-regulatory cytokines within the mucosal immune system. Much of the inflammation is mediated via cytokines released by activated Th1/Th17 lymphocytes. In addition, TNF (TNF)-α has been shown to play a key role, exerting its effects via stimulation of other proinflammatory cytokines such as IL (IL)-1, IL-6 and IFN-γ. Each of these proinflammatory cytokines have been shown to be elevated during active IBD⁽Reference Neuman¹⁷⁵⁾, and biological therapies such as anti-TNF-α-antibodies directly target this immunological cascade. Other proinflammatory cytokines include IL-12 and IL-18, both of which are involved in IFNγ production. In contrast, the immuno-regulatory response is mediated by cytokines such as IL-10, which down-regulates IFNγ production⁽Reference Lindsay and Hodgson¹⁷⁶⁾. Furthermore, some animal studies have indicated immuno-regulatory roles for IL-4 and transforming growth factor (TGF)-β in IBD⁽Reference Brown and Mayer¹⁷⁷⁾.

There is convincing evidence that the inflammation observed in IBD is driven by the GI microbiota. For example, it has been shown that animal models of IBD do not develop inflammation when reared in germ-free conditions, whereas they subsequently develop inflammation once transferred to non-sterile conditions or are artificially colonised with bacteria⁽Reference Sellon, Tonkonogy and Schultz¹⁷⁸⁾. Similar observations have been described in human subjects with IBD. In patients with colonic CD, formation of an ileostomy, which diverts the faecal stream away from the site of inflammation, results in disease remission in 65 % of patients, while reversal of this procedure results in disease relapse in 60 %, implying that the content of the faecal stream is in part responsible for driving inflammation⁽Reference Fasoli, Kettlewell and Mortensen¹⁷⁹⁾. Patients with active IBD also have elevated GI permeability, thereby increasing the exposure of the mucosal immune system to the resident microbiota⁽Reference Chichlowski and Hale¹⁸⁰⁾. An underlying pathogenic mechanism linking CD and the GI microbiota was realised when it was found that mutations in the caspase-activating recruitment domain 15 (CARD15) gene, involved in bacterial recognition, were found to result in a 38-fold increase in risk for CD⁽Reference Hugot, Chamaillard and Zouali¹⁸¹⁾. Interestingly, this mutation does not result in a higher risk of UC and further genome wide association studies have identified numerous other mutations associated with increased risk of either UC or CD but that are unrelated to bacterial recognition or sensing⁽Reference Zhang, Massey and Tremelling¹⁸²⁾. Therefore, there are clearly genetic and environmental triggers related to the onset of IBD other than those involving the GI microbiota.

Despite the evidence that the GI microbiota is necessary to drive the inflammation in IBD, some bacteria may indeed protect the mucosa from such inflammation. Studies in both animals models and patients with IBD have shown that some bacteria decrease abnormal GI permeability⁽Reference Miyauchi, Morita and Tanabe^183, Reference Garcia, De Lourdes De Abreu Ferrari and Oswaldo Da Gama¹⁸⁴⁾, thereby reducing exposure of the mucosal immune system to the GI microbiota. Meanwhile, some probiotics, in particular bifidobacteria, up-regulate immuno-regulatory IL-10 production by dendritic cells⁽Reference Hart, Lammers and Brigidi^185, Reference Ng, Plamondon and Hart¹⁸⁶⁾, the production of which is therapeutic in animal models of IBD⁽Reference Lindsay and Hodgson¹⁷⁶⁾. In view of this, studies have shown some success of both antibiotics and probiotics in the management of IBD and these have been extensively reviewed elsewhere⁽Reference Sartor^187, Reference Hedin, Whelan and Lindsay¹⁸⁸⁾.

Components of the GI microbiota therefore drive proinflammatory and/or immuno-regulatory cytokine production during IBD. Interestingly, numerous studies demonstrate alterations in the GI microbiota of patients. Such studies are varied, utilising a wide variety of microbiological techniques (e.g. traditional culture and molecular microbiology) in different samples (i.e. faeces, inflamed mucosa, non-inflamed mucosa). Comparisons have been made between UC and/or CD and/or healthy controls, and these vary as to whether patients were in relapse or remission. Consequently, studies of the GI microbiota in IBD are too varied to review in detail here. However, some conclusions can be drawn regarding the alterations in GI microbiota in IBD that suggest that ingredients showing a prebiotic effect may be of potential benefit in its treatment or maintenance.

In general, studies adopt two different approaches to investigating the microbiota in IBD. Some investigate differences in concentration, proportion or diversity of microbial communities (i.e. dysbiosis theory), whereas others investigate the presence or absence of selected species (i.e. single strain theory). For example, patients with inactive CD have been shown to have lower proportions of faecal bifidobacteria⁽Reference Seksik, Rigottier-Gois and Gramet^189, Reference Sokol, Seksik and Furet¹⁹⁰⁾, whereas both patients with active UC or active CD have lower faecal bifidobacteria, C. coccoides and C. leptum compared with healthy controls⁽Reference Sokol, Seksik and Furet¹⁹⁰⁾. Lower concentrations of bifidobacteria⁽Reference Macfarlane, Furrie and Cummings^191, Reference Mylonaki, Rayment and Rampton¹⁹²⁾ and higher concentrations of bacteroides⁽Reference Swidsinski, Ladhoff and Pernthaler¹⁹³⁾ have also been found in the mucosa of both patients with UC or CD. Meanwhile, another study has shown that some patients with CD or UC have lower numbers of mucosal Firmicutes and Bacteroidetes⁽Reference Frank, St Amand and Feldman¹⁹⁴⁾. Increased presence of E. coli has been demonstrated in patients with UC or CD⁽Reference Sokol, Lepage and Seksik^195, Reference Martinez-Medina, Aldeguer and Gonzalez-Huix¹⁹⁶⁾ and more recently, lower concentrations of F. prausnitzii were found in the faeces of patients with CD or UC compared with controls⁽Reference Sokol, Seksik and Furet¹⁹⁰⁾. This is important as F. prausnitzii is immuno-regulatory and higher mucosal concentrations are associated with longer maintenance following surgically induced remission of CD⁽Reference Sokol, Pigneur and Watterlot¹⁹⁷⁾.

In view of the role of the certain components of the GI microbiota in driving intestinal inflammation, combined with the apparent dysbiosis in IBD, the use of ingredients showing a prebiotic effect as an approach to modifying the microbiota in order to induce or maintain remission in IBD has been investigated.

The prebiotic concept is defined as the selective stimulation of growth and/or activity of one or a limited number of microbial genera, species or strains in the gut microbiota that confers health benefits to the host. Ingredients showing a prebiotic effect have been shown to increase faecal and mucosal bifidobacteria in healthy subjects⁽Reference Kolida and Gibson^198, Reference Langlands, Hopkins and Coleman¹⁹⁹⁾. This is relevant because bifidobacteria are present in lower concentrations in the faeces and mucosa of patients with IBD⁽Reference Sokol, Seksik and Furet^190, Reference Mylonaki, Rayment and Rampton¹⁹²⁾, while in vitro experiments have shown that some species of bifidobacteria stimulate IL-10 production, potentially via interaction with TLR on lamina propria dendritic cells⁽Reference Hart, Lammers and Brigidi¹⁸⁵⁾. In addition, prebiotic ITF have recently been shown to increase concentrations of F. prausnitzii in healthy subjects⁽Reference Ramirez-Farias, Slezak and Fuller²⁰⁰⁾, although this has not yet been confirmed in patients with IBD. Furthermore, SCFA, produced through the fermentation of such ingredients, modulate inflammation, with cell-culture studies showing that butyrate inhibits pro-inflammatory IL-2 and IFNγ production and acetate and propionate increases immuno-regulatory IL-10 production⁽Reference Cavaglieri, Nishiyama and Fernandes⁹⁵⁾. In addition, mucin production is stimulated by both propionate and butyrate⁽Reference Burger-van, Vincent and Puiman²⁰¹⁾. Mucins (such as MUC2) are required for the maintenance of the mucous layer that enables epithelial protection and which may be lower in patients with IBD⁽Reference Pullan, Thomas and Rhodes²⁰²⁾.

Numerous experiments have been conducted to investigate the impact of these ingredients on chronic intestinal inflammation in animal models of IBD, and these have been reviewed elsewhere⁽Reference Leenen and Dieleman²⁰³⁾. However, at the current time, their use among patients with IBD remains relatively low⁽Reference Hedin, Graczer and Sanderson²⁰⁴⁾. However, over the last decade, there has been an increase in the number of clinical trials investigating their use in inducing or maintaining remission in IBD (Table 11).

Table 11 Clinical trials on the prebiotic effect in inflammatory bowel disease

FOS, fructo-oligosaccharides; DB-RCT, double-blind randomised controlled trial; CO, crossover; UC, ulcerative colitis; CD, Crohn's disease; ESR, erythrocyte sedimentation rate; TLR-2, toll-like receptor 2.

* Numbers recruited to each group.

Prebiotic effects and colon cancer

Prebiotic protective effects and bacterial activities

Summary and conclusion

1. (1) Data from animal models as well as preliminary evidence in human study suggest reduction in the risk of colon cancer development associated with the prebiotic effects.

2. (2) Data from animal models, with endpoints such as DNA damage, AC foci and tumours in the colon, suggest that reduction in the risk of colon cancer development is associated with prebiotic effects.

3. (3) Limited animal studies also indicate that combinations of prebiotics and probiotics may be more effective than either agent alone.

4. (4) A prebiotics and probiotics study in human subjects using putative biomarkers of cancer risk showed improvements in some, including a reduction in DNA damage and cell proliferation in colon biopsies. Further studies are needed.

5. (5) A number of potential mechanisms for reduction in cancer risk by prebiotic effect, including changes in gut bacterial enzyme activities, up-regulation of apoptosis and induction of protective enzymes have been explored in animal models, but currently evidence for such effects in human subjects is lacking.

Rationale behind the prebiotic effects on mineral absorption

Summary of key studies

Table 12 provides the list of review papers dealing with the effect of prebiotics on mineral metabolism.

Table 12 Published reviews on the prebiotic effect on mineral metabolism

FOS, fructo-oligosaccharides; GOS, galacto-oligosaccharides; SOS, soy-oligosaccharides; TOS, transgalacto-oligosaccharides; ITF, inulin-type fructans.

Animal study

Animal studies targeting the effect of prebiotics on Ca absorption are listed in the Tables 13 and 14. The points arising from these studies are the following:

1. (1) Different types of molecules have been studied, including ITF-_{Dpav 3-4}, ITF-_Dpav12, ITF-_Dpav25, ITF-_MIX, GOS, lactulose or resistant starch.

2. (2) Dietary supplementation with ITF enhances the uptake of Ca, improves bone mineral content (BMC) in growing rats and alleviates the reduction in bone mineral content and bone mineral density (BMD) which follows ovariectomy or gastrectomy in rats.

Table 13 The prebiotic effects on bone metabolism in the rat

GOS, galacto-oligosaccharides; FOS, fructo-oligosaccharides; AAS, atomic absorption spectrophotometry; ICPMS, inductively coupled plasma MS; DP, degree of polymerisation; BMC, bone mineral content; BMD, bone mineral density; OF, oligofructose; DEXA, dual-energy X-ray absorptiometry; ITF, inulin-type fructans; DPD, deoxypyridinoline; OVX, ovariectomized. Femoral mechanical testing (three-point bending test).

Table 14 The prebiotic effects on mineral absorption in the rat

FOS, fructo-oligosaccharides; GOS, galacto-oligosaccharides; TOS, transgalactosylated oligosaccharides; OF, oligofructose; AAS, atomic absorption spectrometry; ICPOES, inductively coupled plasma optical emission spectroscopy; DP, degree of polymerisation; ITF, inulin-type fructans; ICPMS, inductively coupled plasma MS; OVX, ovariectomized.

Apparent absorption: Ca intake (I) − Ca fecal excretion (F).

Net retention: Ca intake (I) −  (Ca fecal excretion (F)+Ca urinary excretion (U)).

True intestinal Ca absorption: (Ca⁴⁵ Ca⁴⁴) = (I −  F)+f (endogenous net Ca excretion).

Fractional Ca absorption: Ca⁴⁷, Sc⁴⁹ ratio (I − F).

Ca balance: 4–7 d balance period (I, F, U using metabolic cages) % Ca⁴⁵ absorption: % Ca⁴⁵ oral dose/% Ca⁴⁵ IP dose × 100.

Clinical trials

In infants

The only available study targeting the prebiotic effect on mineral metabolism in infants was conducted in 6–12 months healthy formula-fed babies (Tables 15 and 16). Even though, ITF did not elicit any modulation of faecal SCFA concentration, a beneficial effect on both Fe and Mg absorption and retention was reported. No significant difference was observed for Ca, Cu or Zn⁽Reference Yap, Mohamed and Yazid²⁶⁹⁾.

Table 15 The prebiotic effects on mineral absorption in the human

Sc, short chain; R, randomized; DB, double-blind; AAS, atomic absorption spectrometry; CO, crossover; OF, oligofructose; ICPMS, inductively coupled plasma MS; ScFOS, short-chain fructo-oligosaccharides; ITF, inulin-type fructans; TIMMS, thermal ionisation magnetic sector MS; PC, placebo control; CPP, casein phosphopeptide.

Fractional Ca: (Ca⁴⁴, Ca⁴³) ratio; (Ca⁴⁶, Ca⁴²) ratio.

Table 16 The prebiotic effects on human bone health

ScFOS, short-chain fructo-oligosaccharides; ITF, inulin-type fructans; BMC, bone mineral content; BMD, bone mineral density; DB, double-blind; PC, placebo control; DEXA, dual-energy X-ray absorptiometry; DPD, deoxypyridinoline; CO, crossover; PTH, parathyroid hormone; IRMA, immunoradiometric assay; POM, postmenopausal women; ICPMS, inductively coupled plasma MS; OC-DPD, osteocalcin deoxypyridinoline; PP, caseinophosphopetide; b-ALP, bone specific alkaline phosphatase; ELISA, enzyme-linked immunosorbent assay; CPP, casein phosphopeptide; NTx, N-Telopeptide.

In adolescents

As far as adolescents are concerned, in 1999, van den Heuvel et al. ⁽Reference van den Heuvel, Muys and van Dokkum²⁷⁰⁾ demonstrated that a daily consumption of 15 g of ITF for 9 d stimulated fractional Ca absorption by 10 % in young boys (14–16 years). Later on, Griffin et al. ⁽Reference Griffin, Davila and Abrams²⁷¹⁾ provided the evidence that modest intake of ITF_MIX, corresponding to 8 g/d, stimulated Ca absorption in sixty girls at or near menarche. The increase reached about 30 % after 3 weeks of consumption, when compared with oligofructose only or placebo intakes.

This effect was mostly observed in girls with lower Ca absorption status⁽Reference Griffin, Hicks and Heaney²⁷²⁾. Moreover, when given for 36 d to adolescent girls (12–14 years), 10 g of ITF-_{Dpav 3-4} were able to stimulate Mg absorption (18 %), without affecting Ca absorption, vitamin D or parathyroid hormone (PTH) serum concentration or urine concentration which are used as markers of bone resorption⁽Reference van den Heuvel, Muijs and Brouns²⁷³⁾.

The longest and most compelling study is a 1-year intervention trial on pre-pubertal girls and boys (n 100) that found significantly increased Ca absorption in the group receiving ITF-_MIX (8 g/d) after 8 weeks. The effect lasted throughout the intervention period resulting, after 1 year, in improved whole body bone mineral content and significantly increased BMD, compared with the controls⁽Reference Abrams, Griffin and Hawthorne²⁷⁴⁾. This demonstrates a beneficial effect on long-term use of this particular mixture on Ca absorption and bone mineralisation in young adolescents⁽Reference Cashman²⁷⁵⁾. A further study by Abrams et al. ⁽Reference Abrams, Griffin and Hawthorne²⁷⁶⁾ showed that responders to the ‘treatment’ had greater Ca absorption and increased accretion of Ca to the skeleton, and thus concluded on the importance of such a strategy to enhance peak bone mass, as the extra absorbed Ca is deposited in bones.

In adults

It has been previously shown, using the metabolic balance methodology, that addition of up to 40 g/d of ITF and sugarbeet fibres to a normal mixed diet for 28 d improved Ca balance, without adverse effects on the retention of other mineral⁽Reference Coudray, Bellanger and Castiglia-Delavaud²⁷⁷⁾. However, a study carried out by van den Heuvel et al. ⁽Reference van den Heuvel, Schaafsma and Muys²⁷⁸⁾ in healthy young adults found no significant differences in mineral absorption, irrespective of the treatment (which consisted of a constant basal diet supplemented for 21 d with 15 g/d ITF, or GOS, or not supplemented) followed by a 24 h urine collection. It was hypothesised that a 24 h period of urine collection, used in the study, was too short to include the colonic component of Ca absorption and thus to make up a complete balance necessary to detect the effect of ITF. In a similar way, Teuri et al. ⁽Reference Teuri, Karkkainen and Lamberg-Allardt²⁷⁹⁾ investigated a combination of 15 g of ITF and 210 mg of Ca added to 100 g of cheese given at breakfast to fifteen adult healthy women with an average age of 23 years old. The study failed to show any significant influence of the diet on blood ionised Ca or parathyroid concentration over the 8 h assessment period. Nevertheless, measuring serum parathyroid and ionised Ca do not provide direct information about Ca absorption, as do isotope techniques, and it has been suggested that the length of the trial was probably too short. Moreover, the addition of 1·1 g ITF-_Dpav3-4 or caseinophosphopeptides to Ca-enriched milks, a valuable source of well-absorbed Ca, did non-significantly increase Ca absorption in adults (25–36 years), independently of sex⁽Reference Lopez-Huertas, Teucher and Boza²⁸⁰⁾. Finally, Abrams et al. ⁽Reference Abrams, Hawthorne and Aliu²⁸¹⁾ gave a supplementation containing 8 g of ITF-_MIX for 8 weeks to thirteen young adults (average age of 23 years). Eight of the thirteen volunteers were classified as responders, based on their level of Ca absorption.

In postmenopausal women

Ducros et al. ⁽Reference Ducros, Arnaud and Tahiri²⁸²⁾ carried out a clinical trial in postmenopausal women (age between 50 and 70 years with at least 2 years of menopause). The volunteers were provided with 10 g/d ITF-_Dpav3-4 or a placebo for 5 weeks using a crossover design. They demonstrated that consumption of ingredients showing a prebiotic effect was associated with increased Cu absorption, while no significant effect could be demonstrated on Zn or Se bioavailability.

In a similarly designed double-blind randomised, crossover design, post-menopausal women without hormone replacement therapy were given 10 g of ITF-_Dpav3-4 daily for 5 weeks. Mg absorption and status was determined using mass spectrometer analysis in faeces, urine and blood. Results showed that the ITF-_Dpav3-4-enriched diet increased Mg absorption by 12·3 %, compared to the placebo sucrose control group⁽Reference Tahiri, Tressol and Arnaud²⁸³⁾. In the same experiment, Tahiri et al. ⁽Reference Tahiri, Tressol and Arnaud²⁸⁴⁾ showed that over 5 weeks of a moderate daily dose (10 g) of ITF-_Dpav3-4 failed to modify intestinal Ca absorption in the early postmenopausal phase, while in the subgroup of late phase (women who had been going through menopause for more than 6 years), an increase in Ca absorption was observed.

Twelve older postmenopausal women (of at least 5 years past the onset of menopause) drank 100 ml of water containing 5 or 10 g of lactulose or a reference substance at breakfast for 9 d. True fractional Ca absorption was calculated using Ca isotope ratios, and consumption of lactulose was found to increase Ca absorption in a dose–response way⁽Reference van den Heuvel, Muijs and van Dokkum²⁸⁵⁾.

In a crossover trial, twelve postmenopausal women were given a 200 ml yogurt to drink twice a day (at breakfast and lunch) containing either GOS (20 g) or sucrose for 9 d; a greater true Ca absorption (16 %) was observed after consumption of a product rich in GOS. In addition, no increased urinary Ca excretion was observed, suggesting that GOS could also indirectly increase the uptake of Ca by bones and/or inhibit bone resorption⁽Reference van den Heuvel, Schoterman and Muijs²⁸⁶⁾.

Adolphi et al. ⁽Reference Adolphi, Scholz-Ahrens and de Vrese²⁸⁷⁾ tested the hypothesis that, in postmenopausal women (between 48 and 67 years and who had been postmenopausal for 10·5 (sem 0·7) years consumption of fermented milk (supplemented with Ca) at bedtime could prevent the nocturnal peak of bone resorption by decelerating its turnover and that this effect could be improved by adding Ca absorption enhancers. Actually, they showed that indeed such a practice can reduce the nocturnal bone resorption and that supplementation with Ca had no additional effect unless absorption enhancers such as ITF and caseinphosphopeptides were added.

Kim et al. ⁽Reference Kim, Jang and Lee²⁸⁸⁾ who investigated the effects of ITF supplementation (8 g/d for 3 months) in postmenopausal women (mean age: 60 year) showed that apparent Ca absorption was significantly increased by 42 % in the ITF group, while a 29 % decrease was observed in the placebo group. This was associated with lower alkaline phosphate plasma levels (a parameter which is actually not specific of bone formation) and a trend towards a slight reduction in urinary deoxypyridinolin (a biomarker for bone resorption). As expected, due to the very short length of exposure, BMD was not modified by the treatment.

Finally, fifteen women (who were a minimum of 10 year past the onset of menopause and had taken no hormone replacement therapy for the past years) were treated with 10 g/d of a specific mixture of ITF for 6 weeks, according to a double-blind, placebo-controlled crossover design. True fractional Ca absorption, measured by dual isotopes before and after treatment, was significantly increased (+7 %) in women with lower initial BMD⁽Reference Holloway, Moynihan and Abrams²⁸⁹⁾.

In institutionalised patients

Bone resorption, used as indicator of Ca retention, remained unchanged in institutionalised adults after 3 weeks of treatment with 13 g/day of ITF-fortified beverages⁽Reference Dahl, Whiting and Isaac²⁹⁰⁾.

Outline of general rules

From mineral absorption to health benefits

The key question of whether the extra absorption of minerals may exhibit substantial benefits needs to be addressed.

Key points

1. (1) Ingredients showing a prebiotic effect are able to improve mineral absorption (and especially Ca) in the animals.

2. (2) Most data are available for ITF, in particular ITF-_Dpav3-4 as well asITF-_MIX.

3. (3) ITF have been found to increase Mg absorption in human subjects, nevertheless available data are very limited.

4. (4) These ingredients are able to enhance Ca absorption in human, depending on their physiological status (no effect in early postmenopausal women).

5. (5) The benefits on Ca absorption can be translated into benefits to bone health in animals.

6. (6) More interestingly, ITF-_MIX given for 1 year to adolescents was able to elicit not only an enhancement of Ca accretion in bones but also BMD. In this light, such or similar may have important implications for future preventative startegies for osteoporosis.

7. (7) A combination of molecules with different degrees of polymerisation appears to be more efficient as shown with the research on ITF-_MIX in comparison with the small and high MW fractions given alone.

Recommendations (future targets for research)

1. (1) Further studies are required to investigate the underlying mechanisms of the prebiotic effects on absorption of minerals, with special attention to the role of the specific changes in gut microbiota. Indeed the question still remains open of whether these effects are due to the changes in colonic microbiota composition (prebiotic effect) or any other mechanisms. In this regard, high-throughput methodologies such as metabolomics, for example, are warranted.

2. (2) Results from ITF, in particular ITF-_MIX, need to be confirmed in other ingredients showing a prebiotic effect for a generalisation.

3. (3) Further long-term well-designed clinical trials need to be implemented to prove that the benefits of these ingredients persist in the longer term (because bone strength is the ultimate hallmark of bone quality, the issue of persistence of the effect of ITF-_DPav3-4 on the skeleton is important to consider) to assess their potential in the prevention of the risk of fracture.

4. (4) With regards to the bone target, it is interesting to focus on relevant populations, i.e. during childhood and during ageing.

5. (5) It is still challenging to investigate the potential synergy between the prebiotic effect and other nutrients (such as phytoestrogens) endowed with bone-sparing effect.

Description of the prebiotic effects on obesity and related metabolic disorders

Relation between prebiotic effects and improvement of obesity and associated disorders

Methodological aspects

Key questions remain open concerning the adequacy of the experimental protocol to estimate the relevance of ingredients showing a prebiotic effect in the management of obesity and associated disorders. The choice of a placebo is rather difficult, and the type of placebo compounds is different when experiments are conducted in animals or in human subjects. There may also be differences when considering end points such as fat mass development or satiety, or glucose/lipid homeostasis.

In animal studies, the authors often add ingredients showing a prebiotic effect at a relatively high dose (1–10 % wt/wt in the diet) to compare the data obtained in animals receiving the basal diet alone. The interpretation of results would then require the difference in energy/nutrients intake and/or an experimental group with the same intake of energy upon the treatment (pair-fed animals) to be taken into account. Other authors propose to replace the amount of ingredients showing a prebiotic effect by a non-digestible–non fermentable carbohydrate such as microcrystalline cellulose as placebo. This allows a comparison based on differential fermentation properties.

For human studies, the dose of ingredients showing a prebiotic effect is much lower (from 1 to 30 g/d). The organoleptic and physico-chemical properties of the placebo are very important to take into account. Several placebos are proposed in the literature, e.g. a digestible carbohydrate, such as maltodextrin – i.e. alone⁽Reference Cani, Joly and Horsmans^324, Reference Cani, Lecourt and Dewulf³²⁷⁾, or in combination with aspartame⁽Reference Giacco, Clemente and Luongo³⁴¹⁾ – or saccharose⁽Reference Luo, Rizkalla and Alamowitch^339, Reference Luo, Van Yperselle and Rizkalla³⁴⁰⁾, dietary fibres such as oat fibre⁽Reference Peters, Boers and Haddeman³³¹⁾.

The choice of the adequate placebo is really difficult and will depend on the end point and duration of the treatment. When estimating the influence on glucose/lipid metabolism, one must consider a placebo that does not change postprandial glucose level or has a minor impact as lipogenic substrate, for example.

For studies aiming at controlling appetite and energy, one has to choose an adequate placebo which does not exert an effect per se. When estimating a long-term effect on body weight composition, the consequence of placebo treatment on global energy intake must be taken into account.

There are, therefore, several possibilities and the interpretation and discussion of the results might also take into account the differences that could be due to the placebo effect in a specific context.

Conclusions and future trends

Collectively, these studies provide support for the beneficial effect of prebiotics on energy homeostasis and body weight gain. Only a few human studies are available to date, but some of them support a role of gut peptide modulation by ingredients showing a prebiotic effect as a potential mechanism occurring in the gut, and appetite regulation. The question of the relevance of gut microbiota modulation in these effects remains unexplored in most of the studies performed in human subjects. In mice, an inverse relationship has been established between the level of faecal bifidobacteria and some features of the metabolic alterations linked to obesity (endotoxemia, fat mass, glucose intolerance). Some other non-digestible carbohydrates or dietary fibres (i.e. resistant starch, insoluble fibre form barley) – for which prebiotic effect has not yet been established – would be able to modulate gut peptides production with consequences on appetite, inflammation and other components of the metabolic syndrome. The analysis of the gut microbiota changes will be crucial in further research and clinical approach, in order to clearly relate those changes with the improvement of metabolic alterations of the host. This will be the way to propose a ‘targeted approach in the modulation of gut microbiota by ingredients showing a prebiotic effect’ as relevant in the context of obesity.

Which data to support the hypothesis of a causal relationship between a prebiotic effect and health effects/benefits?

The author of this section is Professor Marcel B. Roberfroid. A prebiotic effect exists and is now a well-established scientific fact. A large number of human intervention studies have demonstrated that dietary consumption of food products/ingredients/supplements results in statistically significant changes in the composition of the faecal (and in some cases, the mucosal) gut microbiota. Most of the available data concern the selective stimulation of bifidobacteria (but also lactobacilli). Other purportedly beneficial genera such as Roseburia and Eubacterium may be more fully investigated in the future – although further evidence of their beneficial effects is required. Some, but not all, studies have reported a reduction in the concentration of pathogenic bacteria such as clostridia and salmonella. The more data are accumulating, the more it will be recognised that such changes in the composition of the faecal microbiota, especially increase in bifidobacteria can be regarded as a marker of intestinal health. This is already supported by scientific publications⁽Reference Gronlund, Gueimonde and Laitinen^376–Reference Salminen, Collado and Isolauri³⁸⁰⁾.

Research on the impact of the prebiotic effect on the activity (metabolic, regulatory and signalling) of the microbiota is ongoing and appropriate relevant methodologies are being developed, validated and applied.

1. (1) Results from experimental models but also in a few human studies, food products/ingredients/supplements with a demonstrated prebiotic effect have been shown to modulate certain immunological biomarkers and affect activity(ies) of the immune system. Whether changes in immune function markers or immune-health benefits are related to a prebiotic-induced change in the composition of the gut microbiota is an area for future investigation. While several studies report changes in the faecal microbial composition alongside changes in immune markers, only one study sofar has correlated these findings. Although these observations make the link between immuno-modulation and microbiota changes likely, convincing evidence needs to be established by further studies showing clear correlations between parameters of immune function and changes in the microbiota. Although a causal relationship is virtually impossible to establish in human subjects, current plausible hypotheses and future correlative findings will help to establish the correlation between prebiotic modulation of the intestinal microbiota and changes in immune function.

2. (2) The effect of breast-feeding on infant gut microbiota composition is well established and mother's milk is known to contain a complex mixture oligosaccharides with prebiotic (especially bifidogenic) effects, therefore, infant formulae/foods have been supplemented with prebiotics. Confirming the studies in adults, it has been demonstrated that such supplementation increases the faecal concentration of bifidobacteria. This concomitantly improves stool quality (soft and loose stools), reduces the risk of gastro-enteritis, improves general well-being and reduces the frequency of atopic eczema. It is plausible that these effects were microbiota-induced changes.

3. (3) Changes in the gut microbiota composition are classically considered as one of the many factors involved in the pathogenesis of either IBD or IBS. The use of particular food products/ingredients/supplements with prebiotic effects has thus been tested in clinical trials with the objective to improve the well-being of patients with such disease states. Promising beneficial effects have been demonstrated in some but still preliminary studies with changes in gut microbiota composition (especially increase in bifidobacteria concentration) being associated. Again, it is feasible to conclude that the mechanism of these effects is linked to the prebiotic effect.

4. (4) Colon cancer is another pathology for which a possible role of gut microbiota composition has been hypothesised. Numerous experimental studies in mice and rats have reported reduction in incidence of tumours and cancers after feeding specific food products/ingredients/supplements with prebiotic effects. Some of these studies (including one human trial) have also reported that, in such conditions, gut microbiota composition was modified (especially due to increased concentration of bifidobacteria), however, role of such changes in the eventual anti-cancer effect of these specific food products/ingredients/supplements remains to be definitively proven.

5. (5) Dietary intake of particular food products/ingredients/supplements with a prebiotic effect has been shown, especially in adolescents, but also tentatively in postmenopausal women to increase Ca absorption as well as bone Ca accretion and BMD. No correlation has been reported between such a beneficial effect and changes in gut microbiota composition – although this is plausible but not exclusive. However, other food products/ingredients/supplements that do not show prebiotic effect (e.g. lactose, miscellaneous dietary fibres) have also been reported to exert similar effects. Moreover, a study in adolescents revealed the existence of a genetic component in response (with one-third of non-responders) to increased Ca absorption. It is thus likely that improved Ca absorption is not uniquely caused by changes in gut microbiota composition and might be a consequence of a combination of different effects. Preliminary data have reported, mainly in experimental models, that specific food products/ingredients/supplements with prebiotic effects could also increase the absorption of other minerals (e.g. Mg, Fe). More research is needed to confirm these data and, eventually, to demonstrate if their mechanism involves changes in gut microbiota composition.

6. (6) Recent data, both from experimental models and from human studies, support the beneficial effects of particular food products/ingredients/supplements with prebiotic properties on energy homeostasis, satiety regulation and body weight gain. Together with data that correlate obesity with differences in gut microbiota composition, these studies have led to hypothesise that gut microbiota composition (especially the number of bifidobacteria) may contribute to modulate metabolic processes associated with syndrome X, especially obesity and diabetes type 2. In a study on the mechanism of action of a prebiotic food ingredient in reducing obesity, an inverse correlation between bifidobacteria faecal concentration, and gut permeability and metabolic endotoxemia (plasmatic LPS), has been reported. However, non-prebiotic dietary fibres have also shown some similar effects, the question of the specific benefits that can specifically be attributed to prebiotic effects remains open.

7. (7) By reference to the present knowledge (mostly based on the data obtained with the various ITFs and the GOS) on the prebiotic effect and its possible multiple physiological consequences, it appears likely that different compounds (food ingredients or food supplements) including chemically identical compounds with e.g. different chain lengths (like in the ITF group) will have:

   1. (a) different prebiotic effects will influence differently the composition of the microflora in the different segments of the intestine, especially in the large bowel;

   2. (b) different physiological effects and thus will not affect similarly the same functions (as this is clearly the case for Ca absorption, a function that is more influenced by ITF-_MIX than by the different ITFs given separately.

Any effect of one particular compound with a prebiotic effect can never be generalised to another compound, unless this has been scientifically substantiated for each particular food ingredient/supplement⁽Reference Roberfroid⁷⁸⁾.

The majority of successful human trials on the prebiotic effects show significantly increased intestinal levels of bifidobacteria. Often, these are associated with improvement in well-characterised and accepted markers of health as shown by the extensive and growing body of evidence, outlined in this report. This strongly associates prebiotic-induced increases in numbers of bifidobacteria in the gut with a range of GI and systemic health benefits. Although it could be argued that these studies alone do not necessarily indicate causality, when considered with the results of trials in human subjects and animals supplemented with live bifidobacteria they do indeed provide compelling evidence that the relationship between intestinal bifidobacteria and health might well be causal⁽Reference Gronlund, Gueimonde and Laitinen^376–Reference Salminen, Collado and Isolauri³⁸⁰⁾.

Even so, key questions still remain such as

1. (1) Which effect(s) (see Table 2) is/are causally linked to selective change(s) in gut microbiota composition?

2. (2) Which of the physiological and/or pathophysiological well-being and health benefits are directly linked with a particular composition of the gut microbiota or (a) selective change(s) therein?

3. (3) Which, among the physiological and/or pathophysiological well-being and health benefits, is (are) not linked to a particular composition of the gut microbiota or (a) selective change(s) therein but is (are) the consequence(s) of other mechanism(s) of the product claimed to have a prebiotic effect?

4. (4) Which protocol(s) is (are) now validated to demonstrate change(s) in microbiota composition?

5. (5) Which protocol(s) and methodology(ies) is (are) now available and validated to demonstrate links between a particular composition of the gut microbiota or a selective change therein and a particular physiological and/or pathophysiological well-being and health benefit?

Over the last two decades, data have and continue to accumulate improving our knowledge of the gut microbiota composition but also, through the metabonomic approaches, gut microbiota activities. It has convincingly demonstrated that particular food products/ingredients/supplements can, upon feeding, selectively modulate that composition and possibly these activities. Dietary consumption of some of these specific food products/ingredients/supplements has also been reported to exert a series of beneficial health effects that may justify improved function and/or reduction of disease risk claims⁽Reference Cummings, Antoine and Azpiroz^21, Reference Bellisle, Diplock and Hornstra³⁸¹⁾. A causal relationship between the induced change(s) in gut microbiota composition and/or activity(ies) and these health effects is more than plausible – given our knowledge that prebiotics are known to be specifically metabolised by the gut microbiota. The more we understand the complexity of the gut microbiota, its interactions with the gut epithelium, its roles in modulating epithelial cell differentiation and epithelial cell functions and, beyond, in the whole body, the more we will be in a position to recommend these food ingredients for their health-promoting values. It is becoming more and more clear that gut microbiota plays key roles in modulating human/animal physiology even far beyond the GI tract. Specific food products/ingredients/supplements with prebiotic properties are unique tools to study such effects but also offer unique opportunity to develop new functional foods/food ingredients/food supplements to improve host health. One major contribution of this review article summarising the state of the art in the research on the metabolic and health effects of these compounds is to recommend where research efforts should be concentrated to improve understanding of the activities and the physiological roles of the gut microbiota and in particular the importance of its qualitative composition and the consequences of that modulation. Through this, it should be possible to better address the continuing burden of gastro-intestinally mediated disorders. Importantly, tools exist to underpin this with mechanistic explanations of effect leading to effective hypothesis-driven research.