The human GI microbiota starts already in the mouth, which harbours a viable count of 10⁸–10¹⁰ colony forming units (CFU) of bacteria per g saliva. These bacteria are constantly fed to the GI channel by the swallowing reflex. The numbers are reduced in the stomach (around 10³ CFU/g gastric juice), duodenum and jejunum (10²–10⁴ CFU/g content), and then increase again in ileum and colon (around 10¹⁰ CFU/g content and 10¹⁰–10¹² CFU/g content, respectively). These bacteria are of different types and, traditionally, attempts to identify them have been done by pure-culture technique, i.e., isolates are cultured at the laboratory and both phenotypic and genotypic characteristics are studied in pure cultures. Current methods are more directed towards direct gene-identification, and mostly towards the 16S ribosomal RNA (rRNA) gene but, lately, “shotgun” Sanger sequencing or massively parallel pyrosequencing have also been used in an attempt to obtain unbiased samples of all genes of a community [38]. The term “metagenomics” is frequently used as a label for studies where more or less all the genetic material is recovered and identified directly from environmental samples [39]. For example, the latter principle was used on faeces of 124 individuals, and each one of the individuals was shown to harbour at least 160 prevalent bacterial species in faeces [40]. Some species are found in many individuals and some are only found in a few. In an attempt to establish the existence of a phylogenetic “core” of the microbiota common for a majority of individuals, Tap et al. [41] obtained 10,456 16S rRNA gene sequences by PCR-amplification and cloning from faeces of 17 individuals. 3180 operational taxonomic units (OTUs) were detected, but most of these only appeared in a few individuals, and only 2.1% of the OTUs were present in more than 50% of the faecal samples. On the other hand, most of the OTUs belonged to the phyla Firmicutes (about 80%), Bacteroidetes (about 20%), Actinobacteria (about 3%), Proteobacteria (1%) and Verrumicrobia (0.1%). Consequently, when bacteria are identified on higher hierarchical levels of taxonomy such as phylum (division) and class, the individual differences between persons appear to be smaller while the differences between habitats within the same individual are more pronounced. For example, there is a significant difference in the composition of the microbiota between the oral cavity and rectum (measured in stool) [42], and between jejunum and colon [43]. Furthermore, the general profile of the GI microbiota of an individual seems to be reasonably stable over time [42]. Frequently dominating genera in the human GI channel are summarised in Table 1.

According to Lazarevic et al. [48], dominating phyla in the oral cavity are Firmicutes, Proteobacteria, Actinobacteria, Fusobacteria, an uncultured group of 16S rRNA gene sequences labelled TM7 (TM for “Torf, mittlere Schicht” = peat, middle layer) [52] and, to a lesser extent, Spirochaetes, while other studies have found Proteobacteria, Firmicutes, Actinobacteria, and Bacteroidetes to be the dominant phyla [53]. The most frequently identified genera were Neisseria and Streptococcus, constituting about 70% of the sequences [48]. However, saliva samples from a larger number of individuals (10 individuals from each of 12 worldwide locations) showed that more than 70% of 16S rRNA gene sequences belonged to the genera Streptococcus, Prevotella, Veillonella, Neisseria, Haemophilus, Rothia, Porphyromonas, and Fusobacterium [50]. A further 93 genera could be identified (known genera), but a phylogenetic analysis suggested that 64 unknown genera were also present in the saliva samples. The most frequent genus of all in the saliva was Streptococcus, which accounted for 23% of the sequences [50].

There is a high bacterial diversity in the mouth and huge differences between people, but mostly there seem to be relatively minor geographic differences [50]. Consequently, there was significantly more variation among sequences from different individuals than among sequences from the same individual, and there was not significantly more variation among individuals from different geographic locations than among individuals from the same location [50]. However, the two genera Enterobacter and Serratia (both belonging to the family Enterobacteriaceae) varied significantly in frequency between locations, e.g., Enterobacter, which accounted for 28% of the sequences obtained in samples from Congo, was completely absent in samples from California, China, Germany, Poland, and Turkey. Serratia occurred at relatively high frequency in several individuals from Bolivia [50].

The stomach has always been considered as a relatively harsh environment for bacteria and due to the low viable counts found there, it can always be debated whether the bacteria found are resident or transient (with the exception of Helicobacter pylori, the causative agent of gastric ulcers). An adult produces about two litres of gastric juice daily and the pH in lumen is below 2 under fasting conditions, but 5–6 close to the epithelial cells due to the mucus layer. Based on 16S rRNA gene identification, Bik et al. [47] found that the dominating phyla on the gastric mucosa were Proteobacteria, Firmicutes, Actinobacteria, Bacteroidetes, and Fusobacteria, with Helicobacter (all sequences were identified as H. pylori), Streptococcus and Prevotella as the most abundant genera. A similar pattern was seen by Li et al. [49] who found that the most common phyla were the same as reported by Bik et al. [47], except in another order with Proteobacteria as the least frequently occurring phylum. Besides the genera Streptococcus and Prevotella, Neisseria, Haemophilus and Porphyromonas also represented a substantial proportion of the identified clones. These five phyla made up about 70% of the total number of clones [49].

In colon and rectum, the mucosal microbiota of a middle-aged, healthy woman from Sweden was dominated by Firmicutes and Bacteroidetetes, the former represented by Clostridium clusters XIVa as defined by Collins et al. [54] (Eubacterium halii, Eubacterium eligens, Dorea formicigenerans, Ruminococcus lactaris, Ruminococcus gnavus, Ruminococcus torques, Clostridium symbiosum, Clostridium boltei and Roseburia intestinalis), IV (Faecalibacterium prausnitzii and Clostridium orbiscindens), IX (Dialister invisus), XIVb (Clostridium lactatifermentans), and the latter by B. vulgatus, B. thetaiotaomicron, Bacteroides ovatus and B. uniformis [43]. Minor proportions of Verrucomicrobia, Proteobacteria (E. coli, Acinetobacter johnsonii, Sutterella wadsworthensis and Neisseria subflava) and Fusobacteria (Fusobacterium varium) were also present. These results can be compared with the microbiota of colonic and rectal content taken at autopsy of three elderly persons from Japan where Firmicutes strongly dominated, followed by Proteobacteria [46]. The former were represented by, for example, subgroups of Streptococcus salivarius and Butyrivibrio fibrisolvens, and the latter by subgroups of Klebsiella and Escherichia. The microbiota from sigmoid colon (biopsies) in nine 60-year-old volunteers, without clinical symptoms or medication, showed that a majority of the individuals had a heterogeneous flora, but in one person, 91% of the clones were related to E. coli [45]. The microbiota differed widely between individuals with regard to both composition and diversity. The largest number of clones identified close to the level of species for the whole cohort was related to E. coli, Bacteroides vulgatus and Ruminicoccus torques. Most frequently distributed between the volunteers were Bacteroides uniformis and B. vulgatus (7 out of 9 individuals). Bacteroides caccae, Bacteroides distasonis, Bacteroides putredinis, B. thetaiotaomicron and R. torques were found in 5 out of 9 individuals. Opportunistic pathogens found in more than one individual were Bacteroides fragilis, Escherichia coli and Bilophila wadsworthia, while Acinetobacter baumannii, Brachyspira aalborgi, Cardiobacterium hominis, Clostridium perfringens, Klebsiella pneumoniae and Veillonella parvula were found in single individuals [45]. In an early report, Hold et al [55] investigated the bacterial flora of colonic tissue from three elderly subjects: the flora was dominated by Bacteroides and Firmicutes, the latter related to either Clostridium coccoides (cluster XIVa as defined by Collins et al. [54]) or Clostridium leptum (cluster IV).

Faeces from nine human, middle-aged subjects were dominated by Firmicutes and Bacteroidetes and with smaller proportions of Proteobacteria, Actinobacteria, Fusobacteria and Verrucomicrobia (sequences of 16S rRNA genes) [56]. Faeces from 156 individuals, 21–32 years old, confirmed the general strong domination of Firmicutes and Bacteroidetes, and the less pronounced proportions of Actinobacteria and Proteobacteria [51]. Examples of frequently occurring and dominating species are B. vulgatus, B. uniformis, B. thetaiotaomicron, Bacteroides ovatus, Bacteroides stercoris, B. caccae, B. putredinis, Bacteroides merdae, Bacteroides capillosus, B. fragilis and Parabacteroides distasonis among the Bacteroidetes, and amongst the Firmicutes: Faecalibacterium prausnitzii, Eubacterium rectale, Eubacterium eligens, Eubacterium ventriosum, Eubacterium siraeum, Ruminococcus obeum, R. torques, Ruminococcus gnavus, Clostridium leptum, Clostridium bolteae, Clostridium scindens, Coprococcus eutactus, Dorea longicatena, and Anaerotruncus colihominis [51]. Other frequently found Firmicutes in faeces are Blautia hansenii, Clostridium scindens, Clostridium asparagiforme, Clostridium nexile, Ruminococcus gnavus, Ruminococcus lactaris, Ruminococcus bromii, Eubacterium hallii, Collinsella aerofaciens, Anaerotruncus colihominis, Butyrivibrio crossotus, Coprococcus eutactus, Coprococcus comes, Holdemania filiformis, Subdoligranulum variabile, Dorea formicigenerans, Dorea longicatena, Streptococcus thermophilus, Enterococcus faecalis. Other frequently found Bacteroidetes are Bacteroides intestinalis, Bacteroides pectinophilus, Bacteroides finegoldii, Bacteroides eggerthii, Bacteroides capillosus, Bacteroides dorei, Bacteriodes xylanisolvens, Parabacteroides johnsonii, Parabacteroides merdae, Roseburia intestinalis and Alistipes putredinis [40].

For some chronic diseases, it has been suggested that the pathologic agent might be the disturbed microbiota rather than a single organism [57], and this presumably means a decreased bacterial diversity and/or different degrees of overgrowth by more aggressive fractions of residential bacteria, i.e., bacteria inducing inflammatory responses by the immune system. A key question is then which bacteria are the most forceful ones in causing inflammation? Naturally, bacterial species known to be pathogenic or opportunistically pathogenic and genera including such species should be more prone to inducing inflammation. Species that are known to include pathogenic or opportunistically pathogenic strains and that also have been found as a substantial part of the gut microbiota of healthy individuals are E. coli and B. fragilis. Increased proportions of E. coli and B. fragilis have also been linked to inflammatory bowel diseases (IBD) [58,59,60].

Gram-negative bacteria contain lipopolysaccharide (LPS) as the major constituent in the outer leaflet of the outer cell membrane. LPS contains large regions of variable polysaccharide and oligosaccharide regions and a relatively conserved lipid region (lipid A), which is the endotoxic and biologically active moiety responsible for septic shock. The interaction of LPS with macrophages results in the release of pro-inflammatory cytokines such as TNF-alpha, IL-6, and IL-1, and can lead to endotoxic shock, which is an often fatal outcome of sepsis. Although several receptors have been reported to bind to LPS, CD14 has been proven able to mediate these responses in vivo [61].

The gram-reaction of different taxa relevant for the GI tract is a factor of importance as Gram-negative bacteria can be expected to contain LPS (Table 1). For example, both the facultatively aerobic E. coli and the strictly anaerobic B. fragilis contain LPS, but the chemical structures are somewhat different and the mammalian immune system reacts differently towards the different LPS types [62]. The endotoxic activity of LPS of B. fragilis is relatively low compared with LPS from E. coli and other Enterobacteriaceae [63] but, nevertheless, LPS from Bacteroides is a potent stimulator of the innate immune system [64]. However, the immune response to LPS can differ between LPS from different species of Bacteroides [64,65].

Gram-negatives that typically contaminate foods, and so are ingested on a more or less regular basis, sometimes in high quantities, are mostly Gammaproteobacteria, e.g., Enterobacteriaceaea and Pseudomonadaceae. However, different diet components can also affect the gut microbiota, e.g., a high-fat diet seems to increase the proportion of Gram-negatives in the gut but also increase the leakage of LPS through the intestinal barrier [66]. A theory of how gram-negatives in the gut can affect fattening has been put forward by Cani et al. [66]. Diabetes type 2 and obesity are characterised by insulin resistance and low-grade inflammation, and Cani et al. [66] showed that LPS in the GI tract was the triggering factor for inflammation and obesity in a mouse model. A high-fat diet increased the LPS concentration in the blood, causing endotoxemia, which induced systemic inflammation, and in turn initiated a process leading to obesity and diabetes in the mouse [67]. It is not known why gram-negative components of the microbiota should be stimulated by a fat-rich diet, or why the barrier function of the mucosa should decrease. However, one speculation could be that a fat-rich diet increases the amount of bile in the gut, and bile has strong antimicrobial effects, but some taxa have higher resistance against bile than others, e.g., Enterobacteriaceae and Bacteroides are known for their comparably high bile resistance. Furthermore, bile is a powerful detergent which might have effects on the permeability of the mucosa and mediate an increased leakage of LPS.

It should be stressed that it is not only gram-negatives and LPS that can induce inflammation; other cell components and metabolites can be involved, and there are also several gram-positive pathogenic and opportunistic pathogenic bacteria that can induce inflammation [68]. One example of the latter is Enterococcus, which is frequently found as a contaminant in foods.

An attempt to look for correlation between systemic inflammation and faecal microbiota showed that about 9% of the total variability of the microbiota was related to the pro-inflammatory cytokines IL-6 and IL-8 [69]. All taxa that showed a slightly positive correlation with either IL-6 or IL-8 belonged to the phylum Proteobacteria [69].

It should be borne in mind that different taxa of the microbiota in combination can enhance pathogenic effects. This can be demonstrated in animal models, e.g., in rat models for intra-abdominal sepsis that cause a two-phase disease process consistent with intra-abdominal sepsis in humans, it has been shown that a combination of obligate anaerobes such as B. fragilis or Fusobacterium varium and facultative aerobes such as E. coli or Enterococcus faecalis cause early peritonitis and mortality, and abscess development [70]. In this case, E. coli was necessary for the mortality, and a combination of E. coli and B. fragilis was needed for the abscess development [70,71]. Neither E. coli nor B. fragilis alone provoked abscess formation. Results along the same lines were found in mice infected with E. coli and B. fragilis in the peritoneal cavity. The co-infection showed an increase in TNF-alpha production in the peritoneal tissues compared with infection by B. fragilis alone [72]. KC mRNA in peritoneal tissues was up-regulated after infection with B. fragilis which was paralleled by increased KC protein secretion and, after intraperitoneal co-infection with E. coli and B. fragilis, a synergistic increase in the expression of KC could be noted [72]. B. fragilis inhibits the phagocytic killing of E. coli [71,73] while E. coli inhibits phagocytosis and intracellular killing of B. fragilis [74]. Also, B. fragilis seems to suppress the E. coli associated LPS-induced human endothelial cell adhesiveness for neutrophils [75].

There are fractions of the resident GI microbiota that are less prone to inducing inflammation, and there may even be certain taxa with the ability to counteract inflammation. This seemingly inflammation-suppressing effect can be a result of different actions. The inflammation-suppressing fractions of the microbiota may: (i) counteract some of the inflammation-aggravating bacteria, which will decrease the inflammatory tone of the system; (ii) improve the barrier effect of the GI mucosa, which allows less inflammation-inducing components in the lumen to translocate out into the body; (iii) more directly interact with inflammation-driving components of the immune system. All three actions may be at work simultaneously.

When the systemic inflammatory tone measured as IL-6 and IL-8 was compared, some members of the Clostridium cluster XIVa (as defined by Collins et al. [54]) were inversely correlated with systemic inflammation [69]. It has also been shown that a low proportion of Faecalibacterium prausnitzii (family Ruminococcaceae; Clostridium cluster IV or the Clostridium leptum group in older vocabulary) on resected ileal mucosa from Crohn’s disease patients is associated with endoscopic recurrence [76]. Furthermore, F. prausnitzii was proved to possess anti-inflammatory effects in model systems: secreted metabolites blocked NF-κB activation and IL-8 secretion in Caco-2 cells, and stimulation of peripheral blood mononuclear cells by F. prausnitzii led to an anti-inflammatory IL10/IL12 ratio. Oral administration of F. prausnitzii also reduced the severity of 2,4,6-trinitrobenzenesulphonic acid colitis in mice [76].

The currently most studied inflammation-suppressing taxa of the GI microbiota are certain species/strains of Lactobacillus and Bifidobacterium, and those are also the fractions that are supported by administering probiotics (living microorganisms that upon ingestion exert health-beneficial effects), or certain dietary fibres that selectively stimulate resident Lactobacillus and Bifidobacterium (prebiotics). Intestinal exposure to specific bacterial strains may either suppress an undesired immune response, for example allergic and autoimmune reactions, or act in a more generalised immune stimulatory way, associated with adjuvanticity and increased intestinal non-specific IgA secretion [77].

Originally, probiotics meant organisms or substances that contribute to intestinal microbial balance, in contrast to antibiotics that counteract microbial activity [78]. However a currently widely accepted definition is that “probiotics are live microorganisms which when administrated in adequate amounts confer a health benefit on the host” [79]. In other words, the designation “probiotics” refers to a function, and not to a taxonomic unit. Humans have always ingested bacteria unintentionally together with food. The bacteria could be adverse, but they could also be harmless “dietary bacteria” when fermented foods were consumed. In particular, lactic acid fermented foods such as yoghurt, cheese, sauerkraut, salted gherkins, olives and capers can contain high amounts of live bacteria and often bacteria of the same Lactobacillus species that are now used for probiotics. Yoghurt was launched in Paris 1906 with reference to the theories of Metchnikoff [80]. In search of strains with better resistance to the low pH of the stomach and the digestive juices of duodenum, Lactobacillus acidophilus was launched in USA in the 1930s, and in Japan during the same period, Lactobacillus casei (should probably be L. paracasei) started to be used as probiotics.

Popular probiotic species used commercially include L. paracasei, L. rhamnosus, L. acidophilus, L. johnsonii, L. fermentum, L. reuteri, L. plantarum, Bifidobacterium longum and Bifidobacterium animalis. However, the phylogenetic differences are extremely wide between Lactobacillus and Bifidobacterium as they belong to different phyla, but there are also great differences between Lactobacillus species such as L. acidophilus, L. fermentum, L. reuteri and L. plantarum. Even within different strains of the same species, the genomic differences can be considerable, which has been clearly demonstrated for L. paracasei [81]. Consequently, with major genetic differences between different probiotics it is also to be expected that the human body will respond differently to different probiotics. This is something that is not always taken into account and it is often neglected in discussions of probiotic effects. Furthermore, it should be stressed that the bacterial species includes considerably genomic heterogenicity. The consensus definition of a bacterial species is that two strains are of the same species if they have a relative ratio of binding of 70% DNA:DNA homology of the genomes at optimal and stringent re-association temperatures (optimal temperature, 25 °C below the melting point of the DNA; stringent temperature, 15 °C below the melting point of the DNA). Consequently, the body can react very differently to different strains of the same species. Unfortunately strain identity is not always given in studies of probiotics administered to humans, e.g., in a failed attempt to improve the clinical outfall in acute pancreatitis where a mixture of strains were given to the patients [82]. The species identity is given in the paper, but no labels on the strains are given. The same is true for a successful attempt to treat acute pancreatitis with a single strain of L. plantarum [83]; no strain identity was given. Examples of different human trials with probiotic treatment, and with use of different species/strains are summarised in Table 2.

The metabolic syndrome is a combination of disorders that increase the risk of developing cardiovascular disease and diabetes. Factors contributing to the syndrome are increased triglycerides in the blood, decreased HDL cholesterol in the blood, increased blood pressure, increasing fasting plasma glucose and central obesity. The metabolic syndrome is characterised by a systemic, low-grade inflammation. LPS leaking out into the body from the gram-negative part of the intestinal microbiota may be the triggering factor for the low-grade inflammation, so probiotics may be a means to improve the gut-barrier and suppress gram-negatives in the GI channel. The ability of many Lactobacillus strains to counteract, for example, E. coli is well known, and the ability of certain probiotic strains, for example, L. plantarum 299v, to mitigate bacterial translocation has been proved in animal models but it has also been tentatively shown in humans [105,106]. Furthermore it has been shown in healthy humans that L. plantarum WCSF1 increased the relocation of occludin and ZO-1 into the tight junction area between duodenal epithelial cells [89]. The ability of different Lactobacillus strains to improve the barrier effect of the mucosa and suggested mechanisms for this has recently been reviewed by Ahrné and Johansson Hagslätt [149].

In connection to the metabolic syndrome, it must also be mentioned that the GI microbiota of mice seems to be essential for the processing of dietary polysaccharides [150], and in humans it has been shown that the relative proportion of Bacteroidetes in comparison with Firmicutes is lower in obese individuals than in lean ones; the increased abundance of Bacteroidetes correlated with percentage loss of body weight [151]. Furthermore, the proportion of Bacteroidetes increased with time in obese individuals put on a low-calorie diet [151]. To certain extent this contradicts the suggestion that the LPS should be the trigger of the metabolic syndrome, as members of the phylum Firmicuses do not contain LPS. On the other hand, the genus Lactobacillus belongs to Firmicutes, and probiotic lactobacilli have been accused of contributing to exaggerated weight-gain [152,153]. The accusation has been turned down most convincingly by Ehrlich [154] and Delzenne and Reid [155]. It must be borne in mind that the phylum Firmicutes is a taxon on a high taxonomic hierarchy and includes an extremely wide genomic variation of bacteria, and that loss of weight in mammals also can be an endpoint for ill-health.

Disorders in the lipid metabolism can cause hypertension, and hypertension is often linked to hypercholesterolemia. Yoghurt supplemented with L. acidophilus and B. longum increased HDL cholesterol [90], and a sour-milk fermented with Lactobacillus helveticus together with the yeast Saccharomyces cerevisiae decreased the blood pressure in elderly hypertensive subjects [91]. Furthermore, in a small but randomised, placebo-controlled and double blind study on men with slightly elevated cholesterol levels, it was shown that the concentrations of total cholesterol and of LDL cholesterol were decreased after consumption of L. plantarum 299v in a beverage containing rosehip and a small quantity of oats (placebo was a similar beverage without probiotics [92]). The fall in cholesterol level was small but statistically significant. Interestingly, the fibrinogen level in serum also decreased (P < 0.001), representing a reduction of 13.5% [92]. Fibrinogen is an acute phase protein, a good marker for systemic inflammation and it is also an independent risk factor for coronary artery disease [156].

Smokers are at increased risk of developing systemic inflammation since tobacco smoke triggers the production of free radicals [157,158]. A controlled, randomised, double-blind trial of smokers consuming L. plantarum strain 299v for 6 weeks affected systemic parameters, i.e., the systolic blood pressure decreased, and so did the concentration in blood of leptin, fibrinogen, F₂-isoprostanes (marker for oxidative stress) and the proinflammatory cytokine IL-6 [93].

The ageing process is known to adversely affect the immune system [159,160]. An association between the inflammatory status and the presence of chronic diseases in elderly has been suggested, but also the interaction of an altered microbiota could contribute to maintaining a low-grade, systemic inflammation [161]. In a recent pilot study of elderly persons, the intestinal load of lactobacilli was linked to the count of white blood cells, blood glucose and content of oxidised low-density lipoprotein (ox-LDL), all risk markers in the pathogenesis of inflammation, metabolic syndrome and cardiovascular disease [162].

Thirty healthy elderly volunteers were given a dietary supplement of a probiotic drink containing Bifidobacterium lactis for three weeks. The proportion of mononuclear leukocytes staining positively for CD3+ (T lymphocytes), CD4+ (MHC II–restricted T cells), CD25+, and CD56+ (NK cells) as well as the phagocytic capacity of mononuclear and polymorphonuclear phagocytes and the tumoricidal activity of NK cells increased significantly in blood after the probiotic administration. The greatest relative increase in immune function occurred in individuals with poor immune responses before the intervention [94].

Ulcerative colitis (UC) is characterised by periods of remission marked by episodes of clinical relapse caused by acute colonic and/or rectal inflammation. Treatment is primarily aimed at reducing inflammation during relapse and secondarily at prolonging the time spent in remission of clinical symptoms [196]. The histopathological features of UC are characterised by architectural distortion of colonic crypts with frequent depletion of mucin from the goblet cells and diffuse infiltration of lymphocytes and plasma cells.

During the acute phase of inflammation, macrophages, neutrophils and eosinophils infiltrate the lamina propria of the colonic mucosa. Aggregating neutrophils, especially near the crypts, lead to the formation of abscesses [197]. Activated dendritic cells (DCs) and macrophages secrete cytokines that trigger and differentiate T cells, and activate the adaptive immune response. Increased populations of CD4-positive and CD8-positive cells have been found in the colonic lamina propria of patients with active UC [198]. Upon antigenic stimulation, naive CD4+ T cells are activated, expand and differentiate into different effector subsets of cells (Th1, Th2 and Th17), that are characteristic for the production of distinct cytokines and effector functions [199]. In both UC and Crohn’s disease, polarised immune activity towards Th1 (marked by up-regulation of TNF-α, IL-1β, IFN-γ, IL-6) and Th17 (marked by IL-17 secretion) response is reported, while UC appears to exhibit an added contribution of Th2 responses (characterised by secretion of IL-4, IL-5, and IL-13) [200]. Cytokines, such as IFN-γ and TNF-α, increase the expressions of TLR4 in intestinal epithelial cells, which results in increased LPS responsiveness [201]. During UC, the expression of TLR4 is increased on mucosal DCs as well as on intestinal epithelial cells in inflamed and non-inflamed mucosa throughout the colon and terminal ileum [202,203]. The CD4+ T cell phenotype expressing CD25^high and fork-head box protein 3 (FoxP3) has been recognised as the functional representative of regulatory T cells (Treg). The Treg is known to down-regulate immune responses to both foreign and self-antigens [204] and a significant number of T-regulatory cells can be found in the inflamed intestine. Their ability to overcome the inflammatory response is hypothesised to be a major reason for remission, so is a major goal of therapies aimed at enabling the regulatory functions of these naturally immunosuppressive cells [199].

UC patients seem to have higher numbers of bacteria associated to the mucosa than healthy subjects, and the difference may reflect the altered nature of the mucus, which appears to be thinner and less sulphated than that of healthy subjects [205,206]. A thin mucus layer containing larger than normal numbers of bacteria might facilitate contact between bacterial antigens and the mucosal immune system.

The intestinal microbiota in patients with active UC has been shown to be less diverse than in healthy subjects [207]. It is not clear whether endogenous intestinal bacteria and/or specific bacterial pathogens are directly or indirectly involved in the initiation and/or maintenance of UC. Neither is it known which bacterial components or antigens can be responsible for the unrestrained inflammatory response. The colonic surface and the inflamed area of UC patients are colonised by a wide variety of organisms [58]. The Clostridium histolyticum/Clostridium lituseburense group made up 21% of the microbiota in UC specimens, while these organisms were not found in controls. These phylogenetic groups contain mainly clostridia belonging to clusters I and II, and part of cluster XI of Collins [54], species such as C. histolyticum, Clostridium beijerinckii, Clostridium perfringens, Clostridium botulinum, Clostridium intestinalis or Clostridium lituseburense, C. difficile, Clostridium bifermentans. Several of these organisms may be pathogenic. Enterobacteriaceae have also been considered as being involved in the pathogenesis of UC, owing to the ability to adhere to the intestinal mucosa and to produce enterotoxins [208]. E. coli and Klebsiella accounted for 25% of the mucosa-associated and 20% of the mucosa penetrating bacteria [58]. High proportions of Enterobacteriaceae and B. fragilis, together with a substantial presence of Pseudomonas aeruginosa were found on the inflamed colonic mucosa taken during surgery from a 12-year-old girl suffering from UC [60]. Sulphate-reducing bacteria have received attention due to their ability to reduce sulphate to sulphide, a by-product of their respiration. Hydrogen sulphide is freely permeable to cell membranes and inhibits butyrate oxidation in colonocytes [209], and hydrogen sulphide has been implicated in the pathogenesis of UC [210].

The number of lactobacilli seems to be relatively low in active UC [211]. However, lactobacilli were predominantly detected in inactive patients, and were suggested to have a role in the induction of remission [212]. It has been hypothesised that the changing condition in the intestine may influence the amount as well as the type of Lactobacillus [213,214].

The use of probiotics for patients suffering from UC has gained attention, and studies to verify the efficiency have been performed for both intervention and maintenance therapy. A meta-analysis to evaluate the induction of remission and maintenance of probiotic therapy was carried out by Sang et al. [215] on thirteen randomised, controlled studies. It was concluded that probiotics were more effective than placebo in maintaining remission [215].

The probiotic mixture, VSL#3 for treatment of mild-to-moderate active UC, was analysed by Sood et al. [123]. Six weeks of probiotic treatment resulted in a significantly higher percentage of patients with more than 50% improvement in UC Disease Activity Index score, and after 12 weeks, significantly more patients achieved remission [123].

Several species and strains of bacteria with claimed probiotic potential have been used in clinical trials, e.g., E. coli Nissle [127], a mixture of L. casei, L. plantarum, L. acidophilus, L. delbrueckii subsp. bulgaricus, Bifidobacterium longum, Bifidobacterium breve, Bifidobacterium infantis and Streptococcus salivarius subsp. thermophilus (VSL#3) [122,123,124], a mixture of Streptococcus faecalis, Clostridium butyricum, and Bacillus mesentericus (BIO-THREE) [126], a mixture of L. rhamnosus and L. reuteri [128], L. rhamnosus GG [129], B. breve strain Yakult, B. bifidum strain Yakult and L. acidophilus [130]. Probiotics have shown effects in treatment of active mild-to-moderate UC by decreasing clinical activity index, preventing relapse, and induction of remission [123,124,126,127,128,129,130]. Also, improvements of histological scores and increases in faecal butyrate, propionate and short-chain fatty acid concentrations have been registered [130]. Consumption of probiotics by UC patients prevented flare-ups and induced depressed activation of the transcriptional factor NF-κB, decreased expressions of TNF-α and IL-1β, while the expression of IL-10 was elevated [131]. The percentage of CD4+CD25^high cells increased in IBD patients after ingestion of a probiotic yogurt [128]. Not much data is available on how probiotics might alter the composition of the resident gut microbiota but it seems that some probiotics can increase the load of lactobacilli and/or bifidobacteria [126,131].

Ileal pouch-anal anastomosis is a surgical procedure for management of UC by making an ileal reservoir, a pouch. Unfortunately, a complication frequently occurring is inflammation in the pouch, pouchitis [216]. A meta-analysis of five randomised, placebo-controlled clinical trials indicated an advantage of probiotic administration in the treatment of pouchitis [217]. Pouchitis disease activity index scores, as well as the inflammatory bowel disease questionnaire score, were improved and remission maintenance fulfilled [122,125]. Furthermore, Pronio et al. [122] found increased percentages CD4+CD25^high cells, CD4+ LAP-positive cells (latency-associated peptide) and a significant reduction in IL-1β mRNA expression in mucosal samples. Since an increase in Foxp3 mRNA expression was also found in mucosal biopsis, this indicates higher numbers of regulatory T cells [122].

The pouch microbiota after probiotic treatment indicated higher bacterial diversity and lower fungal diversity during remission induced by probiotic consumption [218]. The opposite was found for control patients developing pouchitis. A recolonisation and diversification of the lactobacilli and bifidobacteria was shown at remission [218].

A variety of hepatobiliary abnormalities have been described in patients with ulcerative colitis, including fatty changes, cholelithiasis, pericholangitis, primary sclerosing cholangitis, cirrhosis, chronic active hepatitis, amyloidosis, and bile duct cancer, with primary sclerosing cholangitis being the most common form [219]. Because gut-derived components are easily accessible to the liver via the portal vein, it is suggested that increases in the permeability of the intestinal epithelium during inflammation allow bacterial antigens and toxins to enter the lamina propria and cause an inflammatory reaction when the bacterial products reach the liver [220]. However, no clinical trials on the effect of probiotics on ulcerative colitis-associated liver injuries seem to have been done.

Patients with ulcerative colitis also represent a risk group for developing colorectal cancer (CRC). The two most important risk factors seem to be the duration and extent of the disease but the severity of inflammation has also been shown to correlate with an increased frequency of dysplasia and therefore a greater CRC risk [221]. Probiotics have been given preoperatively and postoperatively to CRC patients in order to reduce intestinal pathogens and to modulate immune response. A mixture of B. longum and Lactobacillus johnsonii in a dose of 10⁹ CFU decreased the concentration of Enterobacteriacae in faeces while a dose of 10⁷ CFU failed to do so [222]. The same trend was found for enterococci [222]. The ratio of Bifidobacterium/E. coli increased in patients given probiotics with enteral nutrition before colorectal cancer resection [132]. Both preoperative and postoperative ratios were significantly lower in the control group. Nine days after surgery, faecal concentration of SIgA increased, but serum IgG, IgM, IgA, IL-6, CRP concentrations decreased. Furthermore, the probiotic treatment also caused less postoperative septic complications [132].

Similar symptoms may appear during Crohn’s disease (CD) and UC that can give rise to diagnostic difficulties [223]. However, some specific characteristics reflect the different conditions. CD can affect any part of the gastrointestinal tract and the inflammation is transmural and influences the whole intestinal wall [224] while UC mostly affects the superficial lining mucosa in colon and rectum. UC usually begins in the rectum and extends upwards through colon and rarely affects the small intestine [225]. The inflammation in UC has a continuous distribution while CD has a patchy pattern [224]. Analysis of bacterial 16S rRNA genes revealed no significant differences in mucosal bacterial composition between CD and UC patients [226], but a trend towards a larger reduction of diversity was found in UC patients but it was not significant compared to CD patients [226]. Also, a trend has been seen towards a predominance of Bacteroides and an increase of mucosal bacteria in CD patients [227].

The efficacy of probiotics for induction of remission in Crohn’s disease was evaluated by Butterworth et al. [228] through data bases and register-searching of randomised controlled clinical trials. Twelve potentially relevant studies were identified but eleven were not considered to fit the stated criteria. One study fulfilled the inclusion criteria but it only included 11 patients with moderate- to active-Crohn’s disease [115]. The patients received L. rhamnosus GG for six months and they were also given antibiotics one week before the probiotic/placebo treatment was initiated. Sustained remission was the principal endpoint but no significant difference in median time to relapse was observed between placebo and treatment group [115].

Children with CD in remission were given L. rhamnosus GG in addition to standard therapy in order to try to prolong this state [116]. However, no prolonged remission was obtained [116]. Ineffectiveness of remaining remission by supplementation of probiotics was also found by administration of L. johnsonii LA1 after surgical resection [120,121] or by administration of L. rhamnosus GG to adults [117]. In contrast to these negative results, preliminary data on four paediatric patients showed significant improvement by administration of L. rhamnosus GG [118]. Another small prospective study was performed on four children with Crohn´s disease. The patients were given L. rhamnosus GG, and a significant improvement in clinical activity and improved barrier function of the intestine was found [119]. These small pilot studies are needed in order to find an efficient probiotic strain and to provide a base for the estimation of power, in order to be able to include a reasonable number of patients in an extended study. This extended, hypothetical study ought to be a blinded placebo-controlled study, running over a time period clinically relevant for the disease, i.e., the study will be costly and involve serious ethical considerations as a number of patients in such a study will receive a product with no therapeutic effects, and preferably the patients should give up other therapies during the study period. Consequently, in this case, the probiotics must be regarded as a medical drug and not as a supplement of functional foods. If the intention is to evaluate the probiotic effect for a certain bacterial strain intended as ingredient in functional foods, patients with Crohn’s disease hardly form the best test group.

Intestinal injury from radiotherapy of pelvic malignancies is clinically important, as enteritis symptoms commonly occur and there are few therapeutic options. Moreover, it has been suggested that protection from injury of normal tissues may provide an increase in tumour control, by allowing an increase in the radiation dose [229,230]. Since the GI mucosa contains sensitive regenerative epithelium susceptible to the toxic effects of ionising radiation, injury to the small and large intestine is among the most significant complications encountered in patients receiving radiation directed at the abdominal or pelvic cavity [231,232].

The radiation dose that can be applied in clinical practice is usually limited by the need to restrict the number and severity of side effects in normal tissues surrounding a tumour, which are unavoidably exposed to radiation [233]. Acute radiation enteritis is a potentially serious complication of radiation therapy. Histologically detectable alterations of the intestinal mucosa, like protein and fibrin precipitation, inflammation and oedema of the bowel wall, can be found several days after radiation [234,235]. The villous height and number decreases. The affected functioning of the bowel mucosa leads to the loss of proteins, electrolytes and water. Due to the reduced intestinal surface, conjugated bile salts are not reabsorbed in the small intestine and enter the colon. Local bacterial flora deconjugates the bile salts leading to chologenic diarrhoea [236,237].

An early inflammatory response, beginning a few hours after irradiation, is characterised by leucocyte infiltration into the irradiated organs. Radiation activates various cellular signalling pathways that lead to expression and activation of pro-inflammatory and pro-fibrotic cytokines, vascular injury and activation of the coagulation cascade. Certain mucosal cytokines are activated and the levels of TNF-α, IL-1β, IL-2, IL-6, and IL-8 are significantly higher [238].

Radiation influences and disturbs the mucosal microbiota, leading to a translocation of microorganisms or microbial products through the mucosa into the blood circulation [239]. Mucosal permeability of irradiated colon of patients treated for rectal cancer can be expected to be increased. This difference may be attributed to the mucosal atrophy observed in the irradiated patients and may result in an increased risk of radiation enteritis [232]. Translocation of pathogenic organisms through the intestinal wall into the bloodstream, the peritoneal cavity and abdominal organs is a well-recognised cause of supervening sepsis and life-threatening complications in critically ill patients [240].

Clinical trials have implicated probiotic therapy as beneficial against radiation-induced diarrhoea. A double-blind, placebo-controlled clinical trial including almost 500 patients who underwent adjuvant postoperative radiation therapy after surgery for sigmoid, rectal, or cervical cancer has been performed [133]. The patients were assigned to either a probiotic preparation (VSL#3) or placebo treatment, starting from the first day of radiation therapy. The incidence and severity of radiation-induced diarrhoea, the daily numbers of bowel movements, and the use of loperamide were all reduced [133]. Improved stool consistency and reduced number of bowel movements, less abdominal discomfort, and fewer chemotherapy-dose reductions due to toxicity have also been found by the use of L. rhamnosus [134,135]. Administration of L. acidophilus during irradiation of the pelvic area because of gynaecological malignancies also appeared to prevent radiotherapy-associated diarrhoea [136]. Flatulence was increased though, probably due to lactulose given as substrate for the bacteria [136]. In contrast, it was found for gynaecological malignancies that the incidence of radiation-induced diarrhoea was not reduced by the use of L. casei [137]. However, a significant effect on stool consistency was recorded [137].