Probiotics in inflammatory bowel diseases

As early as 1877, Pasteur and Joubert observed an antagonistic interaction between some bacterial strains. In the early 1900s, the Russian scientist Metchnikoff attributed the longevity of rural Bulgarians to the consumption of fermented milk product. However, during the second half of the twentieth century, the focus was more on antibiotics to interfere with intestinal microbiota composition. The growing awareness of the side effects associated with antibiotic use, an increasing fear of antibiotic resistant microbial strains and the fear that industry would not manage to develop new antibiotics at a sufficient rate resulted in a renewed and more general interest in the concept of probiotics and prebiotics⁽ Reference Bengmark ^{22 )}. The safety and low burden of side effects of probiotic and prebiotic interventions compared with corticosteroids, immunosuppressants and antibiotics, are particularly appealing. The first study, describing probiotic administration with Escherichia coli Nissle 1917 to patients with inactive ulcerative colitis was published in 1997⁽ Reference Kruis, Schutz and Fric ^{23 )}. The probiotic strain was found as effective as the standard therapy mesalazine to keep patients in remission. Since then, probiotics have been mentioned in many review papers as promising new therapy for IBD⁽ Reference Jonkers and Stockbrugger ^{24 –} Reference Sartor ^{26 )}, whereas others were much more sceptical⁽ Reference Atlas ^{27 )}. Nevertheless the number of clinical trials that investigated the efficacy of probiotics in IBD remains relatively limited. The promising results obtained with E. coli Nissle 1917 were later confirmed in two large, randomised controlled trials⁽ Reference Rembacken, Snelling and Hawkey ^{28 ,} Reference Kruis, Fric and Pokrotnieks ^{29 )}. In contrast, a dose-finding study with E. coli Nissle 1997 in ninety patients with mild-to-moderate active UC, only found a higher response rate in the treatment group compared with placebo in the per-protocol analysis but not in the intention-to-treat analysis⁽ Reference Matthes, Krummenerl and Giensch ^{30 )}. The fact that a considerable number of patients was excluded from per-protocol analysis because of major protocol violation or discontinued the study because of lack of efficacy might explain these observations. The most extensively tested probiotic preparation is VSL#3, a highly concentrated mixture of four strains of Lactobacillus (L. casei, L. plantarum, L. acidophilus and L. delbrueckii subsp. bulgaricus), three strains of Bifidobacterium (B. longum, B. breve and B. infantis) and one strain of Streptococcus (S. salivarius subsp. thermophilus). In several trials, the potential of VSL#3 was evaluated to induce remission in patients with active UC⁽ Reference Tursi, Brandimarte and Giorgetti ^{31 –} Reference Tursi, Brandimarte and Papa ^{37 )}, to maintain remission in patients with UC⁽ Reference Miele, Pascarella and Giannetti ^{34 ,} Reference Venturi, Gionchetti and Rizzello ^{38 )}, to maintain remission or prevent postoperative relapse in patients with CD⁽ Reference Peddi, Ombiga and Willert ^{39 ,} Reference Campieri, Rizzello and Venturi ^{40 )} or to prevent or treat pouchitis⁽ Reference Gionchetti, Rizzello and Venturi ^{41 –} Reference Shen, Brzezinski and Fazio ^{44 )}. As the functional properties and benefits of probiotic strains are strain-dependent and cannot be extrapolated to other strains⁽ Reference Bron, van Baarlen and Kleerebezem ^{45 )}, not even strains of the same genus, meta-analysis of trials using different probiotic products should be interpreted very cautiously. However, two recent meta-analyses that performed subgroup analyses per probiotic, indicated a significant benefit of VSL#3 over control in inducing remission in UC (relative risk 1·69 (95 % CI 1·17, 2·43)⁽ Reference Jonkers, Penders and Masclee ^{46 )} and relative risk 1·74 (95 % CI 1·19, 2·55)⁽ Reference Shen, Zuo and Mao ^{47 )}). The trial conducted by Miele⁽ Reference Miele, Pascarella and Giannetti ^{34 )} also suggested a benefit for VSL#3 in maintaining remission in UC in children, whereas meta-analysis of three trials⁽ Reference Gionchetti, Rizzello and Venturi ^{41 –} Reference Mimura, Rizzello and Helwig ^{43 )} suggested that VSL#3 significantly prevented relapse in patients with pouchitis (relative risk 0·18 (95 % CI 0·10, 0·34)⁽ Reference Shen, Zuo and Mao ^{47 )}).

Lactobacillus rhamnosus GG was evaluated in a large randomised controlled trial including 187 patients with inactive UC. After 6 and 12 months treatment, the number of subjects remaining in remission was not different in the group receiving L. rhamnosus GG compared with a control group that received mesalazine (2400 mg/d) or the group that received the combination treatment⁽ Reference Zocco, Dal Verme and Cremonini ^{48 )}. In contrast, the same probiotic did not show clinical benefit over control in the treatment of patients with CD⁽ Reference Prantera, Scribano and Falasco ^{49 –} Reference Bousvaros, Guandalini and Baldassano ^{51 )} and was ineffective as primary therapy in patients with ileal pouch inflammation⁽ Reference Kuisma, Mentula and Jarvinen ^{52 )}. Similarly, Lactobacillus johnsonii was ineffective in preventing relapse in patients with inactive CD⁽ Reference Marteau, Lemann and Seksik ^{53 ,} Reference Van Gossum, Dewit and Louis ^{54 )}.

Although Guslandi et al. ⁽ Reference Guslandi, Mezzi and Sorghi ^{55 )} found a significantly lower relapse rate in thirty-two patients with inactive CD after treatment with Saccharomyces boulardii with mesalazine v. mesalazine alone, a recent randomised controlled trial in 165 patients did not find beneficial effects of S. boulardii over placebo⁽ Reference Bourreille, Cadiot and le Dreau ^{56 )}.

The use of VSL#3 received an ‘A’ recommendation by Floch et al. ⁽ Reference Floch, Walker and Madsen ^{57 )}, meaning ‘strong, positive studies in the literature’ for preventing and maintaining remission in pouchitis as well as for maintaining remission in UC. For the latter indication, also E. coli Nissle received an ‘A’ recommendation.

This differential effect of probiotics in UC v. CD may highlight the fact that IBD is a multifactorial disease with a high variety in phenotypes and severity⁽ Reference Strober, Fuss and Mannon ^{3 )}. Indeed, the notion that IBD is actually a syndrome comprising several disease subtypes, is gaining more and more acceptance⁽ Reference Dotan ^{58 )}.

The proposed mechanisms of action that might explain the benefits of probiotics in IBD have mainly been studied in in vitro and in vivo animal experiments. The effects executed by a certain probiotic depend on its metabolic properties, the molecules presented at its surface or on the components secreted⁽ Reference Oelschlaeger ^{59 )}. Probiotic bacteria may affect the composition of the intestinal microbiota by competition for adhesion sites and nutrients or by secretion of bacteriocins or acids with antimicrobial activity against Gram-positive and Gram-negative bacteria⁽ Reference Servin ^{60 )}. In addition, probiotics can enhance the intestinal barrier function by increasing the production of mucus or the production of anti-microbial peptides, such as β-defensins, lysozyme, lactoferrin or phospholipase by the epithelial cells⁽ Reference Veerappan, Betteridge and Young ^{61 )}. For example, VSL#3 increased the basal luminal mucin content in rats by 60 % and significantly stimulated colonic mucin secretion and MUC2 gene expression in isolated rat colonic loops⁽ Reference Caballero-Franco, Keller and De Simone ^{62 )}. Finally, probiotics may affect the mucosal immune system by increasing the production of intestinal anti-inflammatory cytokines such as IL-10 and transforming growth factor β⁽ Reference Imaoka, Shima and Kato ^{63 –} Reference Barletta, Rossi and Schiavi ^{65 )}, by modulating the NF-κB pathway⁽ Reference Borthakur, Bhattacharyya and Kumar ^{66 )} and by influencing the cross-talk between natural killer cells and dendritic cells⁽ Reference Rizzello, Bonaccorsi and Dongarra ^{67 )}.

Prebiotics in inflammatory bowel diseases

Clinical studies that evaluated the impact of prebiotics in IBD are scarce. In a small-scale uncontrolled study in ten subjects with active ileocolonic CD, fructo-oligosaccharides were administered in a dose of 15 g/d. After 3 weeks, the Harvey–Bradshaw Index, which is a simplified version of the Crohn's Disease Activity Index score, was significantly reduced and faecal numbers of bifidobacteria were increased. In addition, the numbers of IL-10 positive dendritic cells as well as Toll-like receptors-2 and -4 positive dendritic cells were significantly increased⁽ Reference Lindsay, Whelan and Stagg ^{68 )}. Based on these promising results, the same group of researchers performed an adequately powered follow-up study in 103 patients with active CD patients⁽ Reference Benjamin, Hedin and Koutsoumpas ^{69 )}. Patients received the same dose of fructo-oligosaccharides (15 g/d; n 49) or placebo (n 53) for 4 weeks. However, neither in the intention-to-treat nor in the per-protocol analysis, a significant improvement in disease activity was achieved and the levels of faecal bifidobacteria or F. prausnitzii were not modified. In addition, the incidence and severity of gastrointestinal symptoms was significantly increased. A similar study was performed independently in sixty-seven patients with inactive and mild-to-moderate active CD⁽ Reference Joossens, De Preter and Ballet ^{70 )}. Patients received 2 × 10 g/d oligofructose-enriched inulin (n 34) or placebo (n 33) for 4 weeks. Similar to Benajmin et al. and in contrast to the studies in healthy subjects, oligofructose-enriched inulin did not increase faecal numbers of F. prausnitzii or B. adolescentis. However, a significant increase in B. longum and a reduction in R uminococcus gnavus was observed. In the subgroup of patients with active disease (n 8), the increase in numbers of B. longum was positively correlated to an improvement in disease activity (R 0·894; P = 0·02). In addition, metabolomic analysis of the faecal samples confirmed a depletion in butyrate in patients with active CD that was restored to levels comparable with healthy controls after intervention with oligofructose-enriched inulin⁽ Reference De Preter, Joossens and Ballet ^{71 )}.

Once they reach the colon, prebiotics are rapidly fermented by the resident bacteria. The major metabolites include SCFA (acetate, propionate, butyrate) besides some lactate and gasses such as carbon dioxide, hydrogen and methane. These SCFA, and in particular butyrate, have been considered as crucial mediators that might explain the anti-inflammatory activity of prebiotics. Butyrate activates the nuclear peroxisome-proliferator-activated receptor γ⁽ Reference Schwab, Reynders and Ulrich ^{72 )} which antagonises several proinflammatory pathways. Activation of peroxisome-proliferator-activated receptor γ has been shown effective in the prophylaxis and to a lesser extent in the treatment of several animal models of acute or chronic colitis⁽ Reference Viladomiu, Hontecillas and Yuan ^{73 ,} Reference Annese, Rogai and Settesoldi ^{74 )}. Besides, the interaction of butyrate with the G-protein receptor (GPR) 43 also influences the inflammatory response, as GPR43-deficient mice showed exacerbated or unresolving inflammation in models of colitis, arthritis and asthma⁽ Reference Maslowski, Vieira and Ng ^{75 )}. GPR43, also known as free fatty acid receptor 2, is a seven-transmembrane receptor largely expressed throughout the gut that is considered a key player in the cross-talk between the gut microbes and the host⁽ Reference Bindels, Dewulf and Delzenne ^{76 )}. In a mouse model, activation of GPR43 by SCFA regulated the size and function of the colonic regulatory T cell pool and protected in this way against colitis⁽ Reference Smith, Howitt and Panikov ^{77 )}. More recently, the niacin receptor GPR109a (also known as the hydroxycarboxylic acid receptor 2) was identified as another receptor on colonic cells important for colonic health. Singh et al. ⁽ Reference Singh, Gurav and Sivaprakasam ^{78 )} showed that GPR109a signalling by butyrate promoted anti-inflammatory properties in colonic macrophages and dendritic cells and enabled them to induce differentiation of regulatory T cells and IL-10-producing T cells.

Other potential protective mechanisms of prebiotic activity include changes in the intestinal microbiota, improvement of the intestinal barrier and regulation of the mucosal and systemic immune responses⁽ Reference Looijer-Van Langen and Dieleman ^{79 )}.

Faecal microbiota transplantation

Faecal microbiota transplantation (FMT) provides a more drastic strategy to modify one's microbiota composition and involves the transfer of a faecal suspension from a healthy person into the gastrointestinal tract of a person with colonic disease⁽ Reference Borody and Campbell ^{80 )}. In contrast to probiotic administration where the exogenous bacterial strains only transiently populate the intestine, FMT intends to induce a protracted modification of the microbiota. In a study in ten patients treated with FMT for irritable bowel syndrome, constipation or Crohn's colitis, a change in the patients’ bacterial populations of the colon to represent those of the healthy donor's microbiota persisted for at least 24 weeks⁽ Reference Grehan, Borody and Leis ^{81 )}.

The first report on FMT in human subjects was by Eiseman and co-workers in 1958 for the treatment of pseudomembranous colitis⁽ Reference Eiseman, Silen and Bascom ^{82 )}, presumably caused by C. difficile infection (CDI). C. difficile is a Gram-positive spore-forming bacteria that can be a minor normal component of a healthy microbiota⁽ Reference Miyajima, Roberts and Swale ^{83 )} but can become pathogenic when the normal microbiota has been destroyed, typically after the use of a broad spectrum antibiotic. Since that time, >500 additional patients were treated with FMT with success rates of 95 %⁽ Reference Borody and Campbell ^{80 )}. More recently, the efficacy of FMT for treatment of CDI was confirmed in a randomised controlled trial⁽ Reference van Nood, Vrieze and Nieuwdorp ^{84 )}.

For indications other than CDI, the number of reports is more limited. In IBD, only a small number of case reports or case studies have been reported. The majority of the patients suffered from refractory UC and only a few patients with CD were treated with FMT. However, randomised controlled trials are currently underway (www.clinicaltrials.gov).

The exact mechanism of action of FMT in the treatment of active IBD or CDI is not well known. As both CDI and IBD are characterised by depletions in the normal intestinal microbiota including lower numbers of Bacteroidetes and Firmicutes, it is assumed that FMT acts by reintroducing a complete and stable community of microorganisms that repair or replace the disrupted microbiota and corrects the underlying imbalance⁽ Reference Borody and Campbell ^{80 )}. In addition, FMT may also (re)introduce species that produce bacteriocins which eradicate susceptible pathogens. Typically, bacteriocins inhibit the growth of strains closely related to the producer and provide in this way a competitive advantage to strains in a complex ecosystem. An example is the narrow-spectrum Thuricin CD that is produced by Bacillus thuringiensis and has activity against C. diffile ⁽ Reference Rea, Sit and Clayton ^{85 )}.

A meta-analysis including nine reports describing FMT in twenty-six patients with IBD concluded that, although the evidence is limited and weak, the majority of the studies suggest that FMT may be an effective treatment option in IBD⁽ Reference Anderson, Edney and Whelan ^{86 )}. A retrospective analysis of sixty-two patients with UC treated with FMT, reported response to FMT in 91·9 % (fifty-seven/sixty-two) of the patients, of which 67·7 % (forty-two/sixty-two) achieved complete remission and 24·2 % (fifteen/sixty-two) achieved partial response. Eight per cent of the subjects (five/sixty-two) were treatment failures. From the twenty-one patients that underwent repeated colonoscopy, 57·1 % (twelve/twenty-one) had mucosal healing with absence of histological inflammation⁽ Reference Borody, Wettstein and Campbell ^{87 )}. In a paediatric study, ten children and young adults with mild-to-moderate colitis received faecal enemas for five consecutive days. Within 1 week, clinical response was shown in 78 % and clinical remission was achieved in 33 %. After 4 weeks, 67 % of the children maintained clinical response⁽ Reference Kunde, Pham and Bonczyk ^{88 )}. In contrast, in a recent study not included in the earlier mentioned reviews, none of the six UC patients that received a FMT infusion achieved clinical remission, despite a reversal of some of the reported dysbiotic changes in the intestinal microbiota after FMT⁽ Reference Kump, Groechenig and Lackner ^{89 )}. Importantly, whereas treatment of CDI is most often achieved after a single FMT procedure, treatment of UC generally requires multiple FMT infusions indicating that the FMT response in UC is not as robust as in CDI⁽ Reference Borody, Brandt and Paramsothy ^{90 )}. Only exceptionally, patients with UC may achieve cure with a single FMT administration. In addition, FMT may provide greater therapeutic benefit in patients whose onset of ulcerative colitis was associated with an alteration in the faecal microbiota due to antibiotic use or concomitant CDI⁽ Reference Aroniadis and Brandt ^{91 )}. A potential explanation for the less robust FMT response in the treatment of UC might be that the dysbiosis observed in UC is a downstream result of the disease rather than the cause. In that case, modification of the intestinal microbiota might not be the right mechanism for treatment⁽ Reference Rubin ^{92 )}. Alternatively, it is possible that phylogenetic or functional (in)compatibilities between donor and recipient microbiota might govern the outcome of the FMT⁽ Reference Angelberger, Reinisch and Makristathis ^{93 )}. For example, it is not known to date whether (mis)matching of the donor's and recipient's enterotype⁽ Reference Arumugam, Raes and Pelletier ^{94 )} might be relevant.

FMT is generally well tolerated. Most patients treated with FMT experience diarrhoea on the day of infusion, and a small percentage report belching and/or abdominal cramping or constipation⁽ Reference Smits, Bouter and De Vos ^{95 )}. Some recent studies in IBD reported adverse effects during and after FMT⁽ Reference Angelberger, Reinisch and Makristathis ^{93 ,} Reference Vermeire, Joossens and Verbeke ^{96 )}. Patients experienced transient fever, increased C-reactive protein levels and diarrhoea on the day after the procedure. These symptoms also disappeared within a few days. A recent case report describes a flare of UC in a patient that had been in remission for more than 20 years, after FMT procedure for treatment of CDI⁽ Reference De Leon, Watson and Kelly ^{97 )}. In a long-term follow-up study in seventy-seven patients (follow-up varying between 3 and 68 months with a mean of 17 months), four subjects developed an autoimmune disease after the FMT although there was no clear relationship between the new disease and the FMT⁽ Reference Brandt, Aroniadis and Mellow ^{98 )}. These reports indicate that we need to remain vigilant to potential complications and possibly also long-term safety effects of FMT.

The optimal protocol for the use of FMT needs to be determined. Several parameters including the selection of the donor, screening of the donor, storage of the donor sample, type of diluents, volume of stool required, way of administration, need for antibiotics and bowel lavage of the recipient have not been investigated in a systematic way. Based on retrospective review of the literature, it seems that in most studies, 50–60 g stool from a related donor is suspended in about 300 ml non-bacteriostatic saline solution and is administered via colonoscopy⁽ Reference Brandt and Aroniadis ^{99 )}. Recipients have generally received antibiotics before transplantation although animal studies show that antibiotic pretreatment may not be necessary to improve engraftment of the FMT⁽ Reference Damman, Miller and Surawicz ^{100 )}.

In the future, standardisation of the protocol for FMT will likely be increased. It has been shown that stool frozen at −80°C is equally transplantable to fresh stool⁽ Reference Turnbaugh, Ridaura and Faith ^{101 )}. The ability to use frozen material allows us to shift from individual donors selected for each patient to standard volunteer donors and ultimately to banking of frozen processed faecal material that is ready to use when needed⁽ Reference Hamilton, Weingarden and Sadowsky ^{102 )}. In addition, a pilot study successfully used a synthetic stool mixture to treat two patients with CDI. The stool mixture consisted of thirty-three well-characterised isolates representing commensal species that were generally sensitive to a range of antimicrobials and were relatively straightforward to culture⁽ Reference Petrof, Gloor and Vanner ^{103 )}. Such a stool substitute offers the advantage that the exact composition of the administered bacteria is known and that the preparation can be extensively tested and controlled and can be reproduced for future treatment.