Prebiotic effects of chicory inulin in the simulator of the human intestinal microbial ecosystem

Abstract

The prebiotic potential of native chicory inulin was assessed in the Simulator of the Human Intestinal Microbial Ecosystem (SHIME) by monitoring microbial community from the colon compartments, its metabolic activity and community structure. Inulin addition selected for a higher short chain fatty acid production with shifts towards propionic and butyric acid. Conventional culture-based techniques and PCR-denaturing gradient gel electrophoresis analysis showed no remarkable changes in the overall microbial community from the colon compartments of the SHIME, whereas selective effects were seen for lactic acid bacteria. Quantitative PCR with bifidobacteria-specific primers revealed a significant increase with more than 1 log CFU ml⁻¹ from the proximal to distal colon, in contrast to culture-based techniques, which only showed a minor bifidogenic effect in the ascending colon. Our results indicate that inulin purports prebiotic effects from the proximal to distal colon and that real-time PCR is a more precise technique to detect differences in bifidobacterial populations whereas conventional culturing techniques are much more variable.

Introduction

The importance of colon microbiota within the human gastrointestinal (GI) tract has become evident from numerous studies showing the role of intestinal microorganisms in the synthesis of fermentation products that provide energy to the colon epithelium [1], the stimulation of the gut immune system [2], the synthesis of vitamins K and B [3] and the colonisation resistance against exogenous pathogens [4]. The use of functional foods such as probiotics and prebiotics, which modulate the colon microbial community towards a more beneficial composition, has therefore gained much attention [5]. Probiotic food products contain live lactic acid-producing bacteria such as lactobacilli, bifidobacteria and streptococci. If these health-promoting bacteria are still active when reaching the large intestine, they may support growth of indigenous beneficial bacteria from the gut and suppress the colonization of pathogens by lowering the intestinal pH and producing bacteriotoxins [6]. A successful proliferation of probiotic bacteria in the human colon however depends on their survival through the acidic stomach environment, the membrane damaging effects from small intestinal bile salts and substrate- and niche competition with other intestinal microorganisms. The concept of prebiotics has therefore been developed.

Prebiotics are indigestible food ingredients that beneficially affect the host by selectively stimulating growth and/or activity of one or a number of health-promoting colon bacteria [7]. Of all the possible prebiotics, the effects of inulin type fructooligosaccharides (FOS) have been investigated the most. The average dietary intake of inulin by humans is estimated to be 1–4 g d⁻¹ [8]. FOS are a mixture of indigestible but fermentable β-d-fructans with variable degree of polymerization (DP). They are believed to typically alter the composition of the human colon microbiota towards a predominance of bifidobacteria, which can readily metabolise FOS, by which they have a nutritional advantage over other colon microbiota that can not use inulin [9,10,4]. Additionally, supplementing inulin or oligofructose would increase colonic Ca, Mg and Fe absorption and enhance bone calcium stores in rats and humans [11–14].

In vivo experiments are the most representative approach for evaluating the success of pre- and probiotics administration, since physiological parameters and interactions with the host organism are taken into account. However, in vivo experiments are costly and time-consuming and – especially with human trials – they only investigate fecal microbiota that do not represent the microbial community composition from the different parts of the colon. Advanced in vitro reactor systems that mimick both the proximal as distal region of the human colon may therefore be useful for studying human intestinal microbiota [15–17]. Additionally, they give more reproducible results and allow mechanistic studies with several parameters under control. Such in vitro methods are therefore well suited for studying the influence of prebiotics on the intestinal microbial population in terms of fermentation activity and community structure in the ascending, transverse and descending colon, respectively.

The outcome of microbial community studies largely depends on the monitoring method used. Historically, colon microbiota were studied with conventional culture techniques that are biased since roughly 90% are not yet cultured [18]. Culture-independent rRNA based molecular techniques offer an appropriate alternative and have been successfully applied for investigating colon mucosa associated bacteria [19], the composition of fecal bacteria [20,21] and the effects of several funtional foods on intestinal microbiota [22,23].

In this study, we combined molecular analysis techniques with conventional methods to investigate the effects of native chicory inulin on the composition and the fermentation activity of an in vitro cultured colon microbial community. We used the Simulator of the Human Intestinal Microbial Ecosystem (SHIME), which harbours a microbial community resembling that from the human colon both in fermentation activity and in composition [17]. The aims of this study were (i) to evaluate whether shifts in fermentation pattern, more specifically SCFA production, could be attributed to inulin, (ii) assess whether inulin addition gave rise to prebiotic effects applying specific PCR-DGGE for bifidobacteria and lactobacilli since the latter are suggested to exert health-promoting activities in vivo and (iii) apply quantitative PCR for bifidobacteria specifically to validate bifidogenic effects from inulin and compare this with conventional culture-based techniques.

Materials and methods

Culture system

The SHIME is a dynamic model of the human gastrointestinal tract [24,17]. It consists of five double-jacketed vessels maintained at a temperature of 37 °C, respectively simulating the stomach, small intestine, ascending colon, transverse colon and descending colon, with a total retention time of 76 h. The colon vessels harbour a mixed microbial community and pH controllers (pH controller R301, Consort, Turnhout, Belgium) maintain the pH in the range 5.6–5.9, 6.2–6.5 and 6.6–6.9 in the ascending colon, transverse colon and descending colon, respectively. There is no gas exchange between the different vessels and the headspace of the culture system was flushed twice a day for 15 min with N₂ to ensure anaerobic conditions. Growth medium for the microbial inoculum consisted of a carbohydrate-based medium containing arabinogalactan (1 g L⁻¹), pectin (2 g L⁻¹), xylan (1 g L⁻¹), starch (4.2 g L⁻¹), glucose (0.4 g L⁻¹), yeast extract (3 g L⁻¹), peptone (1 g L⁻¹), mucin (4 g L⁻¹) and cysteine (0.5 g L⁻¹). The pH of the medium was 5.5. Detailed information on the medium can be found in Molly and Woestyne [17].

Experimental setup

At the beginning of the experiment, the last three vessels were inoculated with a pooled fecal sample of three adult volunteers who had no history of antibiotic treatment in the last year. During the start-up period, the normal nutritional medium was supplemented to the reactor, which enabled the microbial community to adapt themselves to the nutritional and physicochemical conditions that prevail in the different colon vessels [17]. After two weeks, the treatment period was initiated, which lasted for five weeks. The nutrition for the treatment period consisted of the normal compounds as described above, except that the amount of starch in the medium was completely replaced by native chicory inulin (Fibruline Instant, COSUCRA, Warcoing, Belgium). This commercial inulin product has a dry matter of 96% and contains, on dry matter, 92% FOS with an average polymerisation degree of 10% and 8% free sugars. The dose of inulin to the SHIME reactor was 2.5 g d⁻¹, which was equivalent to a human dose of 5 g inulin d⁻¹. As the inulin replaced the starch concentration in the medium, the amount of available carbohydrates for the microorganisms stayed the same throughout the entire SHIME run. After the treatment period, a final control period concluded the run, to see whether the metabolic parameters and/or microbial concentrations evolved towards their initial values from the start-up period. This period lasted for 2 weeks and the inulin compound was again replaced by starch in the medium.

Metabolic activity analysis

Short chain fatty acids (SCFA). Liquid samples were collected and frozen at −20 °C for subsequent analysis. The SCFA were extracted from the samples with diethyl ether and determined with a Di200 gas chromatograph (GC; Shimadzu, ‘s-Hertogenbosch, The Netherlands). The GC was equipped with a capillary free fatty acid packed column (EC-1000 Econo-Cap column (Alltech, Laarne, Belgium), 25 m × 0.53 mm; film thickness 1.2 μm), a flame ionization detector and a Delsi Nermag 31 integrator (Thermo Separation Products, Wilrijk, Belgium). Nitrogen was used as carrier gas at a flow rate of 20 mL min⁻¹. The column temperature and the temperature of the injector and detector were set at 130 and 195 °C, respectively.

Ammonia. Using a 1026 Kjeltec Auto Distillation (FOSS Benelux, Amersfoort, The Netherlands), ammonium in the sample was liberated as ammonia by the addition of an alkali (MgO). The released ammonia was distilled from the sample into a boric acid solution. The solution was titrated using a 665 Dosimat (Metrohm, Berchem, Belgium) and 686 Titroprocessor (Metrohm).

Lactate. Lactate concentrations were measured by dissolving 10 μL of sample in 990 μL lactate reagent (Sigma, Bornem, Belgium). Absorbance at 540 nm after 10 min incubation was read with an UVIKON 930 spectrophotometer.

Enzyme analysis. The samples were centrifuged at 10 000×g for 10 min. Cell free supernatant (100 μL) was pipetted into a 96-well plate, with 100 μL of a 5.0 mM solution of substrate, prepared in a 0.1 mM phosphate buffer (pH 6.5). The substrates (Sigma, Bornem, Belgium) used were p-nitrophenyl-β-galactopyranoside, p-nitrophenyl-β-glucuronide for β-galactosidase and β-glucuronidase, The plates were incubated at 37 °C and the absorbance at 405 nm was read after 30 min with a Biokinetics EL312e multi-well reader (Bio-Tek Instruments Europe, Spijkenisse, The Netherlands). The amount of p-nitrophenol released was measured based on a standard curve of p-nitrophenol. The results were expressed in μmol p-nitrophenol released.(ml min)⁻¹.

Microbiological analysis

Plate counting. The analysed bacterial groups and the used media, purchased from Oxoid (Hampshire, UK), are indicated in Table 1. Liquid samples were withdrawn from the culture system and serially diluted in saline solution (8.5 g NaCl l⁻¹). Three plates were inoculated with 0.1 ml sample of three dilutions, and incubated at 37 °C (43 °C for Escherichia coli) using conditions given in Table 1. Anaerobic incubation of plates was performed in jars with a gas atmosphere (84% N₂, 8% CO₂, and 8% H₂) adjusted by the Anoxomat 8000 system (Mart, Sint-Genesius-Rode, Belgium).

PCR-denaturing gradient gel electrophoresis (DGGE). The protocol for total DNA extraction from the SHIME samples was based on that of Boon et al. [25]. Three microbial groups of the colon were analysed: general bacteria, bifidobacteria and lactobacilli. A nested PCR approach [26] was used to amplify the 16S ribosomal RNA genes of the bifidobacteria and lactobacilli. In brief, one μl of the DNA was amplified using the primers BIF164f-BIF662r [27] and SGLAB0158f-SGLAB0667 [28], respectively for bifidobacteria and lactobacilli. When the first PCR round produced a clearly visible band, a second amplification round with forward primer P338F (with a GC-clamp of 40 bp) and reverse primer P518r was used [29]. The 16S rDNA of all bacteria was amplified applying primers P338F with GC-clamp and P518r on total extracted DNA.

Denaturing gradient gel electrophoresis based on the protocol of Muyzer et al. was performed using the Bio-Rad D Gene System (Bio-Rad, Hercules, CA, USA). PCR fragments were loaded onto 8% (w/v) polyacrylamide gels in 1 × TAE (20 mM Tris, 10 mM acetate, 0.5 mM EDTA, pH 7.4). On each gel, a home made marker of different PCR fragments was loaded, which was required for processing and comparing the different gels [26]. The polyacrylamide gels were made with denaturing gradient ranging from 45% to 60% [25]. The electrophoresis was run for 16 h at 60 °C and 38 V. Staining and analysis of the gels was performed as described previously [25]. The normalization and analysis of DGGE gel patterns was done with the BioNumerics software 2.0 (Applied Maths, Kortrijk, Belgium). The statistical comparison of the DGGE patterns on the same gel was done with the GelCompar software 4.1 (Applied Maths, Kortrijk, Belgium). The calculation of the similarity matrix was based on the Pearson correlation coefficient. Clustering algorithm of Ward was used to calculate dendrograms [30].

DNA sequencing. 16S rRNA gene fragments were cut out of the DGGE gel with a clean scalpel and added to 50 μl of PCR water. After 12 h of incubation at 4 °C, 1 μl of the PCR water was reamplified with primer set P338F and P518r. Five microliter of the PCR product was loaded on a DGGE gel (see above) and if the DGGE pattern showed only 1 band, it was sent for sequencing. Sequencing of ca. 180-bp DNA fragments was carried out by ITT Biotech-Bioservice (Bielefeld, Germany). Analysis of DNA sequences and homology searches were completed with standard DNA sequencing programs and the BLAST server of the National Center for Biotechnology Information (NCBI) using the BLAST algorithm [26].

Nucleotide sequence accession numbers. Nucleotide sequences for bands Bif1, 2 and 3 and Bac1 have been deposited in GenBank database under accession numbers AY647970, AY647971, AY647972 and AY647973, respectively.

Real-time PCR. The real-time PCR was based on the principle of Heid et al. [31]. For quantification of bifidobacteria by real-time PCR, amplification was performed in 25-μl reaction mixtures by using buffers supplied with the qPCR™ Core Kit for Sybr™ Green I as described by the suppliers (Eurogentec, Liège, Belgium) in MicroAmp Optical 96-well reaction plates with optical caps (PE Applied Biosystems, Nieuwerkerk a/d Ijssel, The Netherlands). Primers BIF164f-BIF163r [27] for 16S ribosomal RNA genes were used for the quantification of bifidobacteria at a concentration of 1 μM. PCR temperature program was as follows: 50 °C for 2 min, 95 °C for 10 min, followed by 40 cycles of 94 °C for 1 min, 62 °C for 1 min and 60 °C for 1 min. The template DNA in the reaction mixtures was amplified (n = 3) and monitored with an ABI Prism SDS 7000 instrument (PE Applied Biosystems, Nieuwerkerk a/d Ijssel, The Netherlands). DNA was extracted from a 6.4 × 10⁷ CFU/ml culture of Bifidobacterium breve (LMG11042) [32]. Standard curves were constructed after real-time PCR amplification of four different DNA concentrations (n = 4) ranging from 1.28 × 10⁷ to 1.28 × 10⁴ cell equivalents/well. The standard curve had a R² value of 0.99 and the slope was −4.3.

Results

Metabolic activity

Replacement of starch by a metabolic equivalent amount of chicory inulin in the nutrition of the gastrointestinal simulator changed the microbial fermentation pattern in all colon vessels towards a more saccharolytic metabolism. This metabolic shift was already observed after 1 week of inulin supplementation (data not shown) and eventually resulted in a significant (p < 0.01) increase in short chain fatty acid (SCFA) production with 44%, 23% and 33% in the ascending, transverse and descending colon at the end of the 5 week treatment period (Table 2). The higher SCFA concentrations primarily originated from an increased production of propionate and butyrate, whereas no significant changes in concentrations of acetate or branced fatty acids were observed. During inulin administration, ammonia concentrations significantly decreased in the ascending colon vessel, whereas no significant changes in ammonia production were observed in the other colon vessels. As for lactate production, no significant changes were detected upon administration of inulin. In the control period, inulin was removed from the nutrition for the SHIME reactor and again replaced by the original amount of starch (4.2 g L⁻¹). SCFA production again decreased with 26%, 30% and 28% in the ascending, transverse and descending colon and approached the initial levels from the start-up period (Table 2). This was primarily due to lower levels (p < 0.01) of propionate and butyrate. The metabolic shift during the control period was also apparent from a higher ammonia production (33%) and a decrease in lactate production in the ascending colon vessel and the increase in lactate concentrations in the transverse colon vessel. Enzymatic activities of β-galactosidase and β-glucuronidase were monitored in the respective colon compartments and did not change significantly during the entire SHIME run (data not shown).

Microbial community analysis

Plate counting. Using selective growth media, analysis of the microbial suspension from the SHIME colon compartments revealed that inulin administration had limited effects on the overall composition of the microbial community. Concentrations of the beneficial microbial groups, bifidobacteria and lactobacilli, increased in all colon vessels during the first weeks of inulin supplementation (data not shown), yet this increase only became significant during the last two weeks of the treatment period. For the transverse and descending colon compartments, lactobacilli concentrations were 1.5 log CFU higher (p < 0.01) than the initial levels from the start-up period (Table 3). Significantly higher bifidobacteria concentrations (p < 0.05) were observed in the ascending colon after 5 weeks of inulin supplementation. A limited decrease in staphylococci concentrations (p < 0.05) was observed in the descending colon whereas E. coli was inhibited both in the transverse colon (p < 0.05) and descending colon (p < 0.01) (Table 3). During the control period, starch again replaced inulin in the nutrition of the SHIME reactor. This resulted in lower lactobacilli concentrations in the transverse colon and lower bifidobacteria concentrations in the ascending colon, whereas no other significant changes compared to the startup period were found.

Microbial population analysis. PCR-denaturing gradient gel electrophoresis was used as a molecular fingerprinting technique to monitor qualitative changes in the composition of microbial community from the three colon compartments throughout the SHIME run. Samples from the start-up period were taken at day 14, just before the supplementation of inulin began. Samples from the treatment period were taken at day 35 and day 49, respectively after 3 and 5 weeks of inulin supplementation. Finally, samples from the control period, during which no inulin was supplemented, were taken at day 54 and 61. Thus, for the three colon compartments, a total of 15 samples was collected and DGGE analysis was performed for general bacteria, bifidobacteria and lactobacilli.

The global fingerprint for general bacteria showed that all samples from the ascending colon clustered in a separate group and that most samples from the transverse and descending colon clustered together in another group (Fig. 1). Both within the ascending colon group as the transverse/descending colon group, the influence of inulin addition was observed by the separate clustering of the treatment samples d35 and d49 and the control samples d54 and d61 (Fig. 1). Although this inulin effect was slightly apparent within each colon compartment, the dominant factor for clustering was the colon compartment itself, from which the samples were taken. This roughly corresponds to the limited variations in microbial populations that were observed using conventional plating techniques. This was in contrast to clustering analysis of DGGE patterns for the bifidobacteria. For this bacterial group, samples did not cluster according to the colon compartment of origin, but to the time point at which they were taken (Fig. 1). All colon samples from the treatment period (d35 and d49) clustered together with the first colon samples from the control period (d54). For all three colon vessels, a new band strongly appeared at day 35, after three weeks of inulin supplementation. Samples from the start-up period (d14) and the second control sample from the descending colon formed a second cluster, whereas the control samples at d61 for the ascending and transverse colon clustered separately. For lactobacilli, the clustering pattern was roughly comparable to that for bifidobacteria. All samples from the start-up period (d14) formed a first cluster together with several samples from the control period (d54 or d61), whereas all samples from the treatment period (d35 and d49) and some of the control period were grouped in a second cluster. At higher similarity values, two additional clusters were distinguished, a first one formed by ascending colon samples from the treatment (d35 and d49) and control period (d54 and d61) and a second one formed by several samples from the transverse and descending colon for the treatment and control period.

Sequencing. Based on the DGGE fingerprint analysis of the colon microbial community, several shifts in bands or changes in band intensity were observed. To identify the bacterial species that were responsible for those changes, DNA fragments from bands of interest were excised from the DGGE gel, isolated and finally sequenced. Four bands were successfully sequenced. DNA fragment bands marked “bif1” on the bifidobacteria DGGE gel showed 95% similarity (141 out 147 bases) to an uncultured bacterium isolated from mucosa associated bifidobacteria (AY267921). These bands strongly appeared during the first weeks of inulin supplementation, but then declined again in intensity. Two other bands – marked “bif2” and “bif3”– also got more intense during inulin supplementation and these bands respectively revealed similarity to Bifidobacterium infantis (96% similarity, 169 out of 176 bases, AY166531.1) and Bifidobacterium longum (96% similarity, 170 out of 177 bases, AY166538.1). For the bacterial DGGE pattern too, a remarkable change was observed during inulin administration, with one band marked “bac1” getting more intense in all colon vessels. The DNA sequence of this band showed 93% similarity (178 out 192 bases, AF371889) to an uncultured bacterium from the GI tract, belonging to the genus of Prevotella and was much less intense during the start-up and control period.

Real-time PCR. From the conventional culture based techniques, we observed a significant bifidogenic effect upon inulin supplementation. However, this effect was only seen in the ascending colon (0.7 log CFU increase). Yet, DGGE analysis showed that samples from all colon vessels clustered together during the treatment period. We therefore used real-time PCR as a cultivation independent method to quantify the number of bifidobacteria in samples from all colon compartments at the end of the start-up, treatment and control periods. We noted an overall higher bifidobacteria concentration with real-time PCR compared to those obtained using conventional plating techniques. The real-time PCR data showed a strong bifidogenic effect in all three colon vessels (p < 0.01), whereas conventional plate count data only showed this effect in the ascending colon (Table 4). The observed increase in bifidobacteria concentrations was not maintained till the end of the control period during which no inulin was supplemented (p < 0.01).

Discussion

We observed beneficial effects of inulin both in terms of metabolic activity as composition of the colon microbial community in the SHIME reactor. The dosed amount of inulin to the SHIME reactor, 2.5 g d⁻¹, corresponds to an equivalent human dose of 5 g d⁻¹, which is a feasible human daily intake and well within the range of earlier reports testing the effects of inulin in vitro and in vivo [33,34]. Inulin generally escapes digestion in the stomach and small intestine. In vivo, they reach the colon for around 90% in an intact way [35,36], which is the reason that they are very suitable prebiotic candidates. The major goal of applying inulin as prebiotic is to support growth of lactic acid producing bacteria, especially bifidobacteria, throughout the entire human colon [37]. Bifidobacteria would have a nutritional advantage compared to other intestinal microorganisms due to their β-1,2-glycosidase activity, allowing them to metabolize FOS compounds from the inulin. Our findings indicate the beneficial influence at realistic doses of 2.5 g d⁻¹ and are of special interest since the beneficial effects were not only observed in the ascending colon vessel, but also in the transverse and descending colon vessels. Moreover, we show that real-time PCR analysis is a more powerful technique to detect changes in bifidobacteria populations, in contrast to the more variable culturing techniques.

Firstly, administration of inulin to the nutrition of the SHIME reactor beneficially influenced the fermentation pattern of the colon microbiota towards a more saccharolytic environment with a significantly higher SCFA production, more in particular propionate and butyrate, and a lower ammonia production (Table 2). The 44% increase in SCFA production in the ascending colon seems uncommon since the treatment period entailed a replacement of starch by an equivalent amount of inulin and not an addition. The inulin degrading capacity from bifidobacteria would normally result in SCFA production levels, comparable to the start-up period, when an equivalent amount of starch was added. The additional SCFA production may possibly be explained by the additional bifidobacterial biomass, created by the bifidogenic effect from inulin. Additionally, other microbial groups in the colon suspension that are used to starch breakdown, may ferment alternative carbon sources from the nutrition medium to SCFA. This extra SCFA production needs clarification in further research. The shift towards propionate and butyrate corresponds to earlier observations by Uehara et al. [38], who supplemented fructooligosaccharides to rats and saw a significant increase in concentration for these two SCFA. These observations do not directly point towards bifidogenic effects since bifidobacteria are acetate and lactate producers. Maybe other microbial groups such as Megasphaera elsdenii or Roseburia sp. can convert lactate or acetate to butyrate [39,40]. Similar processes in the SHIME reactor may explain the constant lactate and acetate concentrations during inulin treatment, whereas specific increases in bifidobacteria concentrations were seen. In general, increased SCFA synthesis creates a more acidic environment in the gut, which is important in vivo in terms of colonization resistance against pathogens. Moreover, SCFA are important energy sources for the colonocytes and influence colonic function by stimulating water and sodium absorption and modulating motility [41]. More specifically, butyrate also induces differentiation and together with propionate, it stimulates apoptosis of cancerous cells in vitro which may thus inhibit cancer development [42,43]. A second effect of inulin administration was the significant decrease of ammonia levels in the colon ascendens. This is considered beneficial since ammonia can alter the morphology and intermediary metabolism of intestinal cells, increase DNA synthesis and promote tumorigenesis [44]. Lower proteolytic activities are therefore related to health-promoting effects.

SCFA and lactate analyses are important measures for intestinal metabolism, but in this research, they gave no direct information on bifidogenic effects. We therefore monitored β-galactosidase in the colon suspension of the SHIME reactor, since this enzyme is often related to the presence of bifidobacteria [45]. Due to a large variability in the measurements, no significant changes in β-galactosidase levels were found in the colon compartments during inulin administration. We also monitored β-glucuronidase production, since the latter is often associated to hazardous processes in the gut. The reason for this is that microbial β-glucuronidase is involved in the intestinal hydrolysis of conjugated xenobiotics and may delay the excretion of harmful exogenous compounds [46]. However, no significant changes were detected in the absence or presence of inulin. Hence, from the metabolic perspective of inulin administration, we concluded that inulin administration had beneficial effects towards more SCFA production, but no direct evidence of bifidogenic effects was obtained.

Significant changes in microbial community composition were observed only after two weeks of inulin supplementation whereas metabolic changes were found within days. This can be explained by the faster adaptation of a microbial population towards their metabolism (RNA-based) than towards their community structure (DNA-based) [47]. Plate count analysis of the microbial community structure revealed that bifidogenic effects in the ascending colon vessel and the increase in lactobacilli in the other colon vessels became significant only after three weeks of supplementation. This shows that the lower dose of 2.5 g d⁻¹ from this study needs to be administered over a longer time frame to effectively induce and maintain beneficial effects. Single doses of inulin are therefore of no use. Besides the nutritional advantage that bifidobacteria may have from inulin compared to other intestinal microbiota, bifidobacteria together with lactobacilli also create a more acidic environment by the production of lactic acid, thus inhibiting the excessive growth of pathogens. Plate count data may support this hypothesis by the decreased concentration of opportunistic pathogens as E. coli and staphylococci (Table 3). Comparable inhibitory effects of inulin towards other pathogens as Clostridium difficile were previously described [33]. The bifidogenic effect from inulin was however only observed in the ascending colon vessel. Since culture-based techniques may give biased results, we investigated the colon microbial community of the SHIME reactor more closely using PCR-DGGE.

Structural analysis of the colon microbiota using PCR-DGGE showed that administration of inulin during the treatment period significantly affected bifidobacteria and lactobacilli populations, whereas the overall microbial community kept relatively unchanged (Fig. 1). Clustering of the Eubacteria DGGE patterns according to the colon vessel shows that the SHIME reactor harbours different microbial communities in the different colon vessels and that these populations are relatively unaffected by inulin. In contrast, clustering of the bifidobacteria and lactobacilli DGGE patterns showed that samples from the treatment period were grouped together. This confirms the data obtained with selective growth media and indicates the selectivity of inulin towards these health-promoting bacteria. Sequencing of the more intense DNA fragment bands on the DGGE gels pointed towards a previously uncultured Bifidobacterium species and B. longum and B. infantis. The latter organisms have been described to be beneficially affected by inulin derived substrates [4,48,49] and are also used as probiotics. In contrast to the plate count data, the DGGE and sequencing analysis also showed that the bifidogenic effect was not restricted to the ascending colon alone, but was also visible in the distal colon vessels. To better elucidate the bifidogenic effect in the different colon vessels, we used real-time PCR for quantifying bifidobacteria.

In general, the higher bifidobacteria concentrations from real-time PCR analysis compared to plate count analysis may be explained by the fact that real-time PCR takes into account all bifidobacteria including viable but non culturable organisms (VBNC) and inactive organisms. The bifidogenic effect of inulin in the SHIME reactor was strongly supported by real-time PCR data that showed a more pronounced increase in bifidobacteria, not only in the ascending colon vessel, but also in the other colon vessels. With conventional plating techniques, no significant bifidogenic effects were observed in the distal colon compartments. Analysis of real-time PCR data needs to carefully consider that changes in data may be attributed to stimulation of bacterial species with different copy number. However, the rrndb (ribosomal RNA operon copy number database) [50] indicates that the copy number of the different bifidobacterial species is fairly constant, 3 or 4. Even if inulin promotes growth of a species with a copy number more than the already present bifidobacteria, the difference attributed to this increase in copy number would be negligable to the difference of 1 log CFU ml⁻¹ as observed in our study. This is the first report indicating that real-time PCR is a precise method by which more subtle differences in bifidobacteria populations can be detected, in contrast to more variable results from conventional plate count techniques.

In summary, our study showed beneficial effects from inulin towards microbial metabolism and community composition. Although metabolic effects were rapidly seen with an increase in SCFA synthesis towards butyrate and propionate, DGGE analysis indicated that a longer supplementation time is needed before these effects are observed at the community level. Moreover, we found that the prebiotic effects from inulin were limited to the period of supplementation. In order to maintain these prebiotic effects and support growth of beneficial bacteria, inulin should thus be continuously dosed. Real-time PCR analysis showed us that non-culturable bifidobacteria were affected by inulin administration and more interestingly, the observed prebiotic effects were not restricted to the ascending colon alone. This study indicates the usefulness of advanced in vitro methods that mimick both the proximal as distal region of the colon, in contrast to in vivo studies that often investigate fecal microbiota that are less active compared to the colon lumen microorganisms.

References