OUP user menu

Composition and temporal stability of gastrointestinal microbiota in irritable bowel syndrome — a longitudinal study in IBS and control subjects

Jaana Mättö, Liisa Maunuksela, Kajsa Kajander, Airi Palva, Riitta Korpela, Anna Kassinen, Maria Saarela
DOI: http://dx.doi.org/10.1016/j.femsim.2004.08.009 213-222 First published online: 1 February 2005


Irritable bowel syndrome (IBS) is a common intestinal disorder that includes continuous or recurrent intestinal pain and discomfort and altered bowel habits. The pathophysiology of IBS is incompletely understood, but it may involve an altered intestinal microbiota. The aim of the present study was to compare the composition and temporal stability of faecal microbiota of IBS patients and healthy controls by applying culture-based techniques and PCR-DGGE analysis. No difference in the prevalence or mean culturable numbers of bacteroides, bifidobacteria, spore-forming bacteria, lactobacilli, enterococci or yeasts were observed between the IBS and control groups, whereas slightly higher numbers of coliforms as well as an increased aerobe:anaerobe ratio was observed in the IBS group. PCR-DGGE revealed more temporal instability in the predominant bacterial population of IBS subjects than in controls. In 9 out of 21 IBS subjects and 5 out of 17 controls the PCR-DGGE profiles obtained from the samples of the same individual on different occasions (sampling points 0, 3 and 6 months) were clearly different. However, the instability in some of the IBS subjects could partly be explained by antibiotic consumption during the study. The present study suggests that instability of intestinal microbiota may be involved in IBS. However, further studies are needed to associate the instability with specific IBS symptoms or with specific bacterial groups and species.

  • Irritable bowel syndrome
  • Faecal
  • Intestinal microbiota

1 Introduction

Irritable bowel syndrome (IBS) is an intestinal disorder that is characterized by intestinal dysfunction and pain [1,2]. IBS is diagnosed by symptom criteria and exclusion of any organic disease; no biological markers are available for IBS [1,3]. The symptoms of IBS include continuous or recurrent intestinal pain or discomfort that is relieved during defecation [2]. In addition, IBS symptoms include altered stool frequency, form or passage, and passage of mucus, and bloating [2]. IBS patients can be divided in three groups, according to the predominant symptom, constipation, diarrhea, and pain types [3,4]. IBS is neither life-threatening nor predisposing to more serious conditions, but it can significantly decrease the quality of life [1,3]. Since the prevalence of IBS is high-upto 22% of general population-it causes a major economic burden on society due to increased absence from work and utilisation of health care resources [1,5]. IBS is the most common diagnosis in the gastroenterologists' practice (approximately 28% of the patients) and accounts for about every tenth of primary care visits in the United States [3].

IBS is a multifactorial state involving several pathophysiological factors [3,4,6]. Psychosocial factors, altered intestinal motility and transit, increased sensitivity of the intestine, neurotransmitter imbalance as well as infection are considered to be involved in the development of IBS [3,6]. Infectious diarrhoea due to Salmonella, Shigella or campylobacter precedes the onset of IBS in 7–30% of patients [3,7]. It is thought that the inflammatory response to the infection causes persistent sensory-motor dysfunction in these patients [8]. Besides the link between intestinal infection and IBS, the existence of abnormal colonic fermentation in IBS [9] and alleviation of IBS symptoms by eracidation of small intestinal bacterial overgrowth by antibiotic therapy [10,11], suggests that intestinal microbiota has a role in IBS. Some early studies have reported differences in faecal bacterial population between IBS and control subjects [1214]. However, the results obtained in previous studies are partly contradictory and are based solely on culture-based analysis on the microbiota. Therefore, the role of intestinal microbiota in IBS is still poorly known.

Treatment of IBS requires a multidisciplinary approach and is typically based on the most bothersome symptom [3,4]. Pharmacological therapy for IBS includes medications targeted for pain alleviation with antispasmodics, management of diarrhea with opioid agonist, and management of constipation with fibers and laxatives. Also antidepressants and serotonin agonists and antagonists have proved efficacious in IBS [15]. In addition to pharmacological therapy successful management of IBS involves exclusion diets and life-style modifications [4,16]. Probiotic therapy using Lactobacillus plantarum 299V [17,18] or a mixture of probiotic lactobacilli and bifidobacteria in the VSL-3 capsule [19] has proved effective in alleviating IBS symptoms especially by reducing intestinal bloating and pain. Since the ability of probiotics to balance intestinal microbiota is well known [20], the positive effect of probiotics in IBS supports the former suggestion of potential involvement of disturbed intestinal microbiota in the etiology IBS.

The aim of the present study was to compare the intestinal microbiota of IBS subjects and healthy controls to reveal the possible differences in the composition or numbers of selected microbial groups. In addition to a culture-based approach denaturing gradient gel electrophoresis (PCR-DGGE) followed by cloning and sequencing of selected bacterial PCR amplicons was applied to study the diversity and temporal stability of the predominant intestinal microbiota.

2 Materials and methods

2.1 Subjects and sampling

The IBS group comprised of 26 subjects, who fulfilled the Rome II criteria for IBS [2], except for 3 subjects who reported slightly less than 12 weeks of abdominal pain during the preceding year. Seven male and 19 female subjects were included in the IBS group (age 20–65 years; mean 46 years). The IBS group consisted of subjects with diarrhoea dominant (12 subjects), constipation dominant (9 subjects) and alternating type (5 subjects) symptoms. All patients had undergone clinical investigation and endoscopy or barium enema of the gastrointestinal tract 0–1 years prior to the study. Exclusion criteria were pregnancy, breast feeding, organic intestinal diseases or other severe systematic diseases, previous major or complicated abdominal surgery, severe endometriosis, antimicrobial medication during the previous two months, and dementia or otherwise inadequate co-operation. Patients with lactose intolerance were included if they reported to follow a low-lactose or lactose-free diet. Some of the patients received medication to their IBS symptoms, and they were allowed to continue their medication throughout the study.

The control group consisted of 25 subjects that were age (age 23–63, mean 43 years) and gender matched (7 males, and 18 females) with the IBS group. The main recruiting criterion for the volunteers of the control group was a good (normal) intestinal balance (absence of repeating and/or persisting gastrointestinal symptoms). The exclusion criteria for the control subjects were regular GI-tract symptoms, lactose-intolerance, celiac disease and antimicrobial therapy during the last two months prior to the study.

The study was approved by the human ethics committee at the joint authority for the hospital district of Helsinki and Uusimaa (HUS), Finland (IBS patients) and by the ethical committee of VTT, Finland (controls), Finland. All subjects gave their written informed consent for participation in the study.

Faecal samples were obtained on three occasions three months apart (0, 3 and 6 months). Samples from all three sampling points were obtained from 21 IBS and 18 control subjects. The subjects defecated into a plastic container, which was then made anaerobic with gas-generators (Anaerocult A mini, Merck, Germany) placed on the lid of the container. The samples were transported to the laboratory and processed in an anaerobic workstation (Don Whitley, UK) within 1–5 h from the defecation. The samples were homogenized by blending with a wooden spatula and divided in separate sub-samples. Culturing was performed for fresh samples, and the other samples were maintained at −70 °C until analysis.

2.2 Faecal moisture content

Faecal moisture content was determined by drying 1–2 g of the sample at 95 °C to constant weight (2–3 days) in pre-weighed aluminium foil cups.

2.3 Culturing

Fresh faecal samples were cultured from 26 IBS and 25 control subjects. From 17 IBS and 17 control subjects culturing was performed also three months later. The samples were serially diluted in pre-reduced peptone-saline containing 0.5 g/l l-cysteine–HCl (Merck, Germany), and plated on several culture media in an anaerobic workstation. The following culture media and incubation conditions were used (incubation at 37 °C unless otherwise stated): supplemented Brucella blood agar (Tammer Tutkan maljat, Finland) for total anaerobes (anaerobic incubation for 7 days), Bacteroides Bile Esculin agar (Tammer Tutkan maljat, Finland) for Bacteroides fragilis group (anaerobic, 4 days), Beerens [21] for bifidobacteria (anaerobic, 4 days), Neomycin Egg Yolk agar (N-EYA) [22] for lecithinase positive clostridia (anaerobic, 3 days), CCFA Cycloserine Fructose agar (Tammer Tutkan maljat, Finland) for Clostridium difficile (anaerobic, 3 days), Rogosa (Oxoid, UK) for lactobacilli (microaerophilic, 3 days), Nutrient agar (Difco, USA) for total aerobes (aerobic, 2 days), MacConkey agar (Tammer Tutkan maljat, Finland) for coliforms (aerobic, 1 days), Bile Esculin Azide agar (Difco, USA) for enterococci (aerobic, 1 days), Saboraud Dextrose agar with penicillin and gentamycin (Tammer Tutkan maljat, Finland) for yeasts (aerobic 30 °C, 4 days). For enumeration of spore-forming bacteria, samples were treated with ethanol (1:1 vol) for 1 h prior to culture on Brucella and N-EYA agars (anaerobic, 3 days). Representative colonies were enumerated, and when necessary, group-level identifications were confirmed by microscopic cell morphology, Gram stain reaction and basic biochemical tests such as, catalase, aerotolerance and indole.

2.4 PCR-DGGE analysis and sequencing of amplicons

Faecal samples from 21 IBS and 17 control subjects from whom samples were available on three sampling occasions were analysed by PCR-DGGE. DNA was extracted from 300 mg of faecal material using the Fast DNAII spin kit for soil (BIO 101, USA) according to manufacturer's instructions using Fast Prep™ cell homogenizer (Labnet, USA). Performance of the kit was evaluated by comparing the PCR-DGGE profiles obtained from DNA extracted from faecal samples of four subjects by using the kit and by the method described by Zoetendal et al. [23] (identical profiles were obtained in all cases). A modification of the PCR-DGGE protocol described in Zoetendal et al. [23] was used. Partial 16S rRNA genes were PCR-amplified using bacterium specific primers U968-GCf (5′-CGC CCG GGG CGC GCC CCG GGC GGG GCG GGG GCA CGG GGG G -AAC GCG AAG AAC CTTA-3′) and U1401-r (5′-CGG TGT GTA CAA GAC CC-3′) [24]. PCR amplifications were performed in a total volume of 50 µl containing 1 µl of appropriately diluted template DNA, 0.2 µM of both primers, 0.2 mM dNTP, 3 units of Dynazyme polymerase (Finnzymes, Finland) in a reaction buffer with 10 mM Tris–HCL (pH 8.8), 50 mM KCl and 1.5 mM MgCl2. The PCR program consisted of the following steps: initial denaturing at 94 °C for 5 min, followed by 35 cycles of denaturing at 94 °C for 30 s, primer annealing at 50 °C for 20 s and elongation at 72 °C for 40 s, and final extension for 7 min at 72 °C. PCR products were separated in polyacrylamide gels with denaturing gradient of 38–60% of urea plus formamide at 85 V at 60 °C for 16 h by using the DCode™ universal mutation detection system (BioRad, CA, USA). The gels were stained with SYBR Green (Molecular Probes, the Netherlands), visualized in UV light and photographed with the GelDoc 2000 (BioRad, USA). Similarity of the PCR-DGGE profiles of the samples of a single subject was compared to evaluate the temporal stability of the predominant faecal bacterial population. The comparison of the profiles was performed by visual inspection of the gels by three researchers and by calculating the similarity percentage using Molecular Analyst Software (see below in the statistical analysis).

Amplification products were purified using Qiaquick PCR purification kit (Qiagen, Germany) according to the manufacturer's instructions. DNA purity and yield were estimated by electrophoresis in 1% (w/v) agarose gels. Amplification products were cloned into Escherichia coli JM109 using pGEM-T vector system (Promega, USA) according to manufacturer's instructions. Clones for sequence analysis were selected based on the position of the cloned PCR fragment in DGGE using the profile of the original faecal sample as a reference. Plasmid DNA from selected transformants was extracted with WizardPlus SV Minipreps (Promega, USA) according to manufacturer's instructions. Sequencing reactions of the cloned PCR amplicons were performed with the ABI PRISM BigDye terminator Cycle sequencing kit v. 3.0 (Applied Biosystems, USA) according to the manufacturer's instructions using purified plasmid DNA and primers U968-f and U1401-r. Sequences were analysed with ABI PRISM 3000100 automated capillary DNA cycle sequencer (Applied Biosystems, USA) and corrected manually (Chromas v. 2.13). DNAMAN was used for sequence alignment. Similarity searches of partial 16S rDNA sequences (containing a sequence between U968-f and U1401-r primers) were performed using BLAST (NCBI) analysis.

2.5 Statistical analysis

ANOVA was used for the statistical analysis of culturable bacterial numbers after log conversion in the samples of IBS and control group subjects. Results with microbial numbers below the detection limit (log 3 for total aerobes, coliforms, enterococci and yeasts; log 3.5 for spore-forming bacteria; log 4 for other microbial groups) were excluded from statistical analysis. Molecular Analyst software (BioRad, USA) was used to calculate the similarity percentage of PCR-DGGE profiles of samples obtained at different occasions from the same individual. Amplicons with the total surface area of at least 1% were included in the analysis of Pearson similarity.

3 Results

3.1 Faecal dry weight

Faecal dry weight varied greatly between individuals in both groups (dry matter content 6.8–42.3% in IBS, 12.6–33.7% in controls) and also between samples obtained from the same individual on different occasions. No difference in the faecal dry matter percentage was observed between the IBS group (mean 21.7 ± 7.2) and the control group subjects (mean 23.0 ± 5.6).

3.2 Prevalence and counts of microbial groups by culturing

Large individual variation in the numbers of aerobic/facultative anaerobic bacteria (coliforms, enterococci and lactobacilli) was observed (Table 1). Higher numbers of coliforms were detected in the IBS group than in the control group: Mean number of coliforms in the IBS group was log 7.1 cfu/wet weight and log 7.7 cfu/g dry weight, whereas in the control group the corresponding figures were 6.3 and 7.0 (p= 0.04 for wet weight and p= 0.03 for dry weight). Statistically significant differences were not observed in the numbers of any other cultured microbial groups (Table 1). Mean numbers of total aerobes were slightly higher in the IBS group than in the control group (Table 1), although the difference was not statistically significant (p-values being 0.18 and 0.14 for aerobes/g wet and dry weight, respectively). C. difficile was not detected in any of the samples.

View this table:
Table 1

Viable microbial numbers and prevalence of microbial groups in faecal samples of IBS subjects (n= 26) and controls (n= 25) (one sample point) as determined by culture method

Microbial groupIBS (n= 26)Control (n= 25)
Prevalence (%)Mean log cfu/g (SD)Prevalence (%)Mean log cfu/g (SD)
Wet weightDry weightWet weightDry weight
Total anaerobes10010.6 ± 0.2811.2 ± 0.3110010.6 ± 0.2211.2 ± 0.20
Bacteroides1009.7 ± 0.3510.4 ± 0.351009.6 ± 0.4510.2 ± 0.49
Bifidobacteria1009.1 ± 0.889.8 ± 0.851009.2 ± 1.009.9 ± 0.97
Spore-forming bacteria1005.8 ± 0.916.5 ± 0.89966.2 ± 1.026.9 ± 1.00
Lactobacilli1005.7 ± 1.536.4 ± 1.49965.7 ± 1.066.4 ± 1.06
Total aerobes1007.7 ± 1.028.4 ± 1.011007.4 ± 0.778.0 ± 0.73
Coliforms967.1 ± 1.137.7 ± 1.131006.3 ± 1.317.0 ± 1.27
Enterococci815.4 ± 2.046.1 ± 2.06724.9 ± 1.415.6 ± 1.42
Yeasts353.6 ± 0.584.2 ± 0.66484.1 ± 1.444.7 ± 1.49
  • Difference in the numbers of coliforms was statistically significant between IBS and control groups (ANOVA p= 0.04 for cfu/wet weight and p= 0.03 for cfu/dry weight).

The proportion of aerobic bacteria vs. the total anaerobically grown bacterial numbers was significantly higher (p= 0.048) in the IBS group (mean 1.1%) than in the control group (mean 0.3%) (Table 2).

View this table:
Table 2

Proportion of bacteroides, bifidobacteria, lactobacilli, aerobes, coliforms and spore-forming bacteria of total culturable bacteria (number of anaerobes) in IBS (n= 26) and control (n= 25) subjects

Microbial group% of culturable bacteria
IBS (n= 26)Control (n= 25)
Total aerobes0.003––1.70.3
Spore-forming bacteria<0.001–0.070.01<0.001–0.20.03
  • Difference between IBS and control group statistically significant (p < 0.05 in ANOVA).

No marked differences in the bacterial numbers in the follow-up samples obtained from the same individual (Table 3) were detected between the IBS and the control group subjects (samples were cultured on two sampling occasions from 17 IBS and 17 control subjects).

View this table:
Table 3

Stability of cultured faecal bacterial groups in samples obtained from IBS (n= 17) and control (n= 17) subjects on two sampling occasions 3 months apart (mean log difference in bacterial counts in samples obtained from the same individual)

Bacterial groupIBS (n= 17)Control (n= 17)
Mean log difference ± SD
Total anaerobes−0.03 ± 0.310.01 ± 0.33
Bacteroides−0.39 ± 0.710.29 ± 0.83
Bifidobacteria−0.04 ± 0.400.34 ± 1.06
Lactobacilli−0.54 ± 2.530.21 ± 1.04
Total aerobes−0.14 ± 0.61−0.27 ± 0.64
Coliforms−0.21 ± 1.080.36 ± 1.08
  • No statistically significant differences were observed between IBS and control groups. (−) = lower mean count in samples obtained on a later point in time, (+) = higher mean count in samples obtained on later point in time.

3.3 Temporal stability and diversity of the predominant bacterial population by PCR-DGGE

PCR-DGGE analysis with universal bacterial primers did not reveal any IBS-specific amplicons. PCR-DGGE profiles were heterogeneous containing about 20–30 amplicons. Each individual had a unique PCR-DGGE profile (Fig. 1).

Figure 1

PCR-DGGE profiles of faecal samples obtained from two IBS (IBS1 and IBS6) and two control (CONTR1 and CONTR3) subjects on three sampling occasions (0, 3 and 6 months). The profiles were rather stable over time in IBS6 and CONTR1 subjects and unstable in subjects IBS1 and CONTR3. M is the marker.

Clear visual intraindividual differences in the profiles of follow-up samples (0, 3 and 6 months) were observed in 9 of 21 IBS subjects and in 5 of 17 control subjects (Table 4, Fig. 1). Twelve IBS subjects had received antibiotic therapy during the study, and the PCR-DGGE profiles of these subjects were “rather stable” in 6 subjects and “unstable” in 6 subjects (Table 4). In seven of the 12 IBS subjects the antibiotic therapy was received more than 2 months prior to next sampling occasion. Also one control subject visually showing unstable PCR-DGGE profile had consumed antibiotics during the study.

View this table:
Table 4

Similarity of PCR-DGGE profiles obtained from faecal samples of IBS and control subjects on three sampling occasions

Group (n)Temporal stability (n)
StableRather stableUnstable
IBS (21)0129
  No antibiotic therapy (9)063
  Antibiotic therapy (12)066
  More than 2 months prior to sampling (7)052
  Less than 2 months prior to sampling (5)014
Control (17)395
  • Similarity in profiles obtained from the same individual on different occasions was evaluated by visual inspection of the gels: stable (S) = no differences; rather stable (R) = minor differences (intensity differences); unstable (U) = clear differences (absence or presence of strong amplicon(s) and often additionally intensity differences) in profiles obtained from the same individual on different occasions.

  • Twelve subjects in the IBS group had antibiotic therapy (one antibiotic therapy in 9 subjects, 2 in one subject and 3 in one subject) during the study period (various antibiotics for various clinical infections). In 7 cases the antibiotics were consumed more than 2 months prior to the next sampling occasion.

  • One subject in the control group had received antibiotic therapy during the study (less than 2 month prior to next sampling occasion).

Similarity analysis by using the Molecular Analyst software showed mean 87.5 ± 11.2 percentage similarity of the PCR-DGGE profiles of the samples obtained from the same individual in the IBS group and 85.7 ± 12.7 percentage similarity in the control group. The similarity percentage was somewhat lower between samples obtained 6 months apart (first vs. third samples 84.5 ± 15.5 in the IBS group and 82.8 ± 16.1 in the control group) than between samples taken 3 months apart (first vs. second samples 89.0 ± 7.6 for IBS and 85.6 ± 13.8 for control group). The similarity percentages of the samples obtained 6 months apart (first vs. third samples) in subjects showing unstable profiles by visual inspection were mean 79.4 ± 21.5 in the IBS group and 75.1 ± 24.9 in the control group. The corresponding figures for the profiles in subjects showing stable/rather stable profiles were 90.1 ± 4.9 for the IBS group and 86.5 ± 7.0 for the control group.

Partial 16S rDNA was sequenced from 45 amplicons; 29 amplicons from 5 IBS subjects and 16 from 4 controls, respectively (Table 5). Comparative sequence analysis showed that only 46% of the amplicons from IBS subjects and 38% from the control subjects had at least 97% similarity to 16S rDNA sequences of known bacterial species (Table 5). The species with closest similarity represented mainly Ruminococcus, Clostridium and Eubacterium spp.

View this table:
Table 5

Closest relatives of the PCR-DGGE amplicons obtained from 5 IBS and 4 control subjects as determined by comparative sequence analysis

Subject, Clone (bp)Closest sequence found in the GenBank database (sequence number)Closest known species (similarity%)
I1.1 (434)Eubacterium rectale (AY169428)Eubacterium rectale (96.8%)
I1.2 (434);I1.3 (357)Ruminococcus gnavus (X94967)Ruminococcus gnavus (98.6%); (98.3%)
I1.4 (434); I15 (434); I1.6 (337)Uncultured bacterium, HuCA2 (AJ408958)Ruminoccus obeum (99.8%); (98.8%); (98.2%)
I2.1 (334)Clostridium sp. cTPY-2–4 (AY239465)Clostridium spiroforme (93.9%)
I2.2 (436)Bacterium mpn-isolate group 27 (AF357575)Pseudobutyrivibrio ruminis (95.0%)
I2.3 (434)Uncultured bacterium, HuC09 (AJ315487)Ruminooccus gnavus (96.1%)
I2.4 (434); I2.5 (343)Ruminococcus hansenii (M59114)Ruminococcus hansenii (97.2%); (96.4%)
I3.1 (250); I3.2 (435)Uncultured equine intestinal bacterium, CL09 (AJ408044)Clostridium clostridioforme (95.5%)
I3.2 (435)Uncultured equine intestinal bacterium, CL09 (AJ408044)Clostridium xylanolyticum (95.0%)
I3.3 (365); I3.4 (339)Butyrate-producing bacterium SSC/2 (AY305320)Clostridium indolis, Hespellia stercorisuis, Ruminococcus gnavus (94.0%); R. gnavus (93.8%)
I4.1 (434)Abiotrophia adiacens (D50540), Granulicatella elegans (Y15408)A. adiacens, G. elegans (94.7%)
I4.2 (294)Clostridium sp. MDA2315 (AY238334)Clostridium glycolicum (99.8%)
I4.3 (430)Ruminococcus sp. (AY367006)Ruminococcus albus (94.9%)
I4.4 (431)R. bromii (X85099)Ruminococcus bromii (99.1%)
I4.5 (437)Uncultured equine intestinal bacterium, DL35 (AJ408122)Ruminococcus lactaris (96.9%)
I5.1 (435)Bifidobacterium sp. PL1 (AF306789)Bifidobacterium adolescentis (99.3%)
I5.2 (380)Butyrate-producing bacterium SSC/2 (AY305320)Clostridium fimetarium (95.5%)
I5.3 (435)Clostridium glycolicum (AY244773)Clostridium glycolicum (98.6%)
I5.4 (434)Bacterium mpn-isolate group 24 (AF357572)Clostridium oroticum (95.2%)
I5.5 (396)Enterobacteriaceae spp.Enterobacteriaceae i.e. E. coli, S. flexneri (100%)
I5.6 (373); I5.7 (393)Eubacterium rectale (AY169428)Eubacterium rectale (99.5%); (100%)
I5.8 (427)Butyrate-producing bacterium A2–165 (AJ270469)Faecalibacterium prausnitzii (92.0%)
I5.9 (389)Butyrate-producing bacterium L2–6 (AJ270470)F. prausnitzii (96.4%)
Control 1
C1.1 (434)Eubacterium rectale (AY169428)Eubacterium rectale (98.4%)
C1.2 (434)Eubacterium ventriosum (L34421)Eubacterium ventriosum (97.9%)
C1.3 (434)Bacterium mpn isolate group 20 (AF357568)Ruminococcus productus (92.2%)
C1.4 (445)Uncultured bacterium, F13 (AJ400237)Ruminococcus bromii (89.9%)
Control 2
C2.1 (384)Enterobacteriaceae spp.Enterobacteriaceae i.e. E. coli, S. flexneri (100%)
C2.2 (388)Eubacterium sp. III-35 (AJ132842)Dorea formicigenerans (93.1%)
C2.3 (368); C2.4 (360)Butyrate-producing bacterium, SL6/1/1 (AY305317)Eubacterium hallii (95.4%); (90.0%)
C2.5 (369); C2.6 (336)Bacterium mpn-isolate group 18 (AF357566)Ruminococcus obeum (96.6%); (94.4%)
Control 3
C3.1 (430)Clostridium sp. 48 (AF191245)Clostridium lituseburense/glycolicum (97.8%)
C3.2 (435)Butyrate-producing bacterium SM6/1 (AY305318)Eubacterium hallii (96.6%)
C3.3 (433)Eubacterium ventriosum (L34421)Eubacterium ventriosum (98.8%)
C3.4 (356)Ruminococcus sp. bln9 (luti) (AJ133124)Ruminococcus luti (93.8%)
Control 4
C4.1 (445)Bacterium mpn isolate group 27 (AF357575)Ruminococcus gnavus (91.2%)
C4.2 (374)Bacterium mpn isolate group 18 (AF357566)Ruminococcus obeum (98.6%)
  • Sequencing performed only for one direction.

  • 100% similarity with several species of Enterobacteriaceae i.e. Escherichia coli, Escherichia fergusonii, Shigella flexnerii, Photorhabdus luminescens.

4 Discussion

The present study compared the composition and temporal stability of intestinal microbiota of IBS and healthy control subjects. No difference in the culturable faecal numbers of bacteroides, bifidobacteria, spore-forming bacteria, lactobacilli, enterococci and yeasts were observed between IBS and control groups in the present study, while numbers of coliforms as well as the proportion of aerobically growing bacteria were higher in the IBS group than in the control group subjects. In an early culture-based study by Balsari et al. [12], lower numbers of coliforms, lactobacilli and bifidobacteria were observed in IBS subjects than in control subjects. The difference between our study and the study by Balsari et al. [12] regarding numbers of coliforms, bifidobacteria and lactobacilli in IBS subjects is likely partly due to differences in the culture-media used for the enumeration of these groups. In other previous studies no difference in the numbers of facultative bacteria has been observed between IBS and control subjects [25,26]. However, in a study by Bayliss et al. [25] the numbers of aerobically growing bacteria were higher in post-hysterectomy IBS patients. Additionally, in one case study, facultative bacteria represented a high proportion of total culturable bacteria in a single IBS subject [13]. In these latter studies either a very limited number of subjects or bacterial groups have been studied. The bacterial numbers detected in the present study correspond well with the previous culture-based studies on adult faecal microbiota [27,28]. The anaerobic:aerobic ratio detected for the healthy controls is also well in line with previous studies reporting 1000-fold higher numbers of anaerobes than aerobes in faecal samples [27]. Although coliforms comprise a predominant fraction of facultatively anaerobic population in faeces, they comprise far less than 1% of the total faecal bacteria [28,29]. However, due to their gas producing ability they may play a role in IBS symptoms in some subjects. Disturbed metabolism of intestinal gases as well as increased sensitivity of the colon has been associated with IBS symptoms [9]. Coliforms were not further characterized in the present study, but the assessment of the strain clonality and existence of virulent species and strains warrants further studies to evaluate their possible association with IBS.

In our study we did not observe a significant difference in the culturable numbers of spore-forming bacteria between IBS and control group subjects. We did not attempt to enumerate the total number of vegetative cells of clostridia since, due to their extreme heterogeneity [30], there are no satisfactory culture media available for their direct enumeration. However, we included selective culturing of C. difficile, but we did not detect this species in faecal samples of any of the subjects. Tvede and Willumsen [31] found C. difficile in 9% of the IBS subjects. In their study the colonization was typically transient, which hampers the detection of C. difficile in faecal samples. Due to the rare occurrence of C. difficile it can be postulated that this bacterium is either not involved in IBS or that its appearance is limited to a small subset of IBS patients.

PCR-DGGE analysis targeted to the predominant bacterial population showed considerable biodiversity as well as uniqueness of the microbiota in each subject in both study groups. Sequence analysis of the partial 16S rDNA amplicons obtained from the PCR-DGGE revealed that a large proportion (58%) of the amplicons had less than 97% similarity with any described bacterial species, suggesting that they represent uncultured bacteria. Similar or even higher proportions (60 up to 80%) of uncultured bacteria have been reported in studies applying direct molecular methods for the assessment of the predominant intestinal microbiota [32]. Most of the amplicons sequenced in the present study showed highest similarity with Clostridium XIVa cluster (C. coccoides E. rectale group; including i.e. species E. rectale, R. gnavus, R. obeum), followed by Clostridium IV cluster (C. leptum group; including i.e. species F. prausnitzii, R. albus, R. bromii) and Clostridium XI cluster (C. lituseburense group; including i.e. species C. glycolicum, C. lituseburense). Clostridium clusters XVIa and IV have reported to comprise 14–30% and 16%, respectively, of the faecal microbiota as determined by hybridisation-based techniques [3335]. None of the sequenced amplicons represented Bacteroides spp. although bacteroides comprise a predominant proportion of human intestinal microbiota [3234]. Zoetendal et al. [36] also failed to detect bacteroides among sequenced clones obtained by PCR amplification of faecal sample of a single subject with the same primers as used in the present study. This supports further the suggestion of Vanhoutte et al. [37] of poor efficacy of these primers in the amplification of bacteroides, which likely resulted from mismatches in the annealing sites. Due to the limited number of amplicons sequenced in the present study, comparison of qualitative differences in the predominant microbiota between IBS and control subjects is difficult. However, IBS subjects tended to have more Clostridium spp. and less Eubacterium spp. amplicons. More targeted analysis of microbial species representing the predominant clostridial groups are needed to reveal their role in IBS.

The assessment of temporal stability of the predominant microbiota revealed more instability in IBS subjects (in 43%) than in control group subjects (in 29%). To our knowledge there are no previous studies on the temporal stability of the intestinal microbiota of IBS subjects by molecular techniques. In an early culture-based case study [13], instability of intestinal microbiota was shown in a single IBS patient. Predominant intestinal microbiota of an adult individual has been reported to be fairly stable under normal circumstances [23,36,37]. However, medication, changes in diet, and various diseases may cause disturbances in the intestinal balance. The effect of disease on the stability of intestinal microbiota is clearly seen in a study by Seksik et al. [35] in which a marked instability of the predominant microbiota was observed using TGGE for the analysis of dominant faecal microbiota of four patients with Crohn's disease, a form of inflammatory bowel disease. In our study temporal instability was commonly observed in IBS subjects, but also in some of the control subjects. In previous studies assessing the long-term temporal stability of the predominant microbiota of healthy subjects by population-based molecular techniques the number of subjects has been limited (one to four individuals) [23,35,37]. In our study lower similarity in the PCR-DGGE profiles was observed in samples taken six months apart than in samples taken three months apart, showing that during a longer time period some fluctuation in the predominant population may occur.

In the present study the instability of the microbiota in some of the IBS subjects can partly be explained by consumption of antibiotics during the study. Antibiotic therapy can cause disturbances of the intestinal microbiota [38,39], and may affect also the predominant bacterial groups [40]. The effect is however highly dependent on the pharmacokinetics and the spectrum of the antibiotic used [38]. Antibiotics were consumed by more than half of the IBS subjects, while only one subject in the control group consumed antibiotics during the study. In these subjects various antibiotics were prescribed for treatment of various infections. IBS subjects usually have higher numbers of annual drug prescriptions including anti-infective drugs than non-IBS subjects [5]. None of the subjects consumed antibiotics two months prior to the beginning of the study, and during the follow-up the antibiotic therapy was received more than two months prior to the sampling occasion in all except six cases. A two month period was considered long enough for stabilization of the intestinal microbiota. Stabilization of the intestinal microbiota within one to two weeks after antibiotic therapy has been reported [38,41], although the period needed for the stabilization is likely to be highly dependent on the antibiotic used.

The subjects in the IBS and control groups were age and gender matched. We did not have a possibility to use fixed diet periods during the study, although it would have further improved the harmonization of the study groups. Food intolerances and sensitivities are common in IBS, and some IBS patients avoid consumption of various food substances to decrease IBS symptoms [3,4]. Therefore, diet modification would have been extremely challenging. Instead, the study subjects were asked to avoid major changes in their dietary habits during the study to minimize diet-related changes in the intestinal microbiota between sampling occasions.

IBS is very heterogeneous condition and includes a variety of subjects with different predominant symptoms. In the present study patients representing all three symptom categories described for IBS (diarrhoea-dominant, constipation-dominant, and alternating type) were included. We observed higher numbers of coliforms as well as a higher proportion of facultatively anaerobic bacteria in IBS than in control subjects. Moreover, indication of instability in the predominant bacterial population in IBS was suggested by PCR-DGGE, a technique for molecular population analysis. All these findings may be associated with a specific subset of IBS subjects rather than with IBS status in general. Dividing the IBS group into specific symptom-based subgroups could perhaps have revealed more differences between the IBS and control subjects. However, it would have resulted in less than 10 subjects in each group, which was not considered adequate. Therefore, further studies are needed to evaluate the association of the present findings with specific IBS subgroups.


This study was supported by the Finnish Technology Centre (Tekes grant 954/401/00). We thank Mrs. Marja-Liisa Jalovaara for excellent technical assistance.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
  26. [26].
  27. [27].
  28. [28].
  29. [29].
  30. [30].
  31. [31].
  32. [32].
  33. [33].
  34. [34].
  35. [35].
  36. [36].
  37. [37].
  38. [38].
  39. [39].
  40. [40].
  41. [41].
View Abstract