SEARCH

SEARCH BY CITATION

Keywords:

  • bifidobacteria;
  • fluorescent in situ hybridization;
  • IBD;
  • rectal mucosa

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Background: Colorectal bacteria may play a role in the pathogenesis of inflammatory bowel disease (IBD). To test the hypothesis that, in affected patients, the numbers of potentially protective mucosal bacteria might be reduced and pathogenic species increased, we compared rectal mucosa-associated flora in patients with IBD and normal controls.

Methods: Snap-frozen rectal biopsies taken at routine diagnostic colonoscopy from 33 patients with ulcerative colitis, 6 patients with Crohn's disease, and 14 controls with normal colonoscopy were processed, and individual bacterial groups were counted using fluorescent in situ hybridization.

Results: Bacteria were mostly found apposed to the epithelial surface and within crypts. Epithelium-associated counts of bifidobacteria in active [median 15/mm of epithelial surface (range, 4-56), n = 14] and quiescent ulcerative colitis [26/mm (range, 11-140), n = 19] were lower than in controls [56/mm (range, 0-144), n = 14; P = 0.006 and P = 0.03, respectively]. Conversely, epithelium-associated Escherichia coli counts were higher in active [82/mm (range, 56-136)] than inactive ulcerative colitis [6/mm (range, 0-136), P = 0.0001] or controls [0/mm (range, 0-16), P < 0.0001]. Epithelium-associated clostridia counts were also higher in active [3/mm (range, 0-9)] than inactive colitis [0/mm (range, 0-9), P = 0.03] or controls [0/mm (range, 0-1); P = 0.0007]. Epithelium-associated E. coli counts were higher in Crohn's disease [42/mm (range, 3-90), n = 6] than controls (P = 0.0006). E. coli were also found as individual bacteria and in clusters in the lamina propria in ulcerative colitis and Crohn's disease but in none of the controls (P < 0.01). Numbers of Lactobacillus and Bacteroides showed no differences between patient groups.

Conclusions: The reduction in mucosa-associated bifidobacteria and increase in E. coli and clostridia in patients with IBD supports the hypothesis that an imbalance between potentially beneficial and pathogenic bacteria may contribute to its pathogenesis.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

The pathogenesis of inflammatory bowel disease (IBD) is unknown. Increasing data suggest that colonic bacterial flora, particularly that which are closely associated with the mucosa, play an important role in modulating the mucosal immune response. Specifically, adherent strains of Escherichia coli may have a proinflammatory role.1,2 In contrast, many of the commensal microflora, such as bifidobacteria and Lactobacillus, may have beneficial effects, either by modifying the actions of pathogenic bacteria or by actions on host mucosal function.3

Previous studies comparing the mucosa-associated flora of patients with IBD with that of normal controls are limited, and the results are often conflicting. Most studies have used cultures of biopsies obtained at colonoscopy,4 whereas some have relied on immunocytochemical, polymerase chain reaction-based techniques5 and electron microscopic examination of mucosal biopsies.6 More indirectly, mucosal adhesiveness of E. coli isolated from stools7 or rectal biopsies8 has been assessed in vitro using human epithelial cell lines. In 2 previous studies, in situ hybridization techniques were used, but in 1, the probes were of broad specificity,9 whereas in the other, only eubacteria, γ-proteobacteria, and Bacteroides were assessed, and quantitation could not be undertaken.6

We have now used a quantitative, genera-specific fluorescent in situ hybridization (FISH) technique to compare mucosa-associated bacterial flora in patients with ulcerative colitis, Crohn's disease, and controls, patients undergoing routine diagnostic colonoscopy and found to have a normal colon. We used FISH probes to identify 3 potentially pathogenic bacterial groups (E. coli, Bacteroides, and Clostridium) and 2 probiotic genera (Bifidobacterium and Latobacillus). We also used a universal bacterial probe (EUB 388) to quantify the total mucosa-associated bacterial flora.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Patients

We studied 33 patients with ulcerative colitis, 6 with Crohn's disease, and 14 with a normal colon at routine colonoscopy. The demographic and clinical characteristics of the patients with ulcerative colitis are shown in Table 1. The 14 controls, 6 men and 8 women, median age 33 years (range, 22-69 yr), underwent colonoscopy for study of altered bowel habits or rectal bleeding. No subjects had received antibiotic therapy within 8 weeks of their colonoscopy.

Table 1. Demographic and Clinical Details of Patients with Ulcerative Colitis and Crohn's Disease
 Ulcerative ColitisCrohn′s Disease
  1. Numbers and, where appropriate, percentages are shown. Note that some patients were on more than one therapy.

Number of patients336
Age (median, range)53 (22–76)51 (19–59)
Sex (M/F)19/141/5
Smokers3 (9%)2 (33%)
Disease duration  
(median, range)16 (0–42)14 (0–37)
Active/inactive14/194/2
Extent/site ofProctitis: 2 (6%)Small bowel
DiseaseLeft sided: 9 (27%)Ileocolitis; 4 (66%)
 Extensive : 4 (12%)Colitis; 2 (33%)
 Total: 18(55%) 
Treatment  
Aminosalicylates29 (88%)2 (33%)
Corticosteroids10 (30%)1 (17%)
Thiopurine2 (6%) 
Loperamide 2 (33%)

All patients were prepared for routine diagnostic colonoscopy with two Picolax (sodium picosulphate with magnesium citrate) sachets and 1 tablet (5 mg) of Bisacodyl. For 3 days before the procedure, the patients were on a low residue diet; before that time, the diet was unrestricted. At colonoscopy, 4 biopsies beyond those needed for routine diagnostic purposes were taken from the rectum, using standard 5-mm gape forceps. In patients with ulcerative colitis, involved mucosa was graded inactive if scored 0 or 1 and active if scored 2 or 3 using Baron's sigmoidoscopic score.10 In addition, in 3 patients with active left-sided ulcerative colitis and 3 patients with Crohn's disease, 4 biopsies were also taken from normal-looking mucosa proximal to the diseased segment.

Patients gave informed consent for the biopsies, permission for which was granted by the East London and City Health Authority Ethics Committee.

Processing of Biopsies

Biopsies were placed on filter paper, coded, and immediately snap frozen in liquid nitrogen before further processing. Cryostat sections (5 μm) were cut, placed onto poly L-(+)lycine-coated glass slides, and fixed in 4% (wt/vol) paraformaldehyde. Sections were stained with hematoxylin and eosin (H&E) for morphologic assessment.

FISH Staining

Cy-5-labeled probes recognizing the 16S rRNA sequence of Bifidobacterium, Lactobacillus, E. coli, Bacteroides, and Clostridium were provided as a gift from Prof. G. Gibson (Reading University) (Table 2). Sections were washed in phosphate-buffered saline (PBS), permeabilized with Triton X100 (Sigma, Poole Dorset, UK), and hybridized overnight in filtered, prewarmed hybridization solution [40 mmol/L Tris-HCl, 1.8 mmol/L NaCl, and 0.5% sodium dodecyl sulphate (SDS)] containing 200 ng of the relevant probe. Sections were incubated in prewarmed hybridization solution without SDS to remove any unbound probe, washed a further 3 times in PBS, mounted in antifade fluorescent mountant, and viewed using a Axioplan 2 (mot) imaging microscope (Carl Zeiss, Germany) equipped with a digital (Axioplan) camera and attached to KS 300 image analysis software. Probe specificity was checked by parallel staining of serial sections with an antisense probe to all species studied (representative examples for Bifidobacteria and E. coli are shown in Figure 1A and B and C and D, respectively) In addition, subtractive staining with EUB 388 probe (Eurogenetec, Liege) confirmed the identity of bacteria in each biopsy. EUB 388 probe was also used to quantify total numbers of mucosa-associated bacteria. Autofluorescence was excluded using an appropriate blue filter (BP510-560; Carl Zeiss).

Table 2. Group-specific Oligonucleotide Probes Used for Fluorescent In Situ-hybridisation
Bacterial GeneraCloneSequence
Bifidobacterium1645′-CAT-CCG-GCA-CCC-3′
Lactobacillus1585′-GGT-ATT-AGC-A(T/C)-TTC-CA-3′
E. coli15315′ -CAC-CGT-AGT-GCC-TCG-TCA-TCA-3′
Bacteroides3035′-CAC-ATG-TGG-GGG-ACC-TT-3′
Clostridium1505′ -TTA-TGC-GGT- ATT-AAT- AAT-CT(C/T)-TT-3′
thumbnail image

Figure 1. Photomicrographs of sections of normal and inflamed rectal mucosa processed by FISH (except F). A, normal mucosa: Bifidobacteria are superficially located, mainly apposed to colonic epithelial cells (×40). B, serial section of A showing lack of staining when probed with an antisense probe for Bifidobacteria (×40). C, ulcerative colitis: large numbers of E. coli are seen within the lamina propria (×40). D, no E. coli are observed when probed with an E. coli antisense probe (×40). E, ulcerative colitis: E. coli are seen in the base of crypts (×40). F, ulcerative colitis: parallel section, stained with H&, of E, showing that the E. coli showed in E are not within crypt abscesses (×40). G, ulcerative colitis: E. coli often form clusters in lamina propria (×63). H, ulcerative colitis: E. coli found associated with inflammatory cells (arrow; ×100). I, Crohn's disease: distribution of E. coli resembles that seen in ulcerative colitis (×40).

Download figure to PowerPoint

Histologic Morphometry

Accurate morphometric measurements were made of each biopsy, recording the length of apical mucosal surface and the depth of the lamina propria. From these 2 measurements, the total mucosal area was calculated. Individual bacterial counts were expressed as number of bacteria per millimeter of epithelial surface or as bacteria per square millimeter of lamina propria. Bacteria-forming clusters within the lamina propria were counted separately and recorded as number of clusters per square millimeter. A cluster was defined as being a tightly packed mass of bacteria where individual numbers could not be counted accurately.

Bacteria were counted by 2 independent observers blind to the clinical diagnosis of the patients from whom the specimens were obtained.

Statistics

All data were expressed as median value (range). The Kruskal-Wallis test was used to assess whether there were significant differences across the bacteria and patient groups studied. Individual differences between bacterial populations and between patient groups were tested using the Mann-Whitney U test. All P values were 2 tailed. A P value of <0.05 was considered significant.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Control Mucosa

When present, the majority of bacteria in control mucosa were seen within the mucus overlying the epithelium or closely apposed to the colonic epithelial surface (Table 3).

Table 3. Epithelium-associated Counts of Bifidobacterium, Lactobacillus, E. coli, Bacteroides, and Clostridium in Rectal Biopsies Taken from Patients with Ulcerative Colitis (US), Crohn's Disease, and Normal Controls. Total Numbers of Epithelium-associated Bacteria, Assessed with EUB Probe, are Also Shown
 ControlsInactive UCActive UCCrohn's Disease
Bacterial Speciesn = 14n= 19n = 14n = 6
  1. Results for individual species are shown as the median counts/mm of epithelial surface (range). N = median (range)% of total bacterial count assessed with EUB probe_ In control patients, *P≤/⩽ 0.02 from each of the other species; #P≤/⩽0.02 from each of the other species except Bacteroids (NS) and ¶(P⩽/⩽ 0.001 from each of the other species except Lactobacillus (NS). For patients with UC and Crohn's disease, ‡,P⩽/⩽ 0.05 compared with inactive ulcerative colitis. Precise values are given in the text.

Total bacterial count607 (412–692)635(484–774)661 (555–816)640(544–714)
Bifidobacterium56 (0–144)*26(11–140)‡15 (4–56)‡15 (9–60)
N9 (0–21)4(1–18)2 (0–7)2(1–8)
Lactobacillus16 (0–126)#3 (0–66)5 (0–26)11 (4–20)
N3 (0–19)1(0–8)0 (0–3)0(0–3)
E. coli0 (0–16)6(0 136)‡82 (56–136)‡†42 (3–90)‡
N0(0–2)1 (0–18)12 (7–16)9 (0–13)
Bacteroides22 (5–45)129 (12–55)29 (12–44)30 (5–36)
N4(1–4)5 (2–7)4 (1–5)4(1–5)
Clostridium0(0–1)0(0–9)‡3 (0–9)‡†2 (0–6)
N00(0–1)0(0-I)0(0–1)

There were substantial differences in epithelium-associated counts of the different bacterial groups in normal control subjects (P < 0.002; Table 3). Bifidobacterium was the predominant genus identified (Fig. 1A and B). The second most abundant bacterial groups were Lactobacillus and Bacteroides, whereas E. coli and clostridia were found in only small numbers and in a minority of subjects (Table 3). The sum of the 5 groups specifically probed (median, 108; range, 94-166; counts per millimeter epithelial surface) represented 16% (range, 0-24%) of the total bacterial epithelium-associated population (607 counts/mm; range, 412-692 counts/mm), as assessed with EUB 388 probe (Table 3).

Ulcerative Colitis

As in normal mucosa, bacteria were observed within the mucus overlying the epithelium or closely apposed to the epithelial surface in biopsies from patients with ulcerative colitis. However, there were also substantial numbers of E. coli seen within the lamina propria of these biopsies (Fig. 1C and D). These subepithelial bacterial populations were located at the base of the crypts (Fig. 1E) or scattered within the lamina propria, where they appeared both as individual organisms (Table 4) and in clusters, either in the extracellular space or inside macrophages (Figs. 1G, H). Parallel H&E sections showed that E. coli were located in the lamina propria itself and not in crypt abscesses (Fig. 1E).

Table 4. Counts of Individual E. coli, Bacteroides, and Clostridium in the Lamina Propria of Rectal Biopsies Taken from Patients with Ulcerative Colitis (UC), Crohn's Disease, and Normal Controls. No Bifidobacteria or Lactobacilli Were Seen in the Lamina Propria in Any Patients. Total Numbers of Lamina Propria-associated Bacteria, Assessed with EUB Probe, Are Also Shown. (Bacteria Seen as Clusters Are Not Included in These Counts)
 ControlsInactive LCActive UCCrohn's Disease
Bacterial Speciesn= 14n= 19n= 14n = 6
  1. Results for individual genera are shown as the median counts/mm2 of lamina propria (range).

  2. N = median (range)% of total bacterial count assessed with EUB probe.

  3. †P = 0.01 compared with inactive UC.

  4. ‡.P≤/⩽ 0.008 compared with controls.

Total Bacterial counts0(0–1)8 (0–56)22(0–47)16 (0–36)
E. coli08 (0–55)‡22 (0–4i)‡‡12 (0–35)‡
N0100 (0–100)‡100 (0–100)‡100 (0–100)‡
Bacteroides110 (0–2)0 (0–3)0 (0–1)
N00 (0–1)0(0–6)0(0–2)
Clostridium0(0–1)0 (0–1)0(0–3)0(0–3)
N00(0–2)0(0–6)0 (0–8)

The total number of epithelium-associated bacteria was similar in active ulcerative colitis (Table 3; median, 661 counts/mm), inactive ulcerative colitis (635 counts/mm), and in controls. However, there were significant reductions in the epithelium-associated populations of bifidobacteria in patients with active and inactive ulcerative colitis compared with normal controls (P = 0.006 and P = 0.03, respectively). In contrast, patients with active ulcerative colitis showed an increase in epithelium-associated E. coli in comparison with inactive disease (P = 0.0001) and controls (P < 0.0001). The difference in epithelium-associated E. coli counts between patients with inactive disease and controls was also significant (P = 0.04). Epithelium-associated counts of clostridia were also higher in patients with active ulcerative colitis compared with inactive ulcerative colitis (P = 0.03) and normal controls (P = 0.0007; Table 3).

Very few bacteria of any genera, as indicated by the EUB probe, were found in the lamina propria in controls (0-1 counts/mm2). Lamina proprial counts of all bacteria were, however, increased in both active (median, 22 counts/mm2; P = 0.005) and inactive ulcerative colitis (8 counts/mm2; P = 0.007) compared with controls (Table 4).

There were no E. coli, either as individual bacteria or clusters, found in the lamina propria in normal controls. The counts of individual E. coli in the lamina propria in active and inactive ulcerative colitis were higher than in controls (P < 0.0001 and P = 0.006, respectively). There were also significantly more individual E. coli in active than in inactive ulcerative colitis (P = 0.01; Table 4). There was no significant difference in the number of clusters of E. coli in the lamina propria in active (6/mm2; range, 0-14/mm2) compared with inactive disease (2/mm2; range, 0-8/mm2), but there were significantly more such clusters in the lamina propria of patients with active and inactive colitis than in controls (0/mm2; range, 0-0/mm2; P = 0.009 and P = 0.0002, respectively).

In 3 patients with active left-sided ulcerative colitis, biopsies taken from inflamed mucosa showed lower surface counts of bifidobacteria (11/mm, 19/mm, and 31/mm) than from more proximal normal-looking mucosa (85/mm, 88/mm, and 104/mm); conversely, counts of E. coli were increased on the epithelial surface (36/mm, 48/mm, and 64/mm) compared with normal-looking mucosa (0/mm, 2/mm, and 2/mm). Individual E. coli were also more numerous in the lamina propria (4/mm2, 8/mm2, and 14/mm2) of inflamed compared with uninvolved mucosa (0/mm2, 0/mm2, and 1/mm2); the same was true of clusters of E. coli (0/mm2, 1/mm2, and 3/mm2 and 0/mm2, 0/mm2, and 0/mm2, respectively).

The number of epithelium-associated lactobacilli and bacteroides was similar in patients with ulcerative colitis and controls (Table 3). Within the lamina propria, no bifidobacteria or lactobacilli were seen, but Bacteroides and clostridia were present in small numbers in a minority of patients with ulcerative colitis (Table 4).

Crohn's Disease

In colorectal biopsies taken from patients with Crohn's disease, localization mimicked that found in patients with ulcerative colitis; bacteria were apposed to mucosal surfaces as well as distributed within the lamina propria (Fig. 1I). Collectively, in the small number of patients studied (4 active and 2 inactive), the number of epithelium-associated E. coli was increased, but with bifidobacteria, showed no significant difference compared with controls (Table 3). However in the 3 patients with active disease, in whom biopsies were taken from involved and uninvolved mucosa, the numbers of epithelium-associated bifidobacteria were decreased (9/mm, 11/mm, and 23/mm) in involved mucosa compared with those from uninvolved colon (44/mm, 54/mm, and 60/mm). E. coli were increased (35/mm2, 42/mm2, and 90/mm2) in involved compared with uninvolved mucosa (3/mm2, 14/mm2, and 9/mm2).

As was observed in ulcerative colitis, individual E. coli were also more numerous in the lamina propria of inflamed (14/mm2, 28/mm2, and 35/mm2) than of uninvolved mucosa (0/mm2, 0/mm2, and 1/mm2), and cluster counts also differed (0/mm2, 2/mm2, and 1/mm2 and 0/mm2, 0/mm2, and 0/mm2, respectively). No overt differences were seen in the other bacterial groups probed.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

In this quantitative study, we used group-specific FISH to compare the mucosa-associated flora in the colorectal mucosa of patients with IBD and controls.

In normal control mucosa, of the 5 groups probed, bifidobacteria were the most abundant microorganisms; there were also a significantly higher number of lactobacilli and Bacteroides than E. coli and clostridia. Previous studies, using mainly culture techniques, have suggested that Bacteroides is the predominant genus in the colorectal mucosa of healthy controls.4,11 Our observation that bifidobacteria were the most common of the genera probed is likely to reflect differences in the patient populations studied and particularly in the methods used to detect and quantify mucosa-associated flora.

In patients with ulcerative colitis, we found a reduction in the number of bifidobacteria and an increase in E. coli counts in rectal mucosa in comparison with normal controls. Indeed, the ratio between epithelium-associated counts of these organisms was approximately 50:1 in controls and 1:5 in patients with active colitis (Table 3), a 250-fold change in the balance between them. Epithelium-associated clostridia counts were also significantly higher in ulcerative colitis than in controls.

Colorectal biopsies were also taken from inflamed mucosa in 6 patients with Crohn's disease. Localization resembled that found in patients with ulcerative colitis. The number of epithelium-associated bifidobacteria in Crohn's disease showed no significant difference compared with controls, but the numbers of E. coli were significantly increased both on the surface of the epithelium and, as individual organisms and as clusters, in the lamina propria.

Previous data on mucosa-associated bifidobacteria in IBD, using culture techniques, showed no differences between patients with active ulcerative colitis and controls,4 although a reduced fecal content of bifidobacteria12,13 and an increase in fecal Enterobacteria14 has been reported in Crohn's disease patients compared with controls.

Localization of the bacterial species studied was similar in all patients with IBD. E. coli was seen in much greater numbers than other species within the lamina propria (confirmed by subtractive staining with EUB 388 probe). H&E staining of parallel sections indicated that these bacteria seemed to be associated with cells of the inflammatory infiltrate, and further preliminary immunocytochemical staining identified these cells as being CD68+ cells of the monocyte/macrophage lineage.15 Although Schultsz et al9 and Swidsinski et al6 were unable to detect bacteria within the lamina propria, we have included a permeabilization step in the in situ protocol, which in our experience, allows greater accessibility of the probe to aldehyde-fixed tissue16; this step may explain the apparent discrepancy between these results and earlier reports.6,9

Earlier reports on mucosa-associated E. coli in IBD, as with bifidobacteria, are conflicting. Using an immunostaining technique, Walmsley et al5 reported the presence of E. coli in a minority of patients with ulcerative colitis and Crohn's disease, but in no healthy controls; the organisms were found both on the surface epithelium and in ulcers. In another study, using cultures and PCR in colonic biopsies, E. coli was found to be increased in patients with ulcerative colitis and Crohn's disease.6 Earlier studies suggested a possible role for pathogenic adhesive strains of E. coli in ulcerative colitis1,2,17; however, in another report, there was no difference in adhesive subtypes between patients with ulcerative colitis and controls.18

Our finding of increased mucosa-associated clostridia in patients with active ulcerative colitis confirms an earlier report using culture methods.19 Although we found no differences in mucosa-associated populations of the other bacterial groups probed, there are reports of reduced numbers of lactobacilli in colonic biopsies4,19 and feces12 of patients with ulcerative colitis and Crohn's disease, respectively. There are also studies reporting increased fecal sulfate-reducing bacteria and increased mucosa-associated Bacteroides in patients with ulcerative colitis.6,20,21 Discrepancies between reports are again likely to reflect different patient populations studied, differences in their treatment, and particularly, methods used to detect bacterial flora.

It is not yet clear whether the observed alterations in mucosa-associated flora in patients with IBD are of pathogenic significance. In interleukin-10 gene-deficient mice, an increase in colonic adherent and translocated bacteria and reduction in mucosal lactobacilli preceded histologic injury.22 Repopulation of the colonic lumen with lactobacilli normalized the profile of adherent and translocated bacteria and reduced the histologic severity of the colitis. These observations indicate a pathogenic role, at least in this model, for aerobic bacteria and a protective one for lactobacilli.

There are at least 6 diarrheagenic subtypes of E. coli,23 but our probe for this genera was unable to distinguish pathogenic from nonpathogenic subtypes. Previous reports have suggested a specific pathogenic role for adherent E. coli1,2 and a verocytotoxin-producing subtype24 in ulcerative colitis, and 1 study suggested that tobramycin improved outcome in acute ulcerative colitis.25 On the other hand, 2 studies suggest that administration of a nonpathogenic E. coli may help maintain remission in ulcerative colitis,26,27 whereas in at least 1 animal model of IBD, E. coli seemed to have no pathogenic role.28 Generalizations about host-bacterial interactions in relation to E. coli are therefore inappropriate, and the clinical significance of the increased mucosa-associated E. coli flora that we have found must await specific subtyping of the organisms.

It is more tempting to speculate that the deficiency in mucosal bifidobacteria that we have found in patients with ulcerative colitis could contribute to the pathogenesis of this disease. In animal models, administration of bifidobacteria and Lactobacillus has been found to reduce colonic inflammation,29,30 whereas administration of VSL#3 sachets, which contain 3 strains of Bifidobacterium, 4 strains of Lactobacillus, and 1 strain of Streptococcus thermophilis, is effective in preventing relapse and onset of pouchitis31,32 and possibly as maintenance therapy for ulcerative colitis.33 How such effects might be mediated is not yet clear, but possibilities include competitive interactions with pathogens, reduction in bacterial translocation, production of antibacterial products, and alterations in the mucosal immune response and mucin secretion.23,34 Indeed, even if the observed changes in mucosa-associated flora are secondary to inflammation-induced changes in permeability and/or mucin, rather than its primary cause, they could modulate its amplification and perpetuation through such actions.

In conclusion, we have shown that in patients with IBD, particularly when it is active, the mucosa-associated population of bifidobacteria is reduced and E. coli and clostridia are increased. An imbalance between the potentially protective and pathogenic bacterial flora apposed to and in the colorectal mucosa could play a role in the pathogenesis of ulcerative colitis and Crohn's disease.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

We thank Prof. G. Gibson for providing the FISH probes and for helpful advice during the preparation of this paper. M.M. was kindly supported by the Hellenic Society of Gastroenterology and B.N.H. and N.B.R. by the Christopher Reeves Charitable Trust.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References