Previous studies have suggested that abnormalities of the colonic microbiota occur in both UC and IBS [30
]. However the putative ability of certain bacterial species to enhance gut health, and the practicality of intervention strategies such as the use of prebiotics to achieve this, remains uncertain and poorly understood. Previous investigations of faecal or of biopsy-associated bacteria have failed to identify particular pathogens or bacterial groups linked to the altered microbial composition in UC patients relative to control subjects [30
]. However the application of molecular profiling techniques allows characterisation of the unculturable bacteria in the human colon, and in the present study has enabled us to address this issue in greater detail. Using this approach we have confirmed the existence of significant abnormalities in the predominant faecal bacteria of UC and IBS patients in relation to controls, using qualitative and semi-quantitative DGGE analysis, isolation and identification of the missing phylotypes.
DGGE is a semi-quantitative technique which provides a snapshot of the predominant bacterial species in a complex ecosystem such as human colon. Many of the studies to date have monitored changes in DGGE patterns by comparing the presence and absence of bands, or changes in the intensity of a single band on the same gel [32
]. Other studies have used pair-wise similarity coefficients, based on common bands shared between different samples [33
]. The total band numbers in a given sample, and their intensities, were combined to assess differences between DGGE profiles [34
]. However, human colon has a very complex bacterial community, often resulting in large number of bands, and furthermore the co-migration of the same bands on different gels can often make accurate analysis of DGGE data difficult [36
]. In our study we have applied a careful alignment method using the synthetic standards to accurately align bands between gels to mitigate experimental variation. The duplicate runs on different gels have given us additional confidence in the alignment procedure. To obtain a semi-quantitative DGGE analysis, the total number of bands and their intensities were included in our analysis after normalization.
Clinical and experimental observations have suggested the involvement of intestinal bacteria in the pathophysiology of UC; however the mechanisms by which bacteria or disease-specific pathogen are involved in UC have yet to be determined [37
]. Our data showed significant differences in the predominant faecal bacteria in UC patients relative to controls, suggesting a loss or reduction of bacterial classes, leading to the possibility of phylogenetic changes in these patients. This observation is consistent with other recent studies which reported a 40% decrease of dominant bacteria in UC patients [38
]. Such a decrease might be a response to environmental influences such as changes in diet or drugs or to endogenous signals derived from inflammatory changes or the immune status.
In a previous study, the predominant faecal bacteria of UC patients under remission were shown to be less stable than those of healthy controls, and the results revealed a host-specific pattern in relation to predominant bacteria in UC individual [39
]. Studies based on DGGE have shown small fluctuations of around 20% in predominant bacteria in healthy controls [40
]. Our results indicated that healthy controls share around 78% of predominant faecal bacteria, suggesting that this stability is a feature of a healthy gut. In contrast, UC individuals shared only around 70% of their predominant bacteria, which indicates that a particular shift of bacterial community might be a feature of UC patients. IBS patients shared around 68% of gut predominant species, and thus also displayed an even more diverse bacterial community amongst individuals, compared to controls. This observation indicates that despite the absence of inflammatory activity, this patient group is also disposed to changes affecting the composition of their faecal bacterial community. The reduced microbial diversity seen here with UC and IBS patients is, in agreement with the findings of previous studies. Martinez et al [39
] demonstrated reduced bacterial diversity of faecal bacteria from patients with active UC using a similar approach to that reported here. Ott et al [28
] also reported a reduction in bacterial diversity in tissue biopsies from UC and Crohn's disease patients using polymorphism analysis of amplicons of 16S rRNA. Our study has expanded our understanding of this phenomenon by undertaking qualitative analysis of the missing genera.
The GI microbiota is also thought to play an important role in the development and persistence of IBS; however, the numbers of detailed studies using culture-independent profiling methods are limited. Using classical methods, Balsari et al. [41
] reported qualitatively similar faecal bacteria in IBS patients and healthy controls, with some quantitative differences in the predominant bacteria. In contrast, our results show that the microbiota of IBS patients is quantitatively different compared to those of UC or the healthy controls, and that the inter-individual variation in predominant microbiota composition was greatest amongst the IBS patients. This observation might reflect the fact that IBS is a poorly defined condition not associated with any single, well defined biochemical, structural, or serologic abnormality. Thus although we have based our diagnoses on Rome II (a symptom-based criterion) to identify IBS patients, we are still likely to be sampling IBS patients from different sub-groups. Typical symptoms of IBS vary from constipation to diarrhoea, and both may be present at different times in the same patient. Therefore, it is sensible to assume that differences in intestinal bacteria within the global IBS group could be associated with the IBS subgroups. Such an interpretation is consistent with the observations of Malinen et al [30
] who used RT-PCR to analyse faecal bacteria of IBS individuals and reported changes in different bacterial spp. related to IBS subgroups exhibiting different symptoms. We did not attempt to differentiate the gut flora in IBS subgroups since the total numbers of volunteers was relatively small.
It has been shown by others that 3% of patients with enteric infections can develop UC [42
]. However, this figure is much lower than the 10% who develop IBS after infection [43
]. Hence, it is interesting to note that in the present study the predominant faecal bacterial population from some IBS patients more closely mirrors those of UC patients, and it would therefore be of interest to assess whether such patients might represent a post-infection IBS sub-group, and if they are at a higher risk of developing UC.
The members of the genus Bacteroides are one of the most frequently represented bacterial species in the human colon [44
]. A few species, such as B. fragilis
, are considered to be human pathogens, but most are thought to be normal commensal bacteria. In animal models of IBD, supplementation of the normal intestinal bacteria with a dose of Bacteroides vulgatus
has been reported to worsen the condition [45
]. Furthermore, it has been suggested that Bacteroides
is involved in the reoccurrence of UC after surgery [47
], although this study was small scale and so further evidence is needed. In the present study, despite the fact that there was no specific band found to be associated with UC or IBS patients, we were able to establish that Bacteroides
spp. were most responsible for group discrimination. Surprisingly, and contrary to our expectations, our data revealed that Bacteroides vulgatus
, probably Bacteroides ovatus, Bacteroides uniformis
, and Parabacteroides sp
. were more commonly present at higher levels in healthy controls than in UC or IBS patients.
In this context it is interesting that an earlier study demonstrated that B. vulgatus
can protect against E. coli
-associated colitis [48
]. More recent observations by Sydora et al. suggested that B. vulgatus
plays a role against the development of colitis [49
]. Conte et al (2006) found that amongst mucosal associated bacteria, B. vulgatus
occurrence was particularly low in patients with UC, Crohn's disease or with indeterminate colitis [50
]. A recent report by Takaishi et al (2008) also indicates that that the level of B. ovatus
was reduced in UC patients with active colitis [51
]. B. ovatus
has also been reported to protect germ-free and conventional mice, exposed chronically to dextran-sodium sulphate (DSS), from bleeding, development of intestinal inflammation and death [52