The study reported herein demonstrates that the microbiota associated with the duodenal mucosa of CD patients has a characteristic deviation from the normal microbiota structure, which may characterize the disease. The alterations reported in the present study are partly consistent with those previously detected by molecular techniques using specific primers or probes (19
). Thus, our results support the hypothesis that normal components of the microbiota are excluded and replaced by others that could act as pathobionts in this specific disease environment. Although such associations do not demonstrate causality between the altered microbial groups and the disease, they provide a rationale for further studies on the possible pathogenic modes of action of such alterations and specific bacteria in CD.
To obtain bacterial isolates that are representative of those inhabiting the duodenal mucosa in both numbers and diversity, four different culture media previously described in the literature (34
) were used. In general, the greatest species diversity and quantitative recovery of mucosa-associated bacteria were obtained using the BH and YFCA culture media. These differences could be linked to the high nutritional requirements of duodenal bacteria, which are better met by the compositions of these media; incubation conditions may also have been more appropriate, as they were more anaerobic than those used for PCA and Wilkins-Chalgren agar. The diverse morphology of the small intestine favors a precise spatial relationship for strains within particular intestinal nutritional and microaerobic environments (42
), and therefore, it is rather complicated to completely reproduce the in vivo
environmental conditions. Also, even though the duodenum environment is not strictly anaerobic, the possibility that some anaerobic bacteria were lost due to oxygen exposure during sample manipulation cannot be disregarded.
We also analyzed whether some of the media used proved better at isolation of specific bacteria. In this regard, PCA and Wilkins-Chalgren agar seemed to favor the growth and isolation of members of the family Staphylococcaceae but hindered the growth of members of the family Streptococcaceae. BH medium favored the growth of members of the family Streptococcaceae but hampered that of members of the family Enterobacteriaceae. Wilkins-Chalgren agar also favored the recovery of members of the family Clostridiaceae compared to the other media. We confirm that none of the media or incubation conditions tested were suitable for the recovery of all viable bacteria detected in the samples analyzed when used alone, and therefore, various media must be used to improve the recovery of bacteria that are representative of the live bacteria inhabiting the duodenum.
We observed an increased diversity of the cultivable mucosa-associated bacteria recovered from CD patients compared to the diversity of bacteria recovered from the controls, and these differences were restored after adherence to a gluten-free diet. In concordance with this finding, denaturing or temperature gradient gel electrophoresis (DGGE and TGGE, respectively) analysis of duodenal samples showed a higher bacterial diversity associated with the small intestinal microbiota of CD patients (13
). However, several recent molecular studies (43
) have reported that reduced mucosal bacterial diversity is associated with inflammatory bowel disease (IBD), although the conditions and techniques used were not comparable to those used in the present study and the section of the intestinal tract studied was not the same.
Considering the isolates from all subject groups under study, our results show that the most abundant were those belonging to the phylum Firmicutes
, followed by those of the phyla Proteobacteria
. This is in concordance with the findings of a previous culture-independent study, where the same three phyla dominated the proximal small intestine of CD patients, followed by other phyla, such as Bacteroidetes
). Although our previous culture-independent studies also detected increased numbers of duodenal and fecal Bacteroides
spp. in CD patients compared with controls (19
), this bacterial group was not isolated with the culture conditions applied, probably due to exposure to oxygen during the process of homogenization of biopsy specimens and the use of nonselective media for Bacteroides
, which could have helped to limit the growth of less anaerobic and less nutritionally demanding bacteria. Culture-independent studies indicate that the members of the normal human gut microbiota mainly belong to two phyla, Firmicutes
, with a smaller number of bacteria belonging to the Proteobacteria
, although these conclusions are mainly based on analyses of the fecal microbiota composition (45
). Previous data also suggest that only 12% of the total species richness was detected by applying both molecular and cultivation-based approaches (50
). Remarkably, with both approaches, Firmicutes
represented the most abundant group, Proteobacteria
were relatively poorly detected by molecular approaches, and Bacteroidetes
were less abundant when they were assessed with cultivation-based approaches than with molecular techniques (49
). In relation to CD, differences in phylum representation were identified, and in particular, isolates belonging to the Proteobacteria
were more abundant in active CD patients than in nonactive CD patients and controls. In this context, other studies have also associated an increase in the Proteobacteria
and, in particular, an increase in adherent-invasive Escherichia coli
, Campylobacter concisus
, and enterohepatic Helicobacter
with IBD (52
In addition, active and nonactive CD seemed to be associated with a decreased abundance of members of the family Streptococcaceae
, specifically, the S. anginosus
and S. mutans
groups. The active phase of the disease was also associated with increased proportions of Enterobacteriaceae
and, in particular, the species Klebsiella oxytoca
, S. epidermidis
, and S. pasteuri
. In concordance with these observations, recent culture-independent studies indicate that the duodenal and fecal microbiotas of CD patients are characterized by higher numbers or proportions of Escherichia coli
). Furthermore, previous studies using cultured-dependent techniques have shown increased levels of S. epidermidis
) in feces from both active and nonactive CD patients in comparison with healthy controls and a lower prevalence of salivary S. mutans
in association with CD (53
). It seems that dominant genera in the normal microbiota of healthy individuals, which may act as symbionts, like Streptococcus
spp., are replaced in the CD patient microbiota by potential pathobionts, like Staphylococcus
spp. (S. epidermidis
) and enterobacteria, which could contribute to breaking down the normal dynamics and balance of the ecosystem.
To our knowledge, this is the first time that cultivable mucosa-associated bacteria of patients with active and nonactive CD have been studied, because previous studies were focused on the characterization of CD microbiota using molecular tools, such as DGGE and TGGE (13
), fluorescence in situ
hybridization (FISH) (20
), or real-time PCR (19
). Culture-dependent studies are intrinsically biased by the culture media used, the impact of potential oxygen exposure, and the inability to detect viable but noncultivable bacteria present in biological samples; notwithstanding these limitations, the results obtained in the present study are coherent with those of previous studies based on molecular techniques, which overcome these limitations. Therefore, the use of culture-dependent techniques has allowed the characterization of the active fraction of the mucosal microbiota of CD patients and will facilitate future investigation into the possible pathogenic role that isolated bacteria play in the development of CD.
This study demonstrates that the duodenal-mucosal microbiota of CD patients presents alterations in the diversity and abundance of different cultivable bacterial taxa, which could be a consequence of the pathogenesis of CD, which involves massive destruction of the small bowel mucosa and the consequent release of intracellular contents and serum into the gut. In the active phase of the disease, the mucosa-associated microbiota was characterized by a higher abundance of members of the phylum Proteobacteria and the families Enterobacteriaceae and Staphylococcaceae, apparently excluding members of the phylum Firmicutes and the family Streptococcaceae, which are normal inhabitants of the healthy small intestine. These alterations are attenuated after long-term adherence to a gluten-free diet, but the microbiota is not completely restored; in particular, a reduced abundance of specific species of Streptococcus (S. anginosus and S. mutans) also characterizes the microbiota of CD patients with active and nonactive disease. These findings also suggest their potential use as hallmarks of CD, regardless of inflammatory status.