Microbial Eco-Physiology of the human intestinal tract: a flow cytometric approach

This thesis describes a multifaceted approach to further enhance our view of the complex human intestinal microbial ecosystem. This approach combines me advantages of flow cyrometry (FCM), a single cell and high-throughput technology, and molecular techniques that have proven themselves to be invaluabIe tools in studying the microbial diversity and structure of the intestinal microbiota. The ultimate aim was to relate the genetic biodiversity of the intestinal microbiota with their in situ physiological activity. The major findings of this thesis are summarized below and discussed in a broader perspective:
In chapter 1, the insights and newly developed molecular memods to study the ecophysiology of the gastrointestinal (GI)-tract are presented with a brief outline of the thesis.
In chapter 2, the principle of FCM was reviewed as weil as its applications in microbial ecology and physiology in different environmental settings. The power ofFCM sterns essentially from the ability to perform multiparametric analysis at the single celllevel, the high-throughput capacity, and the option of cell sorting.
In chapter 3 the technique to detect fecal bacteria using FCM and fluorescent in situ hybridization (FISH) was reported for the first time. The initial focus was to design and validate a 16S rRNA oligonudeotide probe (Urobe63), targeting uncultured Ruminnococcus obeum-like bacteria whose presence was only suggested by 16S rDNA doning studies of fecal samples. Subsequently, a flow cyrometric-based method was developed to ascertain the specificity of the probe, and to determine the abundance of this bacterial group in fecal samples using FISH. A hierarchical set of probes was used induding the general bacterial probe Eub338, the Clostridium coccoides - Eubacterium rectale group-specific probe Erec482, that encompasses the uncultUred R. obeum-like bacteria, and the newly designed pro be Urobe63. Combining the scatter parameters and the probe-conferred fluorescence of the detected cells revealed that a group of morphologically similar bacteria in feces hybridized with the Urobe63 probe. Quantification of the uncultured R. obeum-like bacteria by FCM-FISH, demonstrated that this phylogenetic group comprises a significant fraction of the fecal community (2.2%). Statistical analysis showed that FCM and fluorescent microscopy counts were in good agreement. These FISH results substantiate previous studies of phylogenetic investigation of fecal clone libraries that foreshadowed the high prevalence of this uncultured group of bacteria in fecal samples of different individuals (8,9). The present study demonstrates that combination of FCM and 16S rRNA-targeted pro bes can indeed be successfully applied to enumerate those microorganisms that have proven difficult to culture and which presumably play an important role on the gut physiology.
The knowledge about the mucosa-associated bacterial communities in different parts of the colon is limited, since most attention has been focused on bacteria present in feces. In chapter 4, the spatial distribution of the intestinal microbiota was investigated by comparing the bacterial communities in feces and biopsy samples taken from the ascending, transverse, and descending colon using a 16S rRNA approach. Flow cyrometric analysis indicated that 105 CO 106 bacteria were present in the biopsy samples (- 0.5 mg) with a detection limit of 104 cells/sample. These results were in contrast co a previous report using a microscopic method whereby hardly any bacteria were detected in biopsy specimens of healthy persons (6). Differences to these results could be a result of the detection limit of both methods or to the biopsy preparation. The numbers of bacteria counted in the ascending colon were slightly lower than those from other locations of the intestine. Denaturing gradient gel electrophoresis (DGGE) analysis of 16S rRNA gene amplicons and similarity index comparisons demonstrated, that the predominant mucosa-associated bacterial community was host-specific and differed significantly from the fecal community. However, no association of biopsy localization with bacterial diversity was observed demonstrating that the mucosa-associated microbiota is uniformly distributed along the colon. The cluster analysis of the PCR-DGGE profiles pointed out a potential difference between the bacterial communities of healthy and diseased individuals. However, it was not clear whether the differences observed were in fact the consequence of the health status of the individuaIs, due co the low number of samples studied. On the other hand, recent studies demonstrated that indeed the fecal and mucosa-associated microbiota significantly differed from that of patients with ulcerative colitis (UC) and Crohn's disease (2, 7) see also Chapter 5.
It has been well established that the predominant fecal bacterial communities of healthy individuaIs is host-specific, relatively stable in time, differs from mucosa-associated bacteria and not significantly altered following consumption of certain probiotic strains. In Chapter 5 the structure and population dynamics of the intestinal microbiota of UC patients over time and as a response co the intake of probiotics (Lactobacillus salivarius subsp. salivarius UCCl18 or Bifidobacterium infantis 35624), was monitored using FISH-FCM and PCR-DGGE analysis. The results obtained by both molecular approaches indicated that the microbiota of UC patients is host specific, but revealed marked alterations in the major bacterial groups of each patient. During the course of the probiotic trial, a significant increase in both bifidobacterial and Lactobacillus populations and a decrease in the Atopobium group were observed. However, these changes were not attributable co the type of treatment since they were observed in the three cohorts: placebo, probiotic-1 and probiotic- 2. On the other hand, the between group statistical analysis revealed that Bacteroides counts decreased significantly in the probiotic-2 group, while clostridia decreased in the placebo (a fermented milk containing Streptococcus thermophilus) cohort. The observed placebo effect may suggest that products of fermentation play a role in in vivo effects, and that consideration must be given to the delivery vehicle in comparisons to the efficiency between probiotic strains. Interestingly, bifidobacteria were detected at the pre-treatment period in all patients at a relatively high level. The PCRDGGE profiles did not show any explicit band that could indicate the presence of specific bacteria that may be related co the disease. Statistical analysis of the DGGE profiles showed that UC patients clustered separately from healthy controls. This suggests that the dominant microbiota of UC patients is atypical and that an imbalance of the microbial ecosystem rather than specific bacteria might play a role in the aetiology of the disease. Interestingly, the bacterial profiles of a number of patients tended to stabilize at the end the trial as revealed by the fingerprints. These observations suggest that the administration of live bacteria may have resulted in a "normalization" of the total microbiota.
The application of 16S rRNA-based techniques has greatly contributed to our understanding of the diversity of the intestinal microbiota, however a pertinent question still remains to be answered: which members of the ecosystem are functionally active at a particular time or at a specific site? It is therefore a major challenge to develop methods that aIlow monitoring of microorganisms according to their eco-physiological traits in situ. FCM has been viewed as a possible alternative to current assay techniques to monitor the viability of stressed and starved bacteria. In chapter 6 the efficiency of an FCM-based approach to monitor the effect of bile salt stress on the viability of Bifidobacterium adolescentis, a common member of the GItract, and Bifidobacterium lactis a widely used probiotic strain, was demonstrated. Carboxyfluorescein diacetate, propidium iodide (PI) and oxonol (DiBAC (3)) were used to assess the esterase activity, membrane integrity and membrane potentiaI, respectively, as indicators of bacterial viability. FCM analysis revealed a striking heterogeneity within the stressed bifidobacterial population and allowed to discriminate between viable, injured and dead sub-populations based on their differential uptake of the probes. One interesting finding was that the plasma membrane of the injured cells was apparently permeable to PI. An early stage of the recovery-resuscitation process involved repair of the damaged membrane, these cells could actually be scored dead by traditional criteria. Bile salt may have induced a sublethal-injury within bifidobaterial population possibly through a reversible and transient membrane permeabilization which resulted in a loss of viability, as defined by plate counts but could regain growth after being sorted and resuscitated. Preconditioning with bile salts showed a remarkably increased resistance of B. adolescentis NCC251 to lethal concentrations of bile salts, most likely through the induction of a mechanism allowing the cells to build up a protection against the solubilization of their membrane proteins (5). The findings of this study showed that multiparametric FCM is a very effective means for quantifying physiological heterogeneity of stressed bacteria, and thus provides a sensitive tooI to assess the viability and stability of probiotics.
The FCM approach has been extended and complemented with 16S ribosomal RNA gene analysis to study the relation between diversity and activity of the fecal microbiota (chapter 7). Simultaneous staining with PI and SYTO BC provided a dear differentiation between viabIe, injured, and dead fecal cells. Following sorting, the three subpopulations were characterized by DG GE of 16S rRNA gene amplicons obtained from the total and bifidobacterial communities. This analysis revealed that not only the total community but also the distinct subpopulations are characteristic for each individuaI. Phylogenetic analysis demonstrated that a number of amplicons, dosely related to likely butyrate-producing bacteria, were obviously viabIe at the time and site of sampling. However, amplicons affiliated with Bacteroides, Ruminococcus obeum-, Eubacterium biforme-like bacteria were especially obtained from the dead population at the time of sampling. Furthermore, some bacterial cIones were recovered from all sorted fractions, this was especially noticeable for the C/ostridium feptum cluster. This may indicate that nutrient availability and environmental conditions help determine the niche that a given microbe is able to occupy. Likewise, phylogenetic analysis of the live, dead and injured cells of bifidobacteria revealed a remarkable physiological heterogeneity within these bacterial populations. Basically, bacteria related to the B. longum B. infantis group were retrieved from all sorted fractions, indicating that these bacteria are active in the lower part of the GI-tract. Recently, the genome sequence of B. /ongum indicated its huge capability for untilization of complex carbohydrates derived from plants and which are believed to be a major component of the lower part of the colon (4). However, B. adolescentis related bacteria were mostly recovered from the sorted-dead fraction and these may have been active in the upper part of the colon. Finally, B. pseudocatenulatum, and B. bifidum-like bacteria were identified only in the live fraction. The differences observed between the individuaIs in terms of genetic diversity of the totaI, live and dead fecal bacterial populations might be attributable to not only the host-genotype effect but also to other factors, such as diet and host health.
In conclusion, the combination of FACS and 16S rRNA approaches that have been used in this thesis, contributed to consolidate our understanding of the large spatial diversity, structure, and the contribution to the global microbial activity to the intestinal ecosystem, as related to the host-genotype as well as the host-health status. The striking eco-physiological heterogeneity observed within the total and the bifidobacterial communities, may indicate that the source and availability for nutrients, metabolic interdependencies of microbes, as well as the challenging condition of the gut may play an essential role in shaping this complex ecosystem. The challenge set before us now, is to figure out how these microbes interact with each other and with the host to make their living and how the host designs its own community structure in order to maintain the homeostasis of the ecosystem. Novel approaches are being developed to answer such questions and to better grasp the function of gut-related bacteria in the intestinal tract and help for a rational design of probiotics and other functional foods. These include the use of omics-based approaches, i.e. global investigations of gene expression profiles or in-vivo gene expression identification strategies, and the combination of molecular techniques with substrate labeling such as stabIe isotope probing (SIP) 0, 3). These novel approaches will obviously shed light on the metabolic functions of the gut microbiota and its relation to the host health. With the expansion of bacterial genomes and the innovation of omics-based approaches, FCM will have also an important role to play in resolving the interactions between microbial genetics and physiology.