With 10^14 bacteria and 200 species, the digestive microbial ecosystem
is an extremely complex microcosm. The various species do not occur in the same
proportions, varying from one person to the next and according to location within
the digestive tube. In a given individual, however, they do remain stable through
time. Analysis of the normal flora shows the populations to exist in an equilibrium
state. Certain bacteria are extremely abundant: 10^11 cfu/g. These dominant
bacteria are strict anaerobes (Eubacterium, Bacteroides, Clostridium,
Peptostreptococcus, etc.). Others, facultative anaerobes such as E.
coli, Streptococcus, etc., are less common and are thus sub-dominant.
Lastly, there are the exogenous, transit bacteria which, although just "passing
through", may still, under certain circumstances, invade the host and prove
to be pathogenic (Clostridium difficile, etc.).
Only the dominant bacteria can play a role in the host's physiology. In
man, the sub-dominant population of facultative anaerobic bacteria has no effect
as long as the population level is kept in check.
A bacterial strain does not develop and proliferate unchecked in the intestinal
ecosystem as it does in a pure culture. Interactions between the various species
in the lumen can result in two strains mutually promoting one another, a given
bacterial population being totally or partially inhibited, and genetic information
being exchanged by DNA transfer. Interactions are of vital importance to keeping
the ecosystem stable.
MUCUS - MICRO-ORGANISM ASSOCIATION
Comparison of the mucus in axenic and conventional animals
- Physical aspect:
When the caecum of axenic (germ-free) rodents is compared with that of conventional
rodents, one striking characteristic is its larger size.
Along with larger size there is the secretion of a considerable surplus of mucin
in the faeces. According to some authors, the excess secretion is due
to the absence of intestinal microflora breaking down the mucin. In the conventional
animal, mucin synthesis and secretion is balanced by loss due to lubrication
and bacterial degradation.
Disagreeing with this, in 1990 Szentkuti et al. demonstrated that the absence
of intestinal microflora in the axenic rat reduced the thickness and density
of the pre-epithelial mucous layer.
Given the absence of mucolytic enzymes breaking down the mucous layer, it should,
at a constant rate of secretion, be thicker and more compact. However, the authors
state that the mucous layer of the colon wall is thinner in the axenic rat,
that the crypts are shorter, and that certain parts of the distal colon are
even almost devoid of goblet cells.
click to enlarge  Fig. 1: cross section of the gut wall in a conventional (CV) and axenic or germ-free (GF) rat distal colon (350 days). The pre-epithelial mucus layer (PML) is shown at the epithelial surface. Mean values correspond to the thicknesses of the PLM, mucosa, submucosa and muscularis in each group of rats. |
They explain the phenomena by a reduction in mucin secretion. In the axenic
mouse, renewal of goblet cells in the intestine is distinctly slower.
Histochemical aspect :
Bacterial colonisation of the mouse colon epithelium influences the chemical
structure of mucins. Histochemical examination of mucin in sections of young
axenic or conventional mouse colon has shown that all goblet-cell mucins are
sulphated at birth. Fourteen days later, in the absence of colonising microflora,
the goblet-cell mucins in the ascending colon continue to increase in sulphated
groups, whereas in conventional animals sulphated mucins are found only
in the upper part of the ascending colon crypts.
Generally, sulphate is incorporated into the mucin molecule through N-acetylglucosamine
residues. Onset of sulphatation seems to be influenced by either the presence
of one or more bacteria representing the normal colon flora, or hyperplasia
of the crypt resulting from the natural colonisation process.
Bacterial use of mucus
One of the main roles of gastrointestinal mucus is to act as substrate for intestinal
lumen proteases and bacterial glycosidases; 25% of mucin peptides are released
following the action of pepsin. Volatile fatty acids and sulphates delay this
proteolysis. After depolymerisation, the mucin becomes resistant to proteolytic
degradation. Bacterial glycosidases may the act.
Degradation of the mucin continues due to the extracellular glycosidases and
sulphatases of certain intestinal bacteria, representing only 1% of the total
flora. Despite this low percentage, these bacteria probably play a part in the
nutrition of other resident bacteria. Bacteria would therefore compete for these
substrates; only those strains using them most efficiently would not be eliminated.
Hence, indirectly, the glycosidase-active bacteria could help to regulate
equilibrium of the intestinal microflora ecosystem.
Other glycoconjugates, such as plant fibres, bacterial walls and cell membranes
of the intestinal mucosa, digested by the anaerobic bacteria, could also play
the role of nutritional substrates. These degradation processes could have
serious consequences on the production or elimination of bacterial or alimentary
mutagens and bacterial toxins.
After attack by the various proteases and glycosidases, only 40% of the sugars
and amino acids derived from mucin breakdown would be used by the bacteria,
the remaining 60% would be re-used by the liver in manufacturing new glycoconjugated
products. Hence, the greater part of the sugar does not simply disappear with
the bacteria in the stools, but is re-absorbed and re-used by the host metabolism.
BARRIER EFFECT OF THE INTESTINAL MICROFLORA
The normal bacterial flora in the mucous layer can ward off colonisation of
the mucus and invasion of the normal mucosa by pathogenic bacteria.
Within the host-flora ecosystem of the digestive tube, certain bacteria or,
more probably, certain bacterial associations have barrier effects which may
be drastic or permissive.
For example, the flora can prevent pathogenic bacteria from colonising by destroying
them during intestinal transit. When, as in this case, elimination is total,
the term used is drastic barrier. In other cases, certain exogenous bacteria
can be held down to a level low enough such that their pathogenic capacity has
no opportunity to express itself; in this case, the term is permissive barrier.
Numerous normal bacteria can bind themselves to the mucin's receptor sugars
and thus inhibit the pathogenic germs from binding. More importantly,
the various enzymes produced by the normal flora can selectively damage pathogenic
germs and reduce their viability by attacking their cell wall.
If the mucous layer is affected or deteriorated by dietary changes or antibiotic
treatments, for example, the protection furnished by the mucus is reduced and
results in the host becoming more susceptible to bacterial superinfection (13).
It is now well established that most C. difficile-related digestive disorders
result from the use of antibiotics, prime examples being pseudomembranous colitis
and post-antibiotic-therapy diarrhoeas. In both cases, the mechanisms given
for their appearance are always the change in composition of the normal microflora
and the creation of an environment favourable for the development of C. difficile
and the production of its toxins: toxin A, or enterotoxin, and toxin B, or cytotoxin.
Experimental model
The microflora of the conventional hamster is resistant to Clostridium difficile
colonisation. By administering clindamycin to the hamster, a pathology resulting
from the development and toxin-production of Clostridium difficile similar
to that observed in man may be reproduced.
Transferred caecal flora from the hamster to the C3H axenic mouse and,
after ensuring the barrier effect was maintained, conducted the rest of
experiments on these hamster-flora mice, kept in isolation equipment. The flora
was simplified by two successive treatments: one with erythromycin to eliminate
the facultative aero-anaerobic strains, then heating at 70°C for 10 minutes
to eliminate most of the non-sporulating bacteria. The simplified flora always
has a drastic effect on Clostridium difficile.
click to enlarge  Fig. 2: flowchart for obtaining gnotobiotic mice with simplified hamster microbial flora. |
Isolating the dominant strains in the simplified flora
Isolation was done using simplified-flora mouse caeca in Freter-type anaerobic
chambers (10% H2, 85% N2, 5% CO2).
Compared to the simplified flora which presents 4 or 5 different bacterial
forms on Gram staining, under these conditions, we were able to isolate :
click to enlarge  Fig. 3: slide appearance of the simplified barrier flora. |
Attempt to reconstitute the drastic flora from strains
isolated from the simplified flora
Seeding C3H axenic mice with strains 1, 2 or 3 did not result in colonisation.
Similarly, combinations 1+2, 2+3, 1+3 or 1+2+3 did not result in colonisation
either.
A triggering factor, possibly in the neonatal mouse's digestive tract, was
therefore missing.
Tried to isolate the bacterial strains which progressively colonised
the digestive tube of neonatal mice born of dams with the simplified flora.
This resulted in strain 4, Clostridium indolis, being isolated. The
strain furnished us with monoxenic animals we then successively inoculated
with Clostridium cocleatum and Eubacterium sp. In these trixenic
mice, strains 1 and 3 were dominant, and strain 4, which enabled the colonisation,
became sub-dominant but did not disappear.
The above series of experiments showed the importance of the first colonising
strain in preparing the local terrain.
It then considered two questions :
click to enlarge  Fig. 4: in vitro enzyme activities of each bacterial strain. |
then in association, and then in vivo in axenic, monoxenic (Clostridium
indolis), dixenic (C. indolis + C. cocleatum) animals,
and finally in our drastic flora mice.
click to enlarge  Fig. 5: in vivo enzyme activities observed in the faeces of various types of C3H gnotoxenic mouse. |
The results obtained in vitro and in vivo were the same.
- The first colonising strain (C. indolis) was not the most efficient
in degrading the mucin, since its sialidase activity was low. C. cocleatum
had the highest and most varied enzyme activity, and was the most efficient
at degrading the mucin. The results were confirmed by the mucin electrophoretic
profiles in agarose gel and polyacrylamide gel. In vitro, C. cocleatum
degraded the mucin oligosaccharide chains satisfactorily by means of its
enzyme apparatus but did not break down the peptide core.
Combining the three strains gave additive results and showed there was sufficient
enzyme activity to break down the mucin.
- Although the caecal mucins of axenic and Clostridium indolis-monoxenic
mice presented the same electrophoretic profiles, Clostridium cocleatum
would only colonise if C. indolis was already present. Given its
low mucinolytic capacity, the role of C. indolis is probably not
to provide nutritional substrates from mucin breakdown. On the other hand,
it could be involved in altering the redox potential or the caecal pH, or
in revealing certain binding sites for C. cocleatum. The two strains
would allow the anaerobic conditions required for colonisation of Eubacterium
sp., the strain indispensable for the drastic barrier effect against
C. difficile.
Location of the barrier flora in the intestinal mucosa
To obtain information allowing the barrier effect mechanism to be explained,
we studied the caecal location of bacteria making up the barrier flora as
colonisation of the mouse progressed, to compare it with that of C. difficile.
To do so, we used the "Swiss rolls" method, a technique allowing
the intestinal mucosa to be observed in section along the entire length
of the caecum, and scanning electronic microscopy.
According to our observations, at each stage of colonisation, the anti-C.
difficile barrier bacteria were located in the mucus at the surface
of the caecal mucosa and opening of the crypts, but never in the mucus inside
the crypts.
click to enlarge  Fig. 6: SEM photograph showing bacteria of the barrier flora (mainly dominant fusiform bacteria) agglomerated at the crypt openings and stuck in the mucus. (scale bar: 10 mm). |
In addition, quantitative differences in mucus colonisation were found according
to colonisation stage. The differences could be related to structural changes
in the mucins induced by bacterial enzymes. According to this hypothesis, through
its sialidase activity, C. indolis could detach the sialic acids from
the mucin oligosaccharide chains and reveal receptor sites for C. cocleatum,
allowing it to colonise the mucus. Through its osidase and osaminidase activity,
C. cocleatum could in turn alter the mucins and allow colonisation by
the third strain.
In the C. difficile-monoxenic mouse, the bacteria colonised the mucus
in the same way as the barrier flora bacteria.
Possible hypotheses on the barrier effect mechanism
All our work led us to assume that the mucins played a key role in the barrier
effect mechanism. The mucous layer covering the intestinal surface played a
protective role against the pathogenic bacteria by forming not only a physical
barrier but also by its mucins interfering with the pathogenic bacteria's adhesion
onto their targets. The barrier effect mechanism would thus be due to competition
between C. difficile and the flora bacteria either for an energetic substrate,
or for a specific receptor site located on the mucin oligosaccharide chains.
For resident bacteria, mucin is a major source of carbohydrates. C. difficile
has no enzyme able to attack the mucin oligosaccharide chains. On the
other hand, due to its hydrolytic enzymes (collagenase and hyaluronidase), it
can use nutrients such as the N-acetylglucosamine in the hyaluronic acid.
There might also be a "second line of defence". This could be related
to a phenomenon of competition in C. difficile attachment to specific
receptors located either on the oligosaccharide chains of the mucin, or on the
saccharide chains of the epithelial cell membranes' glycolipids or glycoproteins
which would act as receptors for bacterial adhesion.
The oligosaccharide chains do have similar structures, especially in type of
component sugars. A pathogenic strain such as C. difficile
should be able to penetrate the mucous layer to reach its target, otherwise
it would be trapped in the mucous gel and eliminated.
Value in microbial ecology and pathogenicity
The model is useful for analysing the mechanisms controlling strains' colonisation
and the close relations between host and microbial flora.
A lot remains to be learned about "flora-mucin" relations. They are
one of the keys to understanding the constitution of the starting resident flora
and, in particular, the dominant strains having the main physiological effects
observed. Lastly, barrier phenomena are of vital importance since they are the
foundation of the host's natural defence against pathogenic agents.