B1 cell

In� uence of Resident Intestinal Micro� ora on theDevelopment and Functions of the Gut-AssociatedLymphoid TissueMarie-Christiane Moreau and Valerie Gaboriau-Routhiau

From the Unite d’Ecologie et de Physiologie du Systeme Digestif, BaÃ timent 440, INRA, 78350 Jouy-en-Josas,France

Correspondence to: Marie-Christiane Moreau, Unite d’Ecologie et de Physiologie du Systeme Digestif,Batiment 440, INRA, 78350 Jouy-en-Josas, France.

Microbial Ecology in Health and Disease 2001; 13: 65–86

ORIGINAL ARTICLE

INTRODUCTION

Gut-associated lymphoid tissue (GALT) is under constantexposure to environmental antigens. The digestive � ora isthe main antigenic stimulus. A huge population of livebacterial cells, estimated at 1014 in number, colonizes thehuman gastrointestinal tract (1). Bacterial numbers andcomposition vary considerably along the gastrointestinaltract, constituting complex ecosystems which depend onthe physiology of the host and on interactions betweenbacteria. It has recently been shown that GALT has theability to develop tolerance towards resident bacterial � ora(2). Conversely, the digestive � ora considerably in� uencesthe development and functioning of GALT. To under-stand the relationships between the digestive � ora andGALT, it is important to consider the evolution of bacte-rial equilibrium during the main biological stages of life,from a digestive point of view, i.e. infancy (up to 2 yearsof age), adulthood and old age, as well as the bacterialcolonization of the different parts of the intestine. In thischapter, we will begin by dealing with the role of theresident digestive � ora on the development and functionsof GALT. Then, we will focus on the neo-natal periodwhich could be of particular importance for protectionagainst some pathologies such as allergy andhypersensitivities.

RESIDENT INTESTINAL FLORA

De×elopment of intestinal � ora in newborns

Digestive � ora in newborns is developed sequen-

tially according to the maturation of intestinal mucosaand dietary diversi� cation.

In healthy conditions, the human baby’s intestine issterile at birth but, within less than 48 h, 108 –109 bacteriacan be found in 1 g of faces (reviewed in (3, 4)). The � rstbacteria colonizing the baby’s intestine come from theenvironment, with maternal fecal � ora representing themost important and obviously best adapted source ofbacterial contamination. However, there are large differ-ences between the paucity of the bacterial species coloniz-ing the intestine of new-borns, and the very complex � oraof the adult. Just after birth, only the facultative anaerobicbacteria Escherichia coli and Streptococcus can colonizethe intestine of a baby. Two to four days later, an anaero-bic � ora can develop but only certain anaerobic bacteria,such as Bi� dobacterium, Bacteroides, and Clostridium, arefound. Although the factors responsible for the growth ofthese � rst selected bacteria are not clearly identi� ed, it ishypothesized that they may be endogenous factors, such asmaturation of the intestinal mucosa, growth promoters orinhibitors present in the meconium, or exogenous factors.The main exogenous factor is the nature of the diet whichplays a crucial role as shown by the differences betweenthe intestinal � ora of breast-fed babies and that of bottle-fed babies. In exclusively breast-fed babies, Bi� dobac -terium, E. coli and Streptococcus compose the predominantfecal micro� ora, while in bottle-fed babies it consists ofvarious bacteria, i.e. Bi� dobacterium, Bacteroides,Clostridia, and other Enterobacteria. Thereafter, accordingto dietary diversi� cation, the digestive � ora is enriched bythe development of other strictly anaerobic bacteria.

Weaning time is a crucial period during which deepchanges in the digestive micro� ora equilibrium occur. Dueto the instability of the baby’s bacterial equilibrium and,consequently, the incapacity of the digestive � ora to resistcolonization, the intestine can be colonized by opportunis-tic pathogens. The intestinal � ora of human infants thenreaches the stage of the adult � ora at an age which has not

Abbre×iations: GALT, gut-associated lymphoid tissue, GF,germ-free, CV, conventional, PP, peyer’s patche, LP, laminapropria, IgA-SC, IgA-secreting cell, sIgA, secretory IgA, LPS,lipopolysaccharide, IEL, intra-epithelium lymphocyte, APC, anti-gen-presenting cell, DC, dendritic cell, Abs, antibodies, DTH,delayed-type hypersensitivity, GvHR, graft-versus-host reaction,SPF, speci� c-pathogen free

© Taylor & Francis 2001. ISSN 0891-060X Microbial Ecology in Health and Disease

M.-C. Moreau and V. Gaboriau -Routhiau66

yet been accurately de� ned, but which is presumably atabout 2 years of age.

Bacterial colonization in mice shows some differenceswith that of the human baby (5, 6). During the � rst weekof life only facultative anaerobic bacteria, such as Lacto -bacillus and Streptococcus, colonize the new-born’s intes-tine. Then, at the beginning of the second week of life, E.coli reaches a high level. The � rst anaerobic bacteriadevelop at around day 15, when mice begin to take in solidfood. The weaning time, at around 3 weeks of age, iscorrelated with the development of a lot of new bacterialstrains, especially extremely oxygen-sensitive anaerobicbacteria which make up most of the � ora at this time.During the weaning period the interactions between nutri-ents, digestive � ora and GALT are closely dependent,making this period crucial to the development of theyoung. The digestive intestinal � ora seems to be stabilizedand representative of adult � ora at around 6 weeks of age(7).

Intestinal � ora in adults

More data is available on the location of bacterial � ora inthe different parts of the gut in human adults. The totalbacterial count in gastric content is usually 103 –104 g, withnumbers being kept low due to pH acid. The situation isdifferent in mice where facultative anaerobic bacteria,Lactobacillus, Streptococcus, are present at high levelsthroughout life. In humans, as in mice, the small intestineconstitutes a transition zone with numbers of bacteriaranging from approximately 104 bacteria per ml of theintestinal content in its proximal part to 107 –108 ml at theileocaecal region. In parallel, human micro� ora, essentiallycomposed of gram-positive facultative anaerobes, progres-sively enriches with gram-negative species and a few strictanaerobes (8). The main factors limiting growth in thesmall intestine are rapid peristalsis (transit of content) andsecretion of bile and pancreatic juice (9).

The large intestine (i.e. caecum and colon) of humansand mice is the most intensely colonized region, essentiallybecause of digestive stasis. It is generally accepted thatfecal micro� ora adequately represents that of the colon.This large population (1010 –1011 bacteria per g of intesti-nal content) is dominated by the strict anaerobes and alsocontains extremely oxygen-sensitive bacteria (10, 11). Thesituation leads to a certain amount of dif� culty in studyingcolonic micro� ora with classical microbiological methodswhich are based almost entirely on phenotypic approachesand the cultivation of bacteria on selective media. Al-though several hundred species of bacteria have beenobserved in the human colon only a few have been iden-ti� ed. The individual bacterial counts range over severalorders of magnitude. Thus, bacterial species established atlevels over 5.107 –108 bacteria per g characterize the pre-dominant micro� ora, whereas those below such athreshold compose the subdominant micro� ora. In fact, it

is believed that only predominant bacteria are able to exerta measurable function. The complex human intestinalmicro� ora is relatively stable in composition and the pre-dominant species commonly isolated from the humancolon belong to the genera Bacteroides, Eubacterium,Bi� dobacterium, Ruminococcus and Clostridium. Subdomi-nant species include enterobacteria, particularly E. coli,and Streptococci (12, 11; reviewed in 13). Whereas Lacto -bacilli are predominant in mice, they are frequently sub-dominant in humans or cannot even be detected.Conversely, Bi� dobacterium, which can be found at highlevels in human � ora, are absent in the mice.

Although the predominant human intestinal micro� orais relatively stable from the perspective of bacterial genera,recent studies using 16S rRNA-based approaches haveshown that each individual harbors a speci� c bacterialcommunity which increases the complexity already de-scribed (14, 15). Nevertheless, the predominant microbialcomposition remains quite constant over a period of atleast 6 months for healthy individuals (15). In addition toindividual differences and the in� uence of age (infant ×sadult), some studies have shown that factors such as stress(16) or antibiotic treatments (17) also induce variations inhuman intestinal bacterial micro� ora, resulting in subdom-inant species or even pathogens becoming more dominant.Very little data exists on the evolution of intestinal mi-cro� ora in the elderly. Nonetheless, bi� dobacteria havebeen reported to decrease at old age, which may be relatedto a reduced adhesion to the intestinal mucus (18).

Indigenous intestinal micro� ora plays several roles, aslisted in Table I. Two of them are especially important forhuman health: colonization resistance to maintain micro-bial composition in an apathogenic and stable state (19)and stimulation of the intestinal immune system, whichwill be developed below.

Table I

Main physiologic functions of the indigenous intestinal micro� ora

Modi� cation of intestinal pH, redox potentialProduction of metabolitescontent

(vitamins, digestive enzymes…)Reduction of metabolites (urea,

cholesterol, triglycerides…)

Anatomic modi� cation Caecal volumeof the digestive tract Enterocytes renewal rate

Villus morphometry

Modi� cation of digestive Transit of gastric and intestinalphysiology content

Absorption of abioticcomponents

Barrier effectImproved resistance togastrointestinal infection Modulation of toxin production

in the intestine

Stimulation of immunefunctions

Resident intestinal micro� ora and intestinal immunity 67

INFLUENCE OF INTESTINAL MICROFLORA ONTHE DEVELOPMENT OF GALT

The � rst year of life is a crucial period for human infants.At birth, the human GALT is poorly developed, while itsfunctionality is essential for survival. The relationshipbetween development and functionality of GALT is notclearly understood. GALT is quickly confronted with alarge quantity of foreign antigens mainly represented bythe digestive micro� ora. This period of time is also themoment when the risk of enteric infections and hypersensi-tivities to food proteins are the highest. Environmentaland behavioral factors (stress, hormones, etc.) in� uencethe development of GALT (20). Many observations sup-port the notion that most of the mucosal immune cells arecompetent even before birth, but they need to undergo anactivation process initiated by environmental signals. Ex-perimental data reported here shows the enormous in� u-ence of the presence of digestive � ora on GALT.

The digestive � ora has a dual function. First of all, it isthe main antigenic stimulus responsible for the activationand development of GALT. Secondly, through regulatorymechanisms, it modulates the GALT functions. Fromthese experimental studies, we can postulate that the diges-tive � ora also plays an important role on the developmentand functions of GALT in humans.

To understand the effects of the digestive � ora on thedevelopment of GALT, it is necessary to know what thestructure of GALT is at birth and what changes appearafter birth. As direct evidence from human subjects isscarce, we will report the possible roles of the digestive� ora on these changes on the basis of experiments usinggerm-free (GF) animals.

De×elopment of GALT in newborns

Peyer’s patches. At birth, developed Peyer’s patches (PP)are present in the human small intestine. They containprimary lymphoid follicles, T-cell dependent areas, a domeregion and follicle-associated epithelium. However, as noantigenic stimulation exists in the foetus, secondary folli-cles with germinal centers are absent. The � rst activatedlymphoid follicles with germinal centers appear severalweeks after birth (reviewed in (21, 22)).

Lamina propria. The lamina propria (LP) of the smallintestine is structurally well formed at birth but the cellu-larity is greatly reduced. There are no IgA-secreting cells(IgA-SC), while the machinery for production of secretoryIgA (sIgA) response is in place before birth (J chain andsecretory component). IgA-SC appear after 2–4 weeks ofage and the number increases with time, reaching adultlevel by 1–2 years of age (21, 22). In contrast, saliva andserum levels of IgA take longer to reach adult concentra-tion (around 6 years of age). In growing conventional (CV)mice, the � rst IgA immunocytes appear in the laminapropria at weaning time and a number equivalent to the

adult stage is reached 3 weeks later, i.e. in 6-week-old mice(23). The development is closely correlated to the increasedlevel of IgA in serum. This difference between humans andmice could be due to the fact that there is only one class ofIgA in mice and two subclasses, IgA1 and IgA2, in hu-mans which seem to develop at different rates and respondto different antigenic stimuli. IgA1 is predominant inhuman serum and the adult level is reached at around 6years of age. In intestinal mucosa, there is a heterogeneousdistribution of the two IgA subclasses of immunocytes(reviewed in (24)). IgA1 cells are in large number in theduodenum and jejunum (approximately 77%), whereasIgA2 cells predominate in the colon (64%). IgA2 are moreresistant to microbial proteases than are IgA1. Moreover,an IgA1 antibody response is obtained after stimulationwith a protein antigen (25), while IgA2 antibodies aredirected towards lipopolysaccharides (LPS) from gram-negative bacteria which are abundant in the colon. Arecent work (26) has shown that an abnormal bacterialovergrowth in human jejunal segments alters the IgAsubclass distribution with a preferential IgA2 production.Taken together, this data leads to the hypothesis that,among the unknown factors which in� uence a preferentialIgA1 or IgA2 response, the presence of the digestive � oramay be responsible for the induction of a preferentialswitch to a antibody response belonging to the IgA2subclass.

There are a few reports on LP T cells. At birth, the LPcontains T cells which are capable of secreting cytokinessuch as IFN-g and IL-2 when stimulated with superanti-gens (27). They are predominantly of the helper phenotype(22). However, the majority of T cells colonize the LP onlyafter antigenic activation (28). It is conceivable that anexpansion of T cells occurs in the LP after bacterialcolonization of the intestine and stimulation of the T cellprecursors present in the PP.

Concerning endothelial adhesion molecules (ICAM-1,VCAM-1 and E-selectin), which play an important role inthe recruitment of immune cells in the intestine, studieshave shown that they are potentially present at birth andthat antigenic stimulation leads to their expression on LPendothelium (29).

Epithelium. There are few intra-epithelial lymphocytes(IEL) in the intestinal epithelium of human new-borns.They are mainly CD3 » CD8» T cells. Their number in-creases slowly throughout pregnancy, but after birth itincreases up to ten-fold by 1–2 years of age, suggesting thestrong in� uence of antigenic stimulation on this post-natalexpansion (30). In mice, the homodimeric aa CD8» sub-population is present in suckling mice. The gd-TCR ap-pears during late fetal and perinatal development, whilethe ab-TCR appear later in adult life (31).

In humans, the expression of the epithelial class IImolecules at birth is the subject of controversy. Accordingto several studies, their expression has been found to be


present both before birth (32) and after birth (33). In thelatter case, the expression begins approximately 1 weekafter birth and reaches the adult level at around 1 monthof age. In growing mice, MHC class II molecules areabsent at birth but gradually increase on small intestinalepithelial cells after weaning (34).

Antigen -presenting cells. Antigen-presenting cells (APC),represented by dendritic cells (DC) and macrophage popu-lations, are located in the LP and PP. Lymph DC mayarise from both the sites. The major function of DC inimmune response is thought to be the acquisition of anti-gen and its transport to draining lymph nodes where it ispresented to T lymphocytes.

For human new-borns, no information is available re-garding the distribution, numbers or function of APC inthe intestine. Little data exists on mucosal DC from exper-imental animals. Measurement of the postnatal develop-ment of the airway intraepithelium DC network in ratsshowed that the number of DC expressing Ia moleculesand the intensity of the Ia expression were low at birth andincreased to adult levels around the time of weaning (35).The same results were reported in neonatal mice where asmall number of peritoneal macrophages and splenic ad-herent cells bear major histocompatibility complex (MHC)class II molecules. However, it was possible to increase therate of postnatal development of the intestinal DC popula-tion in rats by i.p. administration of IFN-g, suggestingthat the maturation of DC could depend on in� ammatorystimuli, and the establishment of the bacterial � ora couldafford such stimuli (35).

The ability of DC to interact effectively with the periph-eral T cells did not occur until 3–4 weeks after birth inmice (36). Due to the fundamental role of the APC indelivering adequate costimulatory signals to naive T cells,a de� ciency in the APC function could be a chief explana-tion for the immunological immaturity of new-borns at thesystemic level (37). This fact may explain why an antigenicstimulus in neonatal life is a tolerogenic rather than animmunogenic event.

Role of intestinal micro� ora on the de×elopment of GALT

According to the substantial experimental data showingthat extensive GALT modi� cation occurs at weaning timein mice, it has often been postulated that such changecould be due to the new antigenic proteins introduced inthe diet. In fact, as previously mentioned, at weaning,important changes in the digestive � ora are directly corre-lated to the dietary changes. To discriminate between theantigenic stimuli afforded by digestive micro� ora and di-etary proteins on the development of GALT, GF rodentswere used. The effect of the antigenic stimulation pro-voked by dietary proteins can be observed in GF animalsfed on an antigen-free diet (38). On the basis of this data,it is obvious that the bacterial � ora, instead of foodantigens, is the major stimulus for the induction of the

sudden increase in intestinal mucosal cellularity. It is nowimportant to learn to what extent these effects are opera-tive in humans.

In experimental studies, the role of the digestive � ora isdetermined by the comparison between GF and CV ani-mals. As new-borns, GF animals exhibit an underdevel-oped intestinal immune system which can be rapidlynormalized by bacterial colonization of the intestine withthe fecal � ora from a CV animal (i.e. oral inoculation witha fresh 1:100 dilution of CV fecal sample). It is, then,interesting to know the bacterial strains responsible for theimmunomodulating effect observed. Gnotobiotic mice, i.e.ex-GF mice colonized with known bacteria (mice or hu-man origin) are used for this purpose. After oral coloniza-tion, the bacteria expand rapidly to colonize the intestineto a very high level within 1 day. A period of 3 weeks isestimated to be the time required for an optimal stimula-tory effect on the digestive � ora.

Peyer’s patches. As in human new-borns, PP are poorlydeveloped in GF mice. The germinal centers are absent,showing that their development is directly under the in� u-ence of the digestive � ora (39, 40). CD4 » T cells are foundin Peyer’s patches of GF mice but, in contrast to CV mice,they are CD45RBlow , indicative of naive or unprimed Tcells. Colonization of GF mice with intestinal � ora shiftsthis population to a majority of CD45RBh igh, within 4–8weeks, suggesting a non-speci� c role for resident bacterial� ora in activating GALT CD4» T cells (41). Changes inbacterial intestinal colonization lead to a modi� cation inthe number of M cells which are known to play a crucialrole in the � rst step of the induction of mucosal immuneresponses (42).

Lamina propria. PP are an important source of progeni-tor cells for intestinal IgA-SC (43, 44). Another source ofprogenitor cells seems to be the CD5» B cells, also calledB-1 cells, which are present in the peritoneal cavity in mice(45). It appears that up to 40% of LP IgA-SC are derivedfrom B-1 cells in mice and humans (reviewed in (46)).Several authors have shown that the intestinal microbial� ora plays a pivotal role on the development of theintestinal IgA-SC number. It is unknown if this effect isoperative on the precursor cells present in the PP or:andon the CD5» B cells. However, according to Crabbe et al.(47), who � rst reported that less than 10% of IgA-SC arepresent in GF mice as compared with the CV mice, it isconceivable that the intestinal � ora exerts its stimulatingeffect on both B cell lineages.

Three weeks after bacterial colonization of the intestine,GF mice have an IgA-SC number equivalent to that foundin the CV mice. As in growing CV mice, the adult numberis only reached at 6 weeks of age (23), it has beensuggested that the immaturity of the immune system ofnew-born and:or the suppressive effect of the mother’smilk could be responsible for this delay. In fact, despite thebene� cial effect of breast-feeding in new-borns, it is be-


lieved that maternal antibodies may have a suppressiveeffect on the development of mucosal immune responses inyoung, leading to a partially developed immune system atweaning (reviewed in (20)). Studies in mice nursing for aprolonged time have shown a reduced quantity of IgA inintestinal washings at 5 weeks of age compared with thenaturally weaned litters, suggesting an active role for ma-ternal antibodies in delaying natural IgA responses (40).As previously described, studies in mice have shown thatthe sequential establishment of the digestive � ora isstrongly in� uenced by the diversi� cation of the diet ininfancy, and that weaning is a crucial period for thedevelopment of new bacteria in the intestine. Thus, thesuppressive effect of the maternal antibody effect on thedevelopment of intestinal immune responses may be re-lated to the paucity of the digestive � ora which can onlyexert a partial stimulating effect on the intestinal IgA-SCnumber. To test this hypothesis, several models of gnotobi-otic mice were created and then colonized by the entiredigestive � ora present in growing CV mice from 1 dayafter birth to 4 days after weaning (25 days of age) (6).IgA-SC numbers were evaluated by immunohistochemicalobservations 4 weeks after bacterial colonization. In theseexperimental models, the effect of maternal milk and theimmaturity of the neonate were discarded and only thestimulating effect of the digestive � ora was tested. Diges-tive � oras of mice 3–21 days old exerted only a partialstimulating effect on the intestinal IgA-SC number ingnotobiotic recipients (Table II). However, gnotobioticrecipients colonized with the digestive � ora of 25-day-oldmice have a similar IgA-SC number to that found in adultCV mice.

These results obviously show the important role playedby the sequential establishment of the digestive � ora in thesequential development of the intestinal IgA-SC and thepivotal role of the bacteria colonizing the intestine afterweaning in this process. Results have now been con� rmedby other studies (48, 40) showing that neither the age ofweaning nor milk antibodies played a direct role in IgA-SC development, but rather that weaning itself has animpact through its effect on the digestive � ora composi-tion. Moreover, taking into account the 3-week delaybetween the bacterial stimulus and the intestinal IgA-SCresponse, these results showed that the neonate is capableof developing an IgA response at birth, the intensity ofwhich depends on the stimulating capacity of the intestinalbacteria present in the intestine.

It is tempting to project such results into infants wherethe full development of the intestinal IgA-SC numberobserved at 2 years of age is correlated to the stabilizationof the digestive bacterial equilibrium.

Attempts have been made to elucidate the role played byindividual bacterial strains present in the digestive � ora ofCV growing mice (49). From several gnotobiotic mice

Table II

Effect of the sequential establishment of the digesti×e � ora ofgrowing con×entional mice on the maturation of IgA plasma cells

measured in gnotobiotic mice

Gnotobiotic mice harboring the digestive IgA plasma cellnumber:villus� ora of:

4191Adult conventional miceAdult germ-free 490.5

1592Growing conventional mice 1–4 days oldGrowing conventional mice 7–23 days old 2391Growing conventional mice 25 days old 4391

Results are expressed in mean numbers9SEM.

colonized with different single digestive bacteria, it wasshown that gram-positive bacteria had only a slightlystimulating effect (Table III). The association of gram-pos-itive bacteria led to a stimulatory effect which could bedue to an additive effect. In contrast, a single gram-nega-tive bacterium, such as E. coli or Bacteroides, provoked apartial increase in the IgA-SC number.

A similar stimulating effect was found with dead E. colior Bacteroides, under the condition that the dead bacteriawere given orally in the drinking water at a concentrationof over 107 cells per ml (50). This effect was probably dueto the LPS present in the cell wall of gram-negativebacteria (51). Unfortunately, because of the dif� culties inisolating and cultivating the EOS bacterial species whichare established after weaning, we failed to learn about thebacteria, or the bacterial equilibrium which are responsiblefor the complete development of the IgA-SC numbers (49).

In contrast to the extensive studies on LP plasma cells,there is considerably less information on the role of intesti-nal � ora on development of LP T lymphocytes. In a study

Table III

Effect of different bacterial strains isolated from the intestinal � oraof growing con×entional mice on the maturation of intestinal IgA

plasma cells in gnotobiotic mice

Bacterial strains IgA plasma cellnumber:villus

Micrococcus 391391Corynebacterium591Eubacterium

Lactobacillus 791Streptococcus 892

1092ActinobacillusEscherichia coli 1893Bacteroides 1893

1792E. coli»Bacteroides

4191Conventional mice � ora391Germ-free mice

Results are expressed in mean numbers9SEM.


using pigs, Rothkotter et al. (52) showed that sub-popula-tions of T cells differ substantially between GF and CVanimals.

EpitheliumIntra-epithelial lymphocytes. IEL are found in both thesmall intestine and the colon in humans and mice. In micethere are marked differences in the distribution and per-centages of IEL expressing ab or gd TCR between thelarge intestine, where the TCR-ab T cells (ab-IEL) pre-dominate, and the small intestine, where TCR-gd T cells(gd-IEL) predominate. In human ab-IEL predominateboth in the small intestine and the colon (53).

The biological roles of IEL are still mysterious. Toclarify the function of IEL, it would be interesting to knowwhat their reactivity towards exogenous antigenic stimuliis. The numbers, phenotypes and cytolytic activities of IELdiffer greatly between GF and CV mice, while ab- andgd-IEL seem to respond to different kinds of antigenicstimulation. There is a marked in� uence of the microbialintestinal colonization on the number of single positiveCD4» or CD8» ab-IEL, but little effect on the pool sizeof gd-IEL (54, 55). Moreover, the thymo-independent ho-modimeric aa CD8 » subpopulation of IEL (all the gd-IEL and part of the ab-IEL) is always present in GF mice(31). Thus, the subsets of IEL responding to the microbialstimulation are thymo-dependent precursor T cells presentin PP. Conventionalization of GF mice has shown that in28 days the same percentage of ab-IEL is reached as thatfound in CV mice (56).

In a recent study, Kawaguchi-Miyashita et al. (57) com-pare the effect of microbial and dietary antigens on thecytolytic activity of ab-IEL and gd-IEL in CV, GF andGF mice fed on an antigen-minimized diet (AgM-GFmice). Results show that the development of cytolyticactivation of ab-IEL is sharply attenuated in GF mice, butthe number and cytotoxicity of gd-IEL are comparablebetween CV and GF mice. In contrast, the cytolytic activ-ities of gd-IEL, as well as those of ab-IEL, decreaseremarkably in AgM-GF mice (Table IV).

Although the function of IEL is a subject of debate,these results showing that the cytolytic activity of gd-IELis only under the in� uence of the antigenic composition of

the diet, suggest a role of the IEL subset in relationshipwith dietary antigens. Microbial antigens are quantita-tively low in the small intestine, which is a relatively sterileenvironment, while food proteins are found in large quan-tities in the upper part of the small intestine. gd-IELrespond to food proteins and proliferate more rapidly inthe duodenum, where food antigens � ow in large quanti-ties, than in the ileum (57). It has been suggested thatgd-IEL are involved in the induction and maintenance oforal tolerance to food-proteins (58, 59). Conversely, theobservation that ab-IEL, abundant in the colon, aredeeply in� uenced in their number and cytolytic activity bythe presence of the digestive micro� ora supports the ideathat this population may act against enteric pathogens inthe colon.

Human and murine ab-IEL have been shown to expressan oligoclonal TCR repertoire. In mice, the same TCRrepertoire appears to predominate in all the parts of theintestine, but it differs from mice with an identical genetic-and environmental background (60). Interestingly, the di-gestive � ora is not responsible for oligoclonality of therepertoire since the TCRb repertoire of the CD8aa orCD8ab IEL populations in GF mice shows the samedegree of oligoclonality as in CV mice (61). Recent studiesin rats show that the microbial colonization of GF ratsin� uences the TCR Vb repertoire of CD8» T cells. Theproportions of Vb8.2 » and Vb10 increase, whereasVb8.5» and Vb16» cells decrease somewhat (62). These� ndings support the hypothesis that the TCR a:b reper-toire of IEL could be directed against bacterial antigens.In contrast, the TCR Vb repertoire in mesenteric lymphnodes was not affected after microbial colonization.Class II molecules. The expression of MHC class IImolecules on the small intestinal epithelial cells is closelyassociated with bacterial colonization of the intestine (63).The authors have shown that in GF mice, although class Imolecules are expressed on the small intestinal epithelium,class II molecules are absent. In their study, conventional-ization of GF mice provoked a gradual expression of classII molecules in the small intestine. However, there weredifferences in the delays and regions of expression of theI-A and I-E molecules. The I-A molecule was induced onthe villus tip and crypt epithelial cells 7 days after conven-

Table IV

Effect of the depri×ation of digesti×e microbial and food antigens on the number and cytolytic acti×ity of IEL (from (57))

Control mice Microbial deprivation Antigen deprivation(GF mice %) (GF mice fed with an Ag-minimized diet %)(CV mice %)

IEL numbers3050ab-IEL 10080 (NS)gd-IEL 100 100

Cytolytic activity1050ab-IEL 100

gd-IEL 100 30100


tionalization, whereas the I-E molecule was induced on themid villus and crypt epithelial cells 14 days after conven-tionalization. Both the I-A and I-E molecules were fullyexpressed 21 days after conventionalization. It has beenshown that the expression of class II molecules is modu-lated by IFN-g secretion (64) and bacterial colonizationmight cause intestinal IEL to release IFN-g.

In another study, using the kidney as the representativeof non-lymphoid tissue, class II molecules were expressedat the same level in GF and CV mice (65). The discrepancybetween the two studies may arise from different epithe-lium studied. Indeed, an equal presence of MHC class II-Awas observed on spleen and lymph node cells from bothCV and GF mice, suggesting the constitutive antigen-inde-pendent expression of these molecules at the peripherallevel.

Antigen -presenting cells. It appears that DC populationsassociated with airway and intestinal surfaces are cyclingthrough these environments at a faster rate than those inother areas or tissues. This fact may be due to the stimu-lating effect of the mucosal bacterial � ora and to a largeantigenic load (66). LPS, present in the cell wall of gram-negative bacteria, has been described to increase the levelsof MHC class II and B7 molecules (67). In an anotherstudy, it has been shown that i.v. administration of LPS inadult rats increases the release of intestinally derived DCinto lymph and that TNF-a could play a role on themechanisms underlying DC release (68). The source ofintestinal TNF-a is still unclear. Mucosal mast cellspresent in the lamina propria are potential candidates forTNF-a release as it is stored in these cells (reviewed in(69)). However, it remains unknown exactly how the mastcells respond to the resident bacterial � ora. On the otherhand, macrophages are also abundant in GALT and it hasbeen shown that the colonization of the intestine of gnoto-biotic mice with E. coli stimulated the release of TNF-a byperitoneal, as well as bone-marrow derived macrophages(70). Unfortunately, it remains unknown whether thisstimulating effect is operative on intestinal macrophages.LP macrophages represent a speci� c subset ofmacrophages. They are derived from circulating bloodmonocytes, but their phenotype differs from that of themajority of blood monocytes. In humans, the most strikingdifference is the absence of CD14 on LP macrophages, areceptor for the LPS-binding protein, a serum protein thatcomplexes with LPS (reviewed in (71)). This absence mayplay an important role in maintaining mucosal homeosta-sis by suppressing local macrophage activation and de-creasing secretion of in� ammatory cytokines. It isunknown whether the presence of bacterial intestinal � orais the environmental factor responsible for the down-regu-lation of CD14 » expression on mucosal macrophages.

From an ecological point of view, we can speculate that,after birth, the colonization of the intestine by E. coli maybe a strong in� ammatory stimulus responsible for the

synthesis of in� ammatory cytokines, IFN-g and TNF-a, atthe intestinal level, with consequences on the functionalityof macrophages and DC. At the adult stage, when thelevel of E. coli decreased, another gram-negative bac-terium, Bacteroides, present in high levels, could be animportant physiological source of LPS.

INFLUENCE OF INTESTINAL MICROFLORA ONTHE MODULATION OF GALT FUNCTIONS

GALT generates two important immune functions. First isthe immune exclusion performed by secretory IgA anti-bodies (sIgA Abs) to protect the mucosa by blocking themicrobial adhesion, microbial translocation and viral mul-tiplication, as well as by neutralizing toxins. Second is asuppressive function, also called oral tolerance, character-ized by regulatory mechanisms avoiding local and periph-eral immune responses to harmless environmental antigenspresent in the intestine, such as bacterial antigens of theresident micro� ora or dietary proteins. Now, it is stillunclear whether suppression of the systemic immunity isaccompanied by local sIgA production or not.

The early postnatal life is a period of high risk forintestinal disorders due to enteric pathogens and:or foodhypersensitivities. During the neonatal period mammalianspecies exhibit some degree of reduced immunocompetencethat could be attributed to functional immaturity in popu-lations involved in immune intestinal responses. It couldbe also attributed to immunoregulatory mechanisms gov-erning the GALT functions. Apart from the role of intesti-nal � ora on the development of GALT, the presence ingreater numbers of gram-positive and gram-negative bac-teria containing immunomodulator components in theircell walls with adjuvant capacities (72) may have modulat-ing effect on the GALT functions. Depending on thebacterial equilibrium, which can differ from one individualto another, immune responses elicited by GALT may bemodulated differently. Gnotobiotic animal models are use-ful in analyzing such modulating effects of digestive bacte-ria on GALT functions.

Secretory IgA antibody responses

The presence of the resident digestive � ora exerts bothdirect and indirect effects on the sIgA responses. It inducessecretion of natural sIgA Abs in response to the bacterialantigens. In addition, it plays a modulating role on thespeci� c sIgA response against some enteropathogens, asrecently described with rotavirus (73, 74). The regulatorymechanisms involved in producing such effects areunknown.

Intestinal micro� ora and natural sIgA antibodies. Sequen-tial intestinal bacterial colonization is responsible for thepresence of natural sIgA Abs. They express polyreactivitiesand are, with other natural factors such as mucus and bile,the � rst line of defense at mucosal surfaces. Their ef� -


ciency is increased by binding an additional intestinalendogenous protein component, the fragment-bindingprotein (pFv), to form large complexes of high-agglutinat-ing activity (75, 76). It has recently been shown thatbacterial colonization of GF rats with a human colonic� ora favored the release of pFv (77). Both the B-1 and B-2cell lineages originating from the peritoneal cavity andPeyer’s patches, respectively, seem to contribute to intesti-nal sIgA Abs responses. Recent studies have shown thatB-1 and B-2 precursors of IgA intestinal B cells havedifferent antigenic repertoires. B-1 cells mainly producepolyspeci� c natural Abs which bind to multiple unrelatedantigens, predominantly microbial T-independent antigens,whereas B-2 cells respond predominantly to T-dependentantigens (reviewed in (46)). T-independent antigens aremainly found in microorganisms and many of the LP IgAplasma cells secreting natural sIgA are postulated to origi-nate from the B-1 cell lineage (78). The site where B-1 cellsare stimulated, their path of migration to seed the LP andhow they differentiate into IgA plasma cells are questionsstill awaiting an answer.

The roles of the natural sIgA Abs are not very wellknown. They have the potential to agglutinate multiplepathogens without delay or prior exposure. In healthyconditions, natural sIgA Abs could be important in in-hibiting the microbial adherence and penetration of resi-dent bacteria in the mucosa. sIgA are present at high levelsin mucus, which is the � rst protective barrier againstmucosal bacterial colonization. Few bacterial species areclosely associated with epithelium. Most of them live in themucus layer. It has been shown in rats that the adherenceof bacteria to the intestinal mucosal surface is an impor-tant factor in bacterial translocation (79). In neonatalrabbits, IgA supplementation abrogated bacterial translo-cation (80). A majority of freshly isolated intestinal bacte-ria from a normal adult mouse are ‘coated’ with sIgA (40).The relevance of this coating in the regulation of theintestinal microbial � ora equilibrium is uncertain (81), butit could be important in protecting the systemic compart-ment from translocation of intestinal resident bacteria.

Intestinal micro� ora and modulation of sIgA anti -ro-ta×irus response. Gastrointestinal infections and their con-sequences are a major clinical and economical problem.

For example, Salmonella typhi, Helicobacter pylori in hu-man adults and rotaviruses in infants cause mortality andmorbidity worldwide. Little information is available re-garding the role of resident intestinal bacteria on themodulation of the speci� c sIgA Ab response against en-teropathogens. The � rst information has emerged fromstudies using lactic-acid producing bacteria as probiotics inmice (82) and humans (83). Although it has not beendemonstrated that probiotics can colonize the intestine ascan resident bacteria, they can exert immunomodulatingeffects during their transit time.

Interestingly, in breast-fed babies a lactic-acid producingbacteria, Bi� dobacterium, is one of the � rst anaerobicbacteria which colonize the baby’s intestine. As it is com-monly observed that breast-fed babies are more resistantto gastrointestinal infection, we hypothesized that the pres-ence of Bi� dobacterium in the resident digestive � ora couldhave a stimulating effect on the sIgA Ab response againstenteropathogens. To test this hypothesis an original modelof adult mice infected with the heterologus simian ro-tavirus strain SA-11 was developed (73, 74). Total andanti-rotavirus sIgA responses were evaluated both in fecesand in small intestine LP by enumerating IgA- (IgA-SC)and anti-rotavirus IgA-secreting cells (ARSC). To assessthe respective immunomodulating role of the two bacteriapresent in the baby’s intestine, Bi� dobacterium (gram-posi-tive bacteria) and E. coli (gram-negative bacteria), twogroups of gnotobiotic mice were created. Bacteria wereallowed to colonize the intestine 3 weeks before viralinfection to permit the development of their immunologi-cal effect on GALT. Results on LP ARSC and IgA-SCnumbers are presented in Table V. They were in goodcorrelation with fecal measurements (73).

Several conclusions have been drawn from the informa-tion found in this work. Firstly, whereas Bi� dobacteriumand E. coli were both established in high numbers in theintestine of gnotobiotic mice, they modulated the IgAanti-rotavirus response in a completely different way. Thepresence of Bi� dobacterium had a strong adjuvant effecton the anti-rotavirus IgA response, whereas E. coli exertedan obvious suppressive effect. It is interesting to note thatthe low sIgA anti-rotavirus response obtained in CV micemay be explained by the presence of another gram-nega-

Table V

Anti -rota×irus secreting cell (ARSC) and IgA -secreting cell (IgA-SC) numbers in small intestine laminapropria of CV, GF and gnotobiotic mice colonized with strains of E. coli or Bi� dobacterium

ARSC:106 cells Ratio of ARSC:IgA-SC½100IgA-SC:106 cellsGroups of mice

0.26 90.15ÀÀ206 000961 000ÀÀConventional 100941ÀÀ

17 000980004479198 3.80 91.74Germ-freeBi� dobacterium 33 000910 000548692670ÀÀ 12.595.65ÀÀ

0.63 90.29ÀÀEscherichia coli 48911ÀÀ 19 00096000

Results are expressed in mean numbers9SEM of secreting-cells. ÀÀ PB0.01 with GF group.


tive bacterium, Bacteroides, which is present in very highlevels in intestinal � ora of adult CV mice. Secondly, in allthe groups of mice, the ARSC number did not depend onthe total IgA-SC number, showing a lack of correlationbetween the modulating effect of bacteria on the sIgAanti-rotavirus response (virus-speci� c sIgA) and on thetotal sIgA response (natural sIgA). Finally, the GF micewere able to mount a strong IgA anti-rotavirus responsewhile GALT was poorly developed.

Other studies have shown an enhancement of serum orintestinal Ab response to orally administered antigens bygram-positive bacteria (84, 85), especially lactic-acid pro-ducing bacteria used as probiotics. The exact mechanismsunderlying such effects at the intestinal level are poorlyunderstood. Bacterial components are known to have im-munomodulatory properties (72). Cell walls of gram-posi-tive bacteria are rich in peptidoglycans, which have beenspeci� cally described as stimulating macrophage functions(reviewed in (86)). In contrast, LPS is particularly abun-dant in gram-negative bacteria and previous studies haveshown its numerous effects on immunity, especially sup-pressive effects on immune responses (51, 87, 88).

According to our results, the modulating effect of thedigestive � ora on natural and speci� c sIgA Ab responseswere surprisingly uncorrelated. To our knowledge, this facthas never been discussed and we do not know the exactcauses explaining such a lack of correlation. We canpropose that it arises from a difference in B-1 and B-2 cellselection upon antigenic stimulation. Interesting resultswere obtained from previous studies measuring intestinalanti-Salmonella IgA plasma cell responses obtained afteroral infection with S. typhimurium in non-responsive irra-diated Xid mice reconstituted with responsive donor cells(89). The authors showed that donor cells from the peri-toneal cavity, enriched mainly with B-1 cells, were capableof giving an intestinal IgA Ab response, in contrast to Bcells isolated from Peyer’s patches, mainly of the B-2 type,which did not contribute to this response. Thus, we canpostulate that thymo-independent bacterial antigens couldact preferentially on B-1 cell lineage to secrete naturalsIgA, whereas thymo-dependent antigens, such as viralproteins, might elicit B-2 cell lineage to secrete virus-spe-ci� c sIgA. Moreover, the direct effect of bacterial compo-nents present in the cell walls and:or secreted by livingbacteria, could also modulate the sIgA anti-rotavirus Abresponse through non-speci� c cellular and molecularevents. Stimulation of APC functions and cytokine synthe-sis may be involved as modulating events between theinduction of the IgA response in Peyer’s patches andsecretion of sIgA in the intestinal lumen (reviewed in (90)).Mechanistic studies are required to clarify the molecularbasis upon which resident bacteria modulate the sIgA Abresponse to enteric pathogens.

The hypothesis of a possible dichotomy in virus-speci� c

sIgA and natural sIgA responses may explain the resultsobtained by Cebra et al. (91) in new-born mice. Theydescribed that 10-day-old suckling mice were as competentas 12-week-old mice at initiating a virus-speci� c sIgAresponse after enteric infection, whereas the complete de-velopment of the intestinal IgA plasma cell number wasreached only 5–6 weeks later. Such results have also beendescribed in humans. One-week-old babies were capable ofdeveloping protective immunity following oral vaccinationwith poliovirus or hepatitis B virus while the completedevelopment of natural sIgA Abs takes several months ininfants (reviewed in (92)). These observations address animportant question about the immunological competenceof new-borns to mount a sIgA Ab response. It is currentlyadmitted that the ability of the immunologicaly naivenew-born to generate a mucosal immune response dependson the functional capacity of GALT at birth but, takentogether, all this data suggests that neonates are probablycapable of mounting an active sIgA Ab response. Conse-quently, the ability to give a high speci� c sIgA anti-ro-tavirus Ab response could be correlated with themodulating effect of intestinal bacteria rather than withthe development of GALT. This data could have impor-tant implications for oral vaccination of human new-borns. Numerous questions remain to be answered. Theyare discussed in Cebra et al. (41).

In conclusion, experimental animal models of gnotobi-otic mice brought to light the immunomodulating proper-ties of two intestinal strains, Bi� dobacterium and E. coli,on intestinal IgA anti-rotavirus response. Our results sug-gest the importance of the presence of Bi� dobacterium inthe baby’s intestine in potentiating the synthesis of IgA Abagainst viral enteropathogens. Foods promoting Bi� dobac -terium in the intestine could be instrumental in promotinga bene� cial effect on health.

Oral tolerance

De� nition. Oral tolerance (OT) is classically de� ned asthe state of antigen-speci� c systemic immunological unre-sponsiveness induced by prior oral administration of theantigen and subsequent systemic exposure to the sameantigen. It has been described in numerous animal modelsby using various antigens, especially particulate ones, suchas sheep red blood cells (SRBC), or soluble ones, such asovalbumin (OVA) (reviewed in (93)). OT has also recentlybeen described in humans (94). Its induction is believed tobe of physiologic importance to avoid hypersensitivityreactions to dietary antigens. Indeed, although smallamounts of orally administered proteins escape enzymaticdigestion in the intestine, they do not induce immuneresponses under healthy conditions (95). Recently, it hasbeen hypothesized that a similar state of tolerance isestablished towards the indigenous gut micro� ora (2).Increasing interest in OT is also related to its potential rolein the treatment of autoimmune diseases (reviewed in (96)).


Orally fed antigens can suppress both speci� c humoral(IgG and IgE antibody responses) and cellular immuneresponses (reviewed in (97)). Cell-mediated immune re-sponses, such as T-cell proliferation, delayed-type hyper-sensitivity (DTH) or contact sensitivity and CD8»

cytotoxic T-cell responses, are generally easier to toleratethan the humoral responses (98), even in primed animals(99, 100). Nevertheless, IgE response seems to be remark-ably sensitive to suppression when antigen is fed before orafter parenteral immunization with the same antigen (101).Considering that IgE- and cell-mediated immune responsesare frequently implicated in human food hypersensitivities(reviewed in (102)), it seems logical that these immunolog-ical responses would be highly susceptible to regulatorymechanisms.

Systemic unresponsiveness is a long-lasting processwhich leads OT to be considered in terms of induction andmaintenance. OT induction can be demonstrated verysoon after feeding a protein, as it can be veri� ed byparenteral immunization with the same antigen within 7days of feeding (103). However, its duration depends onthe immune response studied. Studies on mice have shownthat suppression of DTH lasts up to 17 months after onefeeding of 20 mg OVA, whereas the suppression of the IgGantibody response does not last more than 3–6 months(98).

Mechanisms of oral tolerance. It is now admitted thatmultiple mechanisms are involved in OT. Although thedebate on the relative role of each mechanism in OT stillexists, it seems likely that they are not mutually exclusive.They have been extensively studied, especially in the lastdecade, and most of them have been described in detail inrecent reviews (96, 97, 104, 105). We will focus only ontheir major characteristics and on major questions stillunder debate.

Three principal immunological mechanisms have beenimplicated in OT, antigen-driven active cellular suppres-sion, clonal anergy and clonal deletion of potentially reac-tive lymphocytes.

Active suppression was � rst described as a mechanismmediated by regulatory suppressive CD8» T cells inducedin GALT, which then migrate to the systemic immunesystem (reviewed in (93, 104)). More recently, it has beensuggested that intestinal CD8 » TcR-gd » IEL may beimplicated in OT (58, 59). However, today, the require-ment of CD8» T cells in induction and maintenance ofOT does not seem to be so categorical, as OT can beinduced in mice de� cient or depleted of CD8 » T cells(106–108). In contrast, involvement of regulatory CD4»

T cells in OT appears essential. It has been proposed thatOT may re� ect preferential activation of Th2 suppressiveT cells and down-regulation of Th1 responses by Th2 cellsvia suppressive cytokines, such as Il-4, Il-10 and trans-forming growth factor (TGF)-b (109). In relation to thecytokine-mediated active suppression, the phenomenon of

‘bystander suppression’ has been described. Indeed, sup-pressive cytokines secreted after antigen-speci� c activationof regulatory T cells could suppress the immune responsesto an unrelated antigen anatomically colocalized with thefed antigen. Bystander suppression, therefore, representsan important potential in the treatment of autoimmunediseases where autoantigens are unidenti� ed or available inextremely low quantities (reviewed in (96)). However, OTto OVA can suppress both Th1 and Th2 responses andnormal induction of OT is observed in both Il-4 and Il-10de� cient mice (110, 111). A recent study in a model ofexperimental autoimmune uveitis reported that Il-4 andIl-10 are both required for induction of OT (112), under-lining that the roles of Il-4 and Il-10 are still unclear. Onthe other hand, it has recently been proposed that a newsubset of CD4» T cells, termed Th3 cells, primarily secret-ing TGF-b, may down-regulate properties for Th1 andother immune cells (reviewed in (113)). The importance ofTGF-b is further supported by the prevention of bystandersuppressive effects with anti-TGF-b antibodies (114).

The absence of both active suppression and reactivelymphocytes in ×i×o, in experimental models of OT, wasthought to result from clonal deletion or anergy. Althoughclonal deletion of T cells has been demonstrated in severalstudies (reviewed in (105)), experimental conditions, i.e.ingestion of very high doses of antigen, suggest that suchdeletion probably does not occur after induction of OT in×i×o. In contrast, anergy is considered to be an importantOT mechanism which may preferentially induce unrespon-siveness of Th1 functions. It seems likely that anergyre� ects aberrant presentation of a fed antigen by APC,leading to an absence of Il-2 secretion and T-cellactivation.

Important questions about the way antigens are pro-cessed and presented remain, especially as to the locationof antigen presentation and the APC involved. It is nowproposed that APC may play a crucial role in the induc-tion of OT, especially due to the presentation of theantigen to T cells associated with a failure of appropriatecostimulation (reviewed in (97, 21). Enterocytes expressinglow levels of particular MHC class II molecules with noinvariant chain and no costimulatory molecules, such asB7 or ICAM-1, had � rstly been considered as potentialtolerogenic APC (64, reviewed in (115)). Recent interestingstudies tend to point towards a central role for dendriticcells (DC), thus opening new � elds of investigation (re-viewed in (97)). Expanding mature DC in ×i×o with thegrowth factor Flt3 ligand results in enhanced induction ofOT in treated mice fed with low doses of soluble antigenwhich are inef� cient in control mice (116).

According to the concept that the mechanisms involvedin OT may depend on different doses of fed antigen(reviewed in (104)), a new scheme of intestinal induction ofOT proposes that patterns of tolerance may re� ect the


amounts of peptide:MHC complexes presented to T cellsin the absence of costimulation. More recently, it has beensuggested that CTLA-4, the high-af� nity receptor for B7molecules on T-cells, plays a crucial role in the inductionof high-dose OT (117).

In parallel, other experimental studies have suggestedthat the generation of a tolerogenic form of antigen thatpasses through the gut may also play an important role inthe generation of OT. It has been reported that serumfrom OVA-fed mice, transferred into naive recipients 1 hafter the feeding, induces suppression of systemic DTHresponses, whereas serum from mice injected systemicallywith equivalent doses of OVA has no effect (118, 119).However, the molecular nature of the tolerogen and howthe intestine generates such material are still unclear.Moreover, the importance of intestinal enzymatic digestionof the antigen in induction of OT remains controversial,given that Louis et al. (120) have paradoxically reportedthat speci� c systemic cellular hyporesponsiveness is alsoinduced by one rectocolonic administration of 25 mgOVA. On the other hand, peripheral tolerance in therespiratory tract, especially suppression of the speci� c IgEresponse, has also been demonstrated in response to in-haled antigens (121).

Factors affecting oral tolerance. Although systemic sup-pression after antigen feeding is a general phenomenon, itis also possible to induce humoral and:or cellular systemicimmune responses through the oral route and a number offactors have been reported to affect OT establishment.Moreover, several recent studies provide evidence thatfactors affecting OT and mechanisms governing OT areinterrelated. These observations may have important im-plications for better understanding the development andtreatment of hypersensitivities or autoimmune diseases.Only the major factors which in� uence OT are presentedhere and we particularly highlight the role of the indige-nous gut � ora.Antigenic factors: nature of antigen. Although OT canprobably be induced to all thymus-dependent soluble anti-gens, it cannot be induced to thymus-independent ones(reviewed in (93)). It also appears that particulate antigens,antigens associated with replicating bacteria or immunestimulating complexes (ISCOMs) tend to induce activeimmunity rather than tolerance (122, 123). It is hypothe-sized that such a difference may be related to their presen-tation in the gut and their preferential uptake by M cellsoverlying the Peyer’s patches.

On the other hand, OT cannot be induced to enzymati-caly-, chemically- or heat-denatured soluble antigens (124–126). It has been proposed that this may be related to themodi� cation of the intestinal processing and absorption ofthe antigen and the resulting absence of any tolerogenicform of the antigen (118, 127).Antigenic factors: dose of antigen. The dose of antigenrequired for establishment of OT depends on the antigen

used and on the systemic immune response studied. Forexample, studies on mice have shown that a single highdose (10–20 mg) of OVA induces suppression of bothhumoral and cell-mediated immune responses, whereassmaller doses can either induce suppression or enhancesystemic responses (128). Moreover, suppression of thehumoral response depends on the isotype considered. Oneoral dose of 1–5 mg OVA induces IgE unresponsivenesswhile the same doses leave IgG response unaffected. Incontrast, repeated ingestion of a comparable dose (5½1mg) suppress both IgE and IgG antibody responses (101).However, we have reported that under the latter conditionboth IgE and IgG antibody suppression are of shortduration (129), suggesting for the � rst time that factorswhich do not prevent the establishment of OT can disturbits maintenance.

Cell-mediated immune responses are easily tolerated as100 mg of OVA is suf� cient to suppress systemic DTHreactions in mice. Nevertheless, lower amounts of OVA(10–50 mg) induce priming for DTH response (128). Thisresult suggests that small doses of dietary proteins, such asb-lactoglobulin or a-lactalbumin present in cow’s-milk-based formula, and egg- or cow’s-milk-proteins present inbreast milk, may predispose susceptible neonates to food-hypersensitivity reactions.

Recent studies have shown that low versus high doses ofantigen feeding in� uence the mechanisms involved in OTinduction. Whereas low doses (5½1 mg) of the antigenhen egg lysozyme or autoantigen myelin basic protein(MBP) induce cytokine-mediated active suppression, highdoses (20 mg) of the antigen induce anergy. These twomechanisms are not mutually exclusive (130). This resultcon� rms early studies in the experimental model of au-toimmune uveoretinitis (131). As previously reported, theB7:CTLA-4 interaction at the intestinal level may be cru-cial in OT induction and the high dose of antigen maydirectly in� uence the costimulatory events, thus givingincreasing importance to environmental parameters in thegeneral process of OT.Host factors: genetic background of host. Early studiessuggested the in� uence of the genetic background in thedegree of suppression of both DTH and IgG responsesafter antigen feeding (132). It has been related to H-2haplotype, H-2d mice being particularly sensitive to toler-ance induction (133). However, no direct proof has as yetbeen obtained.Host factors: in� uence of host age. The age at whichantigens are � rst encountered by the host largely in-� uences OT induction. Thus, mice do not become tolerantif fed OVA before 7 days of age and are even sensitized byearly feeding or prenatal treatment (134, 135). Similarobservations have been reported in the experimentalmodels of autoimmune encephalomyelitis. Feeding MBPto rats under 4 weeks of age results in priming insteadof tolerization (136). In contrast, this is not ob-served in guinea pigs, which are more mature at


birth (137, 138). These results suggest that intestinal mu-cosal immaturity may prevent the induction of systemicunresponsiveness. The defect in oral tolerization can bepartially restored by transfer of mature adult splenocytes(139), suggesting that a more complex regulatory systemmay be involved. Interestingly, it has been thought thatperipheral antigenic challenge during neonatal life repre-sents a tolerogenic rather than an immunogenic event (37),highlighting differences between the peripheral and mu-cosal immune systems.

In humans, oral tolerization has also been shown to beage-dependent, despite a more mature intestinal immunesystem at birth than that found in rodents. Neonates aremore vulnerable to food hypersensitivities, especially tocow’s milk proteins. In contrast to children receiving ca-sein hydrolysate formula, it has been shown that bothcellular and IgG antibody responses to cow’s-milk-derivedb-lactoglobulin are signi� cantly increased in infants receiv-ing cow’s milk within the � rst months of life comparedwith the infants receiving it only after the age of 9 months(140). Nonetheless, it appears that these differences do notpersist after 1 year of age. It is also noteworthy thatsystemic antibodies to food proteins are present in mostnormal individuals and do not correlate with any foodhypersensitivity.

A de� ciency in OT induction is also noticed at weaning(134). However, it is still not known if it is related to thefunctioning of GALT or to gastrointestinal changes takingplace during this period, especially modi� cation of the gutmicro� ora. Nevertheless, establishment of OT is crucialbecause of the numerous new dietary antigens encounteredat weaning.

In parallel to host immaturity in the neonatal period, wehave shown that aging also in� uences OT, especially itslong-term maintenance. Comparing 20-month-old and 2-month-old young adult CV mice fed with a single tolero-genic dose of 20 mg OVA, we observed that both IgG andIgE antibody suppression are induced but do not persist inold mice in contrast to young adult mice (129). Thus, aspreviously reported for repeated ingestion of small dosesof antigen, these results con� rm that factors which allowinduction of OT can also prevent its maintenance, suggest-ing that different mechanisms may be involved in theinduction and maintenance of the OT process.

The age at which an antigen is introduced at the mu-cosal level has also been reported to in� uence OT mecha-nisms. Although both young (4 weeks of age) and adult(12 weeks of age) rats fed with OVA had reduced cell-me-diated immune response, active suppression and bystandertolerance are shown in adult rats, whereas anergy is preva-lent in young ones (141). No clear explanation has beenproposed for this dichotomy, and it may be suspected thatcomplex events occur at the mucosal level with the gutmicro� ora acting as a crucial parameter.

Host factors: intestinal permeability. Gut integrity andpermeability of the intestinal epithelial layer seem to bedeterminant parameters in OT induction. In� ammatoryreactions of the small intestinal mucosa and intestinalepithelial lesions are generally associated with increasedintestinal permeability, which may result in abrogation ofOT. Indeed, experimental studies in mice show that bothgraft-versus-host reaction (GvHR) and indomethacin-me-diated increased intestinal permeability and induced in� -ammatory lesions of the intestine are associated with anabrogation of OT to OVA (142, 143). Recently, by testinggut handling and processing of gliadin in mice withGvHR, Troncone et al. (144) have reported that serumcontaining gut-absorbed gliadin fails to suppress systemiccellular immune responses when transferred intraperi-toneally into naive recipients. However, comparable serumlevels of gliadin between GvHR mice and normal oneshave been detected (144). This data � rst suggested theimportance of an intact gut epithelium in generating atolerogenic serum factor. Secondly, one may wonderwhether increased intestinal permeability might be thecause rather than the consequence of OT abrogation.

The hypothesis that the alteration of the intestinalmucosal epithelium and increased intestinal permeabilityto food proteins is probably not the primary causeof allergy but the secondary effect of an abnormal im-munological response to food proteins is supportedby several studies. Heyman et al. (145) reported thatthe in� ammatory cytokine tumour necrosis factor (TNF)-ais abnormally secreted by peripheral blood mononuclearcells (PBMC) taken from infants suffering cow’s milkallergy. Interestingly, it was further established that thisabnormal secretion results from a reduced threshold forPBMC immune reactivity to intact (i.e. non-intestinallyprocessed) cow’s milk proteins (146). TNF-a directly altersthe intestinal epithelial barrier permeability (145). In an-other clinical study, infants with cow’s milk allergy dis-played reduced permeability to b-lactoglobulin at normalvalues after cow’s milk had been withdrawn from the diet(i.e. during the symptom-free period (147)). Taken to-gether, this data would tend to show that increased perme-ability is not constitutive, and the increase in proteintransport seems to be a consequence rather than a cause offood allergy due to in� ammatory reactions at the gut level.

Role of resident intestinal micro� ora. Although indige-nous gut micro� ora has been overlooked for a long time asan environmental parameter in� uencing the GALT func-tions, increasing evidence now supports the idea that it-could be an important environmental factor in modulatingOT to dietary proteins. We will consider the role of gutmicro� ora on both the induction and the maintenance ofOT.In� uence on induction of OT. The in� uence of gram-neg-ative bacteria on OT induction was � rst suggested byWannemuehler et al. (148). Although OT, estimated by theantigen-speci� c IgM, IgG and IgA antibody suppression,


cannot be induced in GF mice fed with SRBC, it can berestored in GF mice fed with 10–100 mg of LPS the daysbefore they are given SRBC. Moreover, in experimentsusing CV mice, it has been shown that LPS given orallywith myelin basic protein may enhance tolerance develop-ment in experimental autoimmune encephalomyelitis (149).However, other studies demonstrated that the indigenousgut � ora is not the basic requirement for OT induction interms of humoral suppression, as it was possible to induceIgG and IgE antibody unresponsiveness in GF mice fedonce with 20 mg of OVA (150, 129). The differencesobserved in these two models may be related to the natureof the antigen, i.e. particulate versus soluble. Nevertheless,Sudo et al. (151) recently brought to light the role of agram-positive bacterium, Bi� dobacterium, in the suppres-sion of humoral antibody responses. Bi� dobacterium canimprove suppression of Th2-mediated immune responseduring the OT process. Comparing speci� c-pathogen free(SPF) mice, GF mice and gnotobiotic mice associated withBi� dobacterium and fed with tolerogenic doses of OVA,the authors show that OVA-speci� c IgE and IgG1 anti-body levels, and Il-4 synthesis, are signi� cantly reduced inSPF and Bi� dobacterium-associated mice compared withGF counterparts. However, Bi� dobacterium exerts its roleonly when associated in mice at the neonatal stage; it doesnot produce the same effect at an older age. The mecha-nisms involved are not clearly identi� ed and further inves-tigations are needed to elucidate the discrepancy betweenthese results. It can, however, be conjectured that intestinalmicro� ora, especially bacterial species present in the diges-tive tract as of the postnatal period, may exert a funda-mental role on the development of normal GALTfunctions, such as OT induction.In� uence on maintenance of OT. Further observations sup-port the idea that indigenous gut micro� ora also stronglyin� uences the long-term duration of OT, humoral unre-sponsiveness being short-lived in GF mice. We have shownin CV mice that both IgG and IgE antibody unresponsive-ness last up to 3 months after a tolerogenic 20 mg OVA-feeding, whereas in the GF mice antibodyunresponsiveness is short-lived, IgG antibody suppressionlasting no more than 20 days (129). Long-term IgE anti-body suppression is not altered in GF mice, con� rmingthat IgE response may be more readily suppressed thanother isotype responses (129).

More recently, experiments performed on gnotobioticmice inoculated with known bacteria at the adult stage,have suggested that gram-negative bacteria are involved inmaintenance of OT to OVA. Gnotobiotic mice harboringE. coli or Bacteroides, showed a long-term tolerance com-parable to that observed in CV mice. In contrast, miceharboring Gram-positive bacteria, such as Bi� dobacteriumor non-enterotoxigenic Clostridium dif� cile, showed toler-ance similar to that observed in GF mice (74).

Our preliminary results indicated that LPS could be

involved in the effect of gram-negative bacteria on OTmaintenance. The in� uence of LPS on antigen presentingcells, such as dendritic cells (DC), may represent a new� eld of investigation. As previously reported, recent stud-ies suggest that LPS affects splenic DC populations andmay regulate DC functions (67, 68), which may haveimportant consequences if con� rmed with mucosal DC. Itmay thus be hypothesized that intestinal in� ammationrelated to bacterial colonization of the gut may stimulateTNF-a secretion and increase DC biological functionswhich may be involved in OT process.

On the other hand, it may be suggested that the in� u-ence of the indigenous gut micro� ora on intestinal perme-ability could be responsible for the effect on OTmaintenance. Indeed, one study in suckling mice, assayingpatterns of protein absorption in the neonatal period, hasreported four-fold reduced intact horseradish peroxidase(HRP) transport in GF mice compared with the CVcounterparts (152). It can be hypothesized that the declinein permeability in GF mice may result from the absence ofa basic physiological in� ammation normally induced bythe indigenous gut micro� ora, which could play a crucialrole in the complete establishment of the OT process.Thus, whereas infancy is generally considered as a periodduring which the gastrointestinal barrier is immature, re-sulting in increased intestinal permeability to macro-molecules, such a permeability may be necessary fornormal OT induction. However, this possibility must becon� rmed in adult mice.Protecti×e role of resident intestinal micro� ora. Clinicalobservations show that food hypersensitivities are mostcommon in human infants, particularly at the time ofweaning. At weaning, the switch of diet generally alterscolonization resistance and predisposes the child to entericinfections and diarrhea. Studies on mice have shown thatboth cholera toxin (CT) and E. coli heat-labile enterotoxin(LT), secreted by Vibrio cholerae and enterotoxigenic E.coli, respectively, abrogate OT induction, in terms of sys-temic humoral immunity, to an antigen given simulta-neously by the oral route (153–156). The gut micro� ora isnot directly involved in the toxin-mediated abrogativeprocess, as shown by the fact that it exists in both CV andGF mice (157). Therefore, the transient presence of entero-toxins in young children, especially at weaning, may inter-fere with the OT process, resulting in foodhypersensitivities. It is interesting to note that not allenterotoxins affect OT. Neither S. aureus enterotoxin Bnor C. perfringens type A enterotoxin prevent inductionand maintenance of OT to OVA in mice (158).

The relationships between gut micro� ora, the in� uenceof enterotoxins and mechanisms governing OT are still notunderstood. The in� uence of enteropathogens, responsiblefor acute diarrhea, on changes in intestinal permeability tointact macromolecules has been studied. However, the


correlation between diarrhea and changes in intestinalpermeability is still controversial. LT has been reported toincrease intestinal permeability to macromolecules in CVmice (159). In contrast, in ×itro studies have shown thatCT does not alter permeability to HRP (160). Therefore, itis not known whether altered permeability can be impli-cated in the toxin-mediated abrogation of OT.

We recently demonstrated another fundamental role ofthe indigenous gut micro� ora on the OT process by show-ing that its presence allows the recovery of suppressivemechanisms after the transient CT- or LT-mediated break-down of OT (157). Kinetics of the anti-OVA IgG antibodyresponses show that a hyporesponsive state occurs in CVmice fed with CT- or LT-plus OVA as compared with thecontrol mice, but that it does not occur in GF mice. Thus,in children, it may be important to preserve the normal gutmicro� ora, as it could, in time, play a crucial role, such asimproving recovery of tolerance.

Investigating the protective effect of the indigenous gutmicro� ora further in ×i×o, we observed that it also de-creases susceptibility of the individual to the LT-mediatedeffect on OT induction (157). However, this was onlyeffective when the micro� ora was associated in neonates,highlighting the importance of its natural establishmentduring the neonatal period (Gaboriau-Routhiau,manuscript in preparation). As previously suggested bySudo et al. (151), it may be supposed that the gut mi-cro� ora is important in contributing to the functioning ofintestinal immunity, as well as to protecting it and that theneonatal period is crucial for the normal establishment anddevelopment of GALT immune functions. Hence, we be-lieve that in the neonatal and weaning periods it is impor-tant to preserve the intestinal micro� ora. Its alteration, e.g.during antibiotic treatment, may impair the intestinal bar-rier to intact proteins which may be related to increasedfood hypersensitivities in children.

In conclusion, although the in� uence of the indigenousgut micro� ora on induction and maintenance of OT andmechanisms involved in this complex phenomenon are notyet fully understood, indications exist that the gut mi-cro� ora plays a critical role. We have reported that theenvironmental parameters, such as the dose of antigen, theage of the host and the composition of the gut micro� ora,can in� uence induction and:or maintenance of OT, sug-gesting that OT may be divided into two stages character-ized by their sensitivity to environmental parameters. Wewould, therefore, hypothesize that OT is a dynamic pro-cess during which sequentially suppressive mechanisms,leading to the induction and the long-term persistence ofsuppression, might be involved. It would now be interest-ing to identify the mechanisms which correlate with short-and long-term persistence of tolerance induced by proteinfeeding.

NEONATAL PERIOD: A CRITICAL STAGE IN THEPREVENTION OF SHORT AND / OR LONG-TERMPATHOLOGIES?

There is increasing evidence that the neonatal period couldbe important in the etiology of some pathologies, such asallergies, coeliac disease, diabetes and in� ammatory boweldiseases (IBD) developing in infancy and:or later. Accord-ing to the importance of the neonatal period for theinteractions between GALT, the digestive � ora and nutri-tion, questions can be asked about environmental factorswhich can affect these interactions, especially early dietarydiversi� cation and antibiotherapies. Through the examplesgiven here we would like to show that the resident diges-tive � ora has a fundamental role in maintaining health andpreventing disease and that it must be considered as a truepart of the body.

Neonatal period and immaturity of GALT

The high prevalence of dietary hypersensitivities and en-teric infections in the neonate is believed to be due to theimmaturity of GALT functions. Hypersensitivities to cow’smilk proteins and egg albumin occur in approximately3–10% of infants during the � rst 2 years of life in contrastto 0.001 –0.5% of adults (reviewed in (102, 161–163)).Food hypersensitivities are mainly characterized by anelevated IgE response (allergy) with respiratory and:orskin symptoms (atopic dermatitis) or by an intestinalcellular immune response with digestive symptoms (di-arrheas, abdominal pain). After 2 years of age, mostinfants no longer develop food allergies. Those who con-tinue to develop allergies do so mainly towards � sh andegg proteins. However, development of food allergies dur-ing early life seems to be correlated with a high risk ofdeveloping allergies to inhalant allergens in adulthood(164).

The paradigm of Th1:Th2 subclasses of CD4» T cells,well established in murine models, has lent insight into oneof the mechanisms involved in allergies. The balance be-tween Th1:Th2 cytokines is considered to be critical forIgE production, even if regulatory mechanisms of allergiesare much more complex than the sole Th1:Th2 balancesince high IgE levels can be found in healthy children.CD4» T cells expressing the Th2 cytokine pro� le releaseIL-4 and IL-5 which are potent inducers for IgE andrecruitment of eosinophils, respectively. It is known thatthere is a mutual opposition between Th1 and Th2 cells,Th1 cytokines, especially IFN-g, down-regulating the Th2function. Neonates generally display polarized ex-pression of Th2-like cytokines as fetal development occursin a Th2 cytokine pro� le to avoid fetal rejection (165).Thus, mechanisms acting on the switch of Th2 to Th1 afterbirth must be induced. IL-12, produced by varioustypes of APC, strongly activates Th1 cells thus pro-ducing IFN-g (166). Arulanandam et al. (167) haveshown that the neonates have a reduced expres-


sion of IL-12 in the spleen and that the administration ofrecombinant Il-12 redirects the new-born immune systemtowards a Th1-type cytokine pro� le with IFN-g produc-tion. An important question is whether the digestive � orais capable of polarizing the Th1:Th2 balance. This factcould be of importance in allergic diseases. Recent excitingexperiments have shown that conventionalization of GFmice by the resident intestinal � ora of CV mice enhancedIL-12 production in the spleen (168). These � ndings sug-gest that bacterial colonization after birth could be effec-tive in polarizing the Th2 pro� le towards a Th1 pro� le. Itcould be of particular interest during the neonatal periodsince, as recently described, IL-12 cannot reverse Th2response into Th1 in adult human atopic patients in con-trast to non-atopic patients (169). In fact, this data ad-dresses the question of the importance of the roles of IL-12and IFN-g during the neonatal period to the reversibilityof Th2 to Th1 cells later in life and, consequently theimportance of the � rst bacteria colonizing the intestine inthis process.

Neonatal period and nutrition

GALT is the primary target of all the types of dietaryconstituents. A majority of human studies conclude thatexclusive breast-feeding for at least 1 month prevents thedevelopment of allergies to cow’s milk proteins and otherallergic manifestations up to 3 years of age (reviewed in(170)). However, the long-term bene� cial effect of breastfeeding and preventive nutrition of new-borns in avoidingdietary hypersensitivities remains controversial. Humanmilk contains components enhancing maturation of theintestinal mucosa of new-born infants (reviewed in (163,20)). It also promotes the Bi� dobacterium establishment inthe intestine. As described above from experimental ani-mal models, Bi� dobacterium could enhance the anti-ro-tavirus sIgA response in babies. On the other hand, fromthe results of Sudo et al. (151), the very early presence ofBi� dobacterium in the intestine may down-regulate thedevelopment of IgE antibody responses and the resultingsusceptibility to food allergic diseases. These data mayunderline the importance of Bi� dobacterium establishmentin the gut during the post-natal period, highlighting breast-milk feeding or use of Bi� dobacterium as probiotic, inatopic neonates.

The moment new food proteins can be introduced in thebaby’s diet is often questioned and exclusively breast-feed-ing during the � rst 5–6 months of life is now recom-mended for infants predisposed to allergy. Exclusivelybreast-feeding for too long can delay the development ofnatural sIgA production by preventing changes in intesti-nal � ora. In contrast, too early an introduction of newdietary antigens can lead to short and:or long term detri-ment including allergies or coeliac disease. Coeliac diseaseor gluten-sensitive enteropathy affects susceptible infantsand adults who develop in� ammatory intestinal symptoms

after gluten ingestion. From interesting experimental stud-ies, it has been demonstrated in rats that the early intro-duction of gliadin is responsible for the coeliac diseaseetiology but, in this case, the presence of the intestinal� ora has no in� uence (55).

In autoimmune diabetes, it has been suggested thatcross-reactions between caseins and unknown pancreaticself-antigens in early life could be the cause of the disease(reviewed in (171)). Cross-reactions between intestinal bac-teria and self-epitopes (mimicry epitopes) have also beensuspected in the development and prevention of the disease(172).

Neonatal period and intestinal � ora

Over the last 40 years, there has been an increase in theprevalence of allergic diseases in western industrializedcountries (reviewed in (163)). Environmental changes mustplay a role since genetic factors have not changed. Amonga long list of factors, disturbances of the intestinal mi-cro� ora due to early dietary diversi� cation and:or antibio-therapies may have important short- and long-termconsequences on infants.

In 1989, Strachan (173) stated the hypothesis that, inwestern countries, the decrease of natural infections ininfants could be a cause of the increase in the prevalenceof allergic diseases. This fact could be due to a disruptionin driving the pro� le of Th2 to Th1. However, Ruuska(174) reported that infants who had a signi� cantly greaternumber of episodes of acute diarrhea than those who didnot, developed food allergy. These contradictory resultscould arise from the different types of food hypersensitiv-ity studied, i.e. allergy versus DTH, where Th2 or Th1unbalanced polarization is involved, respectively. In aninteresting critical review, Wold (175) discusses the Stra-chan’s ‘hygiene hypothesis’. More than infections, Woldargues that the hygienic lifestyle can lead to an alterationof the normal intestinal colonization pattern in infancy,thus disturbing the OT process. Indeed, reduced intestinalcolonization of E. coli in Swedish neonates is reported. Asdescribed above, the presence of intestinal � ora in early lifeplays important inductive and protective roles on the OTprocess, especially with the importance of E. coli.We suggest that the use of antibiotics as current treatmentduring infancy is more responsible for strong modi� cat-ions or destruction of the intestinal � ora than arepostnatal hygiene habits developed in western countriesand that such antibiotic use could have harmful conse-quences on GALT. Establishment of intestinal � ora inthe sterile intestine of a baby at birth and profoundchanges taking place at weaning time can be considered asphysiological stresses leading to in� ammatory cytokinesecretions (176) which might be important in inducingdevelopment and functioning of GALT. Thus, frequentantibiotherapies with the result of successive de-struction, colonization and modi� cation of the intestinal


� ora equilibrium can lead to disturbances inthe regulatory mechanisms with harmful consequences toGALT functions in preventing immunopathologicalreactions.

From the list of questions arising from the etiology ofIBD, the role of the neonatal period in the establishmentof tolerance to its own intestinal � ora has been men-tioned. Recently, Duchmann et al. (2) provided evidencethat, under healthy conditions, GALT does not developimmune responses towards resident bacteria and a break-down of tolerance could be the cause of immune reac-tions to the resident intestinal � ora leading to IBD. Inexperimental studies, Karlsson et al. (177) have shownthat induction of OT to a transgenic E. coli producingOVA was possible if the strain colonized the intestine justafter birth, whereas colonization at adult stage primedmice to the bacterial antigens. In other studies usingspontaneously colitic C3H:HeBr adult mice, a surprisinghumoral immune reactivity in serum has been found di-rected predominantly towards antigens of facultativeanaerobes, which were in low number in faces (178).Facultative anaerobes are present in high levels duringthe neonatal period, and then their numbers decrease inadult fecal � ora. This antibody response may result froma disordered immune response starting very early in lifeand leading to pathology later on.

CONCLUSIONSThe intestinal � ora exerts a strong effect both on GALTactivation and development and on the regulatory mecha-nisms involved in the maintenance of the steady-state atthe intestinal level. This effect is probably different ac-cording to the bacterial equilibrium which is present inthe different parts of the intestine. A delicate balanceexists in the intestine between the bacterial � ora and theimmune status of the host. Aberrations in the dynamicbalance, either at the microbial level (e.g. antibiotic thera-pies) or at the control level of GALT functions (break-down of OT) may have harmful consequences. Thus, alot of questions have to be answered to maintain humanhealth. For instance, the reversibility of the effects of theintestinal � ora on GALT is poorly understood. Thisknowledge is important as regards the consequences exist-ing after long-term antibiotherapies, elemental enteraldiet, or total parenteral nutrition in humans. Experimen-tal studies have brought to light changes that occur inGALT after such diets and their relationship to intestinalbacterial modi� cations (179). In neonates, many observa-tions support the notion that most mucosal immune cellsare competent even before birth, but they need to un-dergo an activation process initiated by environmentalsignals. According to Ridge et al. (37), orientation totolerance or activation to an antigen is not determined bythe self or non-self origin of the antigen but rather by theconditions under which it is introduced. Mature virgin T

cells can be activated, tolerated or switched to Th1 orTh2 responses according to the dose of antigen, the typeof adjuvant and the type of APC. From this point, itfollows that the role of the intestinal bacterial coloniza-tion could be important in controlling the type of im-mune response. Consequently, particular attention has tobe focused on the intestinal � ora development during theneonatal period when the induction of lifelong regulatoryimmune mechanisms could be established.

Few attempts have been made to elucidate the mecha-nisms of intestinal bacteria in modulating the immunesystem, especially at the intestinal level. Small amounts ofLPS and PG derived from the intestinal � ora may beindispensable to the development, maintenance and goodfunctioning of the immune system. However, there is asyet no information regarding the exact role of these bac-terial components when they issue from digestive � ora.Moreover, living bacteria secrete metabolites resultingfrom intestinal fermentation which can have im-munomodulating effects (180; reviewed in (13)). Particularinterest must be paid to the role of the intestinal � ora onthe APC due to the fundamental role they play in innateand adaptive immunity (181).

In addition to its effects on GALT, the intestinal � orain� uences peripheral immunity in a protective manner.Intestinal microbial colonization is responsible for theenhancement of natural IgG1, IgG2a and IgG2b Abs inserum (182), which have been shown to strongly in� uencethe B-cell repertoire (183). The consequences on humanhealth are important as a vast majority of natural IgGdisplay reactivity towards self antigens and could play arole in the regulation of peripheral tolerance (reviewed in(184)). Other experimental studies have demonstrated thatthe presence of intestinal � ora protects experimental ani-mals from pathologies such as arthritis (185) and anemia(186). As well as causing detrimental effects at the intesti-nal level, antibiotherapies may have an impact at theperipheral level. Recent study has shown that B-cell activ-ity measured in ×itro is depressed in infants receivingantibiotic treatments (187).

It is widely recognized that the features of mucosalimmunity described here have mainly been discoveredand investigated in rodents. We need to know to whatextent these phenomena and their regulation are opera-tive in man, with the aim of applications for humanhealth (188). It is necessary to develop appropriatebiomarkers for direct studies in humans, and gnotobio-logical experiments, as described here, are convenient forsuch investigations.

Dedication

We would like to dedicate this chapter to Anne Fergus-son who died on December 1998, at the age of 57 years.An important clinical researcher in the � eld of intestinalimmunity, she was above all else a very nice woman.


ACKNOWLEDGEMENTS

We wish to thank Donald White for the English correction of thechapter.

REFERENCES

1. Luckey TD, Floch MH. Introduction to intestinal microecol-ogy. Am. J. Clin. Nutr. 1972; 25: 1291 –5.

2. Duchmann R, Kaiser I, Hermann E, Mayet W, Ewe K,Meyer zum Buschenfelde KH. Tolerance exists towards resi-dent intestinal � ora but is broken in active in� ammatorybowel disease. Clin. Exp. Immunol. 1995; 102: 448–55.

3. Raibaud P. Factors controlling the bacterial colonization ofthe neonatal intestine. In: Hanson LA, ed. Biology of Hu-man Milk. New York: Raven Press, 1988: 205–19.

4. Hudault S. Microbial colonisation of the intestine of new-born. In: Bindels JG, Goedhart AC, Visser HKA, eds.Recent Developments in Infant Nutrition. Dordrecht:Kluwer Academic Publishers, 1996: 307–17.

5. Schaedler RW, Dubos R, Costello R. The development ofthe bacterial � ora in the gastrointestinal tract of mice. J.Exp. Med. 1965; 122: 59–66.

6. Moreau MC, Raibaud P, Muller MC. Relation entre ledeveloppement du systeme immunitaire intestinal a IgA etl’etablissement de la � ore microbienne dans le tube digestifdu souriceau holoxenique. Ann. Immunol. (Inst. Pasteur)1982; 133D: 29–39.

7. Moreau MC, Corthier G, Muller MC, Dubos F, Raibaud P.Relationships between rotavirus diarrhea and intestinal mi-cro� ora establishment in conventional and gnotobiotic mice.J. Clin. Microbiol. 1986; 23: 863–8.

8. Ducluzeau R. Role of experimental microbial ecology ingastroenterology. In: Bergogne-Berezin E, ed. MicrobialEcology and Intestinal Secretions. Paris: Springer, 1989:7–26.

9. King CE, Toskes PP. Small intestine bacterial overgrowth.Gastroenterology 1979; 76: 1035–55.

10. Moore WEC, Holdeman LV. Human fecal � ora: the normal� ora of 20 Japanese-Hawaiians. Appl. Microbiol. 1974; 27:961–79.

11. Finegold SM, Sutter VL, Mathisen GE. Normal indigenousintestinal � ora. In: Hentges DJ, ed. Human Intestinal Mi-cro� ora in Health and Disease. New York: Academic Press,1983: 3–31.

12. Savage DC. Microbial ecology of the gastrointestinal tract.Annu. Rev. Microbiol. 1977; 31: 107–33.

13. Salminen S, Bouley C, Boutron-Ruault MC, Cummings JH,Franck A, Gibson GR, Isolauri E, Moreau MC, RoberfroidM, Rowland I. Functional food science and gastrointestinalphysiology and function. Br. J. Nutr. 1998; 80 (Suppl. 1):S147–71.

14. MacCartney AL, Wenzhi W, Tannock GW. Molecular anal-ysis of the composition of the bi� dobacterial and lactobacil-lus micro� ora of humans. Appl. Environ. Microbiol. 1996;62: 4608–13.

15. Zoetendal EG, Akkermans ADL, De Vos WM. Temperaturegradient gel electrophoresis analysis of 16S rRNA fromhuman fecal samples reveals stable and host-speci� c commu-nities of active bacteria. Appl. Environ. Microbiol. 1998; 64:3854–9.

16. Holdeman LV, Good IJ, Moore WEC. Human fecal � ora:Variation in bacterial composition within individuals and apossible effect of emotional stress. Appl. Environ. Microbiol.1976; 31: 359–75.

17. Borriello SP. Clostridial disease of the gut. Clin. Infect. Dis.1995; 20 (Suppl 2): S242–50.

18. Ouwehand AC, Isolauri E, Kirjavainen PV, Salminen SJ.Adhesion of four Bi� dobacterium strains to human intestinalmucus from subjects in different age groups. FEMS Micro-biol. Lett. 1999; 172: 61–4.

19. Van Der Waaij D. Mechanisms involved in the developmentof the intestinal micro� ora in relation to the host organism:consequences for colonization resistance. In: Hormaeche CE,Penn CW, Smyth CJ, eds. Molecular Biology of BacterialInfection: Current Status and Future Perspectives. Cam-bridge: University Press, 1993: 1–12.

20. Husband AJ, Gleeson M. Ontogeny of mucosal immunity.Environmental and behavioral in� uences. Brain Behav. Im-mun. 1996; 10: 188–204.

21. Brandtzaeg P. Development and basic mechanisms of hu-man gut immunity. Nutr. Rev. 1998; 56: S5–S18.

22. MacDonald TT. Development of mucosal immune functionin man: potential for G I disease states. Acta Pediatr.Japonica 1994; 36: 532–6.

23. Crabbe P, Nash D, Bazin H, Eyssen H, Heremans JF.Immunohistochemical observations on lymphoid tissuesfrom conventional and germ-free mice. Lab. Invest. 1970; 22:448–57.

24. Brandtzaeg P. Molecular and cellular aspects of the secre-tory immunolobulin system. APMIS 1995; 103: 1–19.

25. Underdown B, Mestecky J. Mucosal immunoglobulins. In:Ogra A, et al., eds. Handbook of Mucosal Immunology.Orlando, FL: Academic Press, 1994: 79–97.

26. Kette K, Baklien K, Bakken A, Kral JG, Fausa O,Brandtzaeg P. Intestinal B-cell isotype response in relation tolocal bacterial load: evidence for immunoglobulin A subclassadaptation. Gastroenterology 1995; 109: 819–25.

27. Lionetti P, Breese E, Spencer J. Activation of V-b3» T cellsand tissue damage in human small intestine induced by thebacterial superantigen, Staphylococcus aureus enterotoxin B.Eur. J. Immunol. 1993; 23: 664–8.

28. Halstensen TS, Scott H, Brandtzeag P. Human CD8»intraepithelial T lymphocytes are mainly CD45RA-RB»and show increased co-expression of CD45R0 in celiac dis-ease. Eur. J. Immunol. 1990; 20: 1825–9.

29. Dohan A, MacDonald TT, Spencer J. The ontogeny ofadhesion molecule expression in the human intestine. Clin.Exp. Immunol. 1993; 91: 532–7.

30. Cerf-Bensussan N, Guy-Grand D. Intestinal intraepitheliallymphocytes. Gastroenterol. Clin. North Am. 1991; 20: 549–76.

31. Guy-Grand D, Cerf-Bensussan N, Malissen B, Malassis-Seris M, Briottet C, Vassali P. Two gut intraepithelialCD8» lymphocyte populations with different T cell recep-tors: a role for the gut epithelium in T cell differentiation. J.Exp. Med. 1991; 173: 471–81.

32. MacDonald TT, Weinel A, Spencer J. HLA-DR expressionin human fetal intestinal epithelium. Gut 1988; 29: 1342–8.

33. Rognum TO, Stoltenberg L, Vege A, Brandtzaeg P. Devel-opment of intestinal mucosal immunity in fetal life and in� rst postnatal months. Pediatr. Res. 1992; 32: 145–9.

34. Hughes A, Bloch KJ, Bhan AK, Gillen D, Giovino VC,Harmatz PR. Expression of MHC class II (Ia) antigen by theneonatal enterocytes: the effect of treatment with interferon-gamma. Immunology 1991; 72: 491–6.

35. MacWilliam AS, Holt PG. Mucosal dendritic cells in therespiratory tract. Mucosal Immunol. Update 1997; 5: 21–5.


36. Lu CY, Calamai EG, Unanue ER. A defect in the antigen-presenting function of macrophages from neonatal mice.Nature 1979; 282: 327–9.

37. Ridge JP, Fuchs EJ, Matzinger P. Neonatal tolerance revis-ited: turning on newborn T cells with dendritic cells. Science1996; 271: 1723–6.

38. Wostmann BS, Pleasants JR. The germ-free animal fedchemically de� ned diet: a unique tool. Proc. Soc. Exp. Biol.Med. 1991; 198: 539–46.

39. Parrott DMW. The gut-associated lymphoid tissue and gas-trointestinal immunity. In: Fergusson A, MacSween NRM,eds. Immunological Aspects of the Liver and Gastrointesti-nal Tract. Lancaster: MTP Press, 1976: 1–32.

40. Kramer DR, Cebra JJ. Early appearance of ‘natural’ mu-cosal IgA responses and germinal centers in suckling micedeveloping in the absence of maternal antibodies. J. Im-munol. 1995; 154: 2051–62.

41. Cebra JJ, Jlang HQ, Sterzl J, Tlaskalova-Hogenova H. Therole of mucosal microbiota in the development and mainte-nance of the mucosal immune system. In: Ogra A, et al., eds.Mucosal Immunology. New-York: Academic Press, 1999:267–80.

42. Smith MW, James PS, Tivey DR. M cell numbers increaseafter transfer of SPF mice to a normal animal house. Am. J.Path. 1987; 128: 385–9.

43. Craig SW, Cebra JJ. Peyer’s patches an enriched source ofprecursors for IgA-producing immunocytes in the rabbit. J.Exp. Med. 1971; 134: 188–200.

44. Guy-Grand D, Griscelli C, Vassali P. The gut-associatedlymphoid system: nature and properties of large dividingcells. Eur. J. Immunol. 1974; 4: 435–43.

45. Kroese FGM, Butcher EC, Stall AM, Lalor PA, Adams S,Herzenberg LA. Many of the IgA producing cells in murinegut are derived from self-replenishing precursors in the peri-toneal cavity. Int. Immunol. 1989; 1: 75–80.

46. Beagley KW, Bao S, Husband AJ. Mucosal IgA responses incytokine knockout mice: Differential cytokine requirementfor IgA secretion by B-1 and B-2 cells. Mucosal Immunol.Update 1998; 6 (4): 15–9.

47. Crabbe P, Bazin H, Eyssen H, Heremans JF. The normalmicrobial � ora as a major stimulus for proliferation ofplasma cells synthetizing IgA in the gut. Int. Arch. Allergy1968; 34: 362–75.

48. Van Der Heijden PJ, Bianchi ATJ, Heidt PJ, Stok W,Bokhout BA. Background (spontaneous) immunoglobulinproduction in the murine small intestine before and afterweaning. J. Reprod. Immunol. 1989; 15: 217–27.

49. Moreau MC, Ducluzeau R, Guy-Grand D, Muller MC.Increase in the population of duodenal IgA plasmocytes inaxenic mice monoassociated with different living or deadbacterial strains of intestinal origin. Infect. Immun. 1978; 21:532–9.

50. Moreau MC, Ducluzeau R, Muller MC, Raibaud P. Effectof Escherichia coli strain on intestinal IgA plasmocyte stimu-lation and serum antibody response in gnotobiotic mice.Progress Clin. Biol. Res. 1984; 181: 391–5.

51. MacGhee JR, Michalek SM, Kiyono H, Eldrigde JH, Col-well DE, Williamson SI, Wannemuehler MJ, Jirillo E, Mos-teller LM, Spalding DM, Hamada S, Gollahon KA,Morisaki I, Gregory RL, Koopman WJ. Mucosal im-munoregulation: environmental lipopolysaccharide andGALT lymphocytes regulate the IgA response. Microbial.Immunol. 1984; 28: 261–80.

52. Rothkotter HJ, Ulbrich H, Pabst R. The postnatal develop-ment of gut lamina propria lymphocytes: number, prolifera-

tion and T and B cell subsets in conventional and germ-freepigs. Pediatr. Res. 1991; 29: 237–42.

53. Beagley KW, Fujihashi K, Lagoo AS, Lagoo-DeenadaylanS, Black CA, Sharmanov AT, Yamamoto M, McGhee JR,Elson CO, Kyiono H. Differences in intraepitheliallymphocyte T cells subsets isolated from murine small versuslarge intestine. J. Immunol. 1995; 154: 5611–9.

54. Bandeira A, Mota-Santos T, Itohara S, Degermann S,Heusser C, Tonegawa S, Coutinho A. Localization of g:d Tcells to the intestinal epithelium is dependant of normalmicrobial colonization. J. Exp. Med. 1990; 172: 239–44.

55. Stepankova R, Sinkora J, Hudcovic T, Kozakova H,Tlaskalova-hogenova H. Differences in development oflymphocyte subpopulations from GALT of germ-free andconventional rats: effect of aging. Folia Microbiol. 1998; 43:531–4.

56. Umesaki Y, Setoyama H, Matsumoto S, Okada Y. Expan-sion of a:b T-cell receptor-bearing intestinal intraepitheliallymphocytes after microbial colonization in germ-free miceand its independence from thymus. Immunology 1993; 79:32–7.

57. Kawaguchi-Miyashita M, Shimizu K, Nanno M, Shimada S,Watanabe T, Koga Y, Matsuoka Y, Ishikawa H, HashimotoK, Ohwaki M. Development and cytolytic function of intes-tinal intraepithelial T lymphocytes in antigen-minimizedmice. Immunology 1996; 89: 268–73.

58. Ke Y, Pearce K, Lake JP, Ziegler HK, Kapp JA. Tlymphocytes regulate the induction and maintenance of oraltolerance. J. Immunol. 1997; 158: 3610 –8.

59. Mengel J, Cardillo F, Aroeira LS, Williams O, Russo M.Anti-gd T cell antibody blocks the induction and mainte-nance of oral tolerance to ovalbumin in mice. Immunol.Lett. 1995; 48: 97–102.

60. Regnault A, Cumano A, Vassali P, Guy-Grand D, Kouril-sky P. Oligoclonal receptor of the CD8aa and the CD8abTCR-ab murine intestinal intraepithelial T lymphocytes: evi-dence for the random emergence of T cells. J. Exp. Med.1994; 180: 1345–9.

61. Regnault A, Levraud JP, Lim A, Six A, Moreau MC,Cumano A, Kourilsky P. The expansion and selection of Tcell receptor a:b intestinal intraepithelial T cell clones. Eur.J. Immunol. 1996; 26: 914–21.

62. Helgeland L, Vaage JT, Rolstad B, Midtvedt T, BrandtzaegP. Microbial colonization in� uences composition and T-cellreceptor Vb repertoire of intraepithelial lymphocytes in ratintestine. Immunology 1996; 89: 494–501.

63. Matsumoto S, Setoyama H, Umesaki Y. Differential induc-tion of major histocompatibility complex molecules onmouse intestine by bacterial colonization. Gastroenterology1992; 103: 1777–82.

64. Vidal K, Samarut C, Magnaud JP, Revillard JP, KaiserlianD. Unexpected lack of reactivity of allogeneic anti-Ia mono-clonal antibodies with MHC class II molecules expressed bymouse intestinal epithelial cells. J. Immunol. 1993; 151:4642 –50.

65. Cock� eld SM, Urmson J, Pleasants JR, Halloran PF. Theregulation of expression of MHC products in mice. J. Im-munol. 1990; 144: 2967–74.

66. Williams NA, Harper HH, Cochrane L. Antigen presentingcells of the small intestinal lamina propria. Mucosal Im-munol. Update 1997; 5: 29–32.

67. De Smedt T, Pajak B, Muraille E, Lespagnard L, Heinen E,De Baetselier P, Urbain J, Leo O, Moser M. Regulation ofdendritic cell numbers and maturation by lipopolysaccharidein ×i×o. J. Exp. Med. 1996; 184: 1413–24.


68. MacPherson GG, Jenkins CD, Stein MJ, Edwards C. Endo-toxin-mediated dendritic cell release from the intestine.Characterization of released dendritic cells and TNF depen-dence. J. Immunol. 1995; 154: 1317 –22.

69. Gordon JR, Burd PR, Galli S. Mast cells as a source ofmultifunctional cytokines. Immunol. Today 1990; 11: 458–64.

70. Nicaise P, Gleizes A, Forestier F, Sandre C, Quero AM,Labarre C. The in� uence of E. coli implantation in axenicmice on cytokine production by peritoneal and bone mar-row-derived macrophages. Cytokine 1995; 7: 713– 9.

71. Smith PD, Meng G. Mucosal macrophages in infection andimmunity. Mucosal Immunol. Update 1997; 5: 32–4.

72. Johnson AG. Molecular adjuvants and immunomodulators:new approaches to immunization. Clin. Microbiol. Rev.1994; 7: 277–89.

73. Moreau MC, Bisetti N, Dubuquoy C. Immunomodulatingproperties of a strain of Bi� dobacterium used as probiotic onthe fecal and cellular intestinal IgA antirotavirus responsesin mice. In: Sadler M, Saltmarsh M, eds. Functional Foods.The Royal Society of Chemistry, 1998a: 47–54.

74. Moreau MC, Gaboriau-Routhiau V, Dubuquoy C, BisettiN, Bouley C, Prevoteau H. Modulating properties of intesti-nal bacterial strains, Escherichia coli and Bi� dobacterium, ontwo speci� c immune responses generated by the gut, i.e. oraltolerance to ovalbumin and intestinal IgA anti-rotavirusresponse, in gnotobiotic mice. In: Talwar GP, Nath I, Gan-guly NK, Rao KVS, eds. The 10th International Congress ofImmunology. Bologna: Monduzzi Editore, 1998b: 407–11.

75. Bouvet JP, Pires R, Iscaki S, Pillot J. Nonimmune macro-molecular complexes of Ig in human gut lumen: probableenhancement of antibody functions. J. Immunol. 1993; 151:2562–71.

76. Bouvet JP, Fischetti VA. Diversity of antibody-mediatedimmunity at the mucosal barrier. Infect. Immun. 1999; 67:3687 –91.

77. Andrieux C, Pires R, Moreau MC, Bouvet JP. Release ofsoluble co-receptor (Protein Fv) of secretory immunoglobu-lins after colonization of axenic rats by the human gut � ora.Scand. J. Immunol. 1998; 48: 192–5.

78. Murakami M, Honjo T. Involvement of B-1 cells in mucosalimmunity and autoimmunity. Immunol. Today 1995; 16:534–8.

79. Katamaya M, Xu DZ, Specian RD, Deitch EA. Role ofbacterial adherence and the mucus barrier on bacterialtranslocation: effects of protein malnutrition and endotoxinin rats. Ann. Surg. 1997; 225: 317–26.

80. Dickinson EC, Gorga JC, Garett M, Tuncer R, Boyle P,Walkins SC, Alber SM, Parizhskaya M, Trucco M, RoweMI, Ford HR. Immunoglobulin A supplementation abro-gates bacterial translocation and preserves the architectureof the intestinal epithelium. Surgery 1998; 124: 284–90.

81. Marcotte H, Lavoie MC. No apparent in� uence of im-munoglobulins on indigenous oral and intestinal microbiotain mice. Infect. Immun. 1996; 64: 4694–9.

82. Perdigon G, Alvarez S, Gobbato N, De Budeguer MV, DeRuiz Holgado AAP. Comparative effect of the adjuvantcapacity of Lactobacillus casei and lipopolysaccharide on theintestinal secretory antibody response and resistance toSalmonella infection in mice. Food Agric. Immunol. 1995; 7:283–94.

83. Kaila M, Isolauri E, Soppi E, Virtanen E, Laine S,Arvilommi H. Enhancement of the circulating antibody se-creting cell response in human diarrhea by a human Lacto -bacillus strain. Pediatr. Res. 1992; 32: 141–4.

84. Herias MV, Midved T, Hanson LA, Wold AE. Increasedantibody production against gut-colonizing E.coli in thepresence of the anaerobic bacterium PeptostreptococcusScand. J. Immunol. 1998; 48: 277–82.

85. Flo J, Goldma H, Roux ME, Massoud E. Oral administra-tion of a bacterial immunomodulator enhances the immuneresponse to cholera toxin. Vaccine 1996; 14: 1167–73.

86. Hamann L, El-Samalouti V, Ulmer AJ, Flad HD, RietschelET. Components of gut bacteria as immunomodulators. Int.J. Food Microbiol. 1998; 41: 141–54.

87. Babb JL, Kiyono H, Michalek SM, McGhee JR. LPS regu-lation of the immune response: suppression of immune re-sponse to orally administered T-independent antigen. J.Immunol. 1981; 127: 1052–7.

88. Kiyono H, Babb JL, Michalek S, McGhee JR. Cellular basisfor elevated IgA responses in C3H:HeJ mice. J. Immunol.1980; 125: 732–7.

89. Pecquet S, Ehrat C, Ernst P. Enhancement of mucosalantibody responses to Salmonella typhimurium and the mi-crobial hapten phosphorylcholine in mice with X-linkedimmunode� ciency by B-cell precursors from the peritonealcavity. Infect. Immun. 1992; 60: 503–9.

90. Krahenbuhl JP, Neutra M. Molecular and cellular basis ofimmune protection of mucosal surfaces. Physiol. Rev. 1992;72: 853–79.

91. Cebra JJ, Bos NA, Cebra ER, Kramer DR, Kroese FGM,Schrader CE. Cellular and molecular biologic approaches foranalyzing the in ×i×o development and maintenance of gutmucosal IgA responses. In: Mestecky, et al., eds. Advancesin Mucosal Immunology. New-York: Plenum Press, 1995:429–34.

92. Bona C, Bot A. Neonatal immunoresponsiveness. The Im-munol. 1997; 5: 5–9.

93. Mowat AM. The regulation of immune responses to dietaryprotein antigens. Immunol. Today 1987; 8: 93–8.

94. Husby S, Mestecky J, Moldoveanu Z, Holland S, Elson CO.Oral tolerance in humans — T cell but not B cell toleranceafter antigen feeding. J. Immunol. 1994; 152: 4663 –70.

95. Husby S, Jensenius JC, Svehag SE. Passage of undegradeddietary antigen into the blood of healthy adults. Quanti� ca-tion, estimation of size distribution and relation of uptake tolevels of speci� c antibodies. Scand. J. Immunol. 1985; 22:83–92.

96. Weiner HL. Oral tolerance: immune mechanisms and treat-ment of autoimmune diseases. Immunol. Today 1997; 18:335–43.

97. Strobel S, Mowat AM. Immune responses to dietary anti-gens: oral tolerance. Immunol. Today 1998; 19: 173–81.

98. Strobel S, Ferguson A. Persistence of oral tolerance in micefed ovalbumin is different for humoral and cell-mediatedimmune responses. Immunology 1987; 60: 317–8.

99. Lamont AG, Bruce MG, Watret KC, Ferguson A. Suppres-sion of an established DTH response to ovalbumin in miceby feeding antigen after immunization. Immunology 1988;64: 135–40.

100. Peng HJ, Turner MW, Strobel S. The kinetics of oralhyposensitization to a protein antigen are determined byimmune status and the timing, dose and frequency of antigenadministration. Immunology 1989; 67: 425–30.

101. Saklayen MG, Pesce AJ, Pollak VE, Michael JG. Kinetics oforal tolerance: study of variables affecting tolerance inducedby oral administration of antigen. Int. Arch. Allergy Appl.Immunol. 1984; 73: 5–9.

102. Sampson HA, Burks AW. Mechanisms of food allergy.Annu. Rev. Nutr. 1996; 16: 161–77.

M.-C. Moreau and V. Gaboriau -Routhiau .84

103. Challacombe SJ, Tomasi TB. Systemic tolerance and secre-tory immunity after oral immunization. J. Exp. Med. 1980;152: 1459–72.

104. Weiner HL, Friedman A, Miller A, Khoury SJ, Al-SabbaghA, Santos L, Sayegh M, Nussenblatt RB, Trentham DE,Ha� er DA. Oral tolerance: immunologic mechanisms andtreatment of animal and human organ-speci� c autoimmunediseases by oral administration of autoantigens. Annu. Rev.Immunol. 1994; 12: 809–37.

105. Garside P, Mowat AM. Mechanisms of oral tolerance. Cri.Rev. Immunol. 1997; 17: 119–37.

106. Lycke N, Bromander A, Ekman L, Grdic D, Hornquist E,Kjerrulf E, Kopf M, Kosco-Vilbois M, Schon K, Vajdy M.The use of knock-out mice in studies of induction andregulation of gut mucosal immunity. Mucosal Immunol.Update 1995; 3: 1–8.

107. Garside P, Steel M, Liew FY, Mowat AM. CD4» but notCD8» T cells are required for the induction of oral toler-ance. Int. Immunol. 1995; 7: 501–4.

108. Barone KS, Jain SL, Michael JG. Effect of in ×i×o depletionof CD4» and CD8» cells on the induction and mainte-nance of oral tolerance. Cell. Immunol. 1995; 163: 19–29.

109. Chen Y, Kuchroo VK, Inobe JI, Ha� er DA, Weiner HL.Regulatory T cell clones induced by oral tolerance: Suppres-sion of autoimmune encephalomyelitis. Science 1994; 265:1237 –40.

110. Garside P, Steel M, Worthey EA, Satoskar A, Alexander J,Bluethmann H, Liew FY, Mowat AM. T helper 2 cells aresubject to high dose oral tolerance and are not essential forits induction. J. Immunol. 1995; 154: 5649–55.

111. Aroeira LS, Cardillo F, De Albuquerque DA, Vaz NM,Mengel J. Anti-Il-10 treatment does not block either theinduction or the maintenance of orally induced tolerance toovalbumin. Scand. J. Immunol. 1995; 41: 319–23.

112. Rizzo LV, Morawetz RA, Miller-Rivero NE, Choi R, Wig-gert B, Chan CC, Morse HC, Nussenblatt RB, Caspi RR.Il-4 and Il-10 are both required for the induction of oraltolerance. J. Immunol. 1999; 162: 2613–22.

113. Thomas MJ, Kemeny DM. Novel CD4 and CD8 T-cellsubsets. Allergy 1998; 53: 1122– 32.

114. Powrie F, Carlino J, Leach MW, Mauze S, Coffman RL. Acritical role for transforming growth factor-b but not inter-leukin 4 in the suppression of T helper type 1-mediatedcolitis by CD45RBlow CD4» T cells. J. Exp. Med. 1996; 183:2669 –74.

115. Mowat AM, Viney JL. The anatomical basis of intestinalimmunity. Immunol. Rev. 1997; 156: 145–66.

116. Viney JL, Mowat AM, O’Malley J, Williamson E, FangerNA. Expanding dendritic cells in ×i×o enhances the inductionof oral tolerance. J. Immunol. 1998; 160: 5815–25.

117. Samoilova EB, Horton JL, Zhang H, Khoury SJ, WeinerHL, Chen Y. CTLA-4 is required for the induction of highdose oral tolerance. Int. Immunol. 1998; 10: 491–8.

118. Bruce MG, Ferguson A. Oral tolerance to ovalbumin inmice: studies of chemically modi� ed and ‘biologically� ltered’ antigen. Immunology 1986; 57: 627–30.

119. Bruce MG, Ferguson A. The in� uence of intestinal process-ing on the immunogenicity and molecular size of absorbed,circulating ovalbumin in mice. Immunology 1986; 59: 295–300.

120. Louis E, Franchimont D, Lamproye A, Van Kemseke C,Schaaf N, Mahieu P, Belaiche J. Systemic immune responseafter rectocolonic administration of ovalbumin in mice. Int.Arch. Allergy Immunol. 1995; 108: 19–23.

121. MacMenamin C, McKersey M, Kuhnlein P, Hunig T, HoltPG. T cells down-regulate primary IgE responses in rats to

inhaled soluble protein antigens. J. Immunol. 1995; 154:4390 –4.

122. Dahlman A, Ahlstedt S, Hanson LA, Telemo E, Wold AE,Dahlgren UI. Induction of IgE antibodies and T-cell reactiv-ity to ovalbumin in rats colonized with Escherichia coligenetically manipulated to produce ovalbumin. Immunology1992; 76: 225–8.

123. Mowat AM, Maloy KJ, Donachie AM. Immune-stimulatingcomplexes as adjuvants for inducing local and systemicimmunity after oral immunization with protein antigens.Immunology 1993; 80: 527–34.

124. Fritsche R, Pahud JJ, Pecquet S, Pfeifer A. Induction ofsystemic immunologic tolerance to b–lactoglobulin by oraladministration of a whey protein hydrolysate. J. AllergyClin. Immunol. 1997; 100: 266–73.

125. Peng HJ, Chang ZN, Han SH, Won MH, Huang BT.Chemical denaturation of ovalbumin abrogates the induc-tion of oral tolerance of speci� c IgG antibody and DTHresponses in mice. Scand. J. Immunol. 1995; 42: 297–304.

126. Peng HJ, Chang ZN, Lin SY, Han SH, Chang CH. Chemi-cal denaturation of ovalbumin abrogates the induction oforal tolerance of mouse reaginic antibody responses. Scand.J. Immunol. 1998; 47: 475–80.

127. Peng HJ, Turner MW, Strobel S. The generation of a‘tolerogen’ after ingestion of ovalbumin is time-dependantand unrelated to serum levels of immunoreactive antigen.Clin. Exp. Immunol. 1990; 81: 510–5.

128. Lamont AG, Mowat A, Parrott D. Priming of systemic andlocal delayed-type hypersensitivity responses by feeding lowdoses of ovalbumin to mice. Immunology 1989; 66: 595–9.

129. Moreau MC, Gaboriau-Routhiau V. The absence of gut� ora, the doses of antigen ingested and aging affect thelong-term peripheral tolerance induced by ovalbumin feedingin mice. Res. Immunol. 1996; 147: 49–59.

130. Friedman A, Weiner HL. Induction of anergy or activesuppression following oral tolerance is determined by anti-gen dosage. PNAS 1994; 91: 6688–92.

131. Gregerson DS, Obritsch WF, Donoso LA. Oral tolerance inexperimental autoimmune uveoretinitis. Distinct mechanismsof resistance are induced by low ×s high dose feeding proto-cols. J. Immunol. 1993; 151: 5751–61.

132. Stokes CR, Swarbrick ET, Soothill JF. Genetic differences inimmune exclusion and partial tolerance to ingested antigens.Clin. Exp. Immunol. 1983; 52: 678–84.

133. Lamont AG, Mowat A, Browning MJ, Parrott DMV. Ge-netic control of oral tolerance to ovalbumin in mice. Im-munology 1988; 63: 737–9.

134. Strobel S, Ferguson A. Immune responses to fed proteinantigen in mice. III. Systemic tolerance or priming is relatedto age at which antigen is � rst encountered. Pediatr. Res.1984; 18: 588–94.

135. Hanson DG. Ontogeny of orally induced tolerance to solu-ble proteins in mice. I. Priming and tolerance in newborns. J.Immunol. 1981; 127: 1518–24.

136. Miller A, Lider O, Abramsky O, Weiner HL. Orally admin-istered myelin basic protein in neonates primes for immuneresponses and enhances experimental autoimmune en-cephalomyelitis in adult animals. Eur. J. Immunol. 1994; 24:1026 –32.

137. Heppell LM, Kilshaw P. Immune responses in guinea pigs todietary protein. I. Induction of tolerance by feeding ovalbu-min. Int. Arch. Allergy Appl. Immunol. 1982; 68: 54–9.

138. Telemo E, Jacobsson I, Westrom B, Folkesson H. Maternaldietary antigens and the immune response of the offspring inthe guinea pig. Immunology 1987; 62: 35–8.


139. Peng HJ, Turner MW, Strobel S. Failure to induce oraltolerance to protein antigen in neonatal mice can be cor-rected by transfer of adult spleen cells. Pediatr. Res. 1989;24: 486–90.

140. Vaarala O, Saukkonen T, Savilahti E, Klemola T, AkerblomHK. Development of immune response to cow’s milkproteins in infants receiving cow’s milk or hydrolyzed for-mula. J. Allergy Clin. Immunol. 1995; 96: 917–23.

141. Lundin BS, Dahlgren UIH, Hanson LA, Telemo E. Oraltolerization leads to active suppression and bystander toler-ance in adult rats while anergy dominates in young rats.Scand. J. Immunol. 1996; 43: 56–63.

142. Strobel S, Mowat AM, Ferguson A. Prevention of oraltolerance induction to ovalbumin and enhanced antigenpresentation during graft-versus-host reaction in mice. Im-munology 1985; 56: 57–64.

143. Louis E, Franchimont D, Deprez M, Lamproye A, SchaafN, Mahieu P, Belaiche J. Decrease in systemic tolerance tofed ovalbumin in indomethacin-treated mice. Int. Arch. Al-lergy Immunol. 1996; 109: 21–6.

144. Troncone R, Caputo N, Zibella A, Russo R, Rossi M,Gianfrani C, Stern M, Wieser H, Auricchio S. Defective ‘gutprocessing’ of gliadin in mice with graft-versus-host en-teropathy. Int. Arch. Allergy Immunol. 1996; 109: 44–9.

145. Heyman M, Darmon N, Dupont C, Dugas B, Hirribaren A,Blaton AM, Desjeux JF. Mononuclear cells from infantsallergic to cow’s milk secrete tumor necrosis factor alpha,altering intestinal function. Gastroenterology 1994; 106:1514–23.

146. Benlounes N, Dupont C, Candalh C, Blaton MA, DarmonN, Desjeux JF, Heyman M. The threshold for immune cellreactivity to milk antigens decreases in cow’s milk allergywith intestinal symptoms. J. Allergy Clin. Immunol. 1996;98: 781–9.

147. Saidi D, Heyman M, Kheroua O, Boudraa G, Bylsma P,Kerroucha R, Chekroun A, Maragi JA, Touhami M,Desjeux JF. Jejunal response to b-lactoglobulin in infantswith cow’s milk allergy. C.R. Acad. Sci. Paris 1995; 318:683–9.

148. Wannemuehler MJ, Kiyono H, Babb JL, Michalek SM,McGhee JR. Lipopolysaccharide (LPS) regulation of theimmune response: LPS converts germfree mice to sensitivityto oral tolerance induction. J. Immunol. 1982; 129: 959–65.

149. Khoury SJ, Lider O, Al-Sabbagh A, Weiner HL. Suppres-sion of experimental autoimmune encephalomyelitis by oraladministration of myelin basic protein. III. Synergistic effectof lipopolysaccharide. Cell. Immunol. 1990; 131: 302–10.

150. Moreau MC, Corthier G. Effect of the gastrointestinal mi-cro� ora on induction and maintenance of oral tolerance toovalbumin in C3H:HeJ mice. Infect. Immun. 1988; 56:2766 –8.

151. Sudo N, Sawamura SA, Tanaka K, Aiba Y, Kubo C, KogaY. The requirement of intestinal bacterial � ora for thedevelopment of an IgE production system fully susceptibleto oral tolerance induction. J. Immunol. 1997; 159: 1739– 45.

152. Heyman M, Crain-Denoyelle AM, Corthier G, Morgat JL,Desjeux JF. Postnatal development of protein absorption inconventional and germ-free mice. Am. J. Physiol. 1986; 14:G326–31.

153. Elson CO, Ealding W. Cholera toxin feeding did not induceoral tolerance in mice and abrogated oral tolerance to anunrelated protein antigen. J. Immunol. 1984; 133: 2892–7.

154. Clements JD, Hartzog NM, Lyon FL. Adjuvant activity ofEscherichia coli heat-labile enterotoxin and effect on theinduction of oral tolerance in mice to unrelated proteinantigens. Vaccine 1988; 6: 269–77.

155. Snider DP, Marshall JS, Perdue MH, Liang H. Productionof IgE antibody and allergic sensitization of intestinal andperipheral tissues after oral immunization with protein Agand cholera toxin. J. Immunol. 1994; 153: 647–57.

156. Pierre P, Denis O, Bazin H, Mbella EM, Vaerman JP.Modulation of oral tolerance to ovalbumin by cholera toxinand its B subunit. Eur. J. Immunol. 1992; 22: 3179–82.

157. Gaboriau-Routhiau V, Moreau MC. Gut � ora allows recov-ery of oral tolerance to ovalbumin in mice after transientbreakdown mediated by cholera toxin or Escherichia coliheat-labile enterotoxin. Pediatr. Res. 1996; 39: 625–9.

158. Gaboriau-Routhiau V, Moreau MC. Oral tolerance to oval-bumin in mice: induction and long-term persistence unaf-fected by Staphylococcus aureus enterotoxin B andClostridium perfringens type A enterotoxin. Pediatr. Res.1997; 42: 503–8.

159. Verma M, Majumdar S, Ganguly NK, Walia BNS. Effect ofEscherichia coli enterotoxins on macromolecular absorption.Gut 1994; 35: 1613– 6.

160. Heyman M, Dumontier AM, Desjeux JF. Intestinal barrierto intact horseradish peroxidase in experimental secretorydiarrhea. J. Pediatr. Gastroenterol. Nutr. 1986; 5: 463–6.

161. Monneret-Vautrin DA, Kanny G. Allergies alimentaires.Rev. Prat. (Paris) 1996; 46: 961–7.

162. Koning H, Baert MRM, Oranje AP, Savelkoul HFJ, NeijensHJ. Development of immune functions related to allergicmechanisms in young children. Pediatr. Res. 1996; 40: 363–75.

163. Bjorksten B. Environmental in� uence on the development ofchildhood immunity. Nutr. Rev. 1998; 56: S106–12.

164. Holt PG, Macaubas C. Development of long term toleranceversus sensitisation to environmental allergens during theperinatal period. Curr. Opin. Immunol. 1997; 9: 782–7.

165. Warner JA, Jones AC, Miles EA, Colwell BM, Warner JO.Maternofetal interaction and allergy. Allergy 1996; 51: 447–51.

166. Trinchieri G. Interleukin-12 and its role in the generation ofTh1 cells. Immunol. Today 1993; 14: 335–8.

167. Arulanandam BP, VanCleave VH, Metzger DW. IL-12 is apotent neonatal vaccine adjuvant. Eur. J. Immunol. 1999;29: 256–64.

168. Nicaise P, Gleizes A, Sandre C, Kergot R., Lebrec H,Forestier F, Labarre C, 1999. The intestinal micro� ora regu-lates cytokine production positively in spleen-derivedmacrophages but negatively in bone marrow-derivedmacrophages. Eur. Cytokine. Net., 10, in press.

169. Hilkens CMU, Messer G, Tesselaar K, Van RietschotenAGI, Kapsenberg M, Wierenga EA. Lack of IL-12 signalingin human allergen-speci� c Th2 cells. J. Immunol. 1996; 157:4316 –21.

170. Vanderplas Y. Myths and facts about breastfeeding: does itprevent later atopic allergy? Nutr. Res. 1998; 18: 1373–87.

171. Kolb H, Pozilli P. Cow’s milk and type I diabetes: the gutimmune system deserves attention. Immunol. Today 1999;20: 108–10.

172. Singh B, Rabinovitch A. In� uence of microbial agents onthe development and prevention of autoimmune diabetes.Autoimmunity 1993; 15: 209–13.

173. Strachan DP. Hay, fever, hygiene and household size. Br. J.Med. 1988; 1259–60

174. Ruuska T. Occurrence of acute diarrhea in atopic andnonatopic infants: the role of prolonged breast-feeding. J.Pediatr. Gastroenterol. Nutr. 1992; 14: 27–33.

175. Wold AE. The hygiene hypothesis revised: is the risingfrequency of allergy due to changes in the intestinal � ora?Allergy 1998; 53: 20–5.


176. Sarandakou A, Giannaki G, Malamitsi-Putchner A, RizosD, Hourdaki E, Protonotariou E, Phocas I. In� ammatorycytokines in newborn infants. Mediators In� amm. 1998; 7:309–12.

177. Karlsson MR, Kabu H, Hanson LA, Telemo E, DahlgrenUIH. Neonatal colonization of rats induces immunologicaltolerance to bacterial antigens. Eur. J. Immunol. 1999; 29:109–18.

178. Brandwein SL, MacCabe RP, Cong Y, Wiates KB, RidwanBU, Dean PA, Ohkusa T, Birkenmeier EH, Sundberg JP,Elson CO. Spontaneously colitic C3H:HeJBir mice demon-strate selective antibody reactivity to antigens of the entericbacterial � ora. J. Immunol. 1997; 159: 44–52.

179. Guihot G, Merle V, Leborgne M, Pivert G, Corriol O,Brousse N, Ricour C, Colomb V. Enteral nutrition modi� esGut-Associated Lymphoid Tissue in rat regardless of themolecular form of nitrogen supply. J. Pediatr. Gastroenter.Nutr. 1997; 24: 153–61.

180. Siavoshian S, Blottiere HM, Bentouimou N, Cherbut C,Galmiche JP. Butyrate enhances major histocompatibilitycomplex class I, HLA-DR and ICAM-1 antigen expressionon differentiated human intestinal epithelial cells. Eur. J.Clin. Invest. 1996; 26: 803–10.

181. Medzitov R, Janeway CA. Innate immunity: impact on theadaptative immune response. Curr.Opin. Immunol. 1997; 9:4–7.

182. Bos NA, Meeuwsen G, Wostman BS, Pleasants JR, BennerR. The in� uence of exogenous antigenic stimulation on the

speci� city repertoire of background immunoglobulin-secret-ing cells of different isotypes. Cell. Immunol. 1988; 112:371–80.

183. Freitas AA, Viale AC, Sunblad A, Heusser C, Coutinho A.Normal serum immunoglobulins participate in the selectionof peripheral B-cell repertoires. PNAS 1991; 88: 5640–4.

184. Kaveri SV, Lacroix-Desmazes S, Mouthon L, KazatchkineMD. Human natural autoantibodies: lessons from physiol-ogy and prospects for therapy. The Immunologist 1998; 6:227–33.

185. Van Den Broek MF, Van Bruggen MCJ, Koopman JP,Hazenberg MP, Van Der Berg WB. Gut � ora induces andmaintains resistance against streptococcal cell wall-inducedarthritis in F344 rats. Clin. Exp. Immunol. 1992; 88: 313–7.

186. Milon G, Moreau MC, Lebastard M, Marshall G. Hemato-poiesis during infection in mice: an inducible, geneticallycontrolled response mediated by CD4» T cells homing intheir bone marrow. In: van Furth R, ed. MononuclearPhagocytes. The Netherlands: Kluwer Academic Publishers,1992: 50–4.

187. Cukrowska, B., Lodinova-Zadnikova, R., Sokol, D.,Tlaskalova-Hogenova, H., 1999. In ×itro immunoglobulinresponse of fetal B-cells is in� uenced by perinatal infectionsand antibiotic treatment: a study in preterm infants. Eur. J.Pediatr., 158, in press.

188. Fergusson A. Mucosal immunology: from bench to thebedside and beyond. Immunology 1996; 89: 475–782.

B1 cell

Documents

newbornsdigestive ora

digestive ora andgalt

maternal fecal ora

anaerobic ora

colonizethe intestine

intestine of newborns

exogenous factors

functioning of galt