INTRODUCTION The microbial community inhabiting the gastrointestinal tract is characterized by its high population density, wide diversity, and complexity of interactions. All major groups of microbes are present in the gut. Bacteria are predominant but a variety of protozoans are commonly found (13). Anaerobic fungi are widely distributed in the gastrointestinal tract of herbivores (4) as are yeasts (5) and bacteriophages (6). It has been estimated that the human colon contains > 1011 bacterial cells/g contents belonging to as many as 400 different species (79). Note that these numbers were derived from fecal samples and may not accurately represent the intestinal microbiota, especially in terms of species abundance and their relative importance. The prominent role played by anaerobic bacteria in this dynamic ecosystem is evident from the finding that > 99% of the bacteria isolated from
1 From the Department of Animal Sciences and the Division of Nutritional Sciences, the University of Illinois at Urbana-Champaign. 2 Address reprint requests to RI Mackie, University of Illinois at UrbanaChampaign, 1207 West Gregory Drive, Urbana, IL 61801. E-mail: r-mackie@uiuc.edu.
Am J Clin Nutr 1999;69(suppl):1035S45S. Printed in USA. 1999 American Society for Clinical Nutrition
1035S
1036S
MACKIE ET AL
FIGURE 1. Flow chart of possible molecular approaches to the phylogenetic identification and detection of microbial groups or species in the complex microbial ecosystem of the gastrointestinal tract. Molecular microbial ecology techniques are generally based on the informational content encoded in nucleic acid (DNA and RNA) molecules as described in the text. Refer to the reviews by Amann et al (16), Olsen and Woese (20), Raskin et al (17), and Stahl and Amann (21) for more detailed information. rRNA, ribosomal RNA; DGGE, denaturing gradient gel electrophoresis; RAPD, rapid analysis of random amplified polymorphic DNA.
tion and development of gastrointestinal microbiota is critical to an ecological understanding of colonization, succession, and mechanisms of interaction between intestinal microbes and their host. This distinction is difficult especially in infants, in whom bacteria are acquired transiently during and immediately after the birth process as well as from the surrounding environment during this important developmental phase. In general, the terms autochthonous and indigenous are considered as synonyms and imply that these microbes are ubiquitous in the gastrointestinal ecosystem and occupy all habitats and niches available (12, 13). On the other hand, allochthonous species found in a habitat would not be established but merely in passage being derived from food or water, from another habitat in the gastrointestinal ecosystem, or from elsewhere on the host (12, 13). Clearly, some pathogens are autochthonous to the gut ecosystem and can live in harmony with their hosts, becoming pathogenic only when the ecosystem is disturbed in some way. Also, a particular microbial species may be autochthonous to one habitat in the intestinal tract but allochthonous to another, through which it normally passes after it is shed from its native habitat (12, 13). The critical distinction to be made is that an autochthonous microbe colonizes the habitat natively, whereas an allochthonous one cannot colonize the same habitat except under abnormal circumstances. Colonization describes a bacterial population in the gastrointestinal tract that is stable in size, over time, without the need for periodic reintroduction of bacteria by repeated oral doses or other means. This implies that colonizing bacteria multiply in a
particular intestinal niche at a rate that equals or exceeds their rate of washout or elimination at that site (14).
MOLECULAR ECOLOGY Current knowledge of gut microbial ecology and diversity is almost exclusively based on the use of classic culture techniques. The 2 major problems faced by gut microbiologists are the inevitable bias introduced by culture-based enumeration and characterization techniques and the lack of a phylogenetically based classification scheme (1519). Modern molecular ecology techniques based on sequence comparisons of nucleic acids (DNA or RNA) can be used to provide molecular characterization while at the same time providing a classification scheme that predicts natural evolutionary relations (Figure 1). In principle, nucleic acid probes can be designed to hybridize with a complementary target sequence and thus provide a complete description independent of the growth conditions and media used (15, 20, 22, 23). An example of the power of molecular ecology techniques is provided by the analysis of 16S ribosomal RNA sequences (average length of 1500 nucleotides). The highly conserved regions of the ribosomal RNA molecule can serve as primer binding sites for in vitro amplification by polymerase chain reaction (24, 25). The more conserved regions are also useful, serving as targets for universal probes that react with all living organisms or for discriminating between broad phylogenetic groups such as the domains Archaea, Bacteria, and Eucarya.
MICROBIAL ECOLOGY OF THE NEONATAL INTESTINE The more variable sequence regions are more appropriate for genus-, species-, and sometimes even strain-specific hybridization probes (21, 26, 27). The use of molecular probe-based techniques is likely to revolutionize our approach to microbial ecology in the gastrointestinal tract, leading to not simply a refinement or increased understanding, but possibly a complete description for the first time of the complex cascade of regulatory mechanisms and factors, both internal and external, that control the gastrointestinal ecosystem (17, 28). Despite the advantages of a phylogenetically based ecological analysis and the specificity and sensitivity of molecular ecology techniques, few studies of this kind have been carried out in neonatal humans during this critical phase of initial acquisition and subsequent succession of intestinal bacteria. Thus, our current understanding of gut microbiology and ecology in infants is based almost entirely on classic enumeration and phenotypic classification techniques, the limitations of which must be taken into account when evaluating and reviewing this critical area of research in gut microbiology. In this article, development of the microbiota in the human infant intestine is discussed in terms of initial acquisition and subsequent succession of bacteria as revealed by conventional cultivation-based techniques. Intrinsic and extrinsic factors that influence succession, and their health effects, are discussed. With a thorough understanding of these factors, combined with novel insight from a molecular ecology approach, it may become possible to manipulate succession for the benefit of the host. This aspect is also considered briefly.
1037S
ACQUISITION OF THE GUT MICROBIOTA BY NEONATES Origin of human intestinal microbiota The processes involved in the establishment of microbial populations are complex, involving microbial succession as well as microbial and host interactions and eventually resulting in dense, stable populations inhabiting specific regions of the gut. Fetuses are sterile in utero. However, during and after the birth process, infants are exposed to microbes that originate from the mother and the surrounding environment. It is well established that the type of birth delivery has a significant effect on the development of the intestinal microbiota. With a vaginal delivery, the longer the birth process the more likely it is to be able to isolate viable microbes from the stomach and mouth of the infant (29, 30). The same Escherichia coli serotypes were found in both the mouths of babies immediately after birth and in their mothers feces, implying that during natural birth microbes from mothers feces contaminate infants (29, 30). The gastric content of 510-min-old babies was similar to that of their mothers cervix (30). Also, immediately after birth, the nasopharynxes of 62% of babies contained bacteria that were consistent with those of their mothers vaginas immediately before delivery (31). Infants born by cesarean delivery can also be exposed to their mothers microbiota, but initial exposure is most likely to environmental isolates from equipment, air, and other infants, with the nursing staff serving as vectors for transfer (29, 32, 33). After birth, environmental, oral, and cutaneous microbes from the mother will be mechanically transferred to the newborn by several processes including suckling, kissing, and caressing. Thus, the proximity of the birth canal and the anus, as well as
parental expression of neonatal care, are effective methods of ensuring transmission of microbes from one generation to the next (34). The few studies that have been performed on infants in developing countries show a pattern characterized by heavy exposure to and acquisition of bacteria very early in life. For example, Indian infants from Guatemala passed bacteria-containing meconium samples 47 h after birth. Enterobacteria and streptococci were the first groups to establish, and all infants were colonized by E. coli within 48 h (35). A pronounced dominance of bifidobacteria was observed over the entire breast-feeding period, with a corresponding reduction in facultative bacteria. A comparison between Pakistani infants from poor areas and Swedish hospital-delivered infants showed that regardless of delivery method (vaginally or abdominally), the Pakistani infants were colonized significantly earlier (36). Similarly, very early colonization with enterobacteria has been described in Nigerian infants regardless of delivery mode (37). These results confirm pronounced exposure of infants to environmental bacteria very early in developing countries. Thus, the pattern and level of exposure during the neonatal period is likely to influence the microbial succession and colonization in the gastrointestinal tract. In industrialized countries, obstetric and hygienic procedures aimed at reducing the spread of pathogenic bacteria in maternity and neonatal facilities may result in a delayed development pattern or even the absence of certain groups of intestinal bacteria during succession. After the birth process, neonates are continuously exposed to new microbes that enter the gastrointestinal tract with food. This begins with breast milk, which contains up to 109 microbes/L in healthy mothers (38). The most frequently encountered bacterial groups include staphylococci, streptococci, corynebacteria, lactobacilli, micrococci, propionibacteria and bifidobacteria. These commensal bacteria originate from the nipple and surrounding skin as well as the milk ducts in the breast (39, 40). The bacterial composition of milk formulas has not been extensively studied but, based on hospital studies, it is likely that they contain organisms that originate from the dried powder, the equipment used for preparation, and the water used for suspension (32, 33, 39, 41). Both adults and neonates are regularly exposed to microorganisms via the diet, but are affected differently. The microorganisms entering newborns via milk are more likely to colonize than are those entering healthy adults with stable climax communities. Thus, factors that regulate the fate of ingested microbes differ between newborns and adults. Because the intestinal tract is a dynamic ecosystem that is influence by host, intrinsic, and environmental factors, opportunities may arise in which colonization by ingested microbes in infants is favored (42). Major groups and species of bacteria Up to 18 different groups of microbes have been found in the lower genital tract of pregnant women, and these are the rst bacteria to which the neonate is exposed (Table 1). E. coli and streptococci are the most commonly isolated organisms from the upper digestive tract immediately after birth. Detection of lactic acid bacteria has been variable. Lactobacilli are numerous throughout the intestinal tract (62). The lactobacillus species usually found in infant feces, Lactobacillus acidophilus, L. salivarius, and L. fermentum, are also present in adults (48, 60). Presence of bidobacteria has been reported to be both high and low (42), with Bido-
1038S
MACKIE ET AL
TABLE 1 Comparison of viable counts of common groups of bacteria found in the lower genital tract of pregnant women, in feces of neonates (< 1 wk of age), and in fecal samples from adults and infants (> 1 mo of age) on the basis of results from conventional culture-dependent techniques1 Bacterial group Aerobes or facultative anaerobes Bacilli Corynebacteria Enterobacteria Enterococci Lactobacilli Micrococci Propionibacteria Staphylococci Streptococci Range (log10CFU) Anaerobes Bacteroides Bidobacteria Clostridia Eubacteria Fusobacteria Peptostreptococci Ruminococci Veillonella Range (log10CFU)
1 2
Lower genital tract of pregnant women2 Moderate Moderate Low Low Moderate Low Moderate Low Moderate 7.78.6 Moderate Moderate Low Moderate Moderate Moderate Moderate Moderate 8.38.8
Neonate Low Low Moderate Moderate Low Low Moderate Low Moderate 8.29.1 High High Low Moderate Moderate Moderate Low Moderate 7.89.3
Feces3 Infant Low Low Moderate Moderate Moderate Low Moderate Low Moderate 8.08.7 High High Moderate High High High Moderate Moderate 9.811.3
Adult Low Low Moderate Moderate Moderate Low Moderate Low Moderate 6.98.8 High High Moderate High High High Moderate Moderate 10.511.5
Viable counts summarized as high [log109 colony forming units (CFU), moderate (log1068 CFU), and low (log105 CFU)]. Numbers reported as log10 CFU/mL or g secretion. Data summarized from studies of Brook et al (30), Goplerud et al (43), Levison et al (44, 45), Onderdonk et al (46), and Larsen and Galask (47). 3 Numbers reported as log10 CFU/g feces (wet wt). Data summarized from studies of Mata and Urrutia (35), Cooperstock and Zedd (48), Stark and Lee (49), Lundequist et al (50), Yuhara et al (51), Benno et al (52), Mitsuoka (53), Mevissen-Verhage et al (54), Finegold et al (55, 56), Finegold and Sutter (57), Bennet and Nord (58), Mata et al (59), Benno and Mitsuoka (60), and Yoshioka et al (61).
bacterium bidum found most commonly in the vagina. The presence of large numbers of lactobacilli inhabiting the proximal intestinal tract correlates with the presence of stratied squamous epithelium and the ability of lactobacillus cells to adhere to the epithelium. The mechanism by which they adhere ie specic binding to identied receptors, has not been elucidated but involves both protein and carbohydrate molecules on the surface of lactobacillus cells and enterocytes (62, 63). Direct evidence for a mannose-sensitive adherence of L. plantarum to the human colonic cell line HT-29 was detected by Adlerberth et al (64). This binding specicity has not been described previously in Gram-positive bacteria, but is widespread among intestinal Gram-negative bacteria such as E. coli and other members of the enterobacteriaceae family (65). Further studies concerning the molecular basis for adhesion will provide critical information on factors that control colonization and persistence as well as contribute to the elucidation of mechanisms by which indigenous bacteria stimulate the development of host immunity in the intestinal tract (11). Plasmid profiling is a useful molecular technique for distinguishing strains of bacteria. Plasmid proles of isolates of enterobacteria, lactobacilli, and bifidobacteria cultured from vaginal, rectal, and oral swabs collected from women soon after admission to a maternity hospital were compared with those of strains detected in the feces of their infants. Lactobacilli inhabiting the vagina of the mothers did not appear to colonize the infant digestive tract, but evidence for transmission of fecal isolates of enterobacteria and bifidobacteria was obtained in 4 of 5 cases (66). From the previous description, it is clear that neonates are exposed to diverse microbial populations. However, not all com-
ponents of these populations will colonize the digestive tract. Specific microbes become established in particular hosts during different phases of development in a process referred to as microbial succession.
MICROBIAL SUCCESSION Microbial succession during the first few weeks of life in the alimentary tracts of humans, chicks, pigs, and calves is remarkably similar even though neonatal animals are exposed to greater numbers of fecal and environmental bacteria than are human neonates (38, 67, 68). Within a few days of birth, coliforms and streptococci dominate the microbiota in all the above species. Obligate anaerobes appear some time later. Clostridia and lactobacilli may also be present in most hosts within a short period of time. The external and host factors controlling which of the ingested bacteria will establish and the order of succession of the colonizing strains is of major importance. However, this is a difficult topic to both study and review because of extreme subject variability, great variation in rearing and feeding practices, and limitations in sampling and analyzing gut microbial communities. Despite these limitations and variations, an attempt has been made to summarize the trends that have emerged from human infant studies. General trends in humans The development of the intestinal microbiota in infants was divided into 4 separate phases by Cooperstock and Zedd (48). Phase 1 is the initial acquisition phase over the first 12 wk,
MICROBIAL ECOLOGY OF THE NEONATAL INTESTINE phase 2 is the remaining period of solely breast-feeding, phase 3 is the time between the beginning of supplementation and the cessation of breast-feeding, and phase 4 is the period of conversion to adult biota patterns beginning after completion of weaning. A consistent finding is that all infants are initially colonized by large numbers of E. coli and streptococci, often reaching 1081010/g feces. It has been proposed that these bacteria are responsible for the creation of a reduced environment favorable for the establishment of the anaerobic genera Bacteroides, Bifidobacterium, and Clostridium by day 47 (49, 69). In breast-fed infants reduction in numbers of E. coli and streptococci, as well as bacteroides and clostridia, can then decrease in relation to dominance of fecal microbiota by bifidobacteria. Once dietary supplementation begins, the bacterial profile of breast-fed infants resembles that of formula-fed infants in which bifidobacteria are no longer dominant. During phases 3 and 4, after introduction of solid food and weaning, respectively, differences between breast-fed and formula-fed infants are lost and the fecal microbiota resembles that of adults by about the second year of life (48, 49). Numbers of bacteroides and anaerobic Gram-positive cocci (peptococci and peptostreptococci) gradually increase during Phases III and IV (48). In addition, many other bacterial groups including eubacteria, veillonella, staphylococci, propionibacteria, bacilli, fusobacteria, and yeasts are present in adult feces and can also be detected in infants (42). Generally, fluctuations in fecal microbial populations are greater for infants than adults with larger day-to-day and dietinduced variations. Relatively small amounts of formula supplementation of breast-fed infants will result in shifts from a breastfed to a formula-fed pattern (42, 49, 70). There is also considerable variation in the detection of bifidobacteria. Some infants have no detectable bifidobacteria until solid food is introduced, with more formula-fed infants than breast-fed infants lacking bifidobacteria. Frequently, infants with no detectable bifidobacteria have high numbers of bacteroides, clostridia, and E. coli (49, 50, 7173). Breast-feeding compared with formula feeding To resolve this controversial topic Tannock (34) proposed that, if differences do exist, then it should be possible to predict whether an infant is breast-fed or formula fed by the amounts various groups of bacteria found in feces. He found that numbers of clostridia were always lower in breast-fed babies and that clostridia were the only group predictive of formula-fed babies when comparing several individual studies. Conway (42) made a systematic attempt to quantify these findings based on extensive analysis of numerous original studies cited. Results of the various studies were grouped as follows: development during the first week, bacterial numbers in 1-wk-old infants, and bacterial numbers in 119-wk-old infants. Also, studies before 1960 were excluded because milk-formula composition and microbiological techniques have improved since. During the first week of life, as measured by mean values for specific groups of bacteria, the initial colonizers were enterobacteria and streptococci in both formula- and breast-fed babies, reaching their highest numbers by age 3.3 d on average (42). The bifidobacteria and bacteroides appeared 1 d later and then reached their highest numbers on days 56. Interestingly, bifidobacteria and bacteroides took a day longer to colonize the breast-fed infants. Conway (42) speculated that by age 3 d breast milk has passed through the tract and that breast milk, but not formula milk, contains antibacterial factors.
1039S
Another possible explanation is that breast milk, but not formula or cow milk, contains specific bifidobacterial growth factors (74), although it takes longer for the neonate to reach an adequate intake of these growth factors because lactation must first be established. This accounts for the longer time for colonization in breast-fed than in formula-fed infants. When viable counts from 1-wk-old infants were compared, bifidobacteria occurred in larger numbers in feces of breast-fed than in with formula-fed infants (52, 54). This was not the case for viable counts averaged for 119-wk-old infants in which fecal bifidobacterial counts of formula-fed infants increased and were comparable with those in breast-fed infants. The higher counts of bidobacteria in breast-fed infants may be present up to age 56 wk. Bacteroides numbers were low in 1-wk-old breast- and formula-fed infants, and increased in formula-fed infants > 1 wk of age. The mean value for clostridia in 119-wk-old formula-fed infants was log106.6 0.8 colony forming units. Counts of enterobacteria and streptococci were slightly lower in breast-fed infants in both age groups. Variations between studies and between individuals and day-to-day sampling were more marked for breast-fed infants than for those receiving formula. This was most likely due to unrecorded additives to the diet of breast-fed infants. Consistent with the nding that numbers of bidobacteria in 1wk-old infants were higher breast-fed infants than in those receiving formula, bidobacteria were more frequently isolated from breast-fed (84%) than from formula-fed infants (62%). A high correlation was found between viable counts and frequency of detection. For example, clostridia counts were higher in formula-fed infants for both age groups, as was the frequency of detection (49). Succession of the microbiota in the intestinal tract at the group level as determined by selective enumeration is complicated by the fact that there is also a succession of bacterial species and strains. Thus, although total numbers of specific bacterial groups may not vary with time or diet, the dominant species and strains are likely to vary (42). Unfortunately, few studies examined the fecal microbiology to the level of species and strain. However, in one example, the species of bifidobacteria most frequently isolated and occurring in highest number was B. breve in both breast-fed and formula-fed infants, with B. adolescentis, B. longum, and B. bifidum occurring less frequently and in lower numbers (51, 52). Others report that B. infantis was dominant in breast-fed infants with B. longum and B. bifidum being the next most common (53, 72, 75). In one European study, B. breve was most dominant in breast-fed infants, whereas B. longum and B. adolescentis were more frequent in formula-fed infants (54). Interestingly, B. longum and B. adolescentis are dominant in adults (76); this is consistent with the concept that supplementation of infants with formula initiates the development of a microbiota more closely resembling the adult profile (42). On the basis of the expression of surface K1 antigens (virulence marker) and serum killing, it is thought that E. coli strains differ in breast-fed and formula-fed infants (77). In addition, enterobacteria (other than E. coli) were more common in formula-fed than in breast-fed infants (77, 78). Of interest is the finding that E. coli strains are continually changing in adults providing further evidence that acquisition of the intestinal microbiota is an ongoing process that is not restricted to neonates (42). Little is known of the bacterial factors that confer an ability to colonize and persist in the intestinal microbiota. Certain E. coli O-group serotypes are more frequently encountered than others in the intestinal population in European and North American
1040S
MACKIE ET AL (94, 95), by using various measures of microbial activity, showed that the effects of breast-feeding on the microbiota extend for a considerable time beyond cessation of exclusive breast-feeding, even though the culturable bacterial groups appeared relatively stable. For example, profiles of short-chain fatty acids for breast-fed and formula-fed infants did not overlap until 12 mo of age and propionic acid levels differed over the entire 2 y of study. Acetic acid levels were comparable in the 2 groups by 12 mo of age, which may reflect the decrease in bifidobacteria as the dominant population. It is worth noting that breast milk contains antimicrobial components that influence the microbiota as well as numerous growth factors that stimulate development and maturation of the intestinal mucosa, both of which promote stability and decrease susceptibility to intestinal disturbances (96, 97). Internal Intrinsic or host-related factors that influence succession include contributions from host physiology, endogenous nutrients, and the microbiota. Numerous factors influence the stability of the microbiota and shifts, or successional changes, in populations. These include intestinal pH and E h, microbial interactions, environmental temperature, physiologic factors, peristalsis, bile acids, host secretions, immune responses, drug therapy, and bacterial mucosal receptors (42). Two examples are presented that show the influence of these factors on microbial succession. Adhesion of a lactobacillus strain and E. coli K88 fimbriated cells to gastric and intestinal mucosa of piglets aged 5, 26, and 47 d were measured (98). The adhesion pattern of the lactobacillus strain was the reverse of that for E. coli with the former reaching a minimum immediately before weaning at 30 d and the latter reaching a maximum at this time. Receptors on the epithelial mucosa that are involved in bacterial colonization differ with age. Thus, successional changes in lactobacillus and E. coli populations in piglets during the weaning period can be related to receptors on the epithelial mucosa that are involved in bacterial colonization. Antibiotic therapy will also have a major effect on succession of the microbiota (58, 99, 100). Three important points concerning microbial succession can be made in relation to antibiotic therapy. First, antimicrobial agents can have specific effects on individual components of the microbiota rather than a general nonspecific suppression of all microbes. Second, that the resultant microbial profile influences the populations that emerge after treatment has stopped. Third, that effects of antibiotic therapy can persist. Because antibiotics have specific modes of action and organisms exhibit differing sensitivities (as reflected by variations in minimal inhibitory concentrations), it is not surprising that particular antibiotics influence individual groups. A comprehensive review of changes in human intestinal microbiota in relation to administration of antimicrobial agents is available (101).
studies (29, 79, 80). These serotypes were also responsible for a high percentage of extraintestinal E. coli infection (81). Uropathogenic O-groups were strongly associated with E. coli persistence in the rectal flora of Swedish schoolgirls (82). Also, a high percentage of E. coli strains from Swedish infants that persisted for > 6 mo belonged to such O-groups (83). However, in the study of home-delivered Pakistani infants there was no association between uropathogenic serotypes of E. coli and persistence in the intestine (84). In contrast, expression of P-fimbrial adhesins was strongly associated with E. coli persistence in the Pakistani infants (84). Thus, persistence of O-group serotypes in the Swedish studies is due to traits commonly possessed by strains of these O-groups, such as expression of certain adhesins (P-fimbria and others) which are likely to be important factors promoting colonization and persistence (83, 85). Species of bacteroides most commonly found in infants are B. distasonis, B. vulgatus, and B. fragilis, with counts normally higher in formula-fed than in breast-fed infants, although these numbers may not differ significantly (49, 52, 54). In adults, B. thetaiotaomicron is the most commonly isolated species, although numbers of B. distasonis, B. vulgatus, and B. fragilis occur in as high numbers as do B. thetaiotaomicron (54, 55). An exception to the apparent similarity between the fecal profile of species isolated from formula-fed infants and adults is Clostridium difficile. The incidence of C. difficile was higher in formula-fed (4966%) than in breast-fed (620%) infants and colonization rates in older children (4%) and adults (0.7%) are considerably lower (49, 52, 55). Healthy newborn infants often harbor C. difficile and its toxins with no apparent consequences. Toxin carriage decreases as the infants grow older (812 mo of age) and the toxin assay is not considered valid for clinical interpretation except in children > 1 y of age. On the other hand, C. difficile is clearly implicated in antibiotic-associated diarrhea and colitis among older children (86). Thus it is not possible, from numbers alone, to assess the ecological significance of an organism in a given habitat.
FACTORS AFFECTING MICROBIAL SUCCESSION External Microbial succession in the gastrointestinal tract is influenced by numerous external and internal host-related factors. Extrinsic factors include the microbial load of the immediate environment, food and feeding habits, and composition of the maternal microbiota. In addition, dietary and temperature stress can influence succession of microbes (87, 88). The most studied external factor that influences establishment of the intestinal microbiota is diet. The effects of specific dietary regimens on the composition of the intestinal microbiota in adults have been reported (56, 57). Specific dietary supplements, such as oligosaccharides, that affect bacterial species as well as strain composition and fermentation in the colon are receiving considerable research attention (8991). Oligosaccharides have also been added to infant formulas to lower fecal pH (92). Nutritional modulation of intestinal microbial ecology is discussed in a related symposium paper by Collins and Gibson (93). Most attention has focused on comparing the effects of breastmilk feeding with formula feeding. Introduction of formula supplements and solid foods correlates strongly with the development of obligate anaerobic populations. Midtvedt and Midvedt
MICROBIOTA-ASSOCIATED CHARACTERISTICS IN HUMANS It is always difficult and not always possible to determine whether a particular microbe is truly indigenous to a particular gut compartment or site in a host species. This limitation reflects the fact that criteria for determining whether an organism is indigenous are largely incomplete because of severe constraints in sampling, culturing, enumerating, and identifying intestinal
MICROBIAL ECOLOGY OF THE NEONATAL INTESTINE microbial species (11, 17). Rather than enumerating individual bacterial groups, there has been increasing interest in using microbiota-associated characteristics (MACs) that represent bacterial metabolites or compounds altered in concentration or composition because of bacterial activity (102). Several of the indicators measured routinely because of ease of sampling and analysis are listed in Table 2, together with the bacterial groups active in these biochemical transformations. The establishment and development of MACs in feces of healthy infants over the first 2 y of life was intensively studied by Midtvedt et al (93, 115119) and has been summarized by Conway (42). Of interest is the finding that at age 2 y some indicators remained undeveloped in relation to the adult, namely conversion of cholesterol to coprostanol, degradation of fecal tryptic activity, conversion of bilirubin to urobilinogen, and concentration of valeric acid. This reinforces the concept that whereas numbers of bacterial groups remain relatively unchanged in the developing infant after introduction of solid food, the dominant bacterial species and strains may shift, resulting in changes in metabolic activity. Other indicators changed continuously during the 2-y period. Acetic, propionic, n-butyric, i-butyric, and i-valeric acid reached adult levels at 1, 6, 9, 18, and 18 mo respectively. i-Butyrate and i-valerate were first detected after 1 mo of age whereas n-valerate was rarely found before 6 mo of age. Acetic, propionic and n-butyric acid accounted for 94% of the total short-chain fatty acids. Fecal tryptic activity increased to 80% of adult levels by 9 mo of age after which time it decreased to 30% of adult levels. Thus, development of a trypsin-inactivating microbiota fluctuates and is prolonged. These results are difficult to interpret because the net amount of fecal tryptic activity is the balance between pancreatic secretion, which increases in response to dietary stimulation, and degradation by microbial activity. Fecal urobilinogen concentration slowly increased to a maximum of 30% of adult concentrations by age 2 y. The only bacterial species, isolated thus far, capable of carrying out this conversion is C. ramosum. Whereas conversion of cholesterol to coprostanol was first detected after 6 mo of age, it reached a maximum of 44% of adult levels at 15 mo of age. This is consistent with earlier findings that microbes capable of reducing cholesterol to coprostanol do not colonize the gut in significant numbers until after age 1 y. To date, all isolates capable of this conversion have been Eubacterium lentum. In general, eubacteria are isolated frequently from adults and less often from infants. Mucin degradation was detectable at 1 mo of age and progressively increased over the first 9 mo reaching adult levels by the end of the first year (119). Glycoproteins present in breast
1041S
milk are similar to intestinal mucin glycoproteins and may result in substrate competition during the breast-feeding period (120). Several species can degrade mucin including peptostreptococci, B. bifidum, B. infantis, Ruminococcus AB, and R. torques (109, 110). These bacterial species are generally associated with the microbiota from older infants and adults with the exception of Bifidobacterium spp. The role of gastrointestinal carbon and energy sources for resident bacteria and their contribution to the stability and structure of the bacterial community remain largely unexplored and warrant closer research focus and effort.
MANIPULATING THE DEVELOPING MICROBIOTA It is widely reported that the intestinal microbiota plays an important role in the health and well-being of the host, although a thorough scientific description and evaluation of this interaction is still lacking. Because microbial composition is influenced by the acquisition and succession of microbes and the factors controlling these processes, the concept of manipulating the development and composition of the intestinal microbiota to improve the health of the host has been promoted. This can be achieved by controlling external factors that have a direct effect on the composition of microbiota characteristics such as diet and medication. The addition of oligosaccharides to infant formulas, which purportedly promotes growth of bifidobacteria in the colon and results in lowered fecal pH (92), is an example. Another approach that has received much attention recently is the concept of orally administering viable preparations of desirable microbes, referred to as probiotics or direct-fed microbes (121). In theory, this concept should be effective when the subject is at risk of colonization by undesirable microbes or when transfer from maternal sources cannot occur (generational discontinuity). For humans, these situations include cesarean-delivered infants, prematurely-born infants, and full-term infants requiring intensive care all of whom are at risk of acquiring hospital-derived microbes and whom may carry antibiotic resistance. Oral ingestion of probiotics may be beneficial to the host and influence succession at different stages because microbial succession is an ongoing process. Opportunities for administration of probiotics are created during times of stress or antibiotic treatment when administration of a desirable organism may prevent colonization by a transient or pathogenic organism. In addition, probiotics may assist or promote reestablishment of the indigenous microbiota with health benefits for the host. However, a serious consideration related to the use of probiotics is whether they are opportunistic pathogens, especially in immunodeficient hosts. This question has particular relevance to new-
TABLE 2 Selected fecal microbiota-associated characteristics used as indicators of colonic microbial activity1 Indicator and reference Conversion of cholesterol to coprostanol (103, 104) Dehydroxylation of bile acids (105, 106) Transformation of conjugated bilirubin (107) Inactivation of tryptic activity (108) Mucin breakdown (109, 110) Dipeptidase activity (111) Presence of short-chain fatty acids (112) Presence of intestinally produced methane in expired air (113, 114)
1
Bacteria Eubacterium Eubacterium and Clostridium Clostridium Bacteroides Ruminococcus, Peptostreptococcus, and Bidobacterium Bacteroides Ubiquitous Methanobrevibacter
1042S
MACKIE ET AL cocci. After the introduction of solid foods, obligate anaerobes increase in numbers and diversity until a pattern similar to that in adult humans is achieved by 2 y of age. Importantly, succession of species and strains, which is apparently ongoing throughout life, has not been well described. This is confirmed by analysis of MACs that has been valuable in identifying differences in the development of breast-fed and formula-fed infants, especially after supplementation; routine microbiological techniques lack the sensitivity required to monitor these population shifts. This emphasizes the importance of modern molecular ecology techniques for studying acquisition and succession of the microbiota in the neonatal intestine. In providing specific, sensitive, and culture-independent evaluation of all members of the gastrointestinal ecosystem, new molecular technologies will change fundamentally our understanding of microbial ecology in the gut and allow a complete description of this key physiologic system for the first time.
We acknowledge the detailed suggestions and comments of an anonymous reviewer that resulted in substantial improvement to the manuscript.
born infants. A recent study that assessed the capacity of probiotic bacterial species to colonize, infect, stimulate immune responses in, and affect the growth and survival of congenitally immunodeficient gnotobiotic mouse strains emphasizes the need for additional research (122). The implication from this work is that although probiotic species may be innocuous for adults, caution and further studies to assess the safety of probiotic bacteria for immunodeficient hosts, especially neonates are required. Lactic acid bacteria are the most frequently used probiotic organisms, although their beneficial effects have not been conclusively proven (123, 124). However, several studies have shown beneficial effects from probiotics containing lactic acid bacteria in humans (126128), pigs (129), chicks (67) and preruminant calves and lambs (130). In addition to lactic acid bacteria, several other microbes have been studied as probiotic organisms to reduce acquisition of undesirable pathogens from the hospital environment by high-risk newborn infants (63). This approach utilizes the competitive exclusion principle where the harmless probiotic organism purportedly occupies the same site as the potential pathogen. An example is the implantation of a streptococcus strain in the pharynx to reduce the incidence of colonization of this site by Staphylococcus aureus (131). Animal experiments have shown successfully that oral administration of E. coli strains that have no K capsular antigens and are nonpathogenic lowered incidence of colonization of newborns by antibiotic-resistant E. coli strains. The strains most likely to colonize are the ones introduced soon after birth, ie, within the first 2 h (132). Further research showed that babies dosed orally with nonpathogenic E. coli had fewer antibiotic-resistant E. coli strains than did a control group in an open study (133). These results indicate that a control mechanism such as competitive exclusion or colonization resistance operates in the neonatal gut with established strains outcompeting introduced strains. Whereas pure bacterial cultures or simple defined mixtures have generally been used, success has also been achieved with complex mixtures of organisms. Complex mixtures of anaerobic bacteria have been introduced into humans in an effort to reestablish desirable gut microbiota in patients colonized by the nosocomial pathogen C. difficile as a result of antibiotic therapy (134). Development of accurate identification procedures to precisely monitor the population of specific probiotic strains during passage through the gastrointestinal tract is essential for a realistic evaluation of attempts to manipulate the development and composition of the intestinal microbiota. Recent research on lactobacilli and bifidobacteria in healthy human adults using molecular fingerprinting techniques highlights individual differences in native populations (135). Variation among individuals in the composition of the intestinal microbiota is certain to effect the efficacy of probiotic strains and must be studied further. Also remaining largely undefined are the principles of microbial community structure, organization, and dynamics that may influence the ability of an introduced strain to colonize or otherwise effect host benefit.
REFERENCES
1. Clarke RTJ. The gut and its micro-organisms. In: Clarke RTJ, Bauchop T, eds. Microbial ecology of the gut. New York: Academic Press, 1977:3571. 2. Dehority BA. Foregut fermentation. In: Mackie RI, White BA, eds. Gastrointestinal microbiology. Vol 1. New York: Chapman and Hall, 1997:3983. 3. Hespell RB, Akin DE, Dehority BA. Bacteria, fungi and protozoa of the rumen. In: Mackie RI, White BA, Isaacson RE, eds. Gastrointestinal microbiology. Vol 2. New York: Chapman and Hall, 1997:59141. 4. Orpin CG, Joblin KN. The rumen anaerobic fungi. In: Hobson PH, ed. The rumen microbial ecosystem. New York: Elsevier Applied Science, 1988:12950. 5. Lund A. Yeasts and molds in the bovine rumen. J Gen Microbiol 1974;81:45362. 6. Klieve A, Bauchop T. Morphological diversity of ruminal bacteriophages from sheep and cattle. Appl Environ Microbiol 1988;54:163741. 7. Conway PL. Microbial ecology of the human large intestine. In: Gibson GR, Macfarlane GT, eds. Human colonic bacteria. Boca Raton, FL: CRC Press, 1995:124. 8. Lee A. Neglected niches: the microbial ecology of the gastrointestinal tract. In: Marshall K, ed. Advances in microbial ecology. New York: Plenum Press, 1984:11562. 9. Savage DS. Microbial ecology of the gastrointestinal tract. Annu Rev Microbiol 1977;31:10733. 10. Berg RD. The indigenous gastrointestinal microflora. Trends Microbiol 1996;4:4305. 11. Gaskins HR. Immunological aspects of host/microbiota interactions at the intestinal epithelium In: Mackie RI, White BA, Isaacson RE, eds. Gastrointestinal microbiology. Vol 2. New York: Chapman and Hall, 1997:53787. 12. Savage DC. Interactions between the host and its microbes. In: Clarke RTJ, Bauchop E, eds. Microbial ecology of the gut. New York: Academic Press, 1977:277310. 13. Dubos R, Schaedler RW, Costello R, Hoet P. Indigenous, normal and autochthonous flora of the gastrointestinal tract. J Exp Med 1965;122:6776. 14. Freter R. Factors affecting the microecology of the gut. In: Fuller R, ed. Probioticsthe scientific basis. New York: Chapman and Hall, 1992:11144. 15. Amann RI, Krumholz L, Stahl DA. Fluorescent-oligonucleotide probing of whole cells for determinative, phylogenetic, and environmental studies in microbiology. J Bacteriol 1990;172:76270.
SUMMARY The gastrointestinal tract is essentially sterile at birth and is subsequently colonized by microbes acquired from the mother and the surrounding environment. The successional pattern is similar for both humans and animals in that the gut is initially colonized by lactic acid bacteria, enterobacteria, and strepto-
1043S
38. Moughan PJ, Birtles MJ, Cranwell PD, Smith WC, Pedraza M. The piglet as a model animal for studying aspects of digestion and absorption in milk-fed human infants. In: Simopoulos AP, ed. Nutritional triggers for health and in disease. Basel, Switzerland: Karger, 1992:40113. 39. Asquith MT, Harrod JR. Reduction in bacterial contamination in banked human milk. J Pediatr 1979;95:9934. 40. West PA, Hewitt JH, Murphy OM. The influence of methods of collection and storage on the bacteriology of human milk. J Appl Bacteriol 1979;46:26977. 41. OFarrell SM, Lennox-King SMJ, Bettelheim KA, Shaw EJ, Shooter RA. Escherichia coli in a maternity ward. Infection 1976;4:14652. 42. Conway P. Development of intestinal microbiota. In: Mackie RI, White BA, Isaacson RE, eds. Gastrointestinal microbiology. Vol 2. New York: Chapman and Hall, 1997:338. 43. Goplerud CP, Ohm MJ, Galask RP. Aerobic and anaerobic flora of the cervix during pregnanacy and the puerperium. Am J Obstet Gynecol 1976;126:85865. 44. Levison ME, Corman LC, Carrington ER, Kaye D. Quantitative miroflora of the vagina. Am J Obstet Gynecol 1977;127:805. 45. Levison ME, Trestman I, Quach R, Sladowski C, Floro CN. Quantitative bacteriology of the vaginal flora in vaginitis. Am J Obstet Gynecol 1979;133:13944. 46. Onderdonk AB, Polk BF, Moon NE, Groven B, Bartlett JG. Methods for quantitative vaginal ora studies. Am J Obstet Gynecol 1977;128:77781. 47. Larsen B, Galask RP. Vaginal microbial flora: composition and influences of host physiology. Ann Intern Med 1982;96:92630. 48. Cooperstock MS, Zedd AJ. Intestinal flora of infants. In: Hentges DJ, ed. Human intestinal microflora in health and disease. New York: Academic Press, 1983:7999. 49. Stark PL, Lee A. The microbial ecology of the large bowel of breastfed and formula-fed infants during the first year of life. J Med Microbiol 1982;15:189203. 50. Lundequist B, Nord CE, Winberg J. The composition of faecal microflora in breast-fed and bottle fed infants from birth to 8 weeks. Acta Paediatr Scand 1985;74:4551. 51. Yuhara T, Isojima S, Tsuchiya F, Mitsuoka T. On the intestinal flora of bottle fed infants. Bifidobacteria Microflora 1983;2:339. 52. Benno Y, Sawada K, Mitsuoka T. The intestinal microflora of infants: composition of fecal flora in breast-fed and bottle-fed infants. Microbiol Immunol 1984;28:97586. 53. Mitsuoka T. Recent trends in research on intestinal flora. Bifidobacteria Microflora 1982;1:324. 54. Mevissen-Verhage EAE, Marcelis JH, de Vos NM, Harmsen-van Amerongen WCM, Verhoef J. Bifidobacterium, Bacteroides and Clostridium, spp. in fecal samples from breast-fed and bottle-fed infants with and without iron supplement. J Clin Microbiol 1987;25:2859. 55. Finegold SM, Sutter VL, Mathisen GE. Normal indigenous intestinal flora. In: Hentges DJ, ed. Human intestinal microflora in health and disease. New York: Academic Press, 1983:331. 56. Finegold SM, Attebery HR, Sutter VL. Effect of diet on human fecal flora: comparison of Japanese and American diets. Am J Clin Nutr 1974;27:145669. 57. Finegold SM, Sutter VL. Fecal flora in different populations, with special reference to diet. Am J Clin Nutr 1978;23(suppl):S11622. 58. Bennet R, Nord CE. The intestinal microflora during the first week of life: normal development and changes induced by caesarean section, preterm birth and antimicrobial treatment. In: Grubb R, Midtvedt T, Norin E, eds. The regulatory and protective role of the normal microflora. London: Macmillan Press, 1989:1934. 59. Mata LJ, Carrillo C, Villatoro E. Fecal microflora in healthy persons in a preindustrial region. Appl Microbiol 1969;17:596602. 60. Benno Y, Mitsuoka T. Development of intestinal microflora in humans and animals. Bifidobacteria Microflora 1986;5:1325. 61. Yoshioka H, Fijita K, Sakata H, Murono K, Iseki K-I. Development
1044S
MACKIE ET AL
intestinal microflora in health and disease. New York: Academic Press, 1983:44879. 87. Tannock GW. Effect of dietary and environmental stress on the gastrointestinal tract. In: Hentges DJ, ed. Human intestinal microflora in health and disease. New York: Academic Press, 1983:51739. 88. Cargill CF. Control of E. coli infections in pigs. Aust Adv Vet Sci 1982;i:2067. 89. Hidaka H, Eida T, Takizawa T, Tokunaga T, Tashiro Y. Effects of fructooligosaccharides on intestinal flora and human health. Bifidobacteria Microflora 1986;5:3750. 90. Ito M, Deguchi Y, Miyamori A, et al. Effect of administration of galacto-oligosaccharides on the human faecal microflora, stool, weight and abdominal sensation. Microbiol Ecol Health Dis 1990;3:28592. 91. Tomomatsu H. Health effects of oligosaccharides. Food Technol 1994;48:615. 92. Tanaka R, Takayama H, Morotomi M, et al. Effects of administration of TOS and Bifidobacterium breve 4006 on human fecal flora. Bifidobacteria Microflora 1983;2:1724. 93. Collins MD, Gibson GR. Probiotics, prebiotics, and synbiotics: approaches for modulating the microbial ecology of the gut. Am J Clin Nutr 1999;69(suppl):1052S7S. 94. Midtvedt A-C, Midtvedt T. Production of short chain fatty acids by the intestinal microflora during the first 2 years of human life. J Pediatr Gastroenterol Nutr 1992;5:395403. 95. Midtvedt A-C. The establishment and development of some metabolic activities associated with the intestinal microflora in healthy children. PhD thesis, Karolinska Institute, Stockholm, 1994. 96. Wold AE, Hanson LA. Defense factors in human milk. Curr Opin Gastroenterol 1994;10:6528. 97. Donovan SM, Odle J. Growth factors in milk as mediators of infant development. Annu Rev Nutr 1994;14:14767. 98. Blomberg L, Conway PL. An in vitro study of colonization resistance to Escherichia coli strain Bd 1107/7508 (K88) in relation to indigenous squamous gastric colonization in piglets of varying ages. Microb Ecol Health Dis 1989;2:28591. 99. Bennet R, Eriksson M, Nord CE, Zetterstrm R. Suppression of aerobic and anaerobic faecal flora in newborns receiving parenteral gentamicin and ampicillin. Acta Paediatr Scand 1982;71:55962. 100. Bennet R, Eriksson M, Nord CE, Zetterstrm R. Fecal bacterial microflora of newborn infants during intensive care management and treatment with five antibiotic regimens. Pediatr Infect Dis 1986;5:5339. 101. Finegold SM, Mathisen GE, George WL. Changes in human intestinal flora related to the administration of antimicrobial agents. In: Hentges DJ, ed. Human intestinal microflora in health and disease. New York: Academic Press, 1983:356446. 102. Midtvedt T. Monitoring the functional state of the microflora. In: Hattori T, ed. Recent advances in microbial ecology. Tokyo: Japan Scientific Societies Press, 1989:5159. 103. Brinkley AW, Gottesman AR, Mott GE. Isolation and characterization of new strains of cholesterol-reducing bacteria from baboons. Appl Environ Microbiol 1982;43:869. 104. Sadzikowski MR, Sperry JF, Wilkins TD. Cholesterol-reducing bacterium from human feces. Appl Environ Microbiol 1977;34:35562. 105. Gustafsson BE, Midtvedt T, Norman A. Isolated fecal organisms capable of 7 -dehydroxylating bile acids. J Exp Med 1966;123:41323. 106. Hirano S, Nakama R, Tamaki M, Masuda N, Oda H. Isolation and characterization of thirteen intestinal microorganisms capable of 7 dehydroxylating bile acids. Appl Environ Microbiol 1981;41:73745. 107. Midtvedt T, Gustafsson BE. Microbial conversion of bilirubin to urobilins in vitro and in vivo. Acta Pathol Microbiol Scand [B] 1981;89:5760. 108. Ramare F, Hautefort I, Verhe F, Raibaud P, Iovanna J. Inactivation of tryptic activity by a human-derived strain of Bacteroides distasonis in the large intestines of gnotobiotic rats and mice. Appl Environ Microbiol 1996;62:14346. 109. Carlstedt-Duke B, Midtvedt T, Nord CE, Gustafsson BE. Isolation and characterization of a mucin-degrading strain of Peptostrepto-
65.
66.
70.
71.
72. 73.
81. 82.
83.
84. 85.
86.
of the normal intestinal flora and its clinical significance in infants and children. Bifidobacteria Microflora 1991;10:117. Tannock GW. The microecology of lactobacilli inhabitating the gastrointestinal tract. Adv Microbiol Ecol 1991;11:14771. Tannock GW. Probiotic properties of lactic-acid bacteria: plenty of scope for fundamental R & D. Trends Biotechnol 1997;15:2704. Adlerberth I, Ativne S, Johansson M-L, Molin G, Hanson LA, Wold AE. A mannose-specific adherence mechanism in Lactobacillus plantarum conferring binding to a human colonic cell line HT-29. Appl Environ Microbiol 1996;62:224451. Duguid JP, Old DC. Adhesive properties of Enterobacteriaceae. In: Beachey EH, ed. Bacterial adherence, receptors and recognition. Vol 6. London: Chapman and Hall, 1980:185217. Tannock GW, Fuller R, Smith S, Hall MA. Plasmid profiling of members of the family Enterobacteriaceae, lactobacilli, and bifidobacteria to study the transmission of bacteria from mother to infant. J Clin Microbiol 1990;28:12258. Barrow PA. Probiotics for chickens. In: Fuller R, ed. Probiotics: the scientific approach. London: Chapman and Hall, 1992:22559. Smith HW. Development of the flora of the alimentary tract in young animals. J Pathol Bacteriol 1965;90:495513. Grutte FK, Horn R, Haenel H. Ernahrung und biochemischmikrookologische vorgange in enddarm von sauglingen. (Nutrition and biochemical microecology processes occurring in the colon of infants.) Z Kinderheilkd 1965;93:2839 (in German). Bullen CL, Tearle PV, Stewart MG. The effect of humanized milks and supplemental breast feeding on the faecal flora of infants. J Med Microbiol 1977;10:40313. Hewitt JH, Rigby J. Effects of various milk feeds on numbers of Escherichia coli and Bifidobacterium in the stools of newborn infants. J Hyg (Lond) 1976;78:8593. Mitsuoka T, Kaneuchi C. Ecology of the bifidobacteria. Am J Clin Nutr 1977;30:1799810. Tomkins AM, Bradley AK, Oswald S, Drasar BS. Diet and the faecal microflora of infants, children and adults in rural Nigeria and urban UK. J Hyg (Lond) 1981;86:28593. Kunz C, Rudloff S. Biological functions of oligosaccharides in human milk. Acta Paediatr 1993;82:90312. Rasic JL, Kurmann JA. Bifidobacteria and their role. Experientia Suppl 1983;39:5180. Mitsuoka T. Taxonomy and ecology of the indigenous intestinal bacteria. In: Hatton T, ed. Recent advances in microbial ecology. Tokyo: Japan Scientific Societies Press, 1989:4938. Orskov F, Sorensen KB. Escherichia coli serogroups in breast-fed and bottle fed infants. Acta Pathol Microbiol Scand 1975;83:2530. Bullen CL, Willis AT. Resistance of the breast-fed infant to gastroenteritis. Br Med J 1971;3:33843. Hartley CL, Richmond MH. Antibiotic resistance and survival of E. coli in the alimentary tract. Br Med J 1975;11:714. Siitonen A. Escherichia coli in fecal flora of healthy adults: serotypes, P and type IC fimbriae, non-P mannose resistant adhesins, and hemolytic activity. J Infect Dis 1992;166:105865. Orskov I, Orskov F. Escherichia coli in extra-intestinal infections. J Hyg (Lond) 1985;95:55175. Wold AE, Caugent DA, Lidin-Janson G, de Man P, Svanborg C. Resident colonic E. coli strains frequently display uropathogenic characteristics. J Infect Dis 1992;165:4652. Tullus K, Kuhn I, Orskov I, Orskov F, Molby R. The importance of P- and type 1-fimbriae for the persistence of Escherichia coli in the human gut. Epidemiol Infect 1992;108:41521. Adlerberth I. Bacterial adherence and intestinal colonization in newborn infants. PhD thesis. University of Gteborg, Sweden, 1996. Adlerberth I, Hanson LA, Svanborg C, Svennerholm AM, Nordgren S, Wold A. Adhesins of Escherichia coli associated with extraintestinal pathogenicity confer binding to colonic epithelial cells. Microb Pathog 1995;18:37385. Bartlett JG. Pseudomembranous colitis. In: Hentges DJ, ed. Human
1045S
122. Wagner RD, Warner T, Roberts L, Farmer J, Balish E. Colonization of congenitally immunodeficient mice with probiotic bacteria. Infect Immun 1997;65:334551. 123. Conway PL. Lactobacilli: fact and fiction. In: Grubb R, Midtvedt T, Norin E, eds. The regulatory and protective role of the normal microflora. London: MacMillan Press, 1989:26382. 124. Marshall VM. Gut-derived organisms for milk fermentation. J Chem Technol Biotechnol 1991;51:54853. 125. Sanders ME. Lactic acid bacteria as promoters of human health. In: Goldberg I, ed. Functional foods. New York: Chapman and Hall, 1994:294322. 126. Goldin BR, Gorbach SL. Probiotics for humans. In: Fuller R, ed. Probiotics, the scientific basis. London: Chapman and Hall, 1992:35576. 127. Sanders ME. Effect of consumption of lactic cultures on human health. Adv Food Nutr Res 1993;37:67130. 128. Lee Y-K, Salminen S. The coming of age of probiotics. Trends Food Sci Technol 1995;6:21951. 129. Jonsson E, Conway PL. Probiotics for pigs. In: Fuller R, ed. Probiotics: the scientific basis. London: Chapman and Hall, 1992:87110. 130. Wallace RJ, Newbold CJ. Probiotics for ruminants. In: Fuller R, ed. Probiotics: the scientific approach. London: Chapman and Hall, 1992:31754. 131. Sprunt K, Leidy G. The use of bacterial interference to prevent infection. Can J Microbiol 1988;34:3328. 132. Duval-Iflah Y, Ouriet MF, Moreau C, Daniel JC, Gabilan JC, Raibaud P. Implantation precoce dune souche de Escherichia coli dans lintestin de nouveau-nes humains: effect de barriere vis-a-vis de souches de E. coli antibioresistantes. Early implantation of a preparation of Eschericia coli into the neonatal intestine: effect of the barrier in relation to antibiotic resistant E. coli. Ann Microbiol 1982;133:393408 (in French). 133. Rastegar LA, Gold LF, Borderon JC, Laugier J, Lafont J-P. Implantation and in vivo antagonistic effects of antibiotic-susceptible Escherichia coli strains administered to premature newborns. Biol Neonate 1990;56:738. 134. Tvede M, Rask-Madsen J. Bacteriotherapy for chronic relapsing Clostridium difcile diarrhea in six patients. Lancet 1989;1:115660.