Colonization of the gut, which is sterile in utero, begins immediately after birth. Several factors influence the immediate and later characteristics of the microbial diversity in the intestinal ecosystem. Interactions between the mode of delivery (1), nutrition (2–4), genetics (5), hygiene (6), and communication between the host cells and the intestinal microbiota (7,5) contribute to the development of the intestinal microbiota, which gradually within the first 2 years of life becomes more stable and later on is unique for each individual (8). The establishment and development of the intestinal microbiota is particularly critical in the days after birth and during weaning (9). Diet seems to be a key factor influencing the intestinal microbial diversity in infancy (2,3,10). A healthy microbiota living in symbiosis with the host acts both as a defense mechanism and as a metabolically active unit. Microbial deviations during early life may therefore affect later health (11,12). An increasing knowledge of the importance of the intestinal microbiota to the maturation of gut-associated lymphoid tissue (13,14) and the establishment of postnatal immune homeostasis (15) has contributed to a growing interest in investigating and understanding the impact of diet on the intestinal microbiota during the first years of life and the potential implications for later health.
The introduction of culture-independent molecular biology–based techniques has provided powerful tools for characterizing and monitoring the development of the intestinal microbiota in infancy and the influence of the diet on this ecosystem. These techniques permit identification of thus far unknown and noncultivable members of intestinal microbiota (16). Polymerase chain reaction (PCR) on a fragment of the 16S rRNA gene followed by denaturing gradient gel electrophoresis (DGGE) and sequence analysis makes it possible to reveal the biodiversity of the intestinal microbiota (3,16–18).
The aim of the present study was to investigate whether cow's milk (CM), infant formula (IF), and fish oil (FO) during the period of complementary feeding influenced the composition of the intestinal microbiota.
PATIENTS AND METHODS
The study was a randomized, nonblinded, 2 × 2 intervention study. The protocol was approved by the Ethical Committee of the Municipalities of Copenhagen and Frederiksberg, Denmark (journal number: 02-014/03). Parents received oral and written information about the study, and they gave their written consent before the study started. The study is registered on ClinicalTrials.gov (KF 02-014/03).
Participants were recruited via mailed invitations sent to 3167 families with 8-month-old infants randomly selected from the Danish Central Personal Register. Infants included in the study were born healthy, at term (>37 weeks' gestation), with normal birth weight for gestational age (10th–90th percentile) (19) and an Apgar score above 7, 5 minutes after delivery. Furthermore, each infant was required to have a habitual intake of at least 250 mL/day of IF or CM at the time of inclusion. Twins and infants with a chronic illness were not included in the study. The 114 included infants were randomized to receive CM and FO (CM+FO), or CM and no FO (CM−FO), or IF and FO (IF+FO), or IF and no FO (IF−FO). Parents of infants randomized to CM were asked to give a daily iron supplement (8.8 mg/day iron) as drops. Depending on group allocation, parents were asked to give the infant 500 mL/day of CM or IF and 5 mL/day of FO (Eskimo-3, Cardinova, Sweden; gift from Anjo A/S, Frederiksberg, Denmark). The intervention period lasted 3 months (from the age of 9–12 months). The objective of the main study was to examine the effect of CM versus IF as primary milk sources with or without supplements of n-3 long-chain (LC) polyunsaturated fatty acids (PUFA) on growth parameters, nutritional status, development, and risk factors for later diseases. Furthermore, a substudy, including 65 of the 114 infants, examined the impact of the intervention on the intestinal microbiota and inflammation. Parents participating in this substudy were instructed to collect a fecal sample from the diaper when the infant was 10 months old. They were instructed not to collect feces if the infant was sick or within 8 days after the infant had received nonsteroidal anti-inflammatory drugs. Infants receiving antibiotics did not participate in this study. Information about the daily average consumption of FO and CM or IF, and whether the infant was still being breast-fed, was collected by questionnaires. Questionnaires and fecal samples were sent to the department, where the fecal samples were frozen (−20°C).
DNA was extracted from feces using the QIAamp DNA stool mini kit (Quiagen, Hilden, Germany) according to the manufacturer's instructions, except that 200 mg of feces was mixed with 1.4 mL ASL kit-buffer, a 1/4-inch ceramic sphere (Bio101, Vista USA), and small acid-washed glass beads with a diameter <106 μm (Sigma Aldrich, Vallensbaek, Denmark) and shaken rigorously (45 seconds, 5.5 m/sec) in a FastPreb instrument (Bio101, Vista USA) to increase the extraction of DNA from the bacterial cells. The DNA extract was stored at −20°C.
Polymerase Chain Reaction
Fragments of the 16S rRNA gene in the extracted DNA were amplified by use of 2 different sets of universal primers in 2 separate reactions. One primer set amplified the V6-8 region with the following primers: 968fGC (5′ GC-clamp-AACGCGAAGAACCTTAC 3′) and 1401r (5′ CGGTGTGTACAAGACCC 3′). Another primer set amplified the V3 region with the primers PRBA338fGC (5′ GC-clamp-TACGGGAGGCAGCAG 3′) and PRUN518r (5′ ATTACCGCGGCTGCTGG 3′). The GC-clamp was (CGCCCGCCGCGCCCCGCGCCCGGCCCGCCGCCCCGCCCC). All of the reactions were carried out in a 50-μL volume containing 1.25 U Taq DNA polymerase (Amersham Biosciences, Piscataway, NJ), 5 μL 10 × PCR reaction buffer (Amersham Biosciences), 200 μmol/L of each deoxynucleotide triphosphate (Amersham Biosciences), 3.0 mmol/L MgCl2 (Amersham Biosciences), 0.1 μmol/L of each primer (DNA Technologies, Aarhus, Denmark), 1% (vol/vol) formamide (Merck), 0.1% (wt/vol) bovine serum albumin (BSA; New England Biolabs, Beverly, MA), 1 μL DNA template, and sterile MilliQ water for adjustment of the volume to 50 μL. The PCR reaction was performed on a Biometra Trio-Thermoblock (Biotron, Göttingen, Germany) under the following thermocycling program: 5 minutes initial denaturation at 95°C; 35 cycles of 95°C for 60 seconds, 52°C for 45 seconds, and 72°C for 60 seconds, followed by a final elongation step of 72°C for 7 minutes. Size and amount of the PCR products were estimated by analysis of 10-μL samples by agarose gel (1.5% wt/vol) electrophoresis.
Denaturing Gradient Gel Electrophoresis
The DGGE analysis was basically performed as first described by Muyzer et al (20) by use of a DCode System apparatus (Bio-Rad, Hercules, CA). Polyacrylamide gels (8% [wt/vol] acrylamide-bisacrylamide [37.5:1] [Bio-Rad]) in 1 × TAE buffer were prepared with a Bio-Rad Gradient Delivery System (Model 475, Bio-Rad) using solutions containing 35% to 60% denaturant (100% denaturant corresponds to 7 mol/L urea [ICN Biomedicals, Aurora, OH] and 40% [vol/vol] formamide [Merck]). Gels were run at 60°C for 16 hours at a constant voltage of 70 V. After electrophoresis, gels were stained with SYBR-GOLD (Molecular Probes, Eugene, OR) for 20 minutes and photographed with UV transillumination (302 nm) by use of a Kodak EDAS 290 system (Eastman Kodak, New Haven, CT).
The fingerprint profiles generated by the PCR-DGGE were compared by use of the Pearson correlation coefficient (21). Similarities were visualized graphically as dendrograms. The unweighted pair group method with arithmetic average was used to calculate the clustering algorithms of the dendrogram (22). Normalization of the gels, cluster analyses, and dendrograms were carried out with BioNumerics version 2.5 (Applied Maths, St Martens, Latem, Belgium). The χ2 test was used to compare the proportion of infants receiving CM or IF with or without FO within cluster groups. Statistically significant difference was assumed when P < 0.05. The statistical analyses were performed with the statistical package SPSS 12.0 (SPSS Inc, Chicago, IL).
Sixty-five infants participated in this part of the main intervention study. Fecal samples were collected from 60 of the infants when they were 10 months old (±1week). One family did not wish to participate, 2 infants were sick, and 2 families never delivered the fecal sample. One infant refused to consume the FO and was therefore transferred from the CM+FO group to the CM−FO group. At the age of 10 months, the 29 infants in the CM groups were reported to consume on average 387 mL/day of CM, and the 31 infants in the IF groups consumed on average 355 mL/day of IF. Nineteen infants were still being breast-fed. The distribution of infants still being breast-fed was as follows: 6 in the CM+FO group, 8 in the CM−FO group, 2 in the IF+FO group, and 3 in the IF−FO group. The average consumption of either CM or IF was reported to be 258 mL/day in breast-fed infants and on average 422 mL/day for infants not being breast-fed.
DGGE Gels with V3 and V6–8 Primers
For unknown reasons it was impossible to extract DNA from 2 of the 60 fecal samples. These 2 samples were then excluded from the PCR-DGGE analysis. The V3-derived DGGE gels showed 2.5 times as many bands per lane as gels with V6–8 primers (mean number of band/lane was 14.7 vs 5.7). Comparison of the V3-derived DGGE gels including all 58 samples showed that the DGGE fingerprint profiles from infants receiving CM clustered together, and profiles from infants receiving IF also clustered together, which resulted in a significantly different distribution of infants from the CM and IF groups between the cluster groups (P < 0.001) (Fig. 1). This pattern was also seen when V6–8 primers were used on the same 58 samples (P = 0.001) (data not shown). Cluster analysis of V3-based DGGE gels, including only samples from infants from either the CM or the IF group, showed a significantly different distribution according to whether infants received FO in the CM group (P = 0.001) but not in the IF group (P = 0.39) (Fig. 2). The same significant pattern was also seen in the V6–8 DGGE gel (data not shown).
The composition of the intestinal microbiota in early life is strongly influenced by nutrition. It is well documented that in formula-fed infants, a complex intestinal microbiota develops with more coliforms, enterococci, bacteroides, and clostridia than in breast-fed infants. By contrast, Bifidobacterium species are usually predominant both in numbers and in frequency in fecal samples from breast-fed infants (2,3,10). In a study by Favier et al (3) the continuously changing composition of the intestinal microbiota caused by changes in nutrition during the first year of life was visualized by analysis of 16s rRNA gene fragments by DGGE. In the present study we have shown that even at the time of complementary feeding, ingestion of either CM or IF creates different DGGE fingerprint profiles, so that samples from infants receiving CM clearly grouped together and samples from infants receiving IF grouped together. Another explanation for the differences in the intestinal microbiota between infants receiving CM and those receiving IF could be the difference in iron intake, with the infants receiving CM getting a higher dose per day because of the iron drop supplementation. Some studies suggest that the intestinal microbiota in infancy is influenced by iron intake (23,24). Furthermore, it was shown that consumption of FO together with CM created a differentiated fingerprint profile depending on whether the infant had consumed FO. This difference was not found in samples from infants who had consumed IF. CM contains considerable less n-3 PUFA than does IF. Therefore, we speculate that the difference in the effect of FO supplementation on the intestinal microbiota in the 2 milk groups could be explained by a dose-response effect of n-3 PUFA.
The influence on the composition of the intestinal microbiota, caused by the intervention, was demonstrated by use of 2 different sets of universal bacterial primers (V3 and V6–8). The V3-derived DGGE gels contained 2.5 times as many bands as the V6–8-derived DGGE gels, but the overall results were the same. It has previously been shown that even though V3-derived DGGE gels on average contain more bands than V6–8-derived DGGE gels, cluster analysis of the profiles reveals comparable results (18). Other new molecular techniques such as terminal restriction fragment length profile or fluorescence in situ hybridization could perhaps have broadened or highlighted other perspectives of how this intervention influences the composition of the intestinal microbiota.
The predominant intestinal microbiota is rather stable in adults and is not easily altered by external changes. However, in children the microbiota is less stable, and in infancy it is highly unstable (25). Bacteria entering the gut in early life may favor their own growth and inhibit the growth of bacteria introduced later by inducing glycosylation in the genetically coded repertoire of adhesion sites in the intestinal mucosa and/or by modulating gene expression in the enterocyte (5). In this way the early colonization of the intestinal tract may be of vital importance for the composition of the stable microbiota later on in life. The gut represents a primary immune organ in the body, and the colonization of the gut is the most important driving force in the maturation of the intestinal mucosa and the immune system, including restoration of the immunological balance after birth (15). Several studies strongly indicate that an aberration from the normal intestinal microbiota may contribute to the pathogenesis of local and systemic immune-related diseases (11,26,27). The equilibrium of the intestinal microbiota is an important factor to maintain good health in the host (5), and it is therefore highly likely that the first 1 or 2 years of life represent an important window for manipulation of the future resident microbiota (eg, through the diet) and thereby maintaining homeostasis between physiological and pathological conditions in childhood and later in life. Immune maturation has also been found to be affected by diet. Preterm infants receiving IF with added LCPUFA were shown to experience a maturation of the immune response that was more consistent with that of human milk-fed infants than with that of infants receiving nonsupplemented IF (28). n-3 PUFA may affect components of both natural and acquired immunity, including the production of inflammatory cytokines (29). Supplementation with FO to lactating mothers has, for example, been shown to increase the in vitro interferon-γ production in children even 2 years after the supplementation was given, which may reflect a faster maturation of the immune system (30) and a faster switch away from the neonatal Th2-dominated cytokine profile. This is consistent with studies showing that FO supplementation may reduce the risk of allergy development (31,32). Few studies have examined the interactions between FO supplementation and the intestinal microbiota, and subsequent potential influence on the immune response. PUFA has been shown to affect both the growth and the bacterial adhesion to the intestinal mucosa, but the results are contradictory and not easy to interpret in terms of health implications (33,34).
In summary, the present study showed that nutrition strongly influences the composition of the intestinal microbiota in 10-month-old infants.
1. Grönlund M, Lethonen O, Eerola E, et al
. Fecal microflora in healthy infants born by different methods of delivery: permanent changes in intestinal flora after caesarean delivery. J Pediatr Gastroenterol Nutr 1999; 28:19–25.
2. Harmsen H, Wildeboer A, Raangs G, et al
. Analyses of intestinal flora development in breastfed and formula-fed infants by using molecular identification and detection methods. J Pediatr Gastroenterol Nutr 2000; 30:61–67.
3. Favier C, Vaughan E, De Vos W, et al
. Molecular monitoring of succession of bacterial communities in human neonates. Appl Environ Microbiol 2002; 68:219–226.
4. Mountzouris K, McCartny A, Gibson G. Intestinal microflora of human infants and current trends for its nutritional modulation. Br J Nutr 2002; 87:405–420.
5. Bourlioux P, Koletzko B, Guarner F, et al
. The intestine and its microflora are partners for the protection of the host: report on the Danone Symposium “the intelligent intestine”. Am J Clin Nutr 2003; 78:675–683.
6. Rook G, Stanford J. Give us this day our daily germs. Immunol Today 1998; 19:113–116.
7. Zoetendal E, Akkermans A, Akkermans-van Vliet W, et al
. The host genotype affects the bacterial community in the human gastrointestinal tract. Microbiol Ecol Health Dis 2001; 13:129–134.
8. Zoetendal E, Akkermans A, de Vos W. Temperature gradient gel electrophoresis analysis from human fecal samples reveals stable and host-specific communities of active bacteria. Appl Environ Microbiol 1998; 64:3854–3859.
9. Edwards C, Parrett AM. Intestinal flora during the first month of life: new perspectives. Br J Nutr 2002; 88:S11–S18.
10. Fanaro S, Chierici R, Guerrini P, et al
. Intestinal microflora in early infancy: composition and development. Acta Paediatr Suppl 2003; 441:48–55.
11. Kalliomäki M, Kirjavainen P, Eerola E, et al
. Distinct patterns of neonatal gut microflora in infants in whom atopy was and was not developing. J Allergy Clin Immunol 2001; 107:129–134.
12. Salminen S, Gueimonde M. Gut microbiota in infants between 6 and 24 months of age. Nestle Nutr Workshop Ser Pediatr Program 2005; 56:43–56.
13. Sudo N, Sawamura S, Tanaka K, et al
. The requirement of bacterial flora for the development of an IgE production system fully susceptible to oral tolerance induction. J Immunol 1997; 159:1739–1745.
14. Guarner F, Malagelada JR. Gut flora in health and disease. Lancet 2003; 361:512–519.
15. Bjorksten B. The intrauterine and postnatal environments. J Allergy Clin Immunol 1999; 104:1119–1127.
16. Tannock G. Analysis of the intestinal microflora using molecular methods. Eur J Clin Nutr 2002; 56:S44–S49.
17. Nielsen DS, Møller PL, Rosenfeldt V, et al
. Case study of the distribution of mucosa-associated Bifidobacterium
species and other lactic acid bacteria in the human colon. Appl Environ Microbiol 2003; 69:7545–7548.
18. Vanhoutte T, Huys G, Brandt E, et al
. Temporal stability analysis of the microbiota in human faeces by denaturing gradient gel electrophoresis
using universal primers and group-specific 16S rRNA gene primers. FEMS Microbial Ecol 2004; 48:437–446.
19. Greisen G, Michaelsen KF. Perinatal growth: a practical perinatal growth curve. Ugeskr Laeger 1989; 151:1813–1815.
20. Muyzer G, De Waal E, Uitterlinden A. Profiling of complex microbial populations by denaturing gradient gel electrophoresis
analysis of polymerase chain reaction
-amplified genes coding for 16S RNA. Appl Environ Microbiol 1993; 59:695–700.
21. Smalla K, Wieland G, Buchner A, et al
. Bulk and rhizosphere soil bacterial communities studied by denaturing gradient gel electrophoresis
: plant-dependent enrichment and seasonal shifts revealed. Appl Environ Microbiol 2001; 67:4742–4751.
22. Fromin N, Hamelin J, Tarnawski S, et al
. Statistical analysis of denaturing gel electrophoresis (DGGE) fingerprinting patterns. Environ Microbiol 2002; 4:634–643.
23. Mevissen-Verhage EA, Marcelis JH, Harmsen-Van Amerongen WC, et al
. Effect of iron on neonatal gut flora during the first three months of life. Eur J Clin Microbiol 1985; 4:273–278.
24. Mevissen-Verhage EA, Marcelis JH, de Vos MN, et al
. Bifidobacterium, Bacteroides
, and Clostridium
spp. in fecal samples from breastfed and bottle-fed infants with and without iron supplement. J Clin Microbiol 1987; 25:285–289.
25. Vaughan EE, de Vries MC, Zoetendal EG, et al
. The intestinal LABs. Antonie Van Leeuwenhoek 2002; 82:341–352.
26. Björkstén B, Sepp E, Julge K, et al
. Allergy development and the intestinal microflora during the first year of life. J Allergy Clin Immunol 2001; 108:516–520.
27. Salminen S, MacCartney A, Gibson G, et al
. Mode of delivery: influence on gut microbiota composition in 7 year old children. Gut 2004; 53:1388–1389.
28. Field CJ, Thomson CA, Van Aerde JE, et al
. Lower proportion of CD45R0+ cells and deficient interleukin-10 production by formula-fed infants, compared with human-fed, is corrected with supplementation of long-chain polyunsaturated fatty acids. J Pediatr Gastroenterol Nutr 2000; 31:291–299.
29. Calder P. Polyunsaturated fatty acids, inflammation, and immunity. Lipids 2001; 36:1007–1024.
30. Lauritzen L, Kjaer TM, Fruekilde MB, et al
. Fish oil
supplementation of lactating mothers affects cytokine production in 2 1/2-year-old children. Lipids 2005; 40:669–676.
31. Kankaanpaa P, Nurmela K, Erkkila A, et al
. Polyunsaturated fatty acids in maternal diet, breast milk, and serum lipid fatty acids of infants in relation to atopy. Allergy 2001; 56:633–638.
32. Dunstan J, Mori T, Barden A, et al
. Fish oil
supplementation in pregnancy modifies neonatal allergen-specific immune responses and clinical outcomes in infants at high risk of atopy: a randomized, controlled trial. J Allergy Clin Immunol 2003; 112:1178–1184.
33. Thompson L, Spiller RC. Impact of polyunsaturated fatty acids on human colonic bacterial metabolism: an in vitro and in vivo study. Br J Nutr 1995; 74:733–741.
34. Kankaanpää P, Salminen S, Isolauri E, et al
. The influence of polyunsaturated fatty acids on probiotic growth and adhesion. FEMS Microbiol Lett 2001; 194:149–153.
Keywords:© 2007 Lippincott Williams & Wilkins, Inc.
Cow's milk; Fish oil; Infant formula; Intestinal microbiota; Polymerase chain reaction; Denaturing gradient gel electrophoresis