Intestinal flora during the first months of life

[Guest guest]

Intestinal flora during the first months of life: new perspectives

 

C. A. Edwards*, A. M. Parrett,

Department of Human Nutrition, Glasgow University, Yorkhill

Hospitals, Glasgow G3 8SJ, UK

 

Abstract

 

Increasing awareness that the human intestinal flora is a major

factor in health and disease has led to different strategies to

manipulate the flora to promote health. The complex microflora of the

adult is difficult to change in the long term. There is greater

impact of diet on the infant microflora. Manipulation of the flora

particularly with probiotics has shown promising results in the

prevention and treatment of diarrhoea and allergy. Before attempting

to change the flora of the infant population in general, a greater

understanding of the gut bacterial colonisation process is required.

The critical stages of gut colonisation are after birth and during

weaning. Lactic acid bacteria dominate the flora of the breast-fed

infant. The formula-fed infant has a more diverse flora. The faeces

of the breast-fed infant contain mainly acetic and lactic acid

whereas the formula fed-infant has mainly acetic and propionic acid.

Butyric acid is not a significant component in either group. The

formula-fed infant also has higher faecal ammonia and other

potentially harmful bacterial products. The composition of the

microflora diversifies shortly before and particularly after weaning.

The flora of the formula-fed infant develops more quickly than that

of the breast-fed infant. Before embarking on any strategy to change

the flora, the following questions should be considered: Should we

retain a breast-fed style flora with limited ability to ferment

complex carbohydrates? Can pro- and prebiotics achieve a flora with

adult characteristics but with more lactic acid bacteria in weaned

infants? Are there any health risks associated with such

manipulations of the flora?

 

Infants: Gut microflora: Breast-feeding: Weaning

 

Introduction

 

The bacterial microflora of the human intestine is complex and

numerous. The bacteria ferment unabsorbed carbohydrate to short-chain

fatty acids (SCFA) which have many health benefits related to heart

disease and cancer prevention. However, they have also been

implicated in the aetiology of colonic disorders such as colitis and

cancer (Onoue et al. 1997; Campieri & Gionchetti, 2001).

 

There has been increasing scientific, medical and commercial interest

in the intestinal microflora. This began with the awareness of the

actions of the SCFA, was increased by the resurgence of the health

benefits of lactic acid bacteria (probiotics) and is maintained by

the emergence of food products containing probiotics and prebiotics

(materials which increase the growth of probiotic bacteria).

 

This increase in awareness has also been fuelled by the development

of new techniques for identifying and investigating the flora. New

molecular techniques based on bacterial RNA and DNA have rekindled

the mystery of the gut flora (Vaughan et al. 2000). These methods

have shown that a significant proportion of bacterial DNA in human

faeces is not accounted for by known and culturable species (Sghir et

al. 2000).

 

The current understanding of the importance of the gut microflora has

led to different strategies to manipulate the bacterial populations

to promote health. These include changes to the diet, inclusion of

unabsorbable carbohydrates and ingestion of live bacteria with

potential health benefits (probiotics). However, the complex

microflora of the adult human is difficult to manipulate in the long

term. The flora is made up of at least 400 different species in the

dominant flora (Moore et al. 1978; Tannock, 2000) which create a

stable ecosystem resistant to new bacterial species.

 

There is a greater impact of diet on the intestinal microflora of the

human infant. Several groups have explored the interactions between

diet, the intestinal bacteria, diarrhoea (Howie et al. 1990;

Pathmakanthan et al. 2000) and allergy (Wold, 1998; Bjorksten et al.

1999; Ouwehand et al. 2001). Manipulation of the gut flora of the

infant has shown promising results (Saavadra et al. 1994; Arvola et

al. 1999; Isolauri, 2001; Kalliomaki et al. 2001). Optimising the

microflora of the infant may also have long-term benefits if the

flora of the adult is determined by events occurring in this critical

period of gut colonisation. However, before the flora of infants in

the general population is altered, for example by the addition of

probiotics or carbohydrates that promote their growth (prebiotics) to

normal infant formula or weaning foods, it must be clear what is

intended and what the long-term consequences might be. To predict the

long-term effects, a much greater understanding is required of the

bacterial colonisation process and the metabolic activity of the

flora.

 

Bacterial colonisation in the newborn

 

The critical stages of gut colonisation are in the days after birth

and during weaning (Fig. 1). At birth the gut is sterile and as the

infant is exposed to bacteria in its environment, the birth canal,

maternal faecal bacteria and other sources (Tannock et al. 1990;

Zetterstrom et al. 1994), the colonisation process begins. In the

first months of life, diet has a crucial influence on the bacteria

and their metabolism. Several studies, using conventional culture

techniques, have shown that the flora of the breast-fed infant is

dominated by bifidobacteria and lactobacilli, whereas the flora of

the formula-fed infant contains more bacteroides, clostridia and

Enterobacteriaceae (Balmer & Wharton, 1989; Fuller, 1991). Some

studies have not seen this difference (Simhon et al. 1982; Lundequist

et al. 1985) which may be due to changes in obstetric practice.

However, the differences have been confirmed by new molecular

techniques (Harmsen et al. 2000; Martin et al. 2000). Most infants in

the UK are fed both breast and formula milk during their first months

of life due either to an inability of the mother to produce

sufficient milk, or cessation of breast-feeding by the mother due to

a return to work or for other reasons. Very little is known of the

bacterial flora of infants fed a mixture of breast and formula milk.

 

 

 

The mechanisms for the difference in the flora of infants fed human

milk and modern formula are numerous and difficult to reproduce

despite great efforts taken by infant formula manufacturers to mimic

human milk as closely as possible. Immunological factors such as

secretory immunoglobulin A and lysozyme in human milk prevents the

growth of some bacteria. The lower faecal pH of breast-fed infants

(Bullen & Willis, 1971) may promote bacteria such as the lactobacilli

and bifidobacteria, which are more tolerant of acid. Fe-related

factors may also be important. The iron content of human milk is low

(about 0·5 mg/l; Bullen et al. 1972), but the Fe present has a high

bioavailibility (Siimes et al. 1979; Calvo et al. 1992). This is

increased by the presence of lactoferrin, which aids Fe absorption

and binds any unabsorbed Fe making it unavailable to bacteria in the

colon. Bifidobacteria and lactobacilli do not need Fe in contrast to

bacteroides and enterobacteria (Archibald, 1983). Fe added to infant

formula has been found to increase the growth of clostridia and

enterococci (Balmer & Wharton, 1991). However, Fe-free formula,

although promoting a flora more like that of the breast-fed infant,

still did not produce a flora dominated by lactic acid bacteria. The

effects of added lactoferrin have been inconsistent perhaps related

to the use of bovine rather than human lactoferrin (Balmer et al.

1989a). Increasing the whey to casein ratio in the protein fraction

of the milk, however, does promote a more breast-fed style flora

although there are still major differences (Balmer et al. 1989b;

Roberts et al. 1992).

 

There are several bifidogenic factors in human milk and these might

be considered natural prebiotics. They include non-absorbable

oligosaccharides (3–6 g/l in mature milk; Kunz & Rudolff, 1993). The

amount of oligosaccharides in colostrum is even higher. These

oligosaccharides mainly consist of a lactose core substituted with N-

acetyl glucosamine, galactose, fucose and sialic acid resulting in

over 100 different compounds (Kunz & Rudolff, 1993). Recently infant

food products, which include oligosaccharides and are aimed at

specific groups of infants with intestinal symptoms, have been

launched in Europe. However, at present there is little published

information on their impact on the faecal flora. Human milk also

contains nucleotides (Gil et al. 1986; Balmer et al. 1994) and

gangliosides (Rueda et al. 1998) which when added to formula milk

have in some studies been shown to increase colonisation by

bifidobacteria.

 

Bacterial metabolism

 

The metabolic activity of the adult microflora is equivalent in size

and diversity to that of the liver but is often involved in reversing

actions of the liver. For example, the liver conjugates toxic

compounds with sugars or amino acids to make them safer; in the colon

these conjugations are broken down and the toxins and other compounds

reactivated. Many of the complex bacterial activities are a function

of bacteroides, clostridia and other bacteria present in only low

concentrations in the infant gut.

 

Short-chain fatty acids

 

The fermentation of carbohydrate is a major function of the

intestinal flora. Carbohydrate provides most of the energy for the

bacteria but the SCFA produced provide the host with an alternative

energy source (Livesey, 1990), a mechanism to absorb water in the

large intestine (Ruppin et al. 1980) and an inhibitory agent against

pathogens (Fay & Faries, 1975).

 

The profile of SCFA in the faeces of infants differs substantially

from that in adult faeces. In meconium, SCFA are very low, about 10 %

of adult values, increasing in the first four days of life. SCFA

concentrations in the breast-fed infant are still low and increase

significantly in the first year of life (Parrett et al. 2000). The

molar proportions of the individual SCFA are also very different. In

adults the major faecal SCFA are acetic, propionic and n-butyric acid

in the proportions 57:22:21 (Szylit & Andrieux, 1993). Acetic and

lactic acids dominate the SCFA profile of the breast-fed infant.

There is little or no propionic or butyric acid (Edwards et al.

1994), whereas the faeces of the formula-fed infant is dominated by

acetic acid and propionic acid with a higher proportion of butyric

acid than breast-fed infants but still much less than adult faeces.

There is generally no lactic acid. Butyric acid is believed to be the

major fuel for the colonocytes (Roediger, 1982) and thought to be

essential for colonic health. Butyrate has also been shown to have

anti-neoplastic properties by stimulating apoptosis (programmed cell

death) of colonic cells (Hague & Paraskeva, 1995), differentiation of

cancer cell lines in vitro (Augeron & Laboisse, 1984), and inhibiting

other important stages in cancer development (Candido et al. 1978;

Borenfreud et al. 1980). The lack of butyric acid in infant faeces

may mean that butyrate is not critical for the infant colon or that

most of the butyrate necessary is absorbed or utilised by the colonic

mucosa before it reaches the faeces. However, studies of fermentation

in vitro, using infant faeces, also show very low butyric acid

production (Parrett et al. 1997). A high production of propionic acid

and its ratio to acetic acid may help promote healthy plasma lipid

levels (Venter et al. 1990; Wolever et al. 1991, 1996; Berggren et

al. 1996). How this relates to the low propionic acid levels in the

breast-fed infant is unclear.

 

As stated earlier, infants in many Western societies are fed a

mixture of breast milk and formula milk. Some are fed breast and

formula milk at different feeds and infants who begin with breast

milk often are given formula milk within a few days or weeks after

birth. In the UK, the number of mothers who are breast-feeding at

discharge from hospital is substantially more than the number 1–6

weeks later (White et al. 1992). The impact of the introduction of

formula milk to an initially breast-fed infant is not well studied.

There is a belief that one bottle of formula milk will convert the

flora of the breast-fed infant to that of an infant fed exclusively

on formula milk. Bullen et al. (1977) reported that faecal pH was

less acidic in mixed-fed infants than those exclusively breast-fed,

but was still lower than in formula-fed infants. In a recent

longitudinal study of nineteen mixed-fed infants, the SCFA in faeces

indicated a distinct pattern of fermentation from either breast-fed

or formula-fed infants before and during the weaning process (Parrett

et al. 2001).

 

The simplicity of the infant flora limits its fermentation capacity

for complex carbohydrates. When the fermentation capacity of the

faecal flora from breast-fed and formula-fed infants for simple and

complex carbohydrates was compared with that of adult faecal flora

using a simple in vitro model, the infant flora from both groups of

infants was much less able to ferment complex carbohydrates (Parrett

& Edwards, 1997a) than the adult flora. The total SCFA produced by

the infants' microflora was not affected by the type of milk fed but

the proportions of individual SCFA (acetate, propionate, lactic) in

the culture supernatant reflected the different patterns of SCFA seen

in infant faeces (Parrett & Edwards, 1997a).

 

Other bacterial activities

 

There are many colonic bacterial activities, not associated with

carbohydrate fermentation, some of which may have undesirable effects

in the human colon. Most of these activities are associated with the

adult colonic flora and do not develop until later in the first or

second year of life.

 

Midtvedt et al. (1988) have studied a range of bacterial activities

in early life. Mucin degradation ability was not established before 3

months of age. This ability was first seen in many children later in

the first year of life and in all children by the second year

(Midtvedt et al. 1994). The ability to deconjugate bile acids was

already present at 1 month (Jonsson et al. 1995) whereas the ability

to convert cholesterol to coprostanol developed in the second half of

the first year and was delayed by breast-feeding (Midtvedt &

Midtvedt, 1993). At 6 months urobilinogen production was very low

with most infants gaining this activity between 11 and 21 months

(Norin et al. 1985).

 

The metabolites of protein degradation such as ammonia, cresol and

paracresol have been associated with harmful effects in the colonic

mucosa and systemically after absorption. Phenol and cresol had

growth-depressing effects in young pigs (Yokoyama et al. 1982).

Before weaning formula-fed infants had higher faecal urease activity

(Gronlund et al. 1999) and faecal ammonia (Heavey et al. 2000) than

breast-fed infants. Phenol and cresol levels were also higher (Heavey

et al. 2000).

 

The enzymes â-glucuronidase and â-glucosidase may be involved in the

activation of carcinogens and other toxins in the large intestine.

The activity levels of these enzymes are low in the breast-fed infant

before weaning but are increased in formula-fed infants possibly

reflecting the more diverse flora. In a recent study â-glucuronidase

activity was significantly higher in formula-fed than in breast-fed

infants (Gronlund et al. 1999; Heavey et al. 2000), â-glucosidase was

also higher in the formula-fed infants but this did not achieve

statistical significance. The level of all these bacterial activities

increased substantially after weaning as the flora diversified

(Heavey et al. 2000).

 

Effect of weaning

 

The next critical stage after birth in the development of the flora

occurs during the slow process of weaning (Fig. 1). There are several

factors which change during weaning, including the addition of non-

milk foods to the diet and the continued development of intestinal

function. The amount of dietary carbohydrate which enters the colon

in the infant depends on the digestive and absorptive ability of the

small intestine (Christian et al. 1999). The intestine of the neonate

is immature with low salivary amylase and low pancreatic function.

Salivary amylase reaches adult levels between six months and one year

(Rossiter et al. 1974). Pancreatic amylase is very low (1·6 % of the

adult level at birth) until four months and does not increase until

the end of the first year, with mature levels only reached by the

fifth year (Gillard et al. 1989). Mucosal enzymes such as

glucoamylase may compensate for this and amylase and lipase in breast

milk may aid digestion and absorption. Pancreatic sensitivity to

cholecystokinin and secretin (Zoppi et al. 1972; Lebenthal & Lee,

1980) is much reduced in infancy and transport of sugars across the

mucosa may also be much lower than in the adult gut (Younoszai, 1974;

McNeish et al. 1983). In neonates, lactose may be a major substrate

for the colonic microflora. Kien et al. (1992) estimated that up to

74 % of lactose might be metabolised in the large intestine of

premature babies but other studies suggest that lactase activity is

already present at adult levels in the small intestine of most term

infants. However, the very acid nature of the breast-fed infant stool

and the high SCFA and lactate content would suggest that a

substantial amount of rapidly fermentable carbohydrate is escaping

digestion in the small intestine.

 

As weaning begins, infants are exposed for the first time to many

different complex carbohydrates. A significant proportion of the

starch will escape digestion because of the lack of chewing ability

and immature pancreatic exocrine function in these young infants.

This will enter the colon along with any dietary fibre. We have shown

that significant amounts of starch are detectable in the faeces of

children up to the age of 3 years (Verity & Edwards, 1994; Parrett et

al. 2000). This starch will have escaped both digestion and

fermentation. If colonic fermentation capacity develops sufficiently

rapidly, the colonic salvage of energy from unabsorbed complex

carbohydrates may contribute significantly to daily energy needs in

these infants. In adults the energy gained from fermentation of

complex carbohydrate is estimated as 8·36 kJ/g (Livesey, 1990). As

weaning progresses, both intestinal function and fermentation

capacity mature and the amount of faecal starch peaks in early

weaning and then decreases (Parrett et al. 2000). However in some

breast-fed infants the level is still significant at 1 year. In

another study using in vitro incubation methods, infants during early

weaning appeared to have a greater and faster capacity for fermenting

starch than adults, perhaps reflecting the greater likelihood of

starch entering the colon of the infants (Christian et al. 2000). In

the developing world, where starchy foods such as cassava with low

digestibility are eaten by pre-school children, faecal starch is high

and may be a cause of high faecal energy losses (Hamaker et al.

1991). However, in the developed world the amount of starch excreted

is too small to have any impact on energy losses (Parrett, 2001).

 

The concentrations and profiles of SCFA in the faeces of human

infants change as weaning progresses (Midtvedt & Midtvedt, 1992). The

rate of change is related to initial feeding practice. In the breast-

fed infant there is a gradual increase in total SCFA concentration

with a decrease in lactic acid production, an increase in acetic and

propionic acid and by late weaning a gradual increase in butyric acid

production. In the formula-fed infant the change is less profound,

mainly because of the more mature SCFA production before weaning. The

proportion of propionic acid decreases as butyric acid increases

(Parrett, 2001).

 

The ability of the colonic flora of the breast-fed infant to ferment

complex carbohydrates appears to develop more slowly than that of the

formula-fed infant. This is understandable in the light of the

greater diversity of species in the formula-fed infant colon and the

greater populations of Gram-negative anaerobes. In a cross-sectional

in vitro study of fermentation capacity of infants at different

stages of weaning (Parrett et al. 1997; Fig. 2), it was found that

before weaning breast-fed infants could ferment the simple sugars

lactose and glucose well. However, raftilose, a fructo-

oligosaccharide from inulin, was less well fermented and the complex

carbohydrate soya polysaccharide was hardly fermented at all. In

early weaning (one month after the introduction of solid foods) the

ability to ferment raftilose increased but the ability to ferment

complex carbohydrates did not increase until late weaning (seven

months to one year). The differences in fermentation capacity of the

formula-fed infants at the different stages of weaning were not

significant indicating a faster maturation of their colonic flora

(Parrett & Edwards, 1997b). In a more recent longitudinal study,

breast-fed infants had a similar slow development of the capacity to

ferment complex carbohydrates although there was much variation in

individual infants (Parrett, 2001). Pectin was more readily fermented

than other complex carbohydrates such as resistant starch and guar

gum at each stage of weaning.

 

 

 

The impact of a weaning diet is likely to persist into adulthood. The

amount and type of dietary fibre in the weaning diet of rats

influenced their fermentation capacity for these fibres and

consequent effects on stool output when the same fibres were fed to

adult rats (Armstrong et al. 1992). More SCFA were present in faeces

when pectin was fed to adult rats for a month if the rats had been

fed this dietary fibre during weaning than if they had been weaned

onto a diet containing other dietary fibre. Moreover, rats produced

more SCFA from a low fibre diet as adults if they were fed dietary

fibre at weaning. In a study in South Africa, faecal bacteria from

children, under the age of three, from Soweto township produced more

butyric acid from different carbohydrates in vitro than the bacteria

from white children in Potchefstroom. This may have been a result of

a higher intake of resistant starch in their diet at weaning (Edwards

et al. 1998).

 

As the gut is developing throughout the first year of life, it is

important to consider if the changes in the intestinal flora occur as

a result of weaning or just at the time that weaning occurs. Recent

longitudinal studies of human infants using rRNA probes show that the

bacterial flora diversifies with greater dominance of bacteroides and

clostridia groups even before weaning in both breast-fed and formula-

fed infants (Martin et al. 2000). The effects of weaning therefore

need to be studied carefully, taking this into account.

 

Conclusion

 

The flora of the infant develops slowly during the first year of

life. The flora of the breast-fed infant has been shown to be

beneficial for the infant reducing the incidence of diarrhoea. It is

possible that manipulation of the flora, by use of food products

containing prebiotics, probiotics and synbiotics (a mixture of pro-

and prebiotics), may allow promotion of a more beneficial flora which

could reduce gastrointestinal problems, allergy and other

debilitating conditions, as well as potentially helping to prevent

chronic adult diseases such as heart disease and colon cancer.

However, there have been very few good studies of the likely impact

of these food products on the infant flora and its metabolism. In

particular, long-term effects are unknown. At this important time in

the infant's development, we must be clear of our intentions. Should

we change the microflora of the infants during weaning? This may have

greater implications than just increasing the numbers of some

bacterial species. If we try and maintain a breast-fed style flora

for as long as possible we may limit the ability of the weaning

infant to cope with the complex carbohydrates and dietary fibres in

their diet. The most sensible manipulation may be to encourage the

development of fermentation capacity for complex carbohydrates but at

the same time increase the dominance of the lactic acid bacteria.

This would promote the production of butyric acid, which may become

more important in the older child despite being low in infancy. The

levels of lactic acid bacteria could then be promoted through

childhood and into adulthood by dietary means. However, much detailed

research is needed. At present most research is aimed at infants with

specific gastrointestinal and allergy problems. We need to know much

more about the normal population and `normal' colonic function.

 

Dr C. Edwards, fax +44 141 201 9275, email cae1n

 

Abbreviations: SCFA; short-chain fatty acids

 

References:

 

 

 

http://journals.cambridge.org/action/displayFulltext?

type=6 & fid=909240 & jid= & volumeId= & issueId=S1 & aid=909236 & fulltextType=RA

& fileId=S0007114502001757