Protein Content of Human Milk, from Colostrum to Mature Milk

Protein Metabolism During Infancy, edited byNiels C. R. Raiha. Nestle Nutrition WorkshopSeries. Vol. 33. Nestec Ltd., Vevey/Raven Press, Ltd.. New York © 1994.

Protein Content of Human Milk, fromColostrum to Mature Milk

Niels C. R. Raiha

Department of Pediatrics, University of Lund, Malmo General Hospital, 21401 Malmo,Sweden

COMPARATIVE ASPECTS OF PROTEIN CONTENT INMAMMALIAN MILKS

By definition, mammals are those animals that suckle their young and they com-prise the highest class of vertebrates. Thus lactation is one important distinguishingcharacteristic of this class of animals. Mammals are born at markedly different stagesof maturity and if it is assumed that their nutritive requirements correspond to theirphysiological maturity, one may reason Ideologically that the milk of a given speciesis best adapted to the nutritional needs of the young of that same species. The nutri-tional adequacy of mother's milk for the young depends not only on the compositionof the milk, but also on the quantity of milk produced. Linzell's comparative studies(1) of milk production in 22 different species indicate a mean daily yield of 0.126 kgmilk/kg body weight and a daily energy output of 140 kcal/kg. The milk yield is closelyrelated to the body size of the lactating mother. In the human, however, there seemsto be a great individual variation in the milk yield and it may therefore be morecommon in humans than in animals for the milk production to be insufficient foroptimal growth.

Protein in milk is the source of amino acids needed for protein synthesis and growthof the young. The protein content of milk produced by various species ranges fromabout 1% in humans to about 20% in rabbits. At present the compositions of milksof about 150 different mammalian species are known. For six species in addition tohumans (horse, goat, cow, sheep, and reindeer) the data are extremely comprehen-sive and detailed (Table 1), due to the fact that the milks of all these animals, particu-larly the cow and goat, have commonly been used for human consumption.

The protein content of milk has been related to the growth rate of the young.Bernhart (2) suggested that there is a direct correlation between milk protein and thetime required to double the weight of the young, based on studies of nine differentspecies: humans, horse, cow, goat, pig, sheep, dog, cat, and rabbit. The data rangefrom humans, with 125 days to double birthweight and 7% of energy intake fromprotein, to the rabbit, with 7 days to double birthweight and nearly 30% of the energy

87

88 PROTEIN CONTENT OF HUMAN MILK

TABLE 1. Protein content of milk from some mammals

Species Protein (%)

HumansHorseGoatCowSheepReindeer

1.02.52.93.45.5

11.5

intake from protein. This generalization, however, does not hold true for the arcticand aquatic mammals, in which milk energy is derived mainly from fat and in whichthe growth of the young is due to a large extent to deposition of fat.

Figure 1 shows that in most mammals the concentration of lactose in milk is in-versely proportional to that of fat and protein (3). It is seen that human milk has the

14. ANTEATER

15. MOUSE

16. REINDEER

17. DEER

18. DOG

19. SHEEP

20. SEA LION

21. DOLPHIN

22. BUFFALO

23. GUINEA PIG

24. HORSE

25. ORANGUTAN

26. MONKEY

2 3 4 5 6 7Percentage of lactose

FIG. 1. Concentrations of lactose, fat, and protein in milk from various animals [From BernhartFW Nature 1961; 191: 358-360.]

PROTEIN CONTENT OF HUMAN MILK 89

highest lactose content and the lowest protein content of all the milks shown in thefigure. In this chapter I shall discuss the quantitative and some qualitative aspectsof the major proteins in human milk during the various stages of lactation.

TOTAL PROTEIN IN HUMAN MILK DURING PHASES OF LACTATION

Colostrum may be defined as "the thin, yellow, milky fluid secreted by the mam-mary gland a few days before or after parturition." In general, the milk excretedduring the first 3-4 days is called colostrum. Transitional milk is that excreted fromday 6 to day 15 and mature milk is the milk excreted from day 15 (4). The meanconcentrations of total protein during these phases of lactation are shown in Fig. 2.Colostrum has a mean protein content of about 20 g/liter, transitional milk about 15g/liter, and mature milk between 10 and 11 g/liter (5). After the first month of lactationthe total nitrogen content decreases somewhat and then fluctuates, following no par-ticular pattern, in the period between 3 and 6 months' postpartum. The changes inconcentration of the major macronutrients of human milk are shown in Fig. 3. Asthe concentration of the two major energy-containing components, fat and lactose,increases, the concentration of protein decreases (5). This leads to a change in theenergy distribution (Table 2). In early lactation during the colostral stage, protein isresponsible for 17% of the energy content, but by the third week of lactation itaccounts for only 7% of the total energy. Thus one may ask from a teleological pointof view whether the infant needs more protein during the first days of life than later.

1 . 0 •

0.5 •

2 4 15 22 29 36day of lactation

Total human milk protein

FIG. 2. Change in total milk protein concentration during lactation. Values are means, bars =SD. [From Harzer G, & Bindels JG New aspects of nutrition in pregnancy, infancy and prematurity.Amsterdam: Elsevier Scientific Publishers, 1987; 83-94.]


ien

6

5

4

FQo•

4

3

2

,1 I i I 1•

* Lactose

i , i ! IIT / -—*\r

• Triglycerides

135 8 15 22 29 36

. Protein

l i • • • —

I"

•

i • -• *

Energy

13 5 8 15 22 29 36

oa• -

en

1.5

1.0

0.5

EOa\"ao

X

65

60

day of lactation

FIG. 3. Lactose, total protein, fat, and energy of human milk during the 5 weeks of lactation. Dailymeans, bars = SD. [From Harzer G, & Bindels JG New aspects of nutrition in pregnancy, infancyand prematurity. Amsterdam: Elsevier Scientific Publishers, 1987; 83-94]

TABLE 2. Distribution of energy in human milk

Fat Protein Lactose Total energyStage of lactation (%) (%) (%) (kcal/100 ml)

ColostralTransitionalMature

444850

1797

394343

6061

In order to answer this question one must examine the functions of the different milkproteins.

CHANGES IN THE PATTERN OF MAJOR MILK PROTEIN FRACTIONSDURING LACTATION

The concentration of the whey protein secretory IgA decreases sharply duringthe first days of lactation, whereas lactoferrin shows only a moderate decrease. Bycontrast, casein, a-lactalbumin, and serum albumin concentrations are more or lessconstant, as shown in Fig. 4. Thus, according to Harzer & Bindels (5), the wheyprotein/casein ratio of human milk decreases from 80:20 in colostrum to 55:45 inmature milk. This has recently been confirmed by Kunz & Lonnerdal (6), who quanti-


i

0 .6

0.4

0 .2

V IgA\\\\\ H . . .

i . ' . \nLactoferrin

100-0

z50

0-100

• -+ • - . .

X

— —^— — —

WHEY PROTEIN

o^ o o o

0 ' CASEIN

1 3 5 8 15 22 20

• - - 4 , Casein

. a-Lactalbumin

-0

38day

.- +

- - *-B

—=¥

1 3 5 8 15 22 29 36day of lactat ion

Protein composition of hunan milk»•* Analyzed by HPLC ••*

FIG. 4. Human milk protein composition during the first 5 weeks of lactation: casein; a-lactalbu-min; lactoferrin; lysozyme; SlgA; serum albumin. [From Harzer G, & Bindels JG (5).]

tated total casein as well as whey proteins in human milk samples during lactationusing two independent methods: precipitation at pH 4.3 in the presence of calciumions followed by Kjeldahl analysis, and polyacrylamide gradient gel electrophoresis(PAGGE) followed by densitometric scanning. Figure 5 shows that casein concentra-tion is low early in lactation and then increases rapidly during the transitional phaseto a subsequent slow decrease during late lactation. In early colostrum, concentra-tions of whey proteins are very high but subsequently decrease significantly. Thesechanges lead to a change in the whey protein/casein ratio during lactation. It wasestimated by Kunz and Lonnerdal to be about 90:10 in early lactation, 55:45 inmature milk, and 50:50 late in lactation (6). Thus the decrease which is seen in thetotal protein content during lactation is due mainly to a decrease in secretory IgAand lactoferrin with lesser changes in the concentrations of the other milk proteins.

NUTRITIONAL VALUE OF PROTEINS IN HUMAN MILK

The nutritional availability of human milk proteins for amino acid metabolism andprotein synthesis of the infant has recently been discussed by several investigators:Hambraeus et al. (7), Raiha (8), and Butte et al. (9). The functions of many major


' 2 1

10 •

E

6-

4"

4 0 6 0 BO 1 CC

Day of Lactation

FIG. 5. Whey protein and casein concentration in milk during lactation. [From Kung C, & Lonner-dal B Acta Paediatr Scand 1992; 81: 107-112.]

whey proteins, secretory IgA, lactoferrin, and lysozyme are to some extent physio-logical rather than nutritional. Secretory IgA is the main immunoglobulin of humanmilk and acts locally in the infant's gut to prevent attachment of microbes to intestinalcells, thereby preventing infection. Lactoferrin is the major iron binding protein ofhuman milk and is believed to prevent microbial proliferation by reducing the avail-ability of iron. Lysozyme attaches to the bacterial cell wall, causing lysis.

These three protective proteins comprise about 30% of the total proteins in maturehuman milk. Recent studies have shown that these proteins are resistant to low pHand to proteolytic enzymes (10,11), which explains why it was possible for Davidson& Lonnerdal (12) to detect large amounts of both secretory IgA and lactoferrin inthe stools of exclusively breast-fed infants. Figures 6 and 7 show that the excretionof these proteins correlates with the concentration in the milk, both of which decreasewith the age of the infant. The amount of secretory IgA excreted is high, 60% ofintake, in the early weeks of lactation and decreases to 10% of the intake during thelater phase of lactation. A similar pattern is found for lactoferrin (Fig. 7).

By calculating the amounts of whey proteins ingested and subsequently excretedit has been estimated that between 3% and 10% of the milk proteins are unavailableto the infants as a nutritional source of amino acids (13,14). The higher percentage


Fecal slgA

(mg/24 h)

1000 -1

800 -

600 -

4O0 -

200 -

r 5

Milk slgAconcentration

- 2 (mg/ml)

: • - •

Infant Age (weeks)

FIG. 6. Comparison of the change in fecal excretion and milk concentration of secretory IgA withinfant's age. [From Davidson LA, et al. Protein and non-protein nitrogen in human milk. Boca Raton,FL: CRC Press, 1989; 161-172.]

of excretion is found in early lactation when the milk content of secretory IgA andlactoferrin is greater. The non-protein nitrogen (NPN) fraction of human milk com-prises 20-25% of the total nitrogen. Of this, urea forms some 50%, but its concentra-tion fluctuates with the state of lactation. The NPN fraction of human milk alsocontains some peptides and free amino acids, of which taurine is the most prominent.The utilization of urea nitrogen for de novo synthesis of amino acids and body proteins

Fecal Lf

(mg/24 h)

160 T

120 -

80 •

40 -

-o- Fecal Lf-+• Milk Lf

1 6— I —20

:

Milk Lfconcentration

(mg/ml)

Infant Age (weeks)

FIG. 7. Comparison of the change in fecal excretion and milk concentration of lactoferrin withinfant's age. [From Davidson LA, et al. (11).]


TABLE 3. Theoretical estimation of the nutritional value of proteins in human milk

g/liter

Total protein (N x 6.38) 11.6Non-protein N (25%) (2% utilized for protein synthesis) 2.8True protein 8.8Non-nutritional proteins (3%) 0.3Nutritional proteins 8.5

in the normal newborn infant has been much debated and the data are conflicting(15). Fomon (16) has estimated that an average of 13% of the ingested urea could beavailable for endogenous synthesis of amino acids in term infants. Most of this synthe-sis may be carried out by intestinal bacteria (17). The total nutritional contributionof the protective whey proteins and the NPN components in human milk for thenormal infant is still not fully understood and needs further elucidation. Thus theminimum nutritionally available protein in mature human milk may be as low as 8.5g/liter (8), as shown in Table 3. From a nutritional point of view, these results alsoimply that the whey/casein ratio of the nutritionally available proteins of maturehuman milk may be 50:50 or perhaps even slightly casein predominant and thusdifferent from that in the total milk proteins.

One of the methods used to estimate protein requirements of infants is based onmeasurements of protein intake from breast milk in healthy breast-fed infants main-taining satisfactory growth (18). The true nutritionally available protein content ofhuman milk must be considered when assessing protein requirements of infants. Onthe basis of the facts presented it is obvious that the requirement may be considerablyless than recommended previously.

AMINO ACID CONTENT OF NUTRITIONALLY AVAILABLE PROTEINSOF HUMAN MILK

Harzer & Bindels (5) have studied the amino acid profile of the major human milkproteins. Figure 8 shows the amino acid concentrations in the various human milkproteins, secretory IgA, lactoferrin, a-lactalbumin, and casein. In comparison to theother milk proteins, secretory IgA is rich in threonine and has more valine than a-lactalbumin. Casein is rich in tyrosine and has a low content of tryptophan andcystine. Due to the change in the nutritional availability of secretory IgA, the threo-nine and valine available in vivo may be less than what would be predicted on thebasis of the amino acid profile of the hydrolyzed total protein of human milk.

The quantities of amino acids ingested by healthy breast-fed infants have beenestimated on the basis of milk consumption and the amino acid concentrations ofhuman milk samples (19). Since not all human milk proteins are fully hydrolyzed itis clear that these estimations have been too high and qualitatively misleading. Harzer& Bindels (5) have measured the nutritionally available amino acid composition of

PROTEIN CONTENT OF HUMAN MILK

c

prot

o

ws120

80

40

.

. 1

I I

Thr

L. r

1V Q I

• 1Met

11Al i e Leu

JELr

PheMHis

•

1• 1JlLy9 Trp

Asp Sar Glu Pro Gly Ala Cys Tyr Arg

Amino acids in human milk proteins• IgA Hi Lactofarrin • Lactalbumin IS Casein

FIG. 8. Amino acid concentrations in various human milk proteins. [From Harzer G, & BindelsJG New aspects of nutrition in pregnancy, infancy and prematurity. Amsterdam: Elsevier ScientificPublishers, 1987; 83-94.]

human milk at 2, 8, and 36 days of lactation (Fig. 9). Their data show that althoughthe amino acid content of human milk decreases considerably from day 2 to day 36of lactation, the pattern of nutritionally available amino acids remains essentially thesame.

Thus it is difficult to design infant formulas on the basis of the amino acid composi-tion of human milk. It is essential to study the composition of the nutritionally avail-able protein amino acids in vivo. By simply changing the quantity or the whey/caseinratio of bovine milk protein it will not be possible to achieve plasma amino acidprofiles identical to those found in breast-fed infants (Fig. 10) (20). If the purpose offormula design is to produce human milk substitutes that will produce plasma aminoacid profiles similar to those found in infants fed human milk, the protein compositionmust be modified by increasing the bovine a-lactalbumin fraction (see the chapterby Heine) or human milk proteins must be produced by the transgenic technique inlarge quantities to be used for formula production.

PROTEIN CONTENT OF PRETERM BREAST MILK

Atkinson et al. 1978 (21) were the first to observe a higher total nitrogen contentin milk from mothers having given birth to preterm infants when compared to that

PROTEIN CONTENT OF HUMAN MILK

100

50 k I |T iThr Val Mat I la Lou Pho His Lys Trp

is;

ISO

Asp Sor Glu Pro Gly Ala Cye Tyr Arg

Amino acids in human milk (days 2.8.36)

FIG. 9. Comparison of the nutritionally available amino acid concentrations with that of totalprotein of human milk at days 2, 8, and 36 of lactation. [From Harzer G, & Bindels JG New aspectsof nutrition in pregnancy, infancy and prematurity. Amsterdam: Elsevier Scientific Publishers, 1987;83-94.]

PHE T R P THR

TYR LYS

MET VAL

CYS

HIS

ARG

I LE

LEU

TAU

ALA ASP

GLY GLN

SER PRO GLU

HUMAN MILK

1

1

1

1

3

5

G/DL

G/DL

G/DL

FIG. 10. Plasma amino acid profiles of infants fed either human milk (circle) or formulas withvarious concentrations of protein. [From Picone TA, et al. (20).]


450

400

350

300

250

200

150

l a > Fu l l - t e r m ( n • 27 )

y - 310 - 4.15(1r • - 0 . 8 0 ( p « 0 . 0 1 )

I b) Premature ( n • 42 )

y • 368 - 4 .36(1r • - 0 . 6 0 ( ( X O . 0 1 )

10 15 20 25Days Postpartum

30 0 10 15 20 25Days Postpartum

-5ir

FIG. 11. Relationship of nitrogen concentration in human milk to day of lactation in mothers offull-term or premature infants. [From Atkinson SA, et al. J Pediatr 1978; 93: 67-69.]

45

43

•O 35

oE

<t

nl

cc

dT

l

cc5

30

25

20

15

10

y = 2 8 . 5 0 7 - O . 3 6 8 O X + 0 . 0 0 1 6 X

r = 0 . 6 7 8 ( p < O . O O O 1 )

O O 0 O

40 60 80 100 120 140 160 110 200 22C

Threonine Intake (mg/kg/day)

FIG. 12. Plasma threonine response in low-birthweight infants to feeding pooled breast milk (O),preterm breast milk (•), or standard infant formula (•). [From Atkinson SA, & Hanning RM Proteinand non-protein nitrogen in human milk. Boca Raton, FL: CRC Press, 1989; 187-209.]


of mothers of term infants (Fig. 11). Subsequent studies (22-24) revealed that it wasthe protein nitrogen that was increased, whereas the NPN fraction was similar tothat of term milk. One study (25), however, using a different method to determinethe milk protein concentration, found no difference in the protein content betweenpreterm and term breast milk.

Most of the increased concentration of protein in preterm milk may be due tohigher concentrations of the protective proteins of breast milk. Considerably higherconcentrations of lactoferrin, secretory IgA, and lysozyme have been documentedin preterm than in term milk (26,27). Lemons et al. (24) have suggested that thehigher protein content of preterm milk may represent a prolonged colostral phase inpremature mothers who are establishing lactation by artificial means during periodsof stress.

It may thus be that some of the protein which makes up the higher concentrationof preterm milk is not nutritionally available to the infant. This is supported by thefact that plasma threonine concentration is not higher in infants receiving pretermmilk containing more total milk proteins, especially more of the threonine-rich secre-tory IgA, than pooled banked human milk (Fig. 12) (28).

SUMMARY

1. Although total protein concentration is high in colostrum and transitional humanmilk, this is due primarily to an increased concentration of the protective proteinssecretory IgA, lactoferrin, and lysozyme, which may not be fully absorbed and arethus partly unavailable to the infant as a protein or amino acid source.

2. The nutritionally available true protein content of mature human milk may beas low as 8.5 g/liter and the whey/casein ratio of those proteins is about 50:50 andthus not whey-predominant. The amino acid profile of the nutritionally availableproteins of human milk does not change much during lactation. Because of the vari-ability in digestion and absorption of human milk proteins, the amino acid composi-tion differs in the hydrolyzed and absorbed protein. Plasma amino acid profiles ininfants must thus be determined when improvements of formulations are tested. Thediscrepancy between threonine concentration of human milk and in the plasma ininfants fed human milk is a typical example of this.

3. For a period of some weeks postnatally the breast milk produced by mothersof preterm infants has a higher total nitrogen content than the milk of mothers ofterm infants. This may reflect a prolonged colostral phase since the increased nitrogencontent is due mainly to an increased concentration of secretory IgA and lactoferrin.The nutritional advantage of preterm milk needs to be studied more thoroughly.

REFERENCES

1. Linzell JL. Milk yield, energy loss in milk and mammary gland weight in different species. Dairy SciAbstr 1972; 34: 351-60.


2. Bemhart FW. Correlation between growth rate of the suckling of various species and the percentageof total calories from protein in the milk. Nature 1961; 191: 358-60.

3. Palmiter RD. What regulates lactose content in milk? Nature 1969; 221: 912-4.4. Geigy scientific tables. Basel: Ciba-Geigy, 1970.5. Harzer G, Bindels JG. Main compositional criteria of human milk and their implications in early

infancy. In: Xanthou M, ed. New aspects of nutrition in pregnancy, infancy and prematurity. Amster-dam: Elsevier Scientific Publishers, 1987; 83-94.

6. Kunz C. Lonnerdal B. Re-evaluation of the whey protein/casein ratio of human milk. Acta PaediatrScand\992;i\: 107-12.

7. Hambraeus L, Fransson G. Lonnerdal B. Nutritional availability of breast milk protein. Lancet 1984;ii: 167.

8. Raiha NCR. Nutritional proteins in milk and the protein requirement of normal infants. Pediatrics1985; 75: 136-41.

9. Butte NF, Goldblum RM. Fehl LM. et al. Daily ingestion of immunologic components in humanmilk during the first four months of life. Acta Paediatr Scand 1984; 73: 296-9.

10. Lindh E. Increased resistance of immunoglobulin A dimers to proteolytic degradation after bindingof the secretory component. J Immunol 1974; 114: 284-9.

11. Davidson LA, Donovan SM, Lonnerdal B. Atkinson ST. Excretion of human milk proteins by termand premature infants. In: Atkinson ST. Lonnerdal B, eds. Protein and non-protein nitrogen in humanmilk. Boca Raton, FL: CRC Press, 1989; 161-72.

12. Davidson LA, Lonnerdal B. Persistence of human milk proteins in the breast-fed infant. Acta PaediatrScand 1987; 76: 733-40.

13. Prentice A, Ewing G, Roberts SB, et al. Nutritional role of breast-milk igA and lactoferrin. ActaPaediatr Scand 1987; 76: 592-8.

14. Picone TA, Benson JD, Maclean WC, Sauls HS. Amino acid metabolism in human milk-fed andformula-fed term infants. In: Atkinson ST. Lonnerdal B, eds. Protein and non-protein nitrogen inhuman milk. Boca Raton, FL: CRC Press, 1989; 173-86.

15. Richards P. Nutritional potential of nitrogen recycling in man. Am J Clin Nutr 1972; 25: 615-9.16. Fomon SJ, Bier OM, Matthews DE, et al. Bioavailability of dietary urea N in the breast-fed infant.

J Pediatr 1988; 113: 515-20.17. Heine W, Tiess M, Stolpe HJ, et al. Urea utilization by the intestinal flora of infants fed mother's

milk and a formula diet, as measured with the 15N-tracer technique. J Pediatr Gastroenterol Nutr1984; 3: 709-12.

18. WHO. Energy and protein requirements. Technical Report Series No 724. Geneva: FAO/WHO/UNU, 1985:64-112.

19. Janas LM, Picciano MF. Quantities of amino acids ingested by human milk-fed infants. J Pediatr1986; 109: 802-7.

20. Picone TA, Benson JD, Moro G, et al. Growth, serum biochemistries, and amino acids of terminfants fed formulas with amino acid and protein concentrations similar to human milk. J PediatrGastroenterol Nutr 1989; 9: 351-60.

21. Atkinson SA, Bryan MH, Anderson GH. Human milk: difference in nitrogen concentration in milkfrom mothers of term and premature infants. J Pediatr 1978; 93: 67-9.

22. Schanler RJ, Oh W. Composition of breast milk obtained from mothers of premature infants ascompared to breast milk obtained from donors. J Pediatr 1980; 96: 679-81.

23. Gross SJ, David RJ, Bauman L, Tomarelli RM. Nutritional composition of milk produced by mothersdelivering preterm. J Pediatr 1980; 96: 641-4.

24. Lemons JA, Moye L, Hall D, Simmons M. Differences in the composition of preterm and term milkduring early lactation. Pediatr Res 1982; 16: 113-7.

25. Sann L, Bienvenu F, Lahet C, et al. Comparison of the composition of breast milk from mothers ofterm and preterm infants. Acta Paediatr Scand 1981; 70: 115-6.

26. Gross SJ, Buckley RH, Wakil SS, et al. Elevated IgA concentration in milk produced by mothersdelivered of preterm infants. J Pediatr 1981; 99: 389-93.

27. Goldman AS, Garza C, Nichols B, et al. Effects of prematurity on the immunologic system in humanmilk. J Pediatr 1982; 101: 901-5.

28. Atkinson SA, Hanning RM. Amino acid metabolism and requirements of the premature infant: ishuman milk the "gold standard"? In: Atkinson SA, Lonnerdal B, eds. Protein and non-proteinnitrogen in human milk. Boca Raton, FL: CRC Press, 1989; 187-209.


DISCUSSION FOLLOWING THE PRESENTATION OF DR. RAIHA

Dr. Guesry: When you say that the entire 15 to 18 g per liter of protein that infant formulamanufacturers put in the formula is available, it is a very nice compliment to the infant milkproducers but we couldn't accept it as true. Bo Lonnerdal yesterday showed that on averageonly 75-80% was available, depending on the processing. We always take this margin ofunavailable protein into account because it is probably less dangerous to incorporate a littlebit more than a little bit less. Second, I should like to thank you for presenting Atkinson'sdata showing the protein content of the milk of mothers of preterm babies compared withmothers of term babies. Various people have advocated the use of preterm milk for feedingpremature babies, but it should be borne in mind from Atkinson's data that by 10 days ofage preterm milk has the same protein content as that of normal term milk. At this age ababy born weighing 1 kg or 1.2 kg is still very small and certainly needs a higher proteinintake than that needed by a term baby.

Dr. Raiha: I agree that the preterm infant of less than 1500 g needs more protein than canbe provided by his mother's milk.

Dr. Cooper: There are very limited data on the amino acid profiles in fetal life. Yousuggested that in the preterm baby fed human milk the amino acid profile should be the modelfor preterm formulas. What I am suggesting is that we should perhaps be looking more atthe normal situation in the fetus, let's say at 26-28 weeks, rather than at the preterm babyfed human milk.

Dr. Raiha: This is a much debated question. We do not have and probably never willhave a gold standard for the plasma amino acid profile in the preterm baby. However, someyears ago we studied preterm babies fed on various amounts of exclusively human milkprotein obtained by supplementing human milk with altered filtrated human milk protein.When very low birthweight infants achieve an intake of about 3.5-3.6 g/kg/d of this protein,their growth rate is maximal. We used the plasma amino acid levels obtained at this intakeas a baseline and compared them with the profiles of formula-fed infants, and with fetal bloodamino acid profiles, umbilical cord amino acid profiles, and amino acid profiles from normalbreast-fed term infants. We found that the closest match was with the breast-fed term infantwho was growing normally. So my view is that the gold standard for the preterm infant shouldbe the amino acid profile of the breast-fed term infant. I don't think we should use the fetalvalues because the fetus in utero is physiologically quite different from the preterm baby.

Dr. Marini: The problem of reference amino acid values is the same for total parenteralnutrition. Some solutions for TPN use mature human milk as the reference composition. Asyou demonstrated, this is by no means the same as the plasma amino acid profile in the infantfed human milk. Cord blood or fetal blood obtained at about 32-24 weeks of gestation wouldprobably be a better reference standard.

Dr. Rassin: Amino acid patterns in fetal blood or cord blood can't be reproduced byfeeding formulas or breast milk because they reflect in part the more aerobic metabolism ofthe baby and several features of the profile are just not matchable. For example, cord bloodhas a much higher glutamate/glutamine ratio than you would ever see in a formula-fed orbreast-fed baby. To feed a formula that would reproduce this ratio would be impossiblebecause you would have to give far too much glutamate. Also, the lysine composition offetal and cord blood is approximately 4 to 5 times higher than you would ever see in a formula-fed or breast-fed baby. We would have to make some very strange changes to formulas toreproduce those profiles, which reflect the transport mechanism of the placenta. For several


amino acids the placenta produces maternal-to-fetal gradients of three- to fourfold, while forothers there is almost no gradient, for example cystine. We are not going to produce theseeffects with any kind of parenteral feeding regimen without causing possible toxicity in theinfant. So I think to suggest that cord blood or fetal blood is a profile to match is just notappropriate.

Dr. Marini: Battaglia carried out several studies in the normal human fetus and in the fetuswith intrauterine growth retardation. He found that serine was quite high in the human fetus.Do you have an explanation?

Dr. Rassin: Each individual amino acid has a transport mechanism in the placenta andthese transports probably determine the fetal amino acid concentrations.

Dr. Marini: But why should the fetus have more serine? Does it depend on muscle metab-olism?

Dr. Rassin: I don't think we are in a position to answer why there are particular concentra-tions of individual amino acids. We just know that under certain circumstances and withparticular transport mechanisms you develop certain concentrations of these compounds.We have tried very hard to determine the implications of this, but we are a long way fromfinding the answers.

Dr. Pettifor: The issue then is why do we use any amino acid profile as a gold standardfor formula-fed or breast-fed infants. If you believe that we can't use fetal amino acid profilesas the gold standard, why should we use postnatal amino acid profiles either?

Dr. Raiha: I can only say that it is because we have done studies on them. I still thinkthat even for the preterm infant who is orally fed the ideal protein is human milk proteinbecause it is of the same species. When we feed very low birth weight babies with adequateamounts of human milk protein, in other words mother's own milk supplemented with ultrafil-trated human milk protein, such that we achieve a growth rate that is actually higher thanthe intrauterine one, we find that the blood amino acid profiles are close to what you findin breast-fed term babies. That is the only reason for thinking that it should be the norm.

Dr. Kashyap: In most studies in preterm infants, the aim has been to try to keep the aminoacid concentrations at least in the safe range. In order to do this, people have come up withstandards. Perhaps the aim should be to ensure that the concentrations are less than whatthe baby is exposed to in utero, on the assumption that they will then be in the safe range.

Dr. Pandit: There is evidence that a good deal of urea is secreted in breast milk. Can thisbe used for protein building?

Dr. Raiha: From studies done so far it seems that under normal conditions, when the babyis getting an adequate supply of nutritional protein, urea is not used extensively for proteinsynthesis. Under pathological conditions, if the baby is starved or when he has an infection,the situation may be different.

Dr. Heine: It is really interesting that human milk contains such a high amount of ureaand the question is for what purpose this high urea concentration is excreted in the milk.Urea is normally an end product which presents a metabolic load for the baby. We carriedout some investigations a few years ago in which we fed [15N]urea together with human milk.We found variable utilization of the urea nitrogen for protein synthesis in the baby. Thelowest utilization was 16% of the administered urea and the highest was 60% (1). WhenFomon repeated these investigations some years later he found that only 13% of the ureanitrogen was used in healthy, well-growing babies. We believe that the variable utilizationhas something to do with catch-up growth and with a low supply of protein in the baby'sdiet. It is quite clear, however, that urea is utilized for protein synthesis, since the I5N labelingwas identifiable in the plasma proteins.


Dr. Pandit: Does this mean that we should consider this urea as a part of the utilizableprotein in the theoretical figure that was presented?

Dr. Heine: This is in part true.Dr. Fazzolari: Could you tell me if the diet of the mother influences the quality and quantity

of human milk protein?Dr. Raih: As far as we know, breast milk protein composition is very resistant to change.

The fat content is most likely to be affected by maternal diet, and after this the volume ofthe milk. The quality of the protein will probably not be much affected, at least from studiespublished so far.

Dr. Fazzolari: The reason for my question is that it has been shown that undernourishedlactating mothers have very low taurine in their plasma amino acid profile.

Dr. Rassin: Please bear in mind that the taurine is not found in protein. It is much moreresponsive to diet than is protein. There are several studies showing that vegetarians will getalmost no taurine in their diet and have low taurine in their milk. You can't really use thatas a marker of protein response.

Dr. Guesry: Philippe Hennart did a study in eastern Zaire 3-4 years ago in severely mal-nourished mothers in which he analyzed their milk. As Niels Raiha has just said, it wasmainly the volume of milk that was severely decreased (2). The malnourished mother couldnot produce more than about 250-400 ml of milk per day, but the protein quality was quitewell preserved. It is also true that the fat was modified.

I should like to comment on Dr. Raiha's recommendation that infant food manufacturersuse transgenic cows to produce so-called human milk. He rightly says that this would bequite expensive, not only because of the technology involved in creating such transgenicanimals, but also because purification and separation of the human protein from the otherproteins produced by the cow will cost a lot of money. However, it is not the price that ismost likely to prevent us from doing it; it is more likely to be the law, since it is currentlyforbidden to give milk from a transgenic animal. It may also prove difficult to get the motherto accept the use of such milk. Public opinion is not yet ready to use transgenic milk forinfaftt feeding. Progress in this will be made initially by the drug industry. Almost everydiabetic now uses insulin produced by genetic engineering and it is accepted because it islife-saving and because it is a drug.

Dr. Raiha: This is what I said really. I personally think that it would be much wiser touse all that money to teach mothers to breast feed, but of course in some countries this isdifficult. I can just say that in Sweden and Finland, almost 100% of mothers are breast feeding,at least when they go home from the maternity ward, and even at 6 months more than 40%are still breast feeding. I can also say that in Sweden and Finland there is not a single pretermbaby who has not been fed exclusively on human milk. We would consider it unethical togive formula.

Dr. Marini: How do you manage if you have a very sick mother? Banked milk is not thesame as milk coming from the baby's own mother. For example, if you don't use fresh humanmilk it is quite impossible to protect from infection.

Dr. Househam: One of the most important causes of low birthweight in South Africa, asin many developing countries, is fetal growth retardation. What information is available onthe composition of the breast milk in such cases?

Dr. Lonnerdal: It seems unlikely that there would be any specific adaptation of themother's milk that would benefit the infant in such cases, but what happened during gestationmay of course affect lactation. Poor nutrition may have been one factor, of course, andanother could be infection. We have done a study in Lima, Peru, looking at women who are


sick during lactation, the reason being that infection is so common in developing countriesthat its effects on lactation need to be examined. We found there was a higher protein contentin the milk of these women, probably due to the increase in the acute-phase reactants thatwe see both in plasma and in some cases in breast milk. Perhaps the reasons for fetal growthretardation may still be present during lactation and therefore affect the milk composition.

Dr. Pettifor: What are the factors that influence the changing concentrations of IgA, lyso-syme, and lactoferrin in breast milk?

Dr. Marini: When you look at concentrations there are big differences between colostrumand mature milk, but when you look at quantities the difference is not so great. The amountof colostrum is probably around 300-350 ml/d. During mature lactation, the mother canproduce almost a liter of milk a day, so the quantity of IgA is not very much reduced interms of the amount consumed by the baby.

Dr. Lonnerdal: Although the volume consideration is important, there is more secretoryIgA early on, even taking volume into account, than there is later. We need to consider milkprotein gene expression. We know too little about this at present. We know in general thathormonal factors and nutritional factors will affect milk protein gene expression but we knowlittle about how nutrition and stress and all the other factors that can affect the hormonalprofile will affect milk protein gene expression.

Dr. Rdihd: I still think that we have missed an important issue in this whole context andthat is the question of the current protein content in standard formulas for normal babies.Nobody has discussed this. It is my personal feeling that most formulas in most countrieshave too much protein.

REFERENCES

1. Heine W, Tiess M, Wutzke KD. I5N tracer investigations of the physiological availability of ureanitrogen in mother's milk. Ada Paediatr Scand 1986; 639-663.

2. Hennart PhF, Brasseur DJ, Delogne-Desnoeck JB, et al. Lysozyme, lactoferrin, and secretory immuno-globulin A content in breast milk: influence of duration of lactation, nutrition status, prolactin status,and parity of mother. Am J Clin Nutr 1991; 53: 32-9. Errata in Am J Clin Nutr 1991; 53: 988.

Protein Content of Human Milk, from Colostrum to Mature Milk

Documents