Main

The immune response is generated following the interaction between T cells, B cells, NK cells, and a number of other accessory cells. Phenotyping of normal peripheral blood cells by flow cytometry has shown considerable age-related variation in lymphocyte subsets in terms of both relative proportions and absolute size(1–4). The percentages of T helper/inducer cells and B cells, as well as the absolute lymphocyte counts, are higher in infancy and decrease with age. On the other hand, the percentages of T suppressor/cytotoxic cells and NK cells are lower in infancy and increase with age. Hence, separate reference ranges are required for different age brackets within the pediatric population(3,5).

The repeated demonstration of a reduction in the incidence of infections and illness in breast-fed infants compared with formula-fed infants leads to speculation that breast milk has an immunoregulatory role that inevitably is absent in those infants consuming only formula. However, there are few data available from well-defined groups of breast-fed and formula-fed infants with regard to their immune status as reflected by lymphocyte immunophenotypic values(6,7).

The aim of this study was to investigate the difference in specific lymphocyte subsets between breast-fed and formula-fed 6-mo-old infants. The percentage and number of the total peripheral blood mononuclear cells expressing each particular cell-surface antigen are reported.

METHODS

Subjects. One hundred and sixty-one healthy 6-mo-old infants, born at term, were enrolled into a study investigating the weaning diet. Infants were recruited from immunization clinics in the metropolitan area of Adelaide. Thirteen infants were excluded because of missed appointments or because of insufficiency of blood collected for analysis. All infants in the study were an appropriate weight for age and had no known history of protein intolerances or allergies. Infants were eligible for entry into the breast-fed group if they received less than 120 mL of formula (or cow's milk) a week, and they were eligible for entry into the formula-fed group if they had been breast-fed for less than 4 wk after birth. Written informed consent was obtained from all participating parents, in adherence with the guidelines of the Flinders Medical Centre Committee on Clinical Investigation (Ethics).

Flow cytometry. Blood samples (2 mL) were taken by venepuncture and stored with lithium heparin at room temperature until staining. All samples were stained within 2 h of collection. Packed cells from whole blood (50 µL) were incubated with fluorochrome-labeled monoclonal antibodies (Immunocytometry Systems SimulSET reagents, Becton Dickinson, San Jose, CA), followed by erythrocyte lysis. Antibodies against CD3 (defining T lymphocytes), CD19 (total B lymphocytes), CD4 (defining a subset that includes cells with helper/inducer function), CD8 (a subset that includes cells with suppressor/cytotoxic function), and CD16 + CD56 (defining NK cells in the absence of CD3 co-expression) were used in this study. We analyzed washed lymphocytes by two-color direct immunofluorescence on a FACScan instrument (Becton Dickinson), using standard instrument parameters and Lysis II software. Lymphocytes were distinguished on the basis of dual scatter parameters and were manually gated to include only lymphocytes in the analysis. The purity of the lymphocyte gate was checked by use of CD14/CD45 double staining and was regarded as adequate if the gate included at least 95% of all lymphocytes and contained less than 5% contamination with monocytes, granulocytes, or debris (SimulSET software, Becton-Dickinson).

A portion of each blood sample was stored in EDTA, and from this the total lymphocyte count was determined with a Coulter Counter, model S+6. The absolute size of each lymphocyte subset was calculated from the relative size of the lymphocyte subset and the total lymphocyte count. The total lymphocyte count could not be obtained from the blood samples of six breast-fed infants and six formula-fed infants because of insufficient sample volume.

Data analysis. The relative size of each lymphocyte subset was analyzed by independent samples t test to determine differences between breast-fed and formula-fed infants, inasmuch as these data were distributed normally. The data are expressed both as mean ± SEM and as median values and the 5th to 95th percentile ranges.

Because the distributions of the absolute values were asymmetric, significant differences between feeding groups were calculated by Mann-Whitney U test, and data are expressed as median values and the 5th to 95th percentile ranges. Subject characteristic variables were analyzed by independent t test to determine differences between breast and formula-fed infants. The association between length of breast feeding and the relative size of lymphocyte subpopulations was determined by one-way analysis of variance and posthoc analysis by Newman-Keuls.

RESULTS

Characteristics of the subjects in the study are shown in Table 1. At venepuncture, the age, length, and head circumference of infants in the breast-fed group was the same as that of infants in the formula-fed group. Breast-fed infants weighed slightly less than their formula-fed counterparts, which is in accordance with previously published data(8). Seventy-two percent of infants in the breast-fed group and 73% of infants in the formula-fed group had received their most recent immunizations [diphtheria/tetanus/pertussis, poliomyelitis (oral), hemophilus influenzae type B] more than 4 wk before the blood sample was taken for this study, and 85% of the infants in each group were having solids daily at the time of venepuncture.

Table 1 Subject characteristics
Full size table

The relative sizes of the T lymphocyte, B lymphocyte, and NK cell subpopulations are shown in Figure 1 and Table 2. The mean percentage of T and B lymphocytes in peripheral blood of 6-mo-old infants was the same, regardless of feeding regimen (65.6% vs 67.9% and 24.1% vs 24.5%, respectively). However, the relative frequency of NK cells was greater in breast-fed than in formula-fed infants (9.7% vs 7.1%; p < 0.001).

Figure 1
figure 1

Relative sizes (mean % ± SEM) of the T, B, and NK cell subpopulations in peripheral blood of 6-mo-old breast-fed and formula-fed infants. The number of individuals per feeding group is indicated.

Full size image
Table 2 Relative (%) and absolute (absolute count × 109/L) size of lymphocyte subpopulations in peripheral blood of 6-mo-old breast-fed and formula-fed infants
Full size table

The relative sizes of the T helper/inducer (CD4+) and T cytotoxic/suppressor (CD8+) cell subpopulations and CD4:CD8 ratio in peripheral blood of infants are shown in Figure 2 and Table 2. The mean percentage of cells expressing CD4 was lower in breast-fed infants than in formula-fed infants (47.3% vs 50.9%; p < 0.005), and that of cells expressing CD8 was greater (18.0% vs 16.4%; p < 0.05). As a result, the CD4:CD8 ratio in breast-fed infants was lower than that in formula-fed infants (2.8 vs 3.3; p < 0.005).

Figure 2
figure 2

Relative sizes (mean % ± SEM) of the T helper/inducer (CD4+) and T cytotoxic/suppressor (CD8+) cell subpopulations and CD4:CD8 ratio in peripheral blood of 6-mo-old breast-fed and formula-fed infants. The number of individuals per feeding group is indicated.

Full size image

The absolute size of lymphocyte subpopulations (expressed as absolute count × 109/L, median and 5th to 95th percentiles) in peripheral blood of 6-mo-old breast-fed and formula-fed infants is shown in Table 2. Differences between lymphocyte subpopulations are not as apparent if absolute values are considered, rather than the relative numbers, although the trends remain the same. The absolute count of total lymphocytes and the subsets T, B, and CD8+ were the same for the two populations of infants. However, breast-fed infants had fewer CD4+ T cells (p < 0.05) and a greater number of NK cells (p < 0.01) than the age-matched formula-fed infants.

Because of the differential between breast and formula-fed infants, data from the 69 infants enrolled in the formula-fed group were divided posthoc into three groups consisting of infants who were never breast-fed (n = 19), infants breast-fed for less than or equal to 2 wk (n = 24) and those breast-fed for between 2 and 4 wk (n = 26). Analysis of variance between these three groups of infants and those breast-fed for 26 wk suggested that the length of breast feeding had an effect on the percentage of T lymphocytes expressing CD4 and CD8 and, hence, the CD4:CD8 ratio (Fig. 3). However, within the formula-fed cohort of infants, the length of breast feeding had no apparent effect on the relative number of T and B cells (data not shown) or NK cells (Fig. 3).

Figure 3
figure 3

Relative sizes (mean % ± SEM) of the T helper/inducer (CD4+), T cytotoxic/suppressor (CD8+) and NK cell subpopulations, and CD4:CD8 ratio in peripheral blood of 6-mo-old infants never breast fed, or breast fed for ≤2 wk, 2-4 wk, or 26 wk after birth. The number of individuals per feeding group is indicated. Within each cell phenotype, there is no difference between groups labeled with the same letter; however, different letters indicate significant differences between feeding groups (p < 0.05).

Full size image

DISCUSSION

It is known that both the relative and absolute sizes of lymphocyte populations vary with age during childhood because of the maturation and expansion of the immune system during this period(1–4). These and other studies have repeatedly shown a wide scatter of values at all age ranges for most lymphocyte subsets. The need for separate reference values at discrete age ranges within the pediatric population has been recognized. However, there has been little consideration of the contributions of confounding factors that may influence these values. Because distributions were symmetric for the relative proportions of lymphocyte subsets, we have presented the data graphically as mean ± SEM. However, because most other authors have described their data as median values and percentiles, we have included this descriptor in Table 2 to facilitate comparisons between our data and previously published works by others. Our infant values fall within the ranges reported elsewhere for the appropriate age group but have shown that there are also significant differences in both the relative number (percentage) and the absolute size of several key lymphocyte subpopulations in the peripheral blood of breast-fed infants compared with formula-fed infants at the same age. The data for T helper/inducer, T suppressor/cytotoxic, and NK cells suggest the accelerated development of these subsets of lymphocytes in the breast-fed infants, inasmuch as median values fall between that of the formula-fed infants in this study and the reference values for older children or adults(1,4). The effect appeared to be independent of age, immunization history, and establishment of weaning foods.

The functional significance of these changes is unclear, because the cell surface markers used in this study do not provide information regarding the level of activity of these cells. However, it may be important to consider that NK cell activity can be significantly enhanced by the lactoferrin(9) and dietary nucleotides(10) that are abundant in breast milk. It is not unreasonable to suggest that these and other components of breast milk could have a role in boosting the NK activity in young infants to promote antimicrobial immunity.

The results presented in this study are supported in part by the findings of Carver et al.(7) who reported that breast-fed infants had a lower percentage of CD4+ cells than did formula-fed infants of a similar age. They also reported lower percentages of CD3+ cells and higher percentages of CD19+ cells in breast-fed infants, compared with formula-fed infants. However, their sample sizes were small (n = 7 and 11, respectively), and the feeding history of the formula-fed infants was unclear.

Although posthoc analyses have limited ability to discriminate true effects, an analysis of our data suggests that early feeding of breast milk and the duration of such breast feeding could have a lasting impact on the maturation of the immune system, as defined by immunophenotypic values of peripheral blood lymphocytes. For example, lower CD4:CD8 ratios were seen in lymphocyte populations from infants who were breast-fed for as little as 2-4 wk after birth, compared with exclusively formula-fed infants, whereas changes in the proportion of NK cells appeared to follow a different time course. There is a clear need to expand the range of ages at which blood samples are taken, for the purpose of obtaining a clearer understanding of the effects of breast milk on the maturing immune system.

The results presented here suggest a relationship between breast milk feeding and the postnatal development of the immune system of the recipient infant. The immunophenotypic differences between infants whose primary source of nutrition was breast milk compared with those who were breast-fed for less than 4 wk are consistent with reported age-related changes, suggesting greater maturity in the development of the immune system of the breast-fed infants. To determine whether these changes are associated with the clinical advantage documented in breast-fed infants will require further investigation.