Main

The causes underlying increased susceptibility of healthy children to infection are not fully understood(1). Numerous studies have demonstrated the relative incompetence of the immature immune system, particularly of humoral responses. IgG (especially IgG2) and IgA production are not fully developed in childhood, and antibody responses to bacterial polysaccharides are inadequate(2). This is reflected in the increased susceptibility of children to particular pathogens, such as Haemophilus influenzae type B, Streptococcus pneumoniae, and other encapsulated organisms(1, 2). Studies addressing this question have demonstrated that inadequate T cell help may be crucial(3). Although cellular immune responses are already functional in the neonate, increased susceptibility to viral and fungal infections suggests that cell-mediated immunity is still not fully developed in childhood(1), a finding corroborated by recent studies demonstrating reduced antigen-specific T cell precursor frequencies in neonates(4). Although numerous factors affect maturation of the immune response, a crucial role for regulatory factors, affecting function and regulation of both humoral and cellular responses, is highly likely. Cytokines function as such regulatory factors(5), and if produced insufficiently or inadequately, could substantially influence the resulting immune response. Although cytokine production has frequently been evaluated in neonates and in children with various diseases(1, 6, 7), surprisingly few comparative studies of cytokine production in healthy children and adults have been conducted. We present our findings of differences in antigen and mitogen-stimulated cytokine production between adults and children unexpectedly encountered during a study of a patient group that necessitated both children and adults as healthy control subjects(7, 8).

METHODS

Subjects. Ten healthy children (six boys and four girls) and 10 adults were studied (mean age 9 and 32 y, ranges 2-15 and 20-44, respectively). The children were undergoing general anesthesia for minor surgery for noninfectious causes (e.g. ingrown toenail and epispadia). All children were otherwise healthy by clinical and routine laboratory criteria, were previously routinely vaccinated against diphtheria, tetanus, pertussis, measles, and had no history of serious or recurrent infections. Adults were healthy, volunteer laboratory personnel. Ethical approval was granted by the Newcastle Ethical Committee, and informed consent of patients or parents was obtained in all cases.

Antigens and mitogen. Two types of Candida antigens were used (kind gifts from Dr L. Morgan, Newcastle General Hospital): a crude preparation of sonicated Candida albicans (CS) and a dialyzed protein precipitate (CP), diluted 1:10 000; PPS (Pneumovax II Vaccine, Merck), 1:5000; TT (tetanus vaccine BP in simple solution, Wellcome), 1:2000; and PWM(Life Technologies, Inc.), 1:1000. Previous exposure to the stimulating antigens was confirmed in proliferation assays for CS and TT (data not shown) and positive serum antibody titers for CS, TT, and PPS(7): PBMC from all subjects proliferated to Candida antigens, two children did not respond in culture to TT but had positive antibody titers, all subjects had positive antibody titers to CS, TT, and PPS, except one child (aged 2) who had borderline antibody titers to PPS.

Cytokine production. In vitro production of IL-2, IL-6, IL-4, IFN-γ, and sIL-6R was performed as previously described(7). Briefly, 1 × 106/mL, PBMC with or without stimulating antigen/mitogen were incubated in culture medium for 24 h in a humidified, 5% CO2 atmosphere, and supernatants were assayed. Optimal concentrations of mitogen, all antigens, cell concentration, and incubation time were all determined in extensive previous experiments (data not shown).

Cytokine assays. IL-2 and IL-6 supernatant levels were determined in bioassays with CTLL-2 and B9 cell lines, respectively. Specificity of assays was confirmed by antibody neutralization studies. IL-4 and IFN-γ supernatant levels, plasma IL-6 levels, and confirmation of IL-6 bioassay levels in selected supernatants as well as sIL-6R levels in supernatants and plasma were determined by ELISA (AMS Biotechnology Oxon UK, Pharmingen Cambridge Biosciences UK and R&D Abingdon UK, respectively). Cytokine levels (pg/mL) were calculated from standard curves calibrated to reference National Institute for Biological Standards and Controls (NIBSC) reagents(7).

Data analysis. Results are presented as individual and median group values with SEM. Statistical significance was calculated with the Mann-Whitney rank sum test, correlation with Spearman's rank correlation coefficient, and analysis of variance. All tests were performed on the statistical software program Minitab 9/w.

RESULTS

IL-2 production. Figure 1a demonstrates antigen and mitogen-stimulated IL-2 production in children and adults. Most children produced lower levels of IL-2 than did adults in response to all antigens used. This was most evident is response to CS (26 ± 12 compared with 74 ± 14 pg/mL, p < 0.03). Some children failed to produce IL-2 to certain antigens (two children did not respond to CS, four to CP and TT), which was not seen in any adult. All children and adults produced significant amounts of IL-2 upon PWM stimulation, although median levels were significantly lower in children (371 ± 202 compared with 1522 ± 299 pg/mL, p < 0.017)(Fig. 1a and Table 1). Several children produced levels of IL-2 that were comparable to adult values (Fig. 1a). As expected, PPS did not stimulate IL-2 production.

Figure 1
figure 1

Production of IL-2 and IFN-γ after stimulation with various antigens and mitogen. PBMC from healthy children and adults were stimulated in culture and supernatants assessed for IL-2 (Fig 1a) and IFN-γ (Fig 1b) levels. Results are presented as levels obtained for each individual and median group values (- - - - -). CS =Candida sonicate, CP = Candid protein, PPS = pneumococcal polysaccharide, TT = tetanus toxoid, PWM = pokeweed mitogen, Bg = background, unstimulated cytokine production.

Full size image
Table 1 Comparative levels of cytokine production following PWM stimulation in individual healthy children
Full size table

IFN-γ production. Median levels of IFN-γ produced by children in response to CS and CP were higher than those in adults, but not significantly (Fig. 1b). This was not seen in response to PPS and TT. However, median levels of IFN-γ produced after PWM stimulation were lower in children, although again not significantly (2650± 860 compared with 4650 ± 937 pg/mL)(Fig. 1b and Table 1).

IL-4 production. Production of this cytokine in response to stimulation with all antigens was either low or absent in both adults and children (Fig. 2a). However, PWM did stimulate IL-4 production in all subjects. Median values in children were significantly lower than in adults (12 ± 4 and 24 ± 8 pg/mL, p < 0.02)(Fig. 2a and Table 1).

Figure 2
figure 2

Production of IL-4 and IL-6 after stimulation with various antigens and mitogen. PBMC from healthy children and adults were stimulated in culture, and supernatants were assessed for IL-4 (a) and IL-6 (b) levels. Results are presented as levels obtained for each individual and median group values (- - - - -). Bg, background, unstimulated cytokine production.

Full size image

IL-6 production. IL-6 production in response to antigen stimulation was higher in children compared with adults for all antigens tested (Fig. 2b). Levels of IL-6 produced were significantly higher upon stimulation with PPS (99 ± 25 and 36 ± 16 pg/mL, p < 0.04) and TT (80 ± 15 and 20 ± 14 pg/mL, p < 0.05). However, children produced significantly lower levels of IL-6 compared with adults when stimulated with PWM (1680 ± 550 and 3916 ± 674 pg/mL, p < 0.05). All individuals were tested for plasma IL-6, and not one was found to have levels above background.

sIL-6R levels. All subjects tested had substantial plasma levels of sIL-6R, ranging from 50 to 108 μg/mL, with negligible differences between children and adults (not shown). Supernatant levels of sIL-6R were low but detectable with all of the antigens and mitogens tested, and no marked differences between children and adults were noted (not shown). Supernatant sIL-6R levels did not correlate with levels of IL-6 produced.

Correlation between levels of different cytokines produced. In children but not in adults levels of IL-2 and IFN-γ produced after stimulation with different antigens and particularly PWM correlated positively to a significant degree (r = 0.802, p = 0.005,Table 1). There was no correlation between levels of IL-4 and IL-6, between IL-2 and either IL-4 or IL-6, or between age and levels of cytokines produced.

DISCUSSION

Our results demonstrate that healthy children differed markedly from adults in cytokine production after stimulation with antigens and mitogen. Overall they had a markedly lower capacity to produce all four cytokines tested (IL-2, IFN-γ, IL-4, and IL-6) when maximally stimulated with mitogen. However, when stimulated with antigens, median levels of IL-2 and IL-4 remained low; levels of IFN-γ were not significantly different compared with adults, although median values were higher in response to two of four antigens tested, whereas levels of IL-6 were higher compared with adults in response to all antigens, albeit reaching statistical significance for two of the four antigens tested. These data suggest that healthy children do not merely have decreased cytokine production but also an altered cytokine response, which could be the result of an immature immune response and/or the result of reduced antigenic experience compared with adults. Previous studies, carried out in neonates but not in older children and assessing only mitogen not antigen stimulation, demonstrated normal IL-2(9) or decreased IL-2(4) and IL-6(1) and markedly decreased IFN-γ and IL-4 production(10). Decreased IL-2 production was shown to be associated with reduced antigen-specific T cell precursor frequencies(4) and reduced protein kinase C isoform-β expression(11). Our data on mitogen-stimulated cytokine production are partially in keeping with these findings (reduced IL-4 and IL-6, markedly reduced IL-2, mildly reduced IFN-γ). The discrepancies are likely the result of different populations studied, i.e. neonates compared with older children, keeping in mind that neonates have an immune response that is less mature and still under the influence of placental/maternal humoral regulatory factors, making comparisons with responses in older children unreliable. One study looking at levels of IL-1β, IL-6, and tumor necrosis factor-α in nasopharyngeal secretions found higher levels in healthy children compared with adults(12), again suggesting differential cytokine responses in children. In our studies the markedly low levels of IL-2 found in children compared with adults could have resulted from high in vitro IL-2 receptor-mediated consumption, although the short incubation time (24 h) and identical culture conditions used for both groups would not favor this possibility.

Our data on antigen-stimulated cytokine production differ markedly from mitogen-stimulated production both in our own and other studies, suggesting that cytokine production in children depends crucially on the type of stimulus used and suggests that antigen-stimulated cytokine production may be more relevant to the situation in vivo than aggressive activation with mitogens. The underlying causes of this altered cytokine production in neonates and children are largely elusive, although a crucial factor is likely to be lack of exposure to antigen. This is supported by findings of a predominantly “naive” T cell phenotype in neonatal peripheral blood (CD45RA+ CD45RO- CD29low), which decreases with age(1, 4). However, other less well defined factors such as inappropriate antigen presentation and/or T cell help may be crucial and could influence the type and level of cytokines produced.

It is interesting that, although IL-2 production is decreased, IFN-γ production is often increased, in spite of the fact that both are Th1-type cytokines, i.e. produced by the same cell type, suggesting a shift in cytokine production. Increased IFN-γ levels could also originate from natural killer cells, although it would not be expected for natural killer cells to be stimulated in an antigen-specific manner. The negative correlation between IL-2 and IFN-γ production was not seen with IL-4 and IL-6, although both are Th2-type cytokines. This could be explained by the fact that, in the culture conditions used in our study, IL-6 is also largely produced by non-T cells (macrophages/monocytes and B lymphocytes) and increased IL-6 production may be reflecting inadequate regulation or control of these cells. Antigen-stimulated IL-4 production on the whole was very low, as has been previously reported(13) and may also account for absence of correlation with IL-6 levels. Levels of sIL-6R were not increased in supernatants in which high IL-6 levels were detected, suggesting that sIL-6Rs were not responsible for increased IL-6 production, as has been described(14). Failure to demonstrate increased sIL-6R levels was not the consequence of sIL-6R binding to IL-6, as the assay is insensitive to IL-6. Because experiments were carried out on unfractionated PBMC without determining the proportion of lymphocyte subsets and monocytes, low levels of cytokines produced could have resulted from low lymphocyte(cytokine-producing cell) numbers; however, all children tested were previously healthy with normal blood counts (i.e. a higher percentage and absolute lymphocyte count is normal in children compared with adults), rendering this possibility unlikely. Furthermore, as mentioned above, levels of some cytokines were found to be increased, arguing against this possibility. However, the fact that the proportion of naive versus mature T cells in peripheral blood decreases with age remains a factor that may have crucially influenced the type and amount of cytokines produced. It must also be borne in mind that cytokine production was evaluated on PBMCs, which represent a restricted compartment of the immune system and because of which conclusions must be subject to the same constraints that apply to all immunologic studies conducted on PBMC.

The effect of altered cytokine production could have far reaching consequences on both the humoral and cellular immune response. As cytokines influence the pattern (Th1/Th2) of other produced cytokines(15) as well as the immunoglobulin isotype secreted(16), changes in cytokine production may result in inadequate immune responses and impaired defense. Our previous findings in patients with chronic mucocutaneous candidiasis(7) suggest that altered cytokine production may underlie the susceptibility to fungal infections in these patients.

Taken together our findings suggest that the altered/immature cytokine production documented in neonates persists during childhood and is present in normal healthy children. As these findings could have both theoretical and practical consequences(17), further studies addressing larger numbers of children, different age groups, and a broader panel of cytokines merit attention.