Progressive polarization towards a T helper/cytotoxic type-1 cytokine pattern during age-dependent maturation of the immune response inversely correlates with CD30 cell expression and serum concentration

Abstract

In order to investigate the T cell cytokine profile during age-dependent maturation of the immune response, we evaluated the cytokine expression of CD4⁺ and CD8⁺ circulating cells by flow cytometric single-cell analysis after non-specific stimulation in vitro in different age groups of normal individuals, from cord blood to adulthood. Moreover, we correlated these lymphocyte cytokine patterns with the expression/release of CD30, a member of the tumour necrosis factor (TNF) receptor superfamily, which has been suggested to be related to the T helper/cytotoxic (Th(c))2-type immune responses, in order to verify this association in vivo, in non-pathological conditions. The results showed a progressive increase of circulating Th(c)1-type, interferon-gamma (IFN-γ)- and/or IL-2-producing T cells along with ageing and, conversely, a stable number, although higher than in cord blood samples, of CD4⁺/IL-4⁺ T cells in the post-natal groups. In addition, serum levels of soluble CD30 (sCD30) and numbers of circulating CD4⁺/CD30⁺ and CD8⁺/CD30⁺ T cells were significantly higher in children aged < 5 years in comparison with those found either in cord blood or in blood from both older children and adults. These data support the concept of a progressive polarization of the Th(c) cell cytokine profile towards the Th(c)1 pattern during age-dependent maturation of the immune response. Moreover, the peak of CD30 expression/release in early infancy before the Th(c)1 shifting occurs, although not associated with a significant increase of circulating IL-4⁺ T cells, raises the question of the possible relationship in vivo between CD30 and Th(c)2-type immune responses.

INTRODUCTION

The T cell compartment includes a variety of subpopulations characterized by different immunophenotypic and functional profiles which play specific roles in immune responses. The pattern of cytokine production is at present largely utilized, both in mice and humans, to define functionally polarized T cell subsets involved in different phases of the effector response. CD4⁺ T-helper (Th) and CD8⁺ T cytotoxic (Tc) cells producing interferon-gamma (IFN-γ), tumour necrosis factor-beta (TNF-β) and IL-2 are referred to as Th(c)1-type cells [1–3]. Cytokines produced by Th(c)1 cells activate macrophages and are involved in DTH reactions [1–3]. T cells producing cytokines mainly involved in antibody responses, such as IL-4, IL-5, IL-10 and IL-13, are identified as Th(c)2-type cells [1–3]. In addition, T lymphocytes expressing less homogeneous sets of cytokines are referred to as Th(c)0 cells [1–3].

Although the immune response is a complex phenomenon which should be considered as a whole, studies which address differential patterns of immune responses, through the characterization of Th(c)1, Th(c)2, or Th(c)0 cells, provide useful information which contributes to a better understanding of the pathogenesis of several diseases [1–3]. In this context, it appears of interest to investigate the patterns of cytokine production during the development of the immune system. A number of observations point to different cytokine patterns by T cell subsets in neonatal compared with adult mice. Newborn mice exposed to antigens eliciting Th1-type responses mainly develop type-2 immune responses in adult animals, a phenomenon which is associated with the acquisition of immune tolerance to these antigens [4–7]. Such a different pattern appears to be modulated by microenvironmental factors rather than to be dependent on intrinsic age-related differences of T cell subsets [4–7].

Most data regarding the identification of T cell subsets on the basis of their cytokine expression are derived from in vitro studies of long-term clones [8,9]. More recently, cytoplasmic cytokine detection at the single-cell level, by flow cytometry, has been used for this purpose [10–15]. Both approaches are currently being applied to investigate the type of T cell responses involved in normal and pathological conditions. In the present study we investigated, by cytoplasmic cytokine detection at the single-cell level, the expression of cytokines accounting for Th(c)1-, Th(c)2- or Th(c)0-type immune responses by T cells obtained from human cord blood (CB) lymphocytes and blood samples of healthy infants, young subjects, and adults. In addition, we correlated the figures observed in the various age groups with those of cell surface expression of CD30 and of soluble CD30 (sCD30) serum concentration.

The reason for this approach is that the expression of CD30 (a molecule belonging to the TNF receptor family [16,17]) by both murine and human T cells has been found to be IL-4-dependent [18,19], and preferentially associated with Th(c)2 and Th(c)0 rather than Th(c)1-type responses [20–23]. Our study would possibly provide further information on the correlation between CD30 and type of immune responses in vivo in a non-pathological condition such as ageing.

The results showed a progressive maturation of the Th(c)1 pattern along with ageing, but a steadily low production, although higher than in CB, of IL-4 in the post-natal life. The age-related Th(c)1 shifting appeared to be inversely correlated with CD30 expression/release, thus further supporting the concept that CD30 is usually not associated in vivo with Th(c)1-type immune responses.

SUBJECTS AND METHODS

Sample collection and preparation

We assembled the cases in four age groups: (i) the prenatal period of life (cord blood), when there are no significant contacts with antigens; (ii) the prescholar period (< 5-year-old infants), when the antigen stimulation progressively occurs (by nursing, food, infectious agents and vaccines) and atopy easily arises; (iii) the scholar period (young subjects, 5–17 years old), when most vaccinations have been already performed but the risk of reciprocal transmission of infectious disease is high; (iv) adult life, when the immune system can be considered mature.

Serum and cells were obtained, after informed consent, from CB of full term normal deliveries (45 serum, and 13 cell samples), peripheral blood of infants < 5 years of age (74 serum, 20 cell samples), young subjects between 5 and 17 years of age (56 serum, 20 cell samples), and adults > 17 years old (146 serum, 25 cell samples). Serum, used for sCD30 detection, was prepared and stored at −70°C until use. Samples for flow cytometry studies were collected by venepuncture into preservative-free heparin. Mononuclear cells (MNC) were separated by density gradient centrifugation (Lymphoprep; Nycomed Pharma, Oslo, Norway) under sterile conditions. Samples of post-natal groups were obtained either from adult blood donors (> 18 years of age), or infants and young subjects selected from those without evidence of atopy, infections or inflammatory diseases, admitted to the hospital for elective surgery.

Cell cultures

Following separation and washing, MNC were counted and incubated in 24-well flat-bottomed plates for 4 h at 37°C, 5% CO₂, in complete culture medium (RPMI 1640; Gibco Labs, Grand Island, NY) containing 10% heat-inactivated fetal calf serum (FCS), 1% l-glutamine and 100 U/ml penicillin plus 100 μg/ml streptomycin (Gibco) at a concentration of 2 × 10⁶/ml. The medium was supplemented with 50 ng/ml phorbol 12-myristate 13-acetate (PMA) and 1 μg/ml ionomycin (both from Sigma Chemical Co., St Louis, MO). After the first hour, 10 μg/ml brefeldin A (Sigma) were added for the remaining 3 h (modified from Jung et al. [10]). Cells were then collected, washed twice with 1% PBS solution and resuspended at a final concentration of 10 × 10⁶ cells/ml.

Immunofluorescence studies

Surface marker staining was performed on heparinized whole blood samples using FITC-conjugated anti-CD30 (Dako, Glostrup, Denmark) and R-PE-conjugated anti-CD4 and anti-CD8 (Becton Dickinson, San Jose, CA) MoAbs. Erythrocytes were lysed by 2 ml of 1:10 diluted FACS Lysing solution (Becton Dickinson) and then samples were washed twice with 2 ml of 1% PBS solution.

Intracellular cytokine detection was performed following immunostaining for surface CD4 and CD8 (Tricolor-conjugated anti-CD4 and anti-CD8; Caltag, S. San Francisco, CA), and subsequent permeabilization of cell membranes (FACS permeabilizing solution; Becton Dickinson; 0.5 ml of 1:10 diluted solution/sample for 10 min, room temperature). After washing with a solution containing 1% PBS, 1% bovine serum albumin (BSA), 0.1% sodium azide, samples were first incubated for 15 min with 10 μg/10⁶ cells mouse immunoglobulins (Sigma) to prevent aspecificity, and then with conjugated anti-cytokine MoAbs (anti-IFN-γ–FITC and anti-IL-2–R-PE from Becton Dickinson; anti-IL-4–R-PE from Pharmingen, San Diego, CA), for 30 min at 4°C in darkness. Specificity controls were performed using isotypic MoAbs (IgG2a–FITC and IgG1–R-PE, both from Becton Dickinson). After washing, immunophenotypic patterns were investigated by a flow cytometer equipped with an argon-ion laser (488 nm, FACScan; Becton Dickinson), using Cell Quest software. At least 50 000 events were acquired. Evaluation of membrane CD30 and cytoplasmic cytokines was performed on the entire CD4⁺ and CD8^bright+ lymphoid population, as determined by side scatter (expressed on a linear scale) and fluorescence properties (expressed on a log scale). CD4⁺ T cells could be distinguished from monocytes on the basis of their higher side scatter and stronger fluorescence patterns. The distinction between CD4⁺ T cell and monocyte populations, as performed on this basis, was proven to be accurate by double immunostaining of some samples with a combination of anti-CD4–FITC and anti-CD14–R-PE (Becton Dickinson) MoAbs. CD8⁺ T cells were distinguished from natural killer (NK) cells on the basis of their stronger expression of CD8 molecule (CD8^bright+) and lack of CD3 (as verified in some cases by double immunostaining with anti-CD3 R-PE-Cy5 MoAb (Dako)).

sCD30 assay

Soluble CD30 concentration was determined on serum samples kept frozen at −70°C using a sandwich enzyme-linked immunosorbent test (Dako CD30 (Ki-1 Ag) ELISA), based on the use of two MoAbs reacting with two different epitopes of the CD30 molecule, as previously described [24].

Statistical analysis

Statistical evaluation of data from intracellular cytokine detection was performed by analysis of variance (anova); when needed, a logarithmic or a square root transformation was carried out. The presence of a linear trend was tested by regression analysis [25]. When data were not consistent with anova assumptions, the Kruskal–Wallis test (non-parametric anova) was performed [25]. The Mann–Whitney test was used for comparisons of group means and significant levels were corrected by the Bonferroni method [25]. Differences were considered as statistically significant when P values were ≤ 0.05.

RESULTS

Intracellular cytokines

A progressive and significant increase in the percentage of IFN-γ⁺ T cells after stimulation was observed along with ageing. In detail, IFN-γ⁺ cells amongst CD4⁺ and CD8⁺ lymphocytes, respectively, were 0.6 ± 0.3% (median 0.7%) and 4.3 ± 5.0% (median 2.0%) in CB samples (n = 13); 2.8 ± 1.6% (median 2.7%) and 15.8 ± 7.0% (median 14.6%) in infants (n = 20); 7.0 ± 3.7% (median 6.0%) and 26.3 ± 12.5% (median 24.0%) in young subjects (n = 20); 15.5 ± 6.6% (median 14.5%) and 41.0 ± 13.4% (median 39.1%) in adults (n = 25). The percentages of both CD4⁺/IFN-γ⁺ and CD8⁺/IFN-γ⁺ T cells were statistically different among all age groups (P < 0.001), with a linear trend (P < 0.001) (Fig. 1a,b). Figure 2 shows the cytofluorometric patterns of IFN-γ immunostaining in some representative cases from the four age groups.

Fig. 1.

Percentage of circulating CD4⁺/IFN-γ⁺ (a), CD8⁺/IFN-γ⁺ (b), CD4⁺/IL-2⁺ (c), CD8⁺/IL-2⁺ (d), CD4⁺/IL-4⁺ (e) and CD8⁺/IL-4⁺ (f) T cells, according to age groups (cord blood (CB): < 5 years, infants < 5 years old; 5–17 years, 5–17-year-old subjects; adults, > 18-year-old subjects). Statistical evaluation showed a significant difference among the four age groups in the median percentage of all the T cell subsets but CD8⁺/IL-4⁺ cells, as indicated. An age-dependent linear trend in the percentage of IFN-γ⁺ and IL-2⁺, but not IL-4⁺, T cells was also observed. Horizontal bars, median values. NS, Not significant.

Fig. 2.

Flow cytometric single-cell analysis of IFN-γ expression by cord blood (CB) and peripheral blood CD4⁺ and CD8⁺ T cells according to age: representative cases.

IL-2⁺ cells, amongst CD4⁺ and CD8⁺ lymphocytes, respectively, were 17.6 ± 14.6% (median 12.0%) and 2.0 ± 2.2% (median 1.0%) in CB; 10.7 ± 9.1% (median 8.5%) and 2.1 ± 2.3% (median 0.9%) in infants; 15.4 ± 9.6% (median 16.0%) and 4.0 ± 3.7% (median 3.2%) in young subjects; 18.5 ± 12.0% (median 14.6%) and 7.2 ± 5.9% (median 6.1%) in adults. Statistical analysis showed a significant difference both for CD4⁺/IL-2⁺ and CD8⁺/IL-2⁺ T cell percentages among the different age groups (P = 0.043 for CD4⁺/IL-2⁺ and P < 0.001 for CD8⁺/IL-2⁺ T cells), with a linear trend (P = 0.05 and P < 0.001, respectively) (Fig. 1c,d).

Percentages of CD4⁺/IL-4⁺ T cells were significantly lower (P = 0.032) in CB samples (0.9 ± 1.0%, median 0.7%) in comparison with the other groups pooled together (1.5 ± 1.0% (median 1.6%) in infants; 2.5 ± 1.8% (median 1.5%) in young subjects; 1.5 ± 0.9% (median 1.4%) in adults). By contrast, no significant differences in CD4⁺/IL-4⁺ T cells were found among the three post-natal groups and in CD8⁺/IL-4⁺ T cells among all the four groups (1.6 ± 1.8% (median 1.8%) in CB; 1.4 ± 2.3% (median 0.6%) in infants; 1.3 ± 1.2% (median 1.0%) in young subjects; and 1.6 ± 1.0% (median 1.5%) in adults). Individual data are reported in Fig. 1e,f. Mean values of double-positive IFN-γ⁺/IL-4⁺ cells among CD4⁺ and CD8⁺ lymphocytes were < 1% in all groups.

CD30 membrane expression

Percentages and absolute numbers of circulating CD30⁺ cells amongst CD4⁺ lymphocytes were 2.1 ± 1.0% (median 2.4%) and 21.1 ± 16.1/μl (median 14.0/μl) in CB (n = 13); 3.5 ± 2.3% (median 2.4%) and 57.1 ± 55.2/μl (median 32.9/μl) in infants (n = 20); 1.9 ± 1.2% (median 1.8%) and 15.2 ± 10.1/μl (median 15.4/μl) in young subjects (n = 20); 1.3 ± 0.7% (median 1.0%) and 12.7 ± 7.2/μl (median 11.8/μl) in adults (n = 25). Circulating CD30⁺ cells amongst CD8⁺ T lymphocytes were 1.0 ± 1.2% (median 0.7%) and 4.1 ± 5.9/μl (median 2.7/μl) in CB; 2.0 ± 2.5% (median 1.2%) and 13.1 ± 13.3/μl (median 7.6/μl) in infants; 1.1 ± 0.7% (median 0.8%) and 4.8 ± 2.6/μl (median 4.7/μl) in young subjects; 1.0 ± 0.7% (median 0.8%) and 4.5 ± 3.6/μl (median 3.1/μl) in adults. CD4⁺/CD30⁺ T cells were significantly higher in infants compared with the other three groups (P < 0.001). CD8⁺/CD30⁺ T cells were significantly higher in infants than in CB (P < 0.001) (Fig. 3).

Fig. 3.

Circulating CD4⁺ and CD8⁺ T cells expressing membrane CD30 according to age groups. Statistical evaluation showed a significant difference between the < 5 years group and the others in the median percentage of CD4⁺/CD30⁺ (P < 0.001) and between the < 5 years group and cord blood (CB) in the median percentage of CD8⁺/CD30⁺ cells (P < 0.001). No age-dependent linear trend was observed. Horizontal bars: median values. NS, Not significant.

Serum sCD30 levels

Mean serum values of sCD30 were: 12.8 ± 17.7 U/ml (median 8.0 U/ml) in CB (sCD30 > 20 U/ml only in 5/45 samples: 30, 32, 44, 76 and 92 U/ml, respectively); 91.7 ± 82.2 U/ml (median 56.5 U/ml) in infants (sCD30 > 20 U/ml in 69/74 samples); 28.1 ± 25.1 U/ml (median 20.0 U/ml) in young subjects (sCD30 > 20 U/ml in 25/56 samples); 5.6 ± 6.5 U/ml (median 4.0 U/ml) in adults (sCD30 > 20 U/ml, but < 30 U/ml, in 5/146 samples) (Fig. 4a). The differences in sCD30 values between infants and the other groups and among all four groups were statistically significant (P < 0.001 in any comparison). The progressive decrease of sCD30 values along with ageing in postnatal groups was clearly associated with the progressive increase of the mean percentages of IFN-γ- and IL-2-producing T cells after stimulation in vitro (Fig. 4b).

Fig. 4.

(a) Soluble CD30 detection in cord blood (CB) and serum samples from different age groups. Statistical evaluation showed a significant difference between the < 5 years group and the others in the median value of sCD30 (P < 0.001), without age-dependent linear trend. These results are compared (b) with the age-dependent trend of median percentages of IFN-γ⁺, IL-2⁺ and IL-4⁺ cells amongst CD4⁺ and CD8⁺ T lymphocytes, as detected by flow cytometry. The peak of CD30 release occurs in early infancy (< 5 years group), when the Th(c)1-type cells are not yet highly represented, although circulating IL-4⁺ T cells do not prevail. Horizontal bars: median values. NS, Not significant.

DISCUSSION

Recent studies on the mechanisms of neonatal tolerance to parenterally administered antigens have provided evidence that the failure of newborn animals to develop adult-equivalent T cell memory following parenteral immunization is due to an intrinsic skewing within the T cell compartment towards the Th2 profile of cytokine production, leading to a failure to prime for Th1 immunity [7,26]. This Th2 skew during infancy is a continuation of the pattern characteristic of the fetal Th cell compartment, which is constitutively polarized towards Th2 cytokines [27]. This observation suggests an innate mechanism to minimize intra-uterine production of Th1 cytokines that have been demonstrated to be highly toxic towards the placenta [27]. The data relating to humans are less clear-cut, but it is evident that the capacity to produce the principal Th1 cytokine IFN-γ is generally low during infancy, as a result of developmental immaturity in both the antigen-presenting cell and T cell compartments [28–32], the latter including hyporesponsiveness to IL-12 [33].

Our study was designed to investigate the pattern of cytokine production of T cells obtained from CB and peripheral blood of normal individuals from different age groups by using flow cytometric analysis after non-specific stimulation in vitro. The results provide clear evidence for a progressive increase in the proportion of circulating Th1- and Tc1-type T cells from infancy to adulthood, as evaluated by single-cell analysis of both IFN-γ and IL-2 production. This is consistent with data obtained in experimental animals [34] showing that adult responses differ qualitatively from neonatal responses by a bias towards a predominant Th1 pattern. This probably reflects the expansion of the pool of memory T cells as a consequence of the interaction of the immune system with many infectious agents, including those of commensal microbial flora, which stimulate both dendritic cells and macrophages to produce IL-12, a powerful Th1-inducing cytokine [35]. However, by using the same single-cell assay to analyse IL-4 production as a main Th(c)2 product, we could not provide evidence for a predominance of circulating Th(c)2-type T cells in early infancy. In fact, although the percentages of CD4⁺/IL-4⁺ T cells were significantly higher in the post-natal groups (pooled together) in comparison with CB, no significant increase of CD4⁺/IL-4⁺ and CD8⁺/IL-4⁺ T cells was found in any of these groups. This finding is apparently at variance not only with the results obtained in newborn animals, but also with several observations in humans which suggest a predominance of Th2-orientated responses in the early phases of life. First, human thymocytes develop in vitro into clear-cut Th2 clones unless exogenous IL-12 is added in bulk culture before cloning [36]. Second, IgE responses to dietary allergens (i.e. production of Th2-type cytokines, IL-4 and/or IL-13) typically peak early during infancy (within 9 months) and with few exceptions are permanently terminated by 2–3 years of age [37]. Finally, higher levels of IL-4 have been detected in serum during infancy in comparison with adulthood and the highest IL-4 levels seem to be associated with an increased risk of atopy [38–40].

We do not know whether the lack of a consistent number of detectable IL-4-expressing T cells in early infancy may simply reflect the fact that IL-4 production in humans is not easily detectable in peripheral blood lymphocytes by our approach, possibly due to the type of anti-IL-4 MoAb used or to the type and timing of stimulation. We feel we can confidently rule out the first possibility, since the results obtained in a number of experiments using anti-IL-4 MoAbs from different sources gave similar figures (data not shown). As for the type and timing of stimulation, it is known that IL-4 is usually demonstrable only following repeated T cell stimulation in vitro or under conditions of persistent and abnormal production in vivo [1–3]. On the other hand, we failed to detect significant percentages of IL-4⁺ T cells in some cases of normal infants, following different stimulation approaches, such as 24-h triggering by mitogenic anti-CD3 MoAb (data not shown). Thus, the failure to detect IL-4 by flow cytometry could simply reflect the non-constitutive expression of this cytokine in normal conditions.

Another point of interest in our results is the demonstration of significantly higher sCD30 values (and of circulating CD30⁺ T cells) in early infancy compared with older ages and their inverse correlation with the Th(c)1 shifting. In fact, previous evidence suggested that CD30 expression and release is preferentially associated with the Th(c)2-type immune response. By examining CD30 expression on CD4⁺ and CD8⁺ human T cell clones with established profiles of cytokine production (Th(c)1, Th(c)0 or Th(c)2), we have found that all Th(c)2 clones and the majority of Th(c)0 clones expressed and released CD30 after activation, whereas Th(c)1 clones showed transient or no CD30 expression [20,41]. The preferential expression of CD30 by Th2 clones was also confirmed by Bengtsson et al. [42] and by Hamann et al. [43]. These authors, however, questioned the possibility that CD30 expression may discriminate between human Th1- and Th2-type T cells [42,43]. In subsequent in vivo experiments, we found high numbers of CD30⁺ T cells in the lymph nodes and skin of three patients with Omenn's syndrome—a severe immunodeficiency characterized by elevated serum IgE levels, eosinophilia and T cell anergy [44], with inefficient and/or abnormal T cell receptor generation [45,46]—and in the skin of patients with systemic sclerosis [23], which are both diseases characterized by strong and persistent activation of Th(c)2 cells [23,44]. By contrast, CD30⁺ T cells were not or rarely found either in the gut of patients with Crohn's disease or in the stomach of patients with Helicobacter pylori-induced peptic ulcer, where activated T cells showing a Th(c)1 profile of cytokine production predominate [23]. In our study we have shown that in a non-pathological condition, which is in early infancy, there is an association among increased sCD30 serum levels, detectable circulating CD30-expressing cells and reduced Th(c)1-type immune response. We found that the progressive increase of circulating Th(c)1-type T cells from infancy to adulthood inversely correlated with serum sCD30 and CD30 expression by T cells. These data further support the concept that CD30 is usually not associated in vivo with a Th(c)1-orientated immune response and that Th(c)1 deficiency could be consistent with the previously reported development of prevalent Th(c)2-type responses in early infancy [36–40].