Intracellular cytokine production by Th1/Th2 lymphocytes and monocytes of children with symptomatic transient hypogammaglobulinaemia of infancy (THI) and selective IgA deficiency (SIgAD)

Abstract

Intracellular expression of several cytokines was assessed in lymphocytes and monocytes of children with transient hypogammaglobulinaemia of infancy (THI) and selective IgA deficiency (SIgAD). THI was characterized by an increased frequency of CD3⁺/CD4⁺ lymphocytes expressing tumour necrosis factor α (TNF-α), TNF-β and interleukin 10 (IL-10), while in SIgAD elevated numbers of these cells containing TNF-α and interferon γ (IFN-γ) were observed. No changes in the number of CD4⁺ T cells expressing IL-4 in both diseases were noted. The proportion of CD33⁺ monocytes containing TNF-α both in THI and SIgAD was unchanged. The secretion of IL-12 by peripheral blood mononuclear cells (PBMCs) of patients with THI and SIgAD was significantly elevated and associated with an increased frequency of IL-12 expressing monocytes in THI but not in SIgAD. IL-18 secretion was slightly, but not significantly, elevated in both diseases. Intracellular Th1 and Th2 type cytokines within CD3⁺/CD4⁺ lymphocytes were also determined in the normal blood donors that showed high or low production of IgG and IgA in vitro. In low producers of IgG an increased proportion of CD3⁺/CD4⁺ cells expressing TNF-α and IFN-γ was found, while in low IgA responders only elevated TNF-α positive CD3⁺/CD4⁺ cells were observed. These results suggest that THI and SIgAD may represent diseases with an excessive Th1 type response that is associated with an up-regulation of IL-12 secretion and, at least in THI, elevated numbers of monocytes expressing intracellular IL-12. Up-regulation of IL-12 may be the essential factor in the patomechanism(s) of these diseases as already described in common variable immunodeficiency (CVID).

INTRODUCTION

Transient hypogammaglobulinaemia of infancy (THI) was first described by Gitlin and Janeway in 1956 [1]. The prevalence of this disorder is uncertain. It has been described as relatively uncommon [2], although in Japan the disorder represents 18·5% of all primary immunodeficiency syndromes [3]. According to Walker et al. [4], THI appears to have a similar prevalence to symptomatic selective IgA deficiency (SIgAD). The controversy concerning the frequency of THI may be explained by the fact that it is not necessarily associated with serious clinical symptoms and such cases remain undetected [4]. The mechanism(s) responsible for the abnormal delay in the onset of immunoglobulin synthesis has not been established and several causes of THI have been postulated. These include delayed maturation of B cell function [5], deficiencies of helper T cells [6] or defect of T cell maturation [7], which may lead to insufficient function of B cells due to defective signalling, the latter being a postulated mechanism in common variable immunodeficiency (CVID) [8]. Clinical manifestations of THI vary from symptomless cases to others with severe recurrent infections [5], unexplained episodes of fever or atopic dermatitis and food allergy [4]. The spectrum of clinical features and range of postulated pathogenetic mechanisms suggest that multiple factors may be involved. Our previous observations indicated aberrant cytokine production by peripheral blood mononuclear cells (PBMCs) of children with THI [9,10], including an enhanced secretion of tumour necrosis factor alpha (TNF-α), TNF-β (lymphotoxin α) and IL-10.

SIgAD is the most common primary immunodeficiency, the clinical manifestations varying from none to recurrent infections, mainly involving the respiratory and gastrointestinal tracts, autoimmune disorders and atopic diseases [11]. A variety of pathological mechanisms has been postulated which include the occurrence of IgA-specific T suppressor cells, inadequate T helper (Th) cell function and an intrinsic B cell defect [12]. Familial inheritance occurs in at least 20% of cases, the major susceptibility locus being within the MHC [13]. Dysregulation of cytokine networks has been suggested as one of the contributing factors to SIgAD [14]. We have recently reported an elevated secretion of TNF-α by PBMCs of children with SIgAD [10].

Multi-parameter flow cytometry is a useful technique to study cytokine production at the single cell level. It enables simultaneous labelling of cells for the expression of surface determinants, thus distinguishing the cytokine responses of phenotypically distinct populations of cells within the same sample. In the present study we have used three-colour flow cytometry to determine the cytoplasmic expression of several cytokines: TNF-α, TNF-β, interferon γ (IFN-γ) (Th1 phenotype), IL-4, IL-10 (Th2 phenotype) and IL-12 by phytohaemagglutinin (PHA) or lipopolysaccharide (LPS)/IFN-γ-stimulated PBMCs of children with THI and SIgAD. Secretion of IL-12 and IL-18 by PBMCs was also studied. Furthermore, the intracellular cytokines were also determined in PBMCs of normal (adult) blood donors with high and low immunoglobulin synthesis in vitro.

Materials and methods

Patients

There were 14 children with abnormal delay in the onset of immunoglobulin synthesis (THI) with total serum IgG level < 2 SD of the normal values for age [10] and normal IgM and IgA levels. We have published previously a definition of THI which, in essence, is characterized by a low serum IgG, the ability to produce IgG antibodies following immunization and recovery of normal IgG levels when follow-up is sufficient [10]. All children in this cohort were later found to have normal IgG levels. The median age was 15 months (range 13–36 months).

Twenty children with SIgAD (IgA serum level below 0·05 g/l) with median age 4 years (range 3–7 years) were also studied. These patients, which were older than those with THI, are presumed to have a permanent defect in IgA production, as their IgA remains undetectable during follow-up (mean observation period over 3 years). Both groups of patients were selected from approximately 600 children with recurrent upper respiratory tract infections referred to our out-patient clinic. None of the children studied had received immunoglobulin substitution therapy.

The control groups were concurrently formed from the examined children with recurrent infections in whom immunodeficiency, as defined by the IUIS Scientific Committee [15], was excluded. There were two control groups: A (n = 19) with median age 19 months (range 15–36 months) and B (n = 10) with median age 3·5 years (range 2–7 years). For these studies samples of blood remaining after completion of routine diagnostic procedures were used. Informed consent was obtained from the parents.

The level of circulating CD3, CD4, CD8, CD19 lymphocytes, T lymphocytes with αβ, γδ receptors, CD3/HLA-DR and NK cells were assessed by flow cytometry after staining with appropriate monoclonal antibodies (MoAbs). All these parameters were within the normal range in both the patients and control groups. Furthermore, no abnormalities in the lymphoproliferative response to mitogens or to anti-CD3 MoAb were observed.

In some experiments blood was also obtained from healthy adult donors (n = 10) and used for determination of immunoglobulin production in vitro.

Antibodies and reagents

The following directly conjugated MoAbs: anti-CD4-fluorescein isothiocyanate (FITC), phycoerythrin (PE)-conjugated MoAb against human: TNF-α, TNF-β, IFN-γ, IL-4, IL-10, IL-12 (p40/p70) and appropriate isotype controls were purchased from Pharmingen/Becton Dickinson (San Diego, CA, USA). Anti-CD3-PE/cyanin 5·1 (Cy5), anti-CD33-PE/Cy5 and appropriate isotype controls were purchased from Immunotech/Coulter (Marseille, France). GolgiStop reagent (containing monensin) and other reagents for cell fixation and permeabilization (Cytofix/Cytoperm and Perm/Wash) were from Pharmingen/Becton Dickinson.

Determination of cytoplasmic cytokine expression

PBMCs were isolated from EDTA-treated peripheral blood by standard Isopaque/Ficoll (Pharmacia, Uppsala, Sweden) density gradient and suspended in RPMI-1640 medium (Biochrom, Berlin, Germany). For generation of cytokine producing cells, PBMCs were cultured in the presence of PHA (2·5 µg/ml, Murex, Dartford, UK) for 6 h (induction of TNF-α, IFN-γ and IL-4) or 48 h (induction of TNF-β and IL-10), as described previously [16]. For stimulation of IL-12, PBMCs were cultured in the presence of recombinant IFN-γ (10 ng/ml, Sigma Chemical Co., St Louis, MO, USA) and LPS (0·1 µg/ml, Sigma) for 24 h. To inhibit cytokine secretion, monensin (GolgiStop, 2 µm) was added at the begining of short-term cultures (6 h) or for the last 12 h in longer cultures. To avoid attachment of monocytes to the tubes, Falcon 2063 non-stick polypropylene round-bottom tubes (Becton Dickinson, San Jose, CA, USA) were used. The cells were harvested, washed, resuspended in a phosphate buffered saline (PBS) and incubated for 30 min on ice with conjugated MoAbs directed against: CD3, CD4 for identification of lymphocytes or CD33 for monocytes. This was done before fixation as antibodies against surface determinants did not stain efficiently after fixation and permeabilization. After washing, the cells were fixed and permeabilized with Cytofix/Cytoperm reagent (20 min at 4°C) and washed twice in Perm/Wash solution. The cells were then stained (30 min, 4°C) for intracellular cytokines using PE-conjugated MoAbs against human cytokines. PE-conjugated isotype controls were used in parallel. After three washes the cells were suspended in PBS with 0·1% bovine serum albumin (Sigma) for flow cytometry analysis.

Flow cytometry analysis

Samples were analysed in FACS Calibur flow cytometer (Becton Dickinson Immunocytometry Systems, Palo Alto, CA, USA) using CellQuest v. 3·1 software. The list mode data of 50 000 events from PBMCs or 20 000 of CD3⁺ or CD33⁺ cells in a ‘live gate’ mode were acquired. The cells were gated on either lymphocytes or monocytes according to forward (FSC) and side-scatter (SSC) gating. Intracellular cytokines were determined in CD4⁺ and CD4⁻ cells within CD3⁺ lymphocytes (three-colour flow cytometry) and CD33⁺ monocytes (two-colour flow cytometry). The frequency of cells containing intracellular cytokines was expressed as their percentage within CD3⁺/CD4⁺ lymphocytes or CD33⁺ monocytes.

Production and measurements of IL-12 and IL-18

For IL-12 and IL-18 production, PBMCs (1 × 10⁶/ml) suspended in RPMI-1640 medium with 10% fetal calf serum (Biochrom) were cultured in microtitre plates (Nunc, Roskilde, Denmark) for 24 h at 37°C in 5% C0₂ in the presence of IFN-γ (10 ng/ml) and LPS (0·1 µg/ml). Supernatants were harvested and kept at −70°C until testing. Concentrations of IL-12 and IL-18 in the culture supernatants were measured by ELISA. IL-12 (p40/p70) production was determined using commercially available matched antibody pair for ELISA (antihuman IL-12 (p40/p70) and biotinylated antihuman IL12 p40/p70) and recombinant human IL-12 p40 protein as the standard (all from Pharmingen) according to manufacturer's protocol (sensitivity 20 pg/ml). IL-18 was measured by human IL-18 ELISA kit (MBL, Naka-ku Nagoya, Japan, sensitivity 12·5 pg/ml).

Synthesis of immunoglobulin in vitro by cells from normal donors

Monocytes were separated from PBMCs obtained from healthy adult donors, by counterflow centrifugal elutriation with a JE-5 elutriation system (Beckman, Palo Alto, CA, USA) equipped with a 5-ml Sanderson separation chamber as described previously [17]. T cells were isolated from elutriated lymphocytes by rosetting with neuraminidase-treated sheep erythrocytes followed by the Isopaque/Ficoll gradient [18], and the remaining lymphocytes were used as the B-cell-enriched population. Monocytes were 90–96% pure as determined by staining with anti-CD14 MoAb (Leu-M3; Becton Dickinson) and flow cytometry analysis. The T cell population contained approximately 97% of CD3⁺ while the B-cell-enriched population approximately 75% of CD19⁺ cells. For in vitro studies of immunoglobulin synthesis, the B-cell-enriched population was mixed with mitomycin C (Sigma)-treated T lymphocytes at a ratio of 1:1 and 5% monocytes. The cells (1 × 10⁶/ml) were stimulated with pokeweed mitogen (PWM, Gibco, Paisley, UK) at the final dilution of 1:1000 and cultured in triplicate in microtitre plates for 7 days. For details see [10].

Determination of immunoglobulin levels in the culture supernatants

The concentration of IgG and IgA in the supernatants were quantified by ELISA as previously described [10]. Briefly, the serial dilutions of the culture supernatants were added (100 µl/well) to the plates coated with goat antihuman IgG or antihuman IgA (Jackson ImmunoResearch Laboratory, Westgrove, PA, USA). After overnight incubation at 4°C, the plates were washed, and alkaline phosphatase-conjugated goat antihuman IgG or IgA antibody (Jackson) was added, followed by 100 µl/well of p-nitrophenyl phosphate.The plates were read with an ELISA reader (Labsystems Multiskan Plus) at 492 nm. Serial dilutions of N protein Standard SY (Behring, Marburg, Germany) were used to obtain the concentration curve.

Statistical analysis

The Mann–Whitney two-tailed test was used. The difference was regarded as significant when the P-value was < 0·05.

Results

Intracellular Th1 and Th2 cytokines in children with THI and SIgAD

Our previous results showed an increased secretion of TNF-α, TNF-β, IL-10 by PBMCs in THI and enhanced production of TNF-α in SIgAD. The enhanced level of cytokines in the culture supernatants may be due to an increased secretion by the same number of cells, elevated number of secreting cells or by different cell subpopulations. To determine the actual number of intracellular cytokines of Th1 (TNF-α, TNF-β, IFN-γ) and Th2 (IL-4, IL-10) types among CD3⁺/CD4⁺ cells, three-colour flow cytometry was used. An increased frequency of TNF-α- and TNF-β-positive CD3⁺/CD4⁺ lymphocytes in THI was observed (Table 1). In SIgAD there was a significantly increased percentage of CD3⁺/CD4⁺ cells expressing TNF-α and IFN-γ(Table 1). No significant changes in the number of CD4⁺ lymphocytes containing intracellular IL-4 and IL-10 were noted though CD4⁺/IL-10⁺ were elevated in THI. The proportion of TNF-α-positive CD33⁺ monocytes in THI and SIgAD was the same as in the control group (data not shown). These results suggested that THI and SIgAD may be associated with an excessive Th1 type response.

Table 1.

The percentage of CD3⁺/CD4⁺ lymphocytes expressing intracellular cytokines

Groups	TNF-α	TNF-β	IFN-γ	IL-4	IL-10
THI	32·3 ± 8·1*	25·2 ± 12·1	23·4 ± 12·7	8·3 ± 4·1	16·4 ± 9·2
SIgAD	48·2 ± 26·2**	11·1 ± 8·4	36·8 ± 20·8*	12·2 ± 7·2	8·2 ± 6·1
Control A	15·4 ± 8·3	10·3 ± 9·6	9·5 ± 8·2	8·2 ± 9·3	10·1 ± 5·5
Control B	17·2 ± 9·8	9·4 ± 4·5	12·1 ± 7·3	9·0 ± 6·3	8·4 ± 4·8

PBMCs from children with THI and SIgAD were stimulated with PHA and cultured for either 6 h (for determination of cytoplasmic expression of TNF-α, IFN-γ and IL-4) or for 48 h (TNF-β and IL-10). Intracellular cytokines were detected by flow cytometry after staining with appropriate MoAbs (see Materials and methods) in CD4⁺ cells within gated CD3⁺ lymphocytes. Data expressed as mean ± s.d. Significantly different from control:

^*P < 0·05

^**P < 0·01.

Production of IL-12 and IL-18 by PBMCs of children with THI and SIgAD

Since IL-12 and IL-18 are required for the expression of the Th1 phenotype, their levels in the culture supernatants of PBMCs stimulated with LPS/IFN-γ were determined. Production of IL-12 (p40/p70) was increased significantly in THI and SIgAD in comparison to the control group (Fig. 1). IL-18 production (Fig. 2) was also elevated in both groups of patients; however, this increase did not reach statistical significance.

Fig. 1.

IL-12 (p40/p70) secretion by PBMCs of children with THI and SIgAD. The cells stimulated with LPS/IFN-γ were cultured for 24 h and IL-12 concentrations were measured in the supernatants by ELISA. The lines represent mean values.*P < 0·0015, **P < 0·0005 (patients versus controls).

Fig. 2.

IL-18 secretion by PBMCs of children with THI and SIgAD. The cells stimulated with LPS/IFN-γ were cultured for 24 h and IL-18 concentrations were measured in the supernatants by ELISA. The lines represent mean values.

During follow-up six children from THI group were examined for IL-12 release by PBMC twice: at the time of diagnosis and when their IgG levels reached normal values. The significant decrease of IL-12 release was observed at the time of normalization of serum IgG levels (Fig. 3).

Fig. 3.

Comparison of IL-12 production by PBMCs of children with THI at the time of diagnosis (low serum IgG) and the time when their serum IgG reached normal values.

Intracellular expression of IL-12 in CD33⁺ monocytes

Finding an enhanced secretion of IL-12 by PBMCs in the patients' groups prompted the study of intracellular expression of this cytokine in monocytes, known to be its main source. PBMCs were stimulated with IFN-γ/LPS and the proportion of CD33⁺ positive monocytes containing IL-12 (p40/p70) was determined. An increased proportion of IL-12 positive CD33⁺ monocytes in children with THI was observed while in patients with SIgAD it was comparable to the control group (Fig. 4). The proportion of CD33⁺ monocytes in stimulated cultures of PBMCs was identical in THI and control groups (not shown).

Fig. 4.

The percentage of CD33⁺ monocytes expressing intracellular IL-12 (p40/p70) in PBMCs cultures stimulated with LPS/IFN-γ. The lines represent mean values.*P < 0·01.

Comparison of intracellular Th1 and Th2 cytokine productionin healthy donors with high and low immunoglobulin production in vitro

The cells (B-enriched lymphocytes, mitomycin C-treated T cells and monocytes) from normal donors were stimulated with PWM and after 7 days of culture IgG and IgA in the supernatants were quantified by ELISA. Donors were divided into two groups according to the level of IgG and IgA production. In low responders the mean level of IgG and IgA in the supernatants was 1051 ± 350 ng/ml and 345 ± 130 ng/ml, respectively. In high responders respective values were 3469 ± 2589 ng/ml and 1500 ± 128 ng/ml.

The proportion of CD3⁺/CD4⁺ lymphocytes expressing intracellular TNF-α, TNF-β, IFN-γ, IL-4 and IL-10 was compared between those two groups. In low producers of IgG the number of CD3⁺/CD4⁺ lymphocytes containing TNF-α and IFN-γ was increased significantly in comparison to high responders (Table 2). Among low producers of IgA, a significantly increased number of TNF-α-positive CD3⁺/CD4⁺ lymphocytes was found (Table 2). Analysis of cytoplasmic expression of Th2 cytokines (IL-4, IL-10) showed no difference between high and low producers of both IgG and IgA.

Table 2.

Synthesis of IgG and IgA in vitro and the percentage of CD3⁺/CD4⁺ lymphocytes expressing intracellular cytokines

	Level of IgG in the supernatants		Level of IgA in the supernatants
Cytokines	< 1500 ng/ml (mean 1051 ± 350)	> 1500 ng/ml (mean 3469 ± 2589)	< 500 ng/ml (mean 345 ± 130)	> 500 ng/ml (mean 1245 ± 280)
TNF-α	15·3 ± 10·2*	5·1 ± 3·5	33·6 ± 15·7*	8·2 ± 3·0
TNF-β	10·2 ± 6·1	7·3 ± 2·0	8·4 ± 4·0	7·3 ± 2·6
IFN-γ	26·3 ± 1·1*	9·4 ± 1·5	5·2 ± 8·0	8·4 ± 3·7
IL- 4	6·3 ± 2·8	3·3 ± 2·6	2·1 ± 1·5	9·5 ± 4·5
IL- 10	6·8 ± 1·6	6·7 ± 3·2	6·2 ± 3·0	3·3 ± 1·5

Intracellular cytokines were induced and determined in PBMCs of normal subjects as described in Table 1. The B cell-enriched lymphocytes, mitomycin C-treated T lymphocytes (ratio 1:1) and monocytes (5%) obtained from the same PBMCs were cultured for 7 days in the presence of PWM. The levels of IgG and IgA were determined by ELISA. Data are expressed as mean ± s.d. and comparisons were made between low and high producers of IgG and IgA.

^*P < 0·05.

Discussion

The reason why immunoglobulin production is low in THI is unknown, although several abnormalities, including delayed maturation of B cell function [5] or a defect of T cell maturation [7], have been described. The present study extended our previous observations demonstrating an increased secretion of TNF-α and TNF-β, but not of IL-1-β, IL-4, IL-6, by PBMCs in THI [10]. This presumptive skewing towards a Th1 type response prompted further analysis of this phenomenon at the single cell level. Cytokines in body fluids or culture supernatants are usually measured by bioassays or immunoassays which do not provide information about the frequency or phenotype of cytokine-producing cells. In the present study, three-colour flow cytometry was used to study the cytoplasmic expression of several cytokines characteristic for Th1 and Th2.

Dysregulation of the Th1/Th2 balance is often observed in human diseases [19]. The differentiation of naive Th0 cells into Th1 or Th2 cells is influenced by several factors such as the nature of antigen, the type of antigen-presenting cells (APC), and cytokine production in the local microenviroment [20]. Only some cytokines, i.e. IFN-γ and IL-12 are involved in the commitment of undifferentiated Th0 cells towards Th1 cells. IL-18 has recently been added to the short list of Th1-promoting cytokines [21].

The present study demonstrated an increased frequency of CD4⁺ lymphocytes producing TNF-α and TNF-β in THI, while the occurrence of CD4⁺/IL-4⁺ T cells (characteristic for Th2) was essentially similar to the control group. The CD4⁺/IL-10⁺ lymphocytes were slightly elevated in THI but the differences were not significant. In SIgAD an increased proportion of CD3⁺/CD4⁺ cells expressing TNF-α and IFN-γ was observed. The present results confirm our previous data based on cytokine secretion by PBMC of children with THI [10], and show additionally an up-regulation of IFN-γ in SIgAD which provides evidence for shifts of cytokine profiles towards Th1 in both disorders. Secretion of both IL-12 and IL-18 by PBMCs was elevated in THI and SIgAD. The role of IL-12 in THI was substantiated further by follow-up observations in six children showing that the increased cytokine production found initially, when serum IgG was low, decreased significantly at the time of normalization of IgG levels. Furthermore, cytoplasmic expression of IL-12 by monocytes was significantly increased in THI. No cytoplasmic IL-18 was determined due to a lack of suitable MoAbs. These results suggest that the shift of cytokine profiles (TNF-α, TNF-β, IFN-γ) towards Th1 may be due to an overproduction of IL-12 by monocytes, and possibly of IL-18 (release only studied). Chronic infections are unlikely to be responsible for the observed changes, as children from the control group had similar clinical symptoms.

Further evidence that excessive Th1 activity is linked to abnormal B cell function was obtained by in vitro observations showing that low production of IgG by PBMCs of normal subjects was associated with a higher proportion of CD4 cells expressing cytoplasmic TNF-α and IFN-γ, while an increased frequency of CD4⁺/TNF-α⁺ cells was observed in low producers of IgA.

These results, including an increased proportion of CD33⁺ monocytes expressing IL-12, are in keeping with the findings of North et al. [22], showing a significantly increased frequency of IFN-γ⁺ T cells in CVID, and of Cambronero et al. [23] suggesting that up-regulation of IL-12 in monocytes is a fundamental defect in CVID. However, in contrast to CVID this was not due to a reduced absolute number of IL-12-negative monocytes. However, our system differed in many ways from that of Cambronero et al. [23]. In particular, the cells studied (whole blood versus PBMCs), MoAbs used for identificatation of monocytes (CD14 versus CD33), stimulators (LPS versus LPS/IFN-γ) and the length of culture were different. Nevertheless, in accord with CVID [23], we also found that the increased frequency of monocytes containing IL-12 was selective, in that TNF-α-positive cells were not increased. Overall, our observations suggest that the mechanism of THI is similar to CVID. Furthermore, as CVID is genetically linked with SIgAD [13,24,25], some of the same susceptibility genes may predispose to THI.

The finding of an enhanced secretion of IL-12 in patients with SIgAD, that in contrast to THI was not associated with an increased proportion of CD33⁺ monocytes expressing cytoplasmic IL-12, is difficult to explain. IL-12 is produced by mainly APCs (monocytes and dendritic cells) and B cells, and IFN-γ up-regulates its synthesis and release [26]. Hence, an increased proportion of CD4⁺/IFN-γ⁺ cells observed in SIgAD suggests that either B cells are responsible for the enhanced secretion of IL-12 or simply that there is increased secretion by similar numbers of monocytes. However, it is unclear where the primary abnormality lies in the IL-12/IFN-γ circuit, or even if this lies downstream of the activated circuit itself. Nevertheless, our data show that THI, SIgAD and CVID are mechanistically linked. An understanding of the mechanism of recovery of immunoglobulin production in THI may help design therapeutic strategies for CVID.