Main

IgA is a major mucosal Ig, and its role in the regulation of dietary antigen uptake by the intestinal mucosa has been extensively studied(1). This isotype Ig is responsible for efficient antigen exclusion and for safe antigen handling and elimination(1, 2).

Intestinal uptake of allergens is a necessary first step in the development of some allergic diseases. Excessive antigen entry, as in cases of mucosal damage or of a defect in IgA synthesis, might lead to increased production of antibodies of other classes, such as IgG and IgE, which might predispose patients to atopic disease. The serum and secretory IgA levels of atopic and food-allergic children have often been reported to be lower than those of healthy infants(37). Increased permeability to ingested antigens was first suspected when a high incidence of cow's milk precipitins was found in IgA-deficient subjects(8). Atopic infants with low serum IgA levels sometimes suffer from eczema. Thus, transient IgA deficiency might play a part in the onset of atopic diseases, such as atopic dermatitis, at the stage of allergen entry(8, 9). Atopic and normal infants might therefore be expected to produce different amounts of IgA, particularly antigen-specific IgA.

In the present study, we used an indirect assay of PFC(10, 11) and found that the production of IgA specific to OVA antigen was below normal in lymphocytes from infants allergic to hen's eggs. We also found that the production of specific IgA was impaired in these patients. We discuss the implications of these findings with regard to the role in the production of specific IgA played by soluble factors derived from antigen-induced T cells.

METHODS

Subjects. Lymphocytes were obtained from peripheral blood of 79 subjects: 54 Japanese patients who were allergic to hen's eggs and had atopic dermatitis or bronchial asthma, or both; 21 healthy children; and 4 children with atopic dermatitis who were not allergic to hen's eggs. The ages of the children ranged from 11 mo to 6 y. The patients had recurrent eczema or intermittent wheezing, or both, without other causes, and had positive immediate skin reactions to egg white (all patients) and to house dust mite(all asthma patients and approximately 1/6 of a topic dermatitis patients). A positive reaction was defined as an immediate wheal response of at least 10 mm to a skin prick test. On an oral provocation test to raw hen's egg, all patients had positive responses, which included urticaria, angioedema, diarrhea, vomiting, coughing, asthmatic attack, anaphylaxis, or exacerbation of atopic dermatitis. The scores on the radioallergosorbent test for egg-white IgE(12) varied from 1 to 4. Children with atopic dermatitis who were not allergic to hen's eggs were used as disease controls. These children had no symptoms in response to an oral provocation with raw hen's egg. Serum IgA levels of the patients, healthy children, and disease controls were 82 ± 60 mg/dL (mean ± SD), 92 ± 61 mg/dL, and 78 ± 44 mg/dL, respectively. The standard deviations of the serum IgA levels were large, and the differences among the groups were not significant. However, the level of OVA-specific IgA in serum from the patients(6.5 ± 1.3 μg/mL, mean ± SD) was lower than that in serum from the healthy children (26.2 ± 10.7 μg/mL, p = 0.00024) and in serum from the disease controls (25.9 ± 13.2 μg/mL,p = 0.00038). These values were measured by ELISA; the level of OVA-specific IgA in serum from a healthy adult, 100 μg/mL, was used as the standard IgA specific for OVA.

No patient had IgA deficiency (defined as an IgA level less than or equal to 5 mg/mL). The patients with hen's egg allergy did not differ from the controls with regard to breast feeding. No patient had been given oral corticosteroids or anti-allergic agents such as ketotifen(13). Some patients were being treated with corticosteroid ointments. All subjects except for four disease controls were included in the study of production of specific IgA at various ages(Fig. 5). Nine patients and seven age-matched healthy children were arbitrarily selected from the total. Data from these subjects and from the four disease controls are shown inFigures 4,6,and7. Other characteristics of these subjects are shown in Table 1. Samples taken from four healthy adults who had none of the symptoms of allergy mentioned above were used in a complementary approach to determine optimal condition for measuring the production of OVA-specific IgA(Figs. 13).

Figure 5
figure 5

Results of a cross-sectional study of OVA-specific IgA production. Data shown are means ± SE. Subjects were divided into groups according to age: 0-1, 2, 3, 4-5, and 6 y. The numbers of healthy children in these age groups were 5, 5, 3, 3, and 5, respectively (□). The numbers of egg-allergic children in these age groups were 23, 15, 6, 4, and 6, respectively (▪). In subjects aged 6 y (**) and older (data not shown) levels of specific IgA did not differ significantly between patients and age-matched normal controls. *p < 0.01 compared with healthy children.

Figure 4
figure 4

Comparison of OVA-specific IgG, IgA, and IgM-PFC between healthy and egg-allergic children (age ≤ 5 y). Data shown are means± SE. *p < 0.01 compared with age-matched healthy controls and to children with atopic dermatitis who were not allergic to hens eggs.

Figure 6
figure 6

Effect of supernatant from OVA-stimulated lymphocytes on generation of OVA-specific IgA-PFC. Data shown are means ± SE(n = 4-6). *p < 0.01 compared with B cells alone.N, normal; P, patients; C, patients not allergic to hen's eggs; B cells, non-T cells; OVA-S, supernatant from OVA-stimulated lymphocytes.

Figure 7
figure 7

Generation of OVA-specific IgA-PFC was increased by supernatant of OVA-stimulated normal T cells, but not by that of non-T cells.OVA-T-S, supernatant from OVA-stimulated T lymphocytes;OVA-B-S, supernatant from OVA-stimulated non-T lymphocytes. Data shown are means ± SE (n = 4). *p < 0.05 compared with data obtained with OVA-B-S.

Table 1 Production of OVA-specific IgA-PFC from patients' and controls' lymphocytes
Figure 1
figure 1

Numbers of PFC generated by lymphocytes from healthy adults exposed to diluted anti-IgA antibody. Data shown are means ± SE(n = 4). *p < 0.01 compared with those not exposed to anti-IgA antibody.

Figure 3
figure 3

Inhibition of IgA-PFC production by OVA antigen. Representative data from four separate experiments on lymphocytes from four healthy adults are shown. Data shown are means ± SE of triplicate experiments. *p < 0.05 compared with data obtained with and without β-LG.

This study was carried out with institutional approval of the protocol. Written informed consent was obtained from the patients' parents.

Measurement of total and OVA-specific IgA. To determine serum levels of total and OVA-specific IgA, flat-bottomed, 96-well microplates(Limbro, Mclean, VA) were used. The wells were coated with 50 μg/mL goat anti-human IgA antibody (Cappel, Malvern, PA) for total IgA, and with 50μg/mL OVA (Sigma Chemical Co., St. Louis, MO) for OVA-specific IgA. Carbonate buffer (60 μL, pH 9.6) was used. Before use, the antibody was absorbed with normal mouse Ig and normal goat Ig to avoid nonspecific binding with the mouse MAb and with the goat antibodies used later. After an overnight coating with antibody at 4°C in a humidified atmosphere, the plates were washed with PBS, and then the uncoated surfaces of each well were saturated with 300 μL of 0.25% gelatin (Wako, Osaka, Japan) for 3 h at 37°C. After another wash, 60 μL of a sample (human sera serially diluted with PBS) were placed in each well, after which the plates were incubated for 2 h at room temperature. After extensive washing of the plates, the samples were allowed to react sequentially with the following three reagents: 60 μL of murine anti-human IgA MAb (Yamasa, Choshi, Japan), 60 μL of biotinylated goat anti-mouse IgG antibody (Tago, Inc., Burlingame, CA), and 50 μL of a 1000-fold dilution of streptavidin-horseradish peroxidase conjugate (Life Technologies, Inc., Gaithersburg, MD). o-Phenylenediamine (Life Technologies) was added, after which the OD at 490 nm was read with an ELISA reader (APR-A4, Tosoh, Tokyo, Japan) and quantified by comparing it with the OD of purified IgA protein (Chemicon International, Inc., Temecula, CA).

The goat anti-mouse IgG antibody used as the second antibody was passed first through a Sepharose 4B column coupled with OVA, gelatin, and goat and human Ig, to prevent nonspecific binding with these substances. Then it was diluted 200-fold for use. The limit of sensitivity of the IgA assay was 0.5μg/mL.

Separation of T cell sand non-T cells. Mononuclear cells were separated from heparinized blood by centrifugation on a Ficoll-Hypaque density gradient(14). The T cells were isolated on a polystyrene resin column(11) rather than by SRBC rosette sedimentation, because in the latter method T cells may be sensitized by SRBC antigen, and the sedimented cells may contain SRBC-binding B cells in addition to T cells which would result in the generation of SRBC-specific PFC(10, 11). Briefly, peripheral mononuclear cells(107) in 0.5 mL of FCS (Life Technologies) were incubated for 1 h on a 2-mL column of polystyrene resin particles (Asahi Chemical Industries, Japan) at 37°C. The resin particles (120-20 μm) consisted of styrene-divinyl benzene copolymer. The T cell-rich fraction was eluted by adding 5 mL of 37°C PBS containing 10% FCS to the column. The remaining cells were recovered by stirring the resin particles and washing out the cells. To remove remaining T cells and SRBC-binding (specific) B cells, the cells were further mixed with SRBC at a ratio of approximately 1:100 in FCS, centrifuged at 200× g for 5 min to obtain tight cell contact, and kept in ice water for 1 h. This procedure allowed SRBC-specific B cells to form antigen-specific rosettes and allowed residual T cells to form spontaneous rosettes with SRBC. The cells were then gently resuspended and centrifuged over a Ficoll-Hypaque density gradient. Rosette-forming cells were separated by sedimentation, and non-T cells free from SRBC-specific B cells were harvested from the interface. This method eliminated SRBC-specific background plaque from the OVA-specific PFC assay.

We did not further separate B cells from monocytes in non-T cell populations, because the function of the non-T cells, in which B cells as well as monocytes were enriched, was not impaired in the patients in the present study, as shown in “Results.” The T cell-rich fraction consisted of 92 ± 3% SRBC rosette-forming cells, 1 ± 1% surface Ig+ cells, and 1 ± 1% peroxidase-positive cells. CD3 (OKT3+) cells made up 90 ± 3% of the cells, which was close to the 89 ± 7% associated with T cells separated by the SRBC rosette sedimentation method. CD4 (OKT4+) and CD8 (OKT8+) cells made up 62 ± 4% and 32± 5%, respectively, which were close to the 59 ± 7% and 30± 6% found with the SRBC rosette sedimentation method; 77 ± 10% of SRBC rosette-forming cells among the lymphocytes loaded onto the resin column were recovered in the T cell-rich fraction. T and B cell populations did not differ between patients and controls (data not shown).

Culture medium. RPMI 1640 tissue culture medium (Life Technologies) was supplemented with gentamicin (0.04 mg/mL), 2 mM L-glutamin, and 50 μM 2-mercaptoethanol. Heat-inactivated pooled stock sera (10%) from normal adult donors with blood type AB were added to the RPMI 1640 tissue culture medium. The sera had been absorbed with SRBC at 0°C to eliminate natural anti-SRBC antibody.

Antigen. OVA and β-LG were purchased from Sigma Chemical Co.

Preparation of OVA-coated SRBC. Two milliliters of 2.5% SRBC suspension in saline were added to 0.2 mL of 1 mg/mL OVA solution and 0.4 mL of 1.25 mM CrCl3 solution(10, 11). The mixture was allowed to stand with occasional shaking for 1.5 h at 32°C. The cells were washed three times with PBS and then resuspended in the medium.

In vitro production of antibodies against OVA. Non-T cells (7× 104) and T cells (14 × 104) were added to different amounts of OVA antigen, suspended in 0.2 mL of the medium in a micro-well plate (Nunclon 1-63320), and cultured at 37°C for 5 d in a 10% CO2 incubator. In most of the experiments we added 10 μg/m OVA to the cultures, because this was found to be optimal for obtaining quantifiable PFC.

PFC assay. The indirect PFC assay was done in Cunningham chambers(10, 11, 15). Cultured lymphocytes (21 × 104) were mixed with 25 μL of 10% target erythrocyte suspension and added to 25 μL of medium or of optimally diluted rabbit anti-human IgA or IgG antibody (Medical and Biological Laboratory Co., Nagoya, Japan), with 10 μL of fresh guinea pig serum as the complement source. The cell suspension was incubated at 37°C for 2 h in a Cunningham chamber. The number of plaques was scored under a light microscope.

Preparation of helper factor. Blood mononuclear cells (2× 106) were added to 10 μg/mL OVA, suspended in 1.0 mL of the medium, and cultured at 37°C for 16 h in a 10% CO2 incubator. Cultured cells were washed four times to remove the antigen and then cultured again without antigen at 37°C for 48 h in a 10% CO2 incubator. Cultured supernatants were collected as a helper factor.

Statistical analysis. Data were analyzed with a t test. All p values of less than 0.05 were accepted as indicators of significant differences. Except as noted, all data are expressed as means± SE.

RESULTS

Detection of cells producing OVA-specific IgA. The cultured cells were mixed with target erythrocyte suspension and then added to diluted rabbit anti-human IgA antibody (×400) and the complement. A total of 214± 56 PFC per 104 non-T cells were generated by lymphocytes from healthy adults, with optimally diluted IgA antibody (Fig. 1). The formation of the PFC was not detected without complement or cultured lymphocytes, and was quite low when the target cells were either SBRC not coated with OVA (CrCl3-treated) or were SBRC coated with the irrelevant protein β-LG (Fig. 2). Formation of PFC was blocked by addition of highly concentrated OVA antigen (1000 μg/mL) to the PFC assay reaction, but not by addition of β-LG at the same concentration, which indicates that the PFC produced by the cultured lymphocytes was specific for OVA antigen (Fig. 3).

Figure 2
figure 2

Formation of PFC was not detected without complement or cultured lymphocytes and was low when the target cells were either SBRC not coated with OVA or were SBRC coated with β-LG. Data shown are means± SE (samples from three healthy adults were used). *p < 0.01 compared with SBRC not coated with OVA or to SBRC coated withβ-LG.

Production of OVA-specific IgA-PFC from the patients' lymphocytes. Characteristics of subjects used in this evaluation are shown in Table 1. The patients had fewer specific IgA-PFC (7± 5 per 7 × 104 non-T cells, n = 9) than did the healthy controls (110 ± 18 per 7 × 104 non-T cells,n = 7) and fewer than did the disease controls (cow's milk allergy)(90 ± 30 per 7 × 104 non-T cells, n = 4). However, the numbers of specific IgG-PFC and IgM-PFC generated from the patients' lymphocytes were equivalent to those of the healthy controls(Fig. 4). In contrast, the number of β-LG-specific IgA-PFC produced by β-LG-stimulated lymphocytes from patients who were allergic only to cow's milk (12 ± 2 per 7 × 104 non-T cells) was lower than that of children with atopic dermatitis who were allergic only to hen's eggs. The combined data indicate that the abnormally low production of antibodies against OVA was isotypespecific in lymphocytes from patients allergic to hen's eggs.

Lymphocytes from patients below the age of 6 y produced significantly less IgA specific to OVA than did lymphocytes from healthy controls(Fig. 5). Production of this IgA by lymphocytes from 6-y-old subjects (Fig. 5) and from those older than 6 y(data not shown) did not differ significantly from the value for healthy controls.

Helper activity of the supernatant of OVA-stimulated lymphocytes. To determine the cause of the abnormally low production of OVA-specific IgA by the patients' lymphocytes, their non-T cells were added to the culture supernatant from OVA-stimulated normal lymphocytes, and then cultured for 5 d with OVA antigen. (Whole T cells were not used, to avoid allogenic reactions.) When the supernatant from normal lymphocytes was added to the autologous B cells, levels of the helper effect (43 ± 5 per 7× 104 non-T cells, n = 6) were comparable to those of OVA-stimulated whole normal T cells (92 ± 9 per 7 × 104 non-T cells, n = 6) (Fig. 6). Therefore, we used the supernatant from OVA-stimulated lymphocytes to evaluate the impaired cellular function in further experiments.

After exposure to culture supernatant from OVA-stimulated normal lymphocytes (56 ± 9 per 7 × 104 non-T cells, n = 6), the patients' lymphocytes were able to produce OVA-specific IgA. In contrast, exposure to the culture supernatant from patients' lymphocytes stimulated with OVA (2 ± 1 per 7 × 104 non-T cells,n = 6) did not affect the production of antibodies by normal non-T cells. This did not occur when the lymphocytes used were from patients with atopic dermatitis who were allergic only to cow's milk (48 ± 6 per 7× 104 non-T cells, n = 6) (Fig. 6). Such helper activity was observed in the supernatant obtained from T cells (82 ± 18 per 7 × 104 non-T cells, n = 4), but not from non-T cells (12 ± 2 per 7 × 104 non-T cells,n = 4) (Fig. 7). Therefore, the patients' T cells could not adequately promote antibody production, although the ability of their B cells to secrete specific antibodies was intact, and the action of the T cells was probably mediated by cytokines.

DISCUSSION

B cells coexpressing sIgM and sIgD migrate from the bone marrow to the circulation and form primary follicles in secondary lymphoid organs including gut-associated lymphoid tissues. After antigen stimulation, primary follicles develop into secondary follicles. The latter are largely composed of the mantle zone, in which sIgM+/sIgD+ naive B cells are located, and the germinal center, in which antigen-dependent maturation occurs(16). The B lymphoblasts that are committed to IgA production usually remigrate from lymphoid tissue such as gut-associated lymphoid tissues to the thoracic duct through lymphatic vessels and lymph nodes, and then into the circulation. Because PWM-stimulated lymphocytes in blood generally produce dimeric IgA(17), IgA production by mononuclear cells in the circulation might reflect immunologic events in the mucosa. Therefore, we used a PFC assay to study the production of IgA specific to OVA antigen by lymphocytes against ingested antigens.

Antigen specificity of the PFC assay was confirmed by the finding that highly concentrated OVA inhibited the formation of plaques, but highly concentrated β-LG did not. Isotype specificity of the PFC was also confirmed by the inhibition in the presence of a high dose of anti-human IgA antibody (data not shown). The number of OVA-specific IgA-PFC from the lymphocytes of the children allergic to hen's eggs was significantly less than that from lymphocytes of healthy controls and of disease controls. The finding that the patients' lymphocytes produced specific IgG-PFC and IgM-PFC(Fig. 4) indicates that the defect in the production of OVA-specific IgA-PFC was isotype-specific.

Until the age of 6 y, the ability of the patients' lymphocytes to produce specific IgA was abnormally low. At the stage of allergen entry, this transiently low production of OVA-specific IgA may contribute to atopic diseases, including allergy to hen's eggs.

To further clarify the cause of the abnormality, the patients' non-T cells were mixed with the culture supernatant from OVA-stimulated normal T cells and then cultured for 5 d. After this treatment the patients' non-T cells produced the specific IgA. In contrast, the supernatant of the patients' T cells did not elicit antibody production in normal non-T cells. Taken together, these data suggest that the patients' T cells were abnormal and that their non-T cells were not. Several B cell helper factors have been reported. These include PWM-induced helper factors(18), mixed lymphocyte culture-derived human helper factors(19), T cell-derived helper factors obtained from phytohemagglutinin-stimulated human T cells(20), concanavalin-A-stimulated supernatant of normal spleen cells(21), and T cell-replacing factor for steroids derived from CD4+ T cells stimulated by monocytes or by small quantities of highly purified IL-1(22). B cells have been induced to secrete Ig after costimulation with anti-Ig reagents and an anti-CD40 MAb in the absence of exogenous factors(23). sIgD+ B cells produced IgM exclusively, whereas sIgD- B cells produced IgM, IgG, and IgA, with a predominance of IgG. Human IL-10(24, 25) strongly enhanced the Ig response elicited in both B cell subsets by dual ligation of sIg and CD40. Addition of TGF-β together with IL-10 and cross-linked anti-CD40 antibodies selectively induced IgA secretion from sIgD+ B cells, but strongly suppressed the IL-10-mediated production of IgG, IgM, and IgA in sIgD- B cells. Thus, IL-10 and TGF-β together cause human B cells activated by anti-CD40 to secrete IgA(23). In the present study, a factor (or factors) from OVA-activated T cells may have been an IgA promoter. This could have been IL-10, TGF-β, a soluble CD40 ligand(26), or other molecules that would together cause OVA-activated B cells to secrete IgA. In this study, we did not evaluate in detail the antigen specificity of the supernatant from OVA-stimulated normal lymphocytes. However, the specificity of antibody responses depends on focusing helper T cells to suitable B cells capable of binding antigen via sIg(27).

The T cell-derived supernatant used in our assay was collected after initial culture together with antigen-stimulated autologous B cells. Thus, the T cells were ready to produce activation molecules promoting the production of specific IgA by antigen-activated allogenic B cells, without restriction by MHC(2731). The supernatant from OVA-stimulated normal lymphocytes increased the number of OVA-specific IgA-PFC from all patients' non-T cells tested and from all allogenic normal non-T cells tested. Therefore, this supernatant might not be restricted by MHC. The activity of the supernatant was partially blocked by treatment with antibodies against TGF-β or IL-10, but not against IL-5 (our unpublished data, 1995).

OVA-specific IgA-PFC from the lymphocytes of patients with hen's egg allergy was significantly less than that of PFC from those of age-matched controls and from children with atopic dermatitis who were not allergic to hen's eggs. Patients' B cells added to the culture supernatant from OVA-stimulated normal T cells were able to produce the specific IgA to levels comparable to those of normal B cells, but the patients' T cells did not cause normal B cells to produce the antibody. This indicates that the ability of the patients' B cells to produce OVA-specific IgA-PFC was intact, but that of the patients' T cells was not; the action of the T cells was probably mediated by cytokines.

The results obtained in the present study indicate possible roles for IgA anti-OVA antibodies. Intestinal secretions of patients allergic to hen's eggs are deficient in OVA-specific secretory IgA antibodies. This deficiency would result in failure to block the uptake of OVA (although the secretory form of IgA was not evaluated directly). These patients' plasma may be deficient in dimeric IgA, such that OVA immune complexes cannot be eliminated via poly-Ig receptor-mediated endocytosis and excretion via the liver, because PWM-stimulated peripheral blood lymphocytes generally produce dimeric IgA(17).

In patients 6 y of age and older, the ability of lymphocytes to produce specific IgA was normal. The patients probably acquired OVA-induced helper activity of T cells and thus may have begun to outgrow their allergy to hen's eggs.

To analyze the onset of food allergy, the helper activity of T cell-derived supernatant should be characterized, and the malfunction of the patients' supernatant must be studied in greater detail.