TY - JOUR
AU - Koike, Takao
AB - Abstract β2‐Glycoprotein I (β2GPI) appears to be the major antigen for antiphospholipid antibodies (aPL) in patients with antiphospholipid syndrome (APS). In early infancy, virtually all children initiate transient immune response to non‐pathogenic nutritional antigens, which fails to terminate in children with atopic diseases. To examine the possibility that a prolonged immune response to β2GPI could also spread to the human protein, antibodies against human β2GPI (anti‐β2GPI) were determined in 93 randomly selected children with different allergic diseases. A high frequency (42%) of IgG anti‐β2GPI was found in children with atopic dermatitis (AD), but not in those with other allergic diseases. Anti‐β2GPI in children with AD were exclusively of the IgG1 subclass and bound to bovine β2GPI as well, but not to either β2GPI combined with the phospholipid cardiolipin. The epitopes were identified in domain V of β2GPI and the antibody binding was abolished upon the specific proteolytic cleavage of the phospholipid‐binding C‐terminal loop in domain V of β2GPI. These results indicated that the epitopes for anti‐β2GPI in children with AD most likely resided in close vicinity of the phospholipid‐binding site of β2GPI. The epitopic difference from anti‐β2GPI in APS may explain presumed non‐thrombogenicity of anti‐β2GPI in children with AD. allergy, antigen, antiphospholipid antibody, autoantibody, autoimmunity, epitope Introduction Antiphospholipid antibodies (aPL) are detected in a number of clinical disorders including autoimmune, infectious, malignant and other diseases. However, thrombosis and recurrent fetal loss associated with aPL, constituting the antiphospholipid syndrome (APS), are reported in autoimmune patients only (1). Originally thought to be directed towards negatively charged phospholipids, aPL in autoimmune patients are now known to recognize certain phospholipid‐binding plasma proteins, thereby differing from aPL in other diseases, which bind directly to phospholipids (2,3). β2‐Glycoprotein I (β2GPI) appears to be by far the commonest and best‐characterized antigenic target for presumably pathogenic aPL (4). β2GPI is a single‐chain (326 amino acids) 50‐kDa plasma protein composed of five homologous short consensus repeat domains (5,6). Domain V of β2GPI differs from the other four domains by two additional cysteins, responsible for the second internal loop and the long C‐terminal tail (7). Domain V also possesses a highly positively charged amino acid sequence Cys281–Cys288, which was shown to be a potent phospholipid‐binding site (8). Recent X‐ray analysis of β2GPI showed that a patch consisted of 14 positively charged amino acid residues on domain V, as well as the flexible and partially hydrophobic C‐terminal loop between Ser311 and Lys317, are involved in the binding to phospholipids (9). The binding of β2GPI to phospholipids could significantly be reduced by the specific proteolytic cleavage of the C‐terminal loop between Lys317 and Thr318 (10,11). The exact epitopic sites for antibodies against β2GPI (anti‐β2GPI) have not yet been revealed, but it seems that domains I and IV are dominant in the binding of anti‐β2GPI from APS patients (12–14). The physiological function of β2GPI is still unclear. However, the interaction of β2GPI and anti‐β2GPI apparently leads to a prothrombotic state, accelerated atherogenesis, and interferes with the clearance and processing of apoptotic cells. Particularly during the first few months of life, intact macromolecules may pass through the epithelium of the gastrointestinal tract in amounts sufficient to effectively engage the T cell system (15). It has been shown that virtually all children initiate IgG1 production against nutritional antigens by 3 months of age, and this production typically peaks during infancy and wanes thereafter, whereas corresponding IgG4 responses continue to develop (16). This early pattern of transient Th2‐dependent antibody production against non‐pathogenic nutritional antigens fails to terminate in children with atopic diseases (17,18). Since β2GPI has been remarkably conserved during the evolution of animal species, showing >80% homology between the amino acid sequences of human, bovine and rat counterparts (19), we have considered the possibility that in children with atopic diseases an exaggerated immune response to β2GPI from milk or meat products could lead not only to the production of antibodies against bovine β2GPI, but also toward human β2GPI. The immunization of experimental animals with heterologous β2GPI was shown to induce the production of antibodies specific for the administered antigen and a subset of antibodies that cross‐reacted with homologous β2GPI (20). To our knowledge, no links between anti‐β2GPI and allergic diseases have been reported yet. We report here a high frequency of anti‐β2GPI in children with atopic dermatitis (AD), but not in those with other allergic diseases examined. Anti‐β2GPI in children with AD were found to recognize epitopes on domain V of β2GPI, most probably in close vicinity of the phospholipid‐binding site. This epitopic difference from anti‐β2GPI in APS and the inability of AD‐associated anti‐β2GPI to bind β2GPI in combination with phospholipids may provide clues to understanding their presumed non‐thrombogenicity. Methods Serum samples Five groups of sera were selected from the serum banks of the Pediatrics Clinic and Department of Rheumatology (University Medical Centre, Ljubljana). The main group comprised 71 consecutive sera from children with different atopic diseases positive for at least one allergen‐specific IgE antibody or having elevated total serum IgE. Among them, 45 children suffered from AD, 36 exhibiting pure‐type AD, while nine also had other types of allergy associated with AD, i.e. mixed‐type AD (27 boys and 18 girls, mean age 3.7 years, range 2 months to 16.8 years). The other 26 children had atopic diseases with respiratory system symptoms only (allergic rhinitis 15, asthma 3, both 8 children; 21 boys and five girls, mean age 11 years, range 1.9–18 years). The second group consisted of 22 children with a non‐atopic allergic disease who had experienced a systemic anaphylactic reaction after stings of Hymenoptera insects (16 boys and 6 girls, mean age 11 years, range 2.1–17.8 years). The normal control group for children comprised 61 sera from apparently healthy children collected at their regular preventive visits (25 boys and 36 girls, mean age 9.4 years, range 4.9–14.2 years). As a positive control group, 26 sera from adult primary or secondary APS patients were randomly selected. The control group for adults consisted of 52 healthy blood donors (31 men and 21 women, mean age 34.0 years, range 18–65 years). mAb Two monoclonal β2GPI‐dependent aPL were used as calibrators to establish standard curves for the anti‐β2GPI and anticardiolipin antibody (aCL) ELISA: HCAL, a chimeric IgG mAb consisting of human κ and γ1 C regions and V regions from the mouse monoclonal aCL WBCAL‐1 (21,22), and EY2C9, an IgM mAb derived from a patient with APS (23). Arbitrary units (AU) were set for the anti‐β2GPI and aCL ELISA by the concentrations of the mAb giving definite absorbance values. The activity of 1 ng/ml of HCAL was defined as 1 IgG AU (AUG) and the activity of 5 ng/ml of EY2C9 as 1 IgM AU (AUM). IgA AU (AUA) were set by the dilutions of a selected IgA anti‐β2GPI+ in‐house standard from a patient with APS. Three anti‐β2GPI mAb (Cof‐19, Cof‐22 and Cof‐23) obtained from BALB/c mice immunized with human β2GPI (12) were applied as positive controls. Anti‐β2GPI ELISA β2GPI was purified from pooled human sera by affinity column chromatography (2) and coated at 10 µg/ml in PBS (pH 7.4, 50 µl/well) on Costar High Binding EIA/RIA plates (cat. no. 3590; Costar, Cambridge, MA) for 2 h at room temperature. After one wash with 300 µl of PBS containing 0.05% Tween 20 (PBS‐T), 50 µl/well of the standards and test sera diluted 1:100 in PBS‐T were applied for 30 min at room temperature. Following four washes, 50 µl/well of alkaline phosphatase‐conjugated goat anti‐human IgG, IgM or IgA (ACSC, Westbury, NY) diluted in PBS were added for 30 min at room temperature. After four washes, 100 µl/well of p‐nitrophenylphosphate (Sigma, St Louis, MO) dissolved at 1 mg/ml in 1 M diethanolamine buffer (pH 9.8) was applied. OD at 405 nm was first measured after 15 min and then every 3 min by a microtiter plate reader (Rainbow Spectra Thermo; Tecan, Salzburg, Austria) until an optimal fitting to the predicted OD of the standards was obtained. All sera were tested in duplicates both in wells with and without β2GPI (sample blank values). Specific absorbance was obtained by subtracting sample blank values from raw data. The cut‐off levels for anti‐β2GPI in children were determined based on the absorbance values of control healthy children. The distribution of absorbance values was significantly skewed on the linear scale and nearly symmetrical on the log scale, the cut‐off points were defined as the mean + 2SD for the log‐absorbance values. Expressed in previously defined AU, the cut‐off points for IgG, IgM and IgA anti‐β2GPI in children were 4.2 AUG, 2.6 AUM and 4.0 AUA respectively (24). Fluid‐phase inhibition test with β2GPI Except for the preparation of test samples, the assay conditions were the same as described for the anti‐β2GPI ELISA. Tested sera or Cof mAb diluted to 50% maximal binding in 1% BSA (essentially fatty acid free, cat. no. A‐6003; Sigma) in PBS‐T were mixed with equal volumes of β2GPI solutions in the same buffer. The final concentrations of β2GPI were 100, 20, 4, 0.8, 0.16, 0.032 and 0 µg/ml. The mixtures were incubated for 45 min at room temperature, then 50 µl/well was applied to plate wells. OD was measured after 30 min of color development. Anti‐β2GPI ELISA for IgG subclasses Coating of β2GPI, washing and application of serum samples were performed as in the anti‐β2GPI ELISA. Secondary mouse anti‐human monospecific antibodies to IgG1 (diluted 1:250; clone HP 6069; Zymed, San Francisco, CA), IgG2 (diluted 1:500; clone MAB 1308; Chemicon, Temecula, CA), IgG3 (diluted 1:1000, clone HP‐6050; Sigma) or IgG4 (diluted 1:1000, Clone HP‐6025; Sigma) were applied in 50 µl/well aliquots for 1 h at room temperature. For color development, alkaline phosphatase‐conjugated goat anti‐mouse IgG (Sigma) and p‐nitrophenylphosphate were used. Anti‐BSA ELISA and fluid‐phase inhibition test with BSA Costar High Binding microtiter plates were coated with 50 µl/well of BSA dissolved at 20 µg/ml in PBS for 2 h at room temperature. All further steps were as described for the anti‐β2GPI ELISA. OD was measured after 45 min of incubation at room temperature. The cut‐off point was defined as for the anti‐β2GPI ELISA. For the inhibition test, β2GPI and BSA were coated on the same plate as described for the anti‐β2GPI and anti‐BSA ELISA respectively. Inhibition mixtures were prepared in duplicates, one containing tested serum and BSA at concentrations up to 1 mg/ml in PBS‐T and the other serum diluted in PBS‐T only. Following incubation for 45 min at room temperature, the mixtures were applied to wells with β2GPI, BSA or without an antigen. All further steps were as in the anti‐β2GPI ELISA. ELISA and fluid‐phase inhibition test with bovine β2GPI The purification of bovine β2GPI, ELISA and inhibition test were performed as described for human β2GPI, which was also used as a control protein. Inhibition mixtures containing either bovine or human β2GPI in fluid phase were applied to parallel duplicate wells coated with either protein so that all four possible antigen combinations were tested on the same plate. The final concentrations of β2GPI in mixtures were 200, 100, 20, 4, 0.8, 0.16 and 0 µg/ml. Standard aCL ELISA Costar Medium Binding EIA/RIA plates (cat. no. 3591) were coated with cardiolipin as in the standard aCL ELISA. After incubation with 120 µl/well of 10% FBS (Sigma) in PBS for 1 h at room temperature, the plates were washed once with 300 µl/well of PBS. Then 100 µl/well of standards and serum samples diluted 1:100 in PBS containing 10% FBS were applied for 2.5 h at room temperature. Plates were washed 4 times with PBS and 100 µl/well of alkaline phosphatase‐conjugated rabbit anti‐human IgG (ACSC) were added. Following 1‐h incubation at room temperature, plates were washed 4 times and 100 µl/well of substrate were applied. OD was measured as described for the anti‐β2GPI ELISA. Modified aCL ELISA Sumilon S microtiter plates (Sumitomo Bakelite, Tokyo, Japan) were coated with 40 µl/well of 50 µg/ml cardiolipin (Sigma) in ethanol by evaporation at 4°C overnight. After blocking with 200 µl/well of 1% BSA in PBS (1% BSA/PBS) for 1 h at room temperature, the plates were washed 3times with 300 µl/well of PBS‐T. Then 50 µl/well of standards and serum samples diluted 1:100 in 1% BSA/PBS containing 10 µg/ml of β2GPI were applied for 2 h at room temperature. Plates were washed 3 times and 50 µl/well of alkaline phosphatase‐conjugated mouse anti‐human IgG (Southern Biotechnology, Birmingham, AL) were added for 1 h at room temperature. After washing, the substrate was added and OD was measured as described for the anti‐β2GPI ELISA. The cut‐off values in the aCL ELISA, determined by the same method as in the anti‐β2GPI ELISA, were 13.9 AUG and 5.6 AUG for the standard and modified method respectively; 13.9 AUG corresponded to 7.6 GPL international standard IgG aCL units (25). ELISA and inhibition tests with domain‐deleted mutants (DM) of human β2GPI Recombinant DM of β2GPI, lacking one or more domains, as well as whole molecule β2GPI (WM) were produced by the baculovirus/Sf 9 insect cell vector expression system (12,26). The ELISA with different DM (DM II‐V, DM III‐V, DM IV‐V and DM V lacking domains at the N‐terminal region or DM I‐III and DM I‐IV lacking domains at the C‐terminal region) as antigens was performed as outlined for the anti‐β2GPI ELISA. Cof mAb and rabbit polyclonal anti‐β2GPI served as positive controls. Inhibition of binding to solid‐phase β2GPI was tested with recombinant domain V of β2GPI (DM V) in fluid phase at its final concentrations of 50, 20, 4, 0.8, 0.16 and 0 µg/ml. ELISA with proteolytically cleaved β2GPI Cleaved β2GPI was prepared by treatment with human plasmin as described (27). The ELISA was performed on Costar High Binding plates using intact serum purified β2GPI on the same plate for comparison. Statistical analysis Statistical tests were performed using subroutines from the statistical analysis package by MS Excel 6.0 for Windows. Chi‐square test, t‐test assuming equal or unequal variances and linear regression analysis were used when appropriate. Differences were considered significant whenever P < 0.05. Results Prevalence and clinical association of anti‐β2GPI in atopic children Anti‐β2GPI of IgG isotype were detected in 20 of 71 (28%) atopic children. Interestingly, 19 of 20 IgG anti‐β2GPI+ children (95%) were diagnosed with AD and the one not suffering from AD had an IgG anti‐β2GPI level of only 5.8 AUG above the cut‐off. Considering the group of 45 children with AD, 19 (42%, 17 pure and two mixed type) had positive IgG anti‐β2GPI, while 26 children (58%, 19 pure and seven mixed type) were anti‐β2GPI–. The difference in anti‐β2GPI positivity between pure and mixed type AD was not significant. A statistically significant difference was found by comparing the age of AD children positive for IgG anti‐β2GPI (mean age 1.8 ± 0.8 years, range 3 months to 3.1 years) and those negative for anti‐β2GPI (5.1 ± 4.4 years, range 2 months to 16.1 years) (P = 0.0003). There was no difference in either the frequency or mean positive IgG anti‐β2GPI values between the children with AD and adult APS patients (Fig. 1). Among children with increased IgG anti‐β2GPI, one also had increased IgM, two IgA, and one child both IgM and IgA anti‐β2GPI. Children negative for IgG anti‐β2GPI were negative for the other two isotypes as well. All 22 children with systemic anaphylaxis after an insect bite were negative for all three isotypes of anti‐β2GPI. Inhibition of anti‐β2GPI with β2GPI in fluid phase The binding of IgG anti‐β2GPI from the children with AD to β2GPI on microtiter plates was inhibited in a dose‐dependent manner by β2GPI in solution. Fifty percent inhibition was achieved by β2GPI concentrations between 15 and 30 µg/ml (0.3–0.6 µM). These concentrations were ∼100‐fold higher than the IC50 of the Cof mAb (0.1–0.35 µg/ml). Anti‐β2GPI from the APS patients and HCAL mAb showed only 15–35% inhibition at 100 µg/ml (2 µM) of β2GPI. Representative inhibition profiles are presented in Fig. 2. Subclass specificity of IgG anti‐β2GPI in atopic children and APS patients IgG1 was the only subclass detected in all eight randomly chosen IgG anti‐β2GPI+ sera from the children with AD, whereas IgG2 was the predominant subclass present in all eight randomly selected sera from the APS patients. Additionally, low levels of IgG1 were detected in seven, IgG3 in one and IgG4 in two of the eight sera from the APS patients. Binding to BSA in and cross‐reactivity of anti‐BSA with β2GPI Because of the high background binding exhibited by many sera of the atopic children in preliminary experiments using BSA as the blocking agent, we presumed that some sera may also contain anti‐BSA. In fact, increased levels of IgG anti‐BSA were found in 24 of 71 sera (34%) from atopic children. The frequency of anti‐BSA was significantly higher in the children with AD (19 of 45 = 42%) than in the other atopic children (five of 26 = 19%, P = 0.038). No correlation between the absorbance values for anti‐β2GPI and anti‐BSA in the children with atopic diseases was observed. There was also no association between the presence of anti‐β2GPI and anti‐BSA in the children with AD. Anti‐BSA in 13 selected atopic children were inhibited in a dose‐dependent manner by fluid‐phase BSA (IC50 = 0.2– 20 µg/ml BSA). BSA had no influence on the detection of anti‐β2GPI, thus excluding cross‐reactivity of anti‐BSA with β2GPI. Binding of anti‐β2GPI from atopic children to bovine β2GPI All 19 sera from the AD children containing antibodies toward human β2GPI bound efficiently to bovine β2GPI, suggesting a cross‐reactivity. Five of 16 randomly selected sera among atopic children negative against human β2GPI expressed a substantial binding to bovine β2GPI. HCAL and EY2C9 mAb bound both antigens similarly to anti‐β2GPI from the children with AD. In addition, anti‐β2GPI from the APS patients bound efficiently both antigens, although some sera expressed a higher binding to human than to bovine β2GPI (Fig. 3A). Human and bovine β2GPI had similar inhibitory effects on the binding to solid‐phase human β2GPI by the sera from the children with AD. However, the binding to solid‐phase bovine β2GPI was less efficiently inhibited by human than bovine β2GPI, indicating a population of anti‐β2GPI specific for bovine β2GPI (Fig. 3B). Binding of IgG anti‐β2GPI from atopic children in aCL ELISA Among 19 anti‐β2GPI+ sera from the children with AD, positive results were obtained in only three and two sera by the standard and modified aCL ELISA, respectively. Among anti‐β2GPI– atopic children, three sera expressed positive values in the standard aCL ELISA and one serum in the modified aCL ELISA. No statistically significant difference in aCL positivity was observed between anti‐β2GPI+ and anti‐β2GPI– sera. Binding to DM of β2GPI Anti‐β2GPI from the children with AD recognized all DM containing domain V and showed no binding to DM without this domain (Fig. 4). Further, all tested anti‐β2GPI+ sera from the children with AD bound to solid‐phase recombinant domain V. There was also a positive correlation between the absorbance values for binding to complete β2GPI and domain V (r = 0.85, P < 0.0001) (Fig. 5A). The binding of anti‐β2GPI from the children with AD to solid‐phase β2GPI was efficiently and in a dose‐dependent manner inhibited by domain V in the fluid phase (IC50 = 6 to 9 µg/ml = 0.4 to 0.7 µM) (Fig. 5B). The molar concentrations of domain V and complete β2GPI giving 50% inhibition were comparable. Anti‐β2GPI from the APS patients bound predominantly to DM I‐IV and to a lower extent to DM I‐III, which was similar to the binding profiles of HCAL and EY2C9 (data not shown). Binding to proteolytically cleaved β2GPI The cleavage of β2GPI practically abolished the binding of anti‐β2GPI from all 17 tested positive children with AD (mean OD 0.386 ± 0.183, range 0.139–0.798 for native purified β2GPI and 0.014 ± 0.018, range 0–0.060 for cleaved β2GPI). Cof mAb, used as controls, bound to both forms of β2GPI efficiently. Discussion The β2GPI requirement for the binding of aPL from patients with autoimmune disorders associated with APS is one of the features that distinguish them from aPL occurring in non‐autoimmune diseases (2,3). Antibodies reacting with β2GPI in the absence of anionic phospholipids may be a more specific marker of thrombotic tendency in APS than aPL (4). Until now, high frequencies and levels of IgG anti‐β2GPI, comparable to those in autoimmune diseases, particularly systemic lupus erythematosus, have not been reported in non‐autoimmune disease. In this study we present the finding of a high frequency (42%) of IgG anti‐β2GPI in 45 children with AD. Furthermore, not only the high frequency but also the levels of IgG anti‐β2GPI were similar to those usually found in APS. Anti‐β2GPI was significantly associated with AD, as a low level of IgG anti‐β2GPI was detected in only one of 26 patients with other atopic diseases and in none of the 22 children who experienced a systemic anaphylactic reaction after stings of Hymenoptera insects. Since the existence of anti‐β2GPI in AD children has not been associated with clinical manifestations of APS, we presumed that those antibodies may differ from anti‐β2GPI in APS patients. A series of experiments was performed to characterize the binding properties of anti‐β2GPI in AD in comparison with those in APS. The specificity of anti‐β2GPI in AD was supported by the efficient inhibition of binding to solid‐phase β2GPI with β2GPI in the fluid phase. Interestingly, >80% inhibition was obtained with 100 µg/ml of β2GPI (Fig. 2), which was approximately half the physiologic concentration in human serum (28). In contrast, anti‐β2GPI from the APS patients was very weakly inhibited by fluid‐phase β2GPI. This difference could be attributed to the presumed conformational specificity of anti‐β2GPI from APS patients, recognizing β2GPI epitopes that are fully expressed only when β2GPI is bound to an appropriate negatively charged surface, or to differing affinities for fluid‐phase β2GPI. Another important distinction between anti‐β2GPI in AD and APS was noted using the standard and modified aCL ELISA. Namely, anti‐β2GPI from the children with AD showed negligible binding in either of these assays, indicating that their target epitopes were not available when β2GPI (from FBS or purified human protein) was combined with cardiolipin. Therefore, anti‐β2GPI in AD could not be classified as aCL. Antibodies reacting in anti‐β2GPI, but not in aCL, ELISA were reported in some APS patients and healthy individuals, but were not further characterized (29,30). The assays with deletion mutants of β2GPI provided further evidence for a distinct fine specificity of anti‐β2GPI in AD. By contrast to antibodies from the APS patients, for which the major epitope proved to reside within domains I–IV (12–14), the sera from the children with AD showed a reactivity restricted to domain V of β2GPI. Furthermore, a high inhibition of binding to β2GPI was obtained by isolated domain V, confirming that this domain possessed main epitope(s) for anti‐β2GPI in AD. Domain V of β2GPI was shown to contain a potent phospholipid‐binding site including the C‐terminal loop (9). The proteolytic cleavage of the C‐terminal loop between Lys317 and Thr318 by plasmin abolished the binding of anti‐β2GPI from the children with AD. The integrity of the phospholipid‐binding site was also critical for the binding of anti‐β2GPI from APS patients in some (8,10,27,31), but not all studies (32). The lack of binding to cleaved β2GPI was attributed to either a conformational change of β2GPI (27) or direct disruption of the epitopic site in domain V (8,10,31). However, anti‐β2GPI from APS patients bound to β2GPI combined with phospholipids (8,10,27,31), which was not the case for anti‐β2GPI in AD. The failure of anti‐β2GPI from children with AD to bind cleaved β2GPI or β2GPI associated with cardiolipin suggested that their epitopes were located at or at least in close vicinity to the phospholipid‐binding site in domain V. We may also assume that a sterical hindrance caused by the binding of such antibodies to β2GPI interferes with the attachment of β2GPI to phospholipids, as already noted for Cof‐18 mAb, which also recognizes domain V of β2GPI (12). Since the adhesion of β2GPI to negatively charged phospholipids and subsequent binding of anti‐β2GPI appears to be the pathogenic sequence of events in APS, the inability of AD‐associated anti‐β2GPI to bind β2GPI in complex with phospholipids may be an explanation for their presumed non‐thrombogenicity. The mechanisms responsible for the occurrence of IgG anti‐β2GPI in AD are not clear. Several interdependent factors are probably involved. In virtually all sera from children with AD having antibodies against human β2GPI we found also a similar reactivity with bovine β2GPI. However, a specific binding to bovine β2GPI was also observed in some sera from children with atopic diseases without antibodies against human β2GPI. The inhibition tests with human and bovine β2GPI implied that children with AD harbored two subsets of anti‐β2GPI: one recognizing specifically bovine β2GPI, found also in other atopic children, and the other cross‐reactive with human and bovine β2GPI, resembling anti‐β2GPI from APS patients in our study and previous reports (33). An important factor responsible for the occurrence of antibodies against bovine β2GPI might be repeated exposure to nutritional β2GPI early in life, when the intestinal mucosa is more permissive for large molecules (15). It is conceivable that in infancy the ingested bovine or some other type of β2GPI (from dietary products) could act as a peroral immunization agent inducing a transient production of species‐specific anti‐β2GPI. In children with AD this humoral immune response may spread also to human β2GPI. An additional explanation for the development of antibodies against human β2GPI in children with AD might also be related to a deficiency of a specific regulatory T cell function in AD (34). It has already been shown that exposure to cow milk during the first few months of life results in the initiation of immune responses towards β‐lactoglobulin, BSA and α‐casein. The specific IgG antibody production typically peaked during early infancy, with particularly high levels in children with AD, and declined thereafter (35,36). In our study as well, the AD children having IgG anti‐β2GPI were of a significantly younger age than those without anti‐β2GPI. Antibodies against BSA were also detected in our atopic children, particularly in AD. Although their frequency was similar to that of anti‐β2GPI, the potential cross‐reactivity was excluded. Anti‐β2GPI in children with AD were restricted to the IgG1 subclass. The IgG subclass distribution of antibodies reactive with dietary proteins, notably β‐lactoglobulin, has already been investigated. In two studies, significantly higher levels of specific IgG1 and IgG4 were found in infants with elevated IgE. This parallelism in IgG1, IgG4 and IgE responses was ascribed to the influence of IL‐4 (37,38). It is tempting to speculate that the elevated levels of IgG1 anti‐β2GPI in children with AD were also a result of a genetically governed Th2 skewing of the response to new nutritional antigens in atopy prone infants. At present we have no evidence that anti‐β2GPI in AD can influence allergic manifestations or induce APS. We are also not aware of any relation between AD and APS, considering patients’ personal and family history. In conclusion, we found a high frequency of IgG anti‐β2GPI in children with AD and no clinical signs of APS. These anti‐β2GPI could not be classified as classic aPL (aCL), as they did not bind to β2GPI associated with cardiolipin. The proteolytic cleavage of the C‐terminal region of β2GPI in close vicinity to the phospholipid‐binding site abolished the binding of anti‐β2GPI from children with AD, implying that the epitope resided near the phospholipid‐binding region in domain V. Further analyses and prospective studies are needed for definite conclusions on the clinical relevance of anti‐β2GPI in AD. The investigation of interactions between dietary β2GPI and the human immune system may provide insights into the regulation of autoimmune anti‐β2GPI response. Acknowledgements We thank Dr Hisao Kato (National Cardiovascular Center Research Institute, Suita, Osaka, Japan) for providing purified bovine β2GPI and Mrs Mirjana Zupančič (Division of Laboratory Diagnostic, Department of Pediatrics, University Medical Centre, Ljubljana, Slovenia) for preparing serum samples from atopic children. This work was supported by the Ministry of Science and Technology of Slovenia (grant no. J3 7924). Abbreviations β2GPI—β2‐glycoprotein I aCL—anticardiolipin antibody AD—atopic dermatitis anti‐β2GPI—antibodies against human β2GPI aPL—antiphospholipid antibodies APS—antiphospholipid syndrome AU—arbitrary units AUA—IgA AU AUG—IgG AU AUM—IgM AU DM—recombinant domain‐deleted mutant of human β2GPI WM—recombinant whole molecule β2GPI View largeDownload slide Fig. 1. Comparison of IgG anti‐β2GPI levels in children with AD and adult APS patients. No statistically significant difference in either frequency or levels of positive IgG anti‐β2GPI was found between the two groups (AD: 19/45 = 42% positive, mean 21.2 ± 18.3 AUG, range 5.5–5.5 AUG; APS: 10/26 = 38% positive, mean 28.2 ± 27.5 AUG, range 5.1–85.8 AUG). Results of IgG anti‐β2GPI ELISA are expressed in AUG defined in Methods. The cut‐off values for AD and APS were 4.2 and 2.9 AUG respectively. Error bars indicate mean ± SEM of positive values. View largeDownload slide Fig. 1. Comparison of IgG anti‐β2GPI levels in children with AD and adult APS patients. No statistically significant difference in either frequency or levels of positive IgG anti‐β2GPI was found between the two groups (AD: 19/45 = 42% positive, mean 21.2 ± 18.3 AUG, range 5.5–5.5 AUG; APS: 10/26 = 38% positive, mean 28.2 ± 27.5 AUG, range 5.1–85.8 AUG). Results of IgG anti‐β2GPI ELISA are expressed in AUG defined in Methods. The cut‐off values for AD and APS were 4.2 and 2.9 AUG respectively. Error bars indicate mean ± SEM of positive values. View largeDownload slide Fig. 2. Inhibition of binding to β2GPI adsorbed on microtiter plates with β2GPI in the fluid phase. Three representative inhibition profiles among IgG anti‐β2GPI+ sera from children with AD and APS are presented. HCAL and Cof mAb were used as controls. Results are expressed as percentage of inhibition (1 – ODinhibited/ODnon‐inhibited). View largeDownload slide Fig. 2. Inhibition of binding to β2GPI adsorbed on microtiter plates with β2GPI in the fluid phase. Three representative inhibition profiles among IgG anti‐β2GPI+ sera from children with AD and APS are presented. HCAL and Cof mAb were used as controls. Results are expressed as percentage of inhibition (1 – ODinhibited/ODnon‐inhibited). View large Download slide View large Download slide Fig. 3. (A) Binding of antibodies against human β2GPI to bovine β2GPI. In children with AD, the correlation between the absorbance values for human and bovine β2GPI was statistically significant (r = 0.92, P < 0.001). Sera from patients with APS, normal human sera (NHS) and HCAL and EY2C9 mAb are shown for comparison. Results are presented as mOD (10–3 × OD) at 405 nm values. (B) Inhibition of binding to human or bovine GPI adsorbed on microtiter plate by either human or bovine β2GPI in the fluid phase. Representative serum inhibition profiles from a child with AD are shown as percentage of inhibition (1 – ODinhibited/ODnon‐inhibited). Symbols indicate different combinations of fluid‐ and solid‐phase antigen (first fluid‐ and second solid‐phase β2GPI). View large Download slide View large Download slide Fig. 3. (A) Binding of antibodies against human β2GPI to bovine β2GPI. In children with AD, the correlation between the absorbance values for human and bovine β2GPI was statistically significant (r = 0.92, P < 0.001). Sera from patients with APS, normal human sera (NHS) and HCAL and EY2C9 mAb are shown for comparison. Results are presented as mOD (10–3 × OD) at 405 nm values. (B) Inhibition of binding to human or bovine GPI adsorbed on microtiter plate by either human or bovine β2GPI in the fluid phase. Representative serum inhibition profiles from a child with AD are shown as percentage of inhibition (1 – ODinhibited/ODnon‐inhibited). Symbols indicate different combinations of fluid‐ and solid‐phase antigen (first fluid‐ and second solid‐phase β2GPI). View largeDownload slide Fig. 4. Binding to serum‐purified human β2GPI (β2GPI), whole molecule recombinant β2GPI (WM) and various domain‐deleted mutants of β2GPI (DM). All nine tested sera from children with AD showed the same binding pattern and results of two sera are presented. Cof‐22 mAb recognizing domain III (50 ng/ml) and Cof‐23 mAb recognizing domain IV (100 ng/ml) were used as positive controls. A normal human serum (NHS) reaching the highest absorbance values is included for comparison. Results are presented as mOD (10–3 × OD) at 405 nm values. View largeDownload slide Fig. 4. Binding to serum‐purified human β2GPI (β2GPI), whole molecule recombinant β2GPI (WM) and various domain‐deleted mutants of β2GPI (DM). All nine tested sera from children with AD showed the same binding pattern and results of two sera are presented. Cof‐22 mAb recognizing domain III (50 ng/ml) and Cof‐23 mAb recognizing domain IV (100 ng/ml) were used as positive controls. A normal human serum (NHS) reaching the highest absorbance values is included for comparison. Results are presented as mOD (10–3 × OD) at 405 nm values. View large Download slide View large Download slide Fig. 5. (A) Binding of anti‐β2GPI to isolated domain V of β2GPI. In 16 tested anti‐β2GPI+ children with AD, a positive correlation was found between absorbance values for complete β2GPI and domain V (r = 0.85, P < 0.0001). Rabbit polyclonal anti‐β2GPI (rabbit) was used as a positive control. HCAL and EY2C9 mAb recognizing domain IV, and normal human sera (NHS) were applied as negative controls. Results are presented as mOD (10–3 × OD) at 405 nm values. (B) Inhibition of binding of sera from children with atopic dermatitis to β2GPI adsorbed on microtiter plates by domain V of β2GPI in the fluid phase. Three representative inhibition profiles are presented. Results are expressed as percentage of inhibition (1 – ODinhibited/ODnon‐inhibited). View large Download slide View large Download slide Fig. 5. (A) Binding of anti‐β2GPI to isolated domain V of β2GPI. In 16 tested anti‐β2GPI+ children with AD, a positive correlation was found between absorbance values for complete β2GPI and domain V (r = 0.85, P < 0.0001). Rabbit polyclonal anti‐β2GPI (rabbit) was used as a positive control. HCAL and EY2C9 mAb recognizing domain IV, and normal human sera (NHS) were applied as negative controls. Results are presented as mOD (10–3 × OD) at 405 nm values. (B) Inhibition of binding of sera from children with atopic dermatitis to β2GPI adsorbed on microtiter plates by domain V of β2GPI in the fluid phase. Three representative inhibition profiles are presented. Results are expressed as percentage of inhibition (1 – ODinhibited/ODnon‐inhibited). References 1 Khamashta, M. A. and Hughes, G. R. 1993. Detection and importance of anticardiolipin antibodies. J. Clin. Pathol.  46: 104. Google Scholar 2 Matsuura, E., Igarashi, Y., Fujimoto, M., Ichikawa, K., Suzuki, T., Sumida, T., Yasuda, T. and Koike, T. 1992. Heterogeneity of anticardiolipin antibodies defined by the anticardiolipin cofactor. J. Immunol.  148: 3885. Google Scholar 3 Hunt, J. E., McNeil, H. P., Morgan, G. J., Crameri, R. M. and Krilis, S. A. 1992. A phospholipid‐beta 2‐glycoprotein I complex is an antigen for anticardiolipin antibodies occurring in autoimmune disease but not with infection. Lupu s 1: 75. Google Scholar 4 Roubey, R. A. 1999. Immunology  of the antiphospholipid syndrome: antibodies, antigens, and autoimmune response. Thromb. Haemost.  82: 656. Google Scholar 5 Steinkasserer, A., Estaller, C., Weiss, E. H., Sim, R. B. and Day, A. J. 1991. Complete nucleotide and deduced amino acid sequence of human beta 2‐glycoprotein I. Biochem. J.  277: 387. Google Scholar 6 Kato, H. and Enjyoji, K. 1991. Amino acid sequence and location of the disulfide bonds in bovine beta 2 glycoprotein I: the presence of five Sushi domains. Biochemistry  30: 11687. Google Scholar 7 Steinkasserer, A., Barlow, P. N., Willis, A. C., Kertesz, Z., Campbell, I. D., Sim, R. B. and Norman, D. G. 1992. Activity, disulphide mapping and structural modelling of the fifth domain of human beta 2‐glycoprotein I. FEBS Lett.  313: 193. Google Scholar 8 Hunt, J. and Krilis, S. 1994. The fifth domain of beta 2‐glycoprotein I contains a phospholipid binding site (Cys281–Cys288) and a region recognized by anticardiolipin antibodies. J. Immunol.  152: 653. Google Scholar 9 Bouma, B., De Groot, P. G., van den Elsen, J. M., Ravelli, R. B., Schouten, A., Simmelink, M. J., Derksen, R. H., Kroon, J. and Gros, P. 1999. Adhesion mechanism of human beta(2)‐glycoprotein I to phospholipids based on its crystal structure. EMBO J.  18: 5166. Google Scholar 10 Hunt, J. E., Simpson, R. J. and Krilis, S. A. 1993. Identification of a region of beta 2‐glycoprotein I critical for lipid binding and anti‐cardiolipin antibody cofactor activity. Proc. Natl Acad. Sci. USA  90: 2141. Google Scholar 11 Ohkura, N., Hagihara, Y., Yoshimura, T., Goto, Y. and Kato, H. 1998. Plasmin can reduce the function of human beta2 glycoprotein I by cleaving domain V into a nicked form. Blood  91: 4173. Google Scholar 12 Igarashi, M., Matsuura, E., Igarashi, Y., Nagae, H., Ichikawa, K., Triplett, D. A. and Koike, T. 1996. Human beta2‐glycoprotein I as an anticardiolipin cofactor determined using mutants expressed by a baculovirus system. Blood  87: 3262. Google Scholar 13 George, J., Gilburd, B., Hojnik, M., Levy, Y., Langevitz, P., Matsuura, E., Koike, T. and Shoenfeld, Y. 1998. Target recognition of beta2‐glycoprotein I (beta2GPI)‐dependent anticardiolipin antibodies: evidence for involvement of the fourth domain of beta2GPI in antibody binding. J. Immunol.  160: 3917. Google Scholar 14 Iverson, G. M., Victoria, E. J. and Marquis, D. M. 1998. Anti‐beta2 glycoprotein I (beta2GPI) autoantibodies recognize an epitope on the first domain of beta2GPI. Proc. Natl Acad. Sci. USA  95: 15542. Google Scholar 15 Kuitunen, M., Savilahti, E. and Sarnesto, A. 1994. Human alpha‐lactalbumin and bovine beta‐lactoglobulin absorption in infants. Allergy  49: 354. Google Scholar 16 Devey, M. E., Beckman, S. and Kemeny, D. M. 1993. The functional affinities of antibodies of different IgG subclasses to dietary antigens in mothers and their babies. Clin. Exp. Immunol.  94: 117. Google Scholar 17 Merrett, J., Barnetson, R. S., Burr, M. L. and Merrett, T. G. 1984. Total and specific IgG4 antibody levels in atopic eczema. Clin. Exp. Immunol.  56: 645. Google Scholar 18 Barnes, R. M., Harvey, M. M., Blears, J., Finn, R. and Johnson, P. M. 1986. IgG subclass of human serum antibodies reactive with dietary proteins. Int. Arch. Allergy Immunol.  81: 141. Google Scholar 19 Matsuura, E., Igarashi, M., Igarashi, Y., Nagae, H., Ichikawa, K., Yasuda, T. and Koike, T. 1991. Molecular definition of human β2‐glycoprotein I (β2‐GPI) by cDNA cloning and inter‐species differences of β2‐GPI in alternation of anticardiolipin binding. Int. Immunol.  3: 1217. Google Scholar 20 Tincani, A., Beltrami, B., Meroni, P. L., Allegri, F., Spatola, L., Cinquini, M. and Shoenfeld, Y. 1998. Immunization of naive BALB/c mice with human beta2‐glycoprotein 1(beta2GPI) breaks tolerance against the murine molecule. Lupus  7: S173. Google Scholar 21 Hashimoto, Y., Kawamura, M., Ichikawa, K., Suzuki, T., Sumida, T., Yoshida, S., Matsuura, E., Ikehara, S. and Koike, T. 1992. Anticardiolipin antibodies in NZW × BXSB F1 mice. A model of antiphospholipid syndrome. J. Immunol.  149: 1063. Google Scholar 22 Ichikawa, K., Tsutsumi, A., Atsumi, T., Matsuura, E., Kobayashi, S., Hughes, G. R., Khamashta, M. A. and Koike, T. 1999. A chimeric antibody with the human gamma1 constant region as a putative standard for assays to detect IgG beta2‐glycoprotein I‐dependent anticardiolipin and anti‐beta2‐glycoprotein I antibodies. Arthritis Rheum.  42: 2461. Google Scholar 23 Ichikawa, K., Khamashta, M. A., Koike, T., Matsuura, E. and Hughes, G. R. 1994. beta 2‐Glycoprotein I reactivity of monoclonal anticardiolipin antibodies from patients with the antiphospholipid syndrome. Arthritis Rheum.  37: 1453. Google Scholar 24 Avcin, T., Ambrozic, A., Kuhar, M., Kveder, T. and Rozman, B. 2001. Anticardiolipin and anti‐beta(2) glycoprotein I antibodies in sera of 61 apparently healthy children at regular preventive visits. Rheumatology  40: 565. Google Scholar 25 Harris, E. N., Pierangeli, S. and Birch, D. 1994. Anticardiolipin Wet Workshop Report. Fifth International Symposium on Antiphospholipid Antibodies. Am. J. Clin. Pathol.  101: 616. Google Scholar 26 Igarashi, M., Matsuura, E., Igarashi, Y., Nagae, H., Matsuura, Y., Ichikawa, K., Yasuda, T., Voelker, D. R. and Koike, T. 1993. Expression of anticardiolipin cofactor, human beta 2‐glycoprotein I, by a recombinant baculovirus/insect cell system. Clin. Exp. Immunol.  93: 19. Google Scholar 27 Matsuura, E., Inagaki, J., Kasahara, H., Yamamoto, D., Atsumi, T., Kobayashi, K., Kaihara, K., Zhao, D., Ichikawa, K., Tsutsumi, A., Yasuda, T., Triplett, D. A. and Koike, T. 2000. Proteolytic cleavage of β2‐glycoprotein I: reduction of antigenicity and the structural relationship. Int. Immunol.  12: 1183. Google Scholar 28 Polz, E. and Kostner, G. M. 1979. The binding of beta 2‐glycoprotein‐I to human serum lipoproteins: distribution among density fractions. FEBS Lett.  102: 183. Google Scholar 29 Alarcon‐Segovia, D., Mestanza, M., Cabiedes, J. and Cabral, A. R. 1997. The antiphospholipid/cofactor syndromes. II. A variant in patients with systemic lupus erythematosus with antibodies to beta 2‐glycoprotein I but no antibodies detectable in standard antiphospholipid assays. J. Rheumatol.  24: 1545. Google Scholar 30 Guerin, V., Couchouron, A., Vergnes, C., Parrens, E., Vernhes, J. P., Constans, J. and Boisseau, M. 1997. Antiphospholipid syndromes with anti‐human beta 2‐glycoprotein I antibodies despite negative reactivity in conventional aPL and LA assays. Thromb. Haemost.  77: 1037. Google Scholar 31 Sheng, Y., Sali, A., Herzog, H., Lahnstein, J. and Krilis, S. A. 1996. Site‐directed mutagenesis of recombinant human beta 2‐glycoprotein I identifies a cluster of lysine residues that are critical for phospholipid binding and anti‐cardiolipin antibody activity. J. Immunol.  157: 3744. Google Scholar 32 Horbach, D. A., van Oort, E., Lisman, T., Meijers, J. C., Derksen, R. H. and De Groot, P. G. 1999. Beta2‐glycoprotein I is proteolytically cleaved in vivo upon activation of fibrinolysis. Thromb. Haemost.  81: 87. Google Scholar 33 Arvieux, J., Darnige, L., Hachulla, E., Roussel, B., Bensa, J. C. and Colomb, M. G. 1996. Species specificity of anti‐beta 2 glycoprotein I autoantibodies and its relevance to anticardiolipin antibody quantitation. Thromb. Haemost.  75: 725. Google Scholar 34 Boguniewicz, M. and Leung, D. Y. M. 1998. Atopic dermatitis. In Middleton, E., Ellis, E. F., Yunginger, J. W., Reed, C. E., Adkinson, N. F. and Busse, W. W., eds, Allergy, Principles & Practise, p. 1123. Mosby, St Louis, MO. Google Scholar 35 Vaarala, O., Saukkonen, T., Savilahti, E., Klemola, T. and Akerblom, H. K. 1995. Development of immune response to cow’s milk proteins in infants receiving cow’s milk or hydrolyzed formula. J. Allergy Clin. Immunol.  96: 917. Google Scholar 36 Jenmalm, M. C. and Bjorksten, B. 1998. Exposure to cow’s milk during the first 3 months of life is associated with increased levels of IgG subclass antibodies to beta‐lactoglobulin to 8 years. J. Allergy Clin. Immunol.  102: 671. Google Scholar 37 Host, A., Husby, S., Gjesing, B., Larsen, J. N. and Lowenstein, H. 1992. Prospective estimation of IgG, IgG subclass and IgE antibodies to dietary proteins in infants with cow milk allergy. Levels of antibodies to whole milk protein, BLG and ovalbumin in relation to repeated milk challenge and clinical course of cow milk allergy. Allergy  47: 218. Google Scholar 38 Lilja, G., Magnusson, C. G., Oman, H. and Johansson, S. G. 1990. Serum levels of IgG subclasses in relation to IgE and atopic disease in early infancy. Clin. Exp. Allergy  20: 407. Google Scholar Author notes Departments of 1Rheumatology and 2Pediatrics, University Medical Centre, Vodnikova 62, 1000 Ljubljana, Slovenia 3Department of Medicine II, Hokkaido University School of Medicine, Sapporo 060‐8638, Japan 4Department of Cell Chemistry, Institute of Cellular and Molecular Biology, Okayama University Medical School, Okayama 700‐8558, Japan
TI - Anti‐β2‐glycoprotein I antibodies in children with atopic dermatitis
JF - International Immunology
DO - 10.1093/intimm/dxf043
DA - 2002-07-01
UR - https://www.deepdyve.com/lp/oxford-university-press/anti-2-glycoprotein-i-antibodies-in-children-with-atopic-dermatitis-Feb61RG5ip
SP - 823
EP - 830
VL - 14
IS - 7
DP - DeepDyve
ER -