Inhibition of allergen‐dependent IgE activity by antibodies of the same specificity but different class

Abstract IgG4 purified from patients undergoing specific allergen immunotherapy inhibits the activities of the serum IgE in in vitro assays and is thought to reduce the symptoms of the disease. However, it is not known whether this is related to an intrinsic property of this subclass or only the allergen specificity. We tested the hypothesis that allergen specificity is the critical determinant for this activity using a panel of antibodies with identical specificity but different subclasses. The different antibodies were all able to inhibit the activity of IgE to the same extent. We demonstrate that specificity is the dominant factor determining the ability of an antibody to block allergen‐dependent IgE activity.

IgG 4 purified from patients undergoing specific allergen immunotherapy inhibits the activities of the serum IgE in in vitro assays and is thought to reduce the symptoms of the disease. However, it is not known whether this is related to an intrinsic property of this subclass or only the allergen specificity. We tested the hypothesis that allergen specificity is the critical determinant for this activity using a panel of antibodies with identical specificity but different subclasses. The different antibodies were all able to inhibit the activity of IgE to the same extent. We demonstrate that specificity is the dominant factor determining the ability of an antibody to block allergen-dependent IgE activity.
Allergy is associated with the excessive production of allergen-specific IgE. However, allergen-specific antibodies of other isotypes are produced both in allergic disease and in states of tolerance, for example in hyperimmune beekeepers and in patients treated by allergen immunotherapy (1,2). The four human IgG subclasses, IgG 1 , IgG 2 , IgG 3 and IgG 4 , differ mainly in the length and rigidity of the hinge region. These differences impart different functional roles, based on their ability to activate the immune system. Similarly, IgA is represented by two subclasses: IgA 1 , which predominates within the serum, and IgA 2 , with a shorter hinge region and more compact structure, which predominates at mucosal surfaces (3) (Fig. 1A).
Serum from patients receiving allergen immunotherapy blocks the activity of IgE, and this inhibitory activity co-elutes with the IgG 4 fraction (4). Furthermore, depletion of IgG 4 from this serum correlates with a decrease in IgE inhibitory activity (5). However, it is unknown whether blocking activity is restricted to the IgG 4 subclasses or whether other subclasses that recognize the same epitope are equally effective in blocking IgE-mediated activity. To test this, we generated a set of recombinant monoclonal antibodies of the same specificity for the grass pollen allergen Phl p 7, with different constant region domains, representing all of the IgG and IgA subclasses. We then measured the affinity of antigen binding and ability of these antibodies to inhibit IgE-mediated activities in in vitro assays.

Methods
Detailed methods are available in the supporting information.

Antibody cloning and expression
Matched heavy-and light-chain variable antibody sequences specific to Phl p 7 allergen were previously isolated from a single B cell derived from a patient undergoing grass pollen immunotherapy (6). These sequences were subcloned into the dual antibody expression vector pVITRO1-102.1F10-IgG4/k (7). Phl p 7-specific human IgG 1, IgG 2, IgG 3 , and IgA 1 and IgA 2 expression vectors were subsequently cloned and expressed using the PIPE method (7).

Characterization of recombinant antibodies
Human 102.1F10 IgG 1 and IgG 2 , and IgG 3 and IgG4 were purified by affinity chromatography with a 5-ml HiTrap Protein-G HP column (GE Healthcare Life Sciences, Amersham, UK). Human 102.1F10 IgA 1 and IgA 2 were purified by affinity chromatography with immobilized SSL7/Agarose (InvivoGen, Toulouse, France). The purified antibodies were analysed by size exclusion chromatography (14), and specificity was confirmed by Phl p 7 allergen ELISA using biotin-labelled isotype-specific antibodies. SPR was performed using a Biacore T200 instrument; antibodies were captured using an immobilized antilambda antibody (Life Technologies Ltd., Paisley, UK), and binding of Phl p 7 (kindly provided by Dr. Rebecca Beavil) was measured using a 3-min association phase followed by 10-min dissociation.

Antibody characterization
Size exclusion chromatography and ELISA confirmed the purified Phl p 7-specific IgG 1 , IgG 2 , IgG 3 , IgG4, IgA 1 and IgA 2 consisted of monodisperse antibodies of the expected size (Fig. 1B) and specificity (Fig. 1C). Changing the constant region had negligible effects on antibody affinities for Phl p 7 (Fig. 1D), with similar K D values obtained for all antibody subclasses tested (Table S1).

IgE blocking activity
As IgG 4 has been previously shown to be an effective blocking antibody for IgE-mediated activity, we wished to determine whether this blocking activity was specific to the IgG 4 subclass. We therefore tested the IgG and IgA subclasses in two independent in vitro assays of IgE activity. Similar to IgG 4 , IgG 1 , IgG 2 , IgG 3 , IgA 1 and IgA 2 were able to inhibit binding of IgE-Phl p 7 complexes to the IgE receptor CD23 (FceRII) on the surface of B cells ( Fig. 2A). In a separate assay, all the monoclonal antibodies tested were able to inhibit IgE-dependent Phl p 7-mediated basophil activation to a similar degree ( Fig. 2B and 2C).

Discussion
The inhibitory activity of non-IgE antibodies in an allergic reaction is thought to be due to their competition with IgE, by masking the epitopes on the allergen. These so-called blocking antibodies represent a potentially valuable but as yet untested therapeutic commodity for use in passive allergen immunotherapy (8). We previously isolated and cloned a monoclonal IgG 4 antibody, specific to the grass pollen allergen Phl p 7, from a single B cell isolated from the peripheral blood of a patient treated by specific allergen immunotherapy (6). This single antibody was able to inhibit Phl p 7-induced IgE activity by up to 60%. This was comparable to the blocking activity of the polyclonal postimmunotherapy serum from the same patient. To test whether this blocking activity was related solely to the specificity or to the (IgG 4 ) subclass of the antibody, we compared the ability of other subclasses to inhibit IgE. Although the affinities for Phl p 7 were similar in all of the recombinant antibodies we generated, subtle but significant differences in binding rates were observed (about threefold differences in both on-and off-rate constants, in the most extreme cases), which tended to cancel out to give very similar overall affinities (ranging from 250-570 pM). These differences may reflect subtle conformational changes that constant region domains impart on the variable region, which have been reported to influence the fine specificity and affinity of isotype-swapped antibodies (9). Nevertheless, we found that the specificity for Phl p 7 was retained and, importantly, isotype exchange had no effect on IgE blocking activity; all subclasses were able to inhibit IgE to nearly the same degree in our in vitro assays. Of course, we cannot exclude the possibility that more subtle effects might be observed by titration of the different antibodies. However, under the conditions used here, blocking activities were dependent only on the ability to bind allergen and not on the constant region effector function.
It is almost certain that the blocking activity of an antibody is dependent on several factors such as epitope specificity, concentration and affinity for antigen (10). The affinity of IgE for allergen is an important determinant of its effector function (11), and a blocking antibody must be of approximately equal or higher affinity to prevent IgE binding. Antibody affinity will therefore be a critical factor for selecting blocking antibodies for passive immunotherapy. It is well established that IgG 4 -expressing B cells secrete the most efficacious blocking antibodies in vivo after specific allergen immunotherapy (2). It follows that they would be the best source of heavy-and light-chain genes from which to derive recombinant blocking antibodies for passive immunotherapy. Indeed, IgG 4 has properties that may favour its use over other IgG subclasses, such as its inability to bind complement, and also the unique property of exchanging one heavy-/light-chain pair with an IgG 4 antibody of a different specificity to generate a bispecific antibody unable to form immune complexes (12).
In summary, our results demonstrate that all IgG and IgA subclasses are capable of inhibiting the activity of IgE in an allergen-specific manner. Further experiments will be required to examine whether antibodies with the same specificity but different isotype are similarly inhibitory in vivo or whether other mechanisms as mentioned above come into play. Phl p 7 is a relatively small allergen, and while our study provides a proof-of-concept for passive immunotherapy, it is likely that combinations of allergen-specific monoclonal antibodies directed against multiple allergens, and possibly multiple epitopes, will be required to ameliorate symptoms depending on the sensitization pattern of each individual (8).
cal Research Centre at Guy's and St Thomas' NHS Foundation Trust and King's College London (HB & TSD). The views expressed are those of the authors and not necessarily those of the NHS, the NIHR or the Department of Health.

Author contributions
TSD, HB and HJB produced and characterized antibodies; MHS performed functional assays; JMM helped to design experiments and analyse data; LKJ, AJB, SRD and BJS helped to design experiments and write the manuscript; and LKJ performed and coordinated experiments. HJG proposed the study and helped to write the manuscript.

Conflict of interest statements
None of the authors have any conflict of interest to declare.

Supporting Information
Additional Supporting Information may be found in the online version of this article: Figure S1. ELISA plates were coated with Phl p 7 and antibody binding was confirmed using subclass-specific monoclonal detection antibodies.
Appendix S1. Methods. Table SI. SPR analysis of the interaction between Phl p 7 and recombinant antibodies (AESE) .