IgE- cross- blocking antibodies to Fagales following sublingual immunotherapy with recombinant Bet v 1

Background: Evidence has accumulated that birch pollen immunotherapy reduces rhinoconjunctivitis to pollen of birch homologous trees. Therapeutic efficacy has been associated with IgE- blocking IgG antibodies. We have recently shown that sera collected after 16 weeks of sublingual immunotherapy with recombinant Bet v 1 (rBet v 1- SLIT) display strong IgE- blocking bioactivity for Bet v 1. Here, we assessed whether rBet v 1- SLIT- induced IgG antibodies display cross- blocking activity to related allergens in Fagales pollen. Methods: IgE, IgG1 and IgG4 reactivity to recombinant Bet v 1, Aln g 1, Car b 1, Ost c 1, Cor a 1, Fag s 1, Cas s 1 and Que a 1 were assessed in pre- and post- SLIT samples of 17 individuals by ELISA. A basophil inhibition assay using stripped basophils re- sensitized with a serum pool containing high Bet v 1- specific IgE levels was es-tablished and used to assess CD63 expression in response to allergens after incubation with pre- SLIT or post- SLIT samples. IgG1 and IgG4 were depleted from post- SLIT samples to assess its contribution to IgE- cross- blocking. Results: Sublingual immunotherapy with recombinant Bet v 1 boosted cross- reactive IgE antibodies and induced IgG1 and IgG4 antibodies with inter- and intra- individually differing reactivity to the homologs. Highly variable cross- blocking activities of post-SLIT samples to the different allergens were found. IgG1 and IgG4 antibodies displayed cross-


| INTRODUC TI ON
Pollen from birch trees is a prominent elicitor of allergic rhinoconjunctivitis in Europe and North America with an estimated prevalence of 20% in adults. 1 Sensitization to the major birch pollen allergen Bet v 1 induces complex patterns of IgE-crossreactivity with homologous proteins in various foods termed birch pollen-related food allergy. 2,3 Bet v 1-homologs also exist in pollen from trees of the order Fagales, for example Alnus glutinosa (alder), Carpinus betulus (hornbeam), Ostrya carpinifolia (European hop-hornbeam), Corylus avellana (hazel), Fagus sylvatica (beech), Castanea sativa (chestnut) and Quercus alba (oak), 4-11 and IgEcross-reactivity within this group of birch homologous trees prolongs the period of respiratory allergic symptoms for most birch pollen-allergic patients. 12,13 Based on IgE competition experiments, Bet v 1 is considered as the primary sensitizer in allergies to Fagales pollen. 5,8,[13][14][15] Along these lines, recombinant (r) Bet v 1 in combination with natural birch extract has been shown to identify patients with allergic rhinitis to homologous trees with a sensitivity of 99.2%. 16 Moreover, birch allergens have been suggested as exclusive tools for the diagnosis of Fagales pollen allergy also in birch-free areas. 17 Beyond diagnostic application, allergen-specific immunotherapy (AIT) with birch pollen has been considered to reduce allergic symptoms to pollen from homologous trees. Indeed, individuals who received AIT with birch pollen showed a significant improvement of symptoms not only during the birch pollen season but also during the pollination of hazel and alder. 12,13,18 A recent double-blind phase II trial demonstrated that 24 weeks of sublingual immunotherapy (SLIT) with birch pollen reduces allergy to both birch and oak. 15 Moreover, birch pollen SLIT induced significant increases of IgG4 antibodies that cross-reacted with allergens from alder, beech, hazel and oak. 15 The clinical success of SLIT has been associated with the induction of blocking IgG antibodies, which competitively prevent IgE binding to the administered allergens and thereby inhibit effector cell activation. 19,20 We recently found that sera collected after treatment with a daily sublingual dose of 25 µg rBet v 1 for 16 weeks potently blocked Bet v 1-induced activation of basophils. 21 Here, we sought to assess whether this functional IgE-blocking activity can be expanded to the Bet v 1-homologs in alder, beech, chestnut, hazel, hornbeam, hop-hornbeam and oak, respectively. We employed single recombinant allergens to characterize and compare the crossreactivity and inhibitory activity of rBet v 1-SLIT-induced IgG1 and IgG4 antibodies. For the first time, the IgE-cross-blocking activity of IgG antibodies following therapy with a reference allergen to a panel of homologous allergens was functionally investigated.

| Allergens
All allergens were recombinant proteins displaying the same IgEbinding capacity as their natural counterparts. 22 Bet v 1.0101 was produced in-house as described. 21 Car b 1.0109 (hornbeam), Cas s 1.0101 (chestnut), Fag s 1.0101 (beech), Ost c 1.0101 (European hop-hornbeam) and Que a 1.0301 (oak) were produced as previously described. 23 Aln g 1.0101 (alder) and Cor a 1.0103 (hazel) were purchased from Biomay, Vienna, Austria.

G R A P H I C A L A B S T R A C T
Bet v 1-allergic individuals show IgE cross-reactivity with allergens from Fagales. Sublingual immunotherapy with the reference allergen Bet v 1 induces individually diverse repertoires of cross-reactive IgG1 and IgG4 antibodies. The cross-blocking bioactivity of IgG1 and IgG4 antibodies is highly variable and not predictable from IgE-cross-reactivity.

| Serum and plasma samples
Serum and plasma samples (diluted 1:2 in PBS) were collected at baseline (PRE) and after 16 weeks (POST) of daily SLIT with 25 µg of rBet v 1 (ClinicalTrials.gov number NCT01449786 and EudraCT no 2011-001221-24). 24 Ethical clearance for this study has been provided by the local ethics committee (EK228/2011). Sera from six untreated birch pollen-allergic individuals, each containing >100 kU A /l of Bet v 1-specific IgE, were collected and pooled. The pool was characterized by ImmunoCAP (Thermo Fisher, Waltham, Massachusetts, USA) for IgE levels specific for rBet v 1 (224 kU A /L), rBet v 2 (0.44 kU A /L), rBet v 4 (<0.35 kU A /L) and common allergen sources in Austria (see Table S1).

| Analysis of allergen-specific antibodies
Bet v 1-specific IgE and IgG4 levels were quantified by ImmunoCAP (Thermo Fisher). ELISA was applied to quantify Bet v 1-specific IgG1 levels using an in-house standard and for the detection of Fagales-

| Basophil inhibition tests
Heparinized blood was collected from untreated birch pollen-allergic individuals and allergic individuals without birch pollen allergy with informed consent and ethical clearance by the local ethics committee (EK1344/2018). Basophil inhibition tests were performed with full blood from untreated birch pollen-allergic individuals as described. 21 To strip basophils, PBMC (70-150 × 10 6 ) from allergic individuals without birch pollen allergy were incubated with 10 ml of a chilled solution of 130 mM NaCl, 5 mM KCl and 10 mM lactic acid (pH 3.9) for 5 min at 4°C. 25

| Depletion of IgG1 and IgG4 antibodies
Plasma samples (2 ml) were incubated with CaptureSelect IgG1 (400 µl) or CaptureSelect IgG4 (200 µl) Human affinity matrix (Thermo Fisher Scientific, Waltham, Massachusetts, USA) for 30 min at RT. After centrifugation, fresh CaptureSelect matrix was added to the supernatant for 30 min at room temperature. Four and two cycles were performed to deplete IgG1 and IgG4, respectively. After centrifugation, the supernatant was assessed for allergen-specific IgG antibodies by ELISA.  (Table 1).

| Statistics
During rBet v 1-SLIT, Bet v 1-specific IgE levels rose significantly (median values pre-SLIT: 11.2 kU A /L and post-SLIT: 36.1 kU A /L, p < .001). 21 Except for Cor a 1, the relative change (RC) of Bet v 1-specific IgE levels before and after SLIT correlated significantly with the RC of IgE reactivity to each homolog (R values ranging from .607 to .896, Figure 1). All samples from non-allergic individuals were negative (data not shown).  Table S2). Figure Figure 2B). Looking at single subjects, diverse inter-and intra-individual patterns of cross-reactive IgG antibodies became evident (see Table S2). Several individuals developed both IgG1 and IgG4 antibodies to particular homologs, whereas some, for example subjects 7, 8, 10, 11, and 14, developed cross-reacting IgG1 but no IgG4 antibodies. Others, like subjects 4, 5, 12 and 13, showed enhanced IgG4 but not IgG1 levels to particular homologs following rBet v 1-SLIT.

| Cross-blocking activity of rBet v 1-SLITinduced IgG1 and IgG4 antibodies
To elucidate the cross-blocking activity of rBet v 1-SLIT-induced IgG antibodies, we chose seven individuals (no. 1-7) whose post-SLIT sera and showed diverse responses of cross-reactive IgG1 and IgG4 antibodies (see Table S2). Their pre-and post-SLIT samples were incubated with Bet v 1-homologs prior to addition to basophils from three untreated birch pollen-allergic individuals. Notably, the same post-SLIT samples showed stark differences of IgE-blocking activity to the same homologs in the three experiments (see Figure S1).
To establish a donor-independent basophil activation test, we stripped basophils from non-Bet v 1-sensitized allergic individuals and incubated them with a serum pool containing high levels of Bet v 1-specific IgE antibodies (see Table S1) and cross-reactive with all homologs (data not shown). Re-sensitization of basophils with a 1:25 dilution of the pool was found to be optimal for activation with Bet v 1 (see Figure S2). Then, re-sensitized basophils were stimulated with titrated amounts of all Bet v 1-homologs in two independent experiments (see Figure S3). All recombinant allergens induced reproducible basophil activation. Cas s 1, which was consistently a weak stimulus, was excluded from further testing. Next, re-sensitized basophils were stimulated with the allergens after incubation with either pre-or post-SLIT samples. The percentages of inhibition of basophil activation are depicted in Figure 3. Overall, very distinct cross-blocking activities were observed and reproduced in three independent experiments (see Figure S4). Only subject 1 displayed a strong cross-blocking activity (>80%) to all homologs, whereas the others showed individual cross-blocking activities to the allergens.
All individuals showed cross-blocking to Aln g 1 and Fag s 1, though to a highly diverse extent.

IgE-cross-blocking
To address the relevance of IgG1 and IgG4 for cross-blocking, we depleted these subclasses from post-SLIT samples of subjects 1-3 who had developed both IgG1 and IgG4 responses to several Bet v

1-homologs. Successful removal of allergen-specific IgG1 and IgG4
and preservation of the respective other subclass were confirmed by ELISA (see Figure S5). Thereafter, the blocking activities of de- The rise of allergen-specific IgG antibodies is the most consistent immunological finding in AIT; however, only their functional IgEblocking activity has been associated with therapeutic efficacy. 21,26 Therefore, this study aimed at the assessment of cross-blocking bioactivity by employing basophil inhibition assays as they in our opinion best mirror prevention of effector cell activation in vivo.
Notably, assays with basophils from three birch pollen-allergic individuals who displayed IgE reactivity to all studied allergens resulted in very discrepant blocking activities of the same post-SLIT sample. We referred the observed inhomogeneity to distinct IgE-epitope recognition patterns of the different basophil donors. To establish a more robust, donor-independent assay, we enhanced the IgE repertoire by re-sensitizing stripped basophils with a pool of Bet v 1-specific IgE Allergens incubated with pre-and post-SLIT samples from 7 subjects were added to stripped and re-sensitized basophils. The percentages of CD63 + basophils with pre-SLIT sera were normalized to 100%, and the percentage of inhibition calculated. rBet v 1-SLIT-induced IgG1 and IgG4 responses to the allergens were classified as described in Material & Methods response but newly evolved antibodies. Such IgG antibodies may recognize epitopes that are more variable among the different homologs as opposed to more conserved IgE epitopes. Accordingly, we have found that SLIT with rMal d 1, the Bet v 1 homolog in apple, induced a de novo IgG response specific for epitopes exclusive of the apple allergen. 21 However, it is conceivable that a broader repertoire of IgG antibodies develops following SLIT for more than 16 weeks.
On the other hand, our findings may hint to the existence of IgE epitopes exclusive for the individual Bet v 1-homologs. Such epitopes have been implied by previous studies reporting Fagales pollen allergy in birch-free areas 17 and demonstrating that IgE antibodies from allergic individuals in birch-free areas reacted stronger to Car b 1 and Que a 1 than to Bet v 1. 23 We also observed that some individ- Further work is required to understand why sensitization to an allergen induces cross-reactive IgE antibodies of clinical relevance whereas desensitization with the same allergen fails to induce functional cross-blocking IgG antibodies.

ACK N OWLED G EM ENTS
The authors wish to thank Birgit Nagl and Sandra Faustmann for excellent help in blood sample processing who have nothing to disclose. We are grateful for helpful discussions with F I G U R E 4 IgE-blocking activity of post-rBet v 1-SLIT samples after depletion of IgG1 and IgG4. Allergens pre-incubated with pre-SLIT or post-SLIT samples with and without IgG1 or IgG4 were used to stimulate re-sensitized basophils. The percentages of CD63 + basophils with pre-SLIT samples were normalized to 100%. n.t., not tested