Persistence and evolution of allergen-specific IgE repertoires during subcutaneous specific immunotherapy

Abstract

Background

Specific immunotherapy (SIT) is the only treatment with proven long-term curative potential in allergic disease. Allergen-specific IgE is the causative agent of allergic disease, and antibodies contribute to SIT, but the effects of SIT on aeroallergen-specific B cell repertoires are not well understood.

Objective

To characterize the IgE sequences expressed by allergen-specific B cells, and track the fate of these B cell clones during SIT.

Methods

We have used high-throughput antibody gene sequencing and identification of allergen-specific IgE using combinatorial antibody fragment library technology to analyze immunoglobulin repertoires of blood and nasal mucosa of aeroallergen-sensitized individuals before and during the first year of subcutaneous SIT.

Results

Of 52 distinct allergen-specific IgE heavy chains from eight allergic donors, 37 were also detected by high-throughput antibody gene sequencing of blood, nasal mucosa, or both sample types. The allergen-specific clones had increased persistence, higher likelihood of belonging to clones expressing other switched isotypes, and possibly larger clone size than the rest of the IgE repertoire. Clone members in nasal tissue showed close mutational relationships.

Conclusion

Combining functional binding studies, deep antibody repertoire sequencing, and information on clinical outcomes in larger studies may in the future aid assessment of SIT mechanisms and efficacy.

Introduction

Type I allergic responses are initiated by the cross-linkage of allergen-specific IgE, bound to the high-affinity IgE receptors (FcεRI) on the surface effector cells, by allergen. This initiates a series of events involving degranulation of mast cells and basophils. The released effector molecules, such as histamines, proteases, chemokines, and cytokines, eventually give rise to the symptoms associated with allergic disease (1, 2). IgE antibodies indisputably play a key role in determining the allergen specificity of allergic disease, a condition that affects the quality of life of as many as a quarter of the population inflicting morbidity in individuals and large costs to society (3, 4).

Allergen-specific immunotherapy (SIT) is currently the only available disease-modifying method used to treat allergic disease with both established long-term clinical (5, 6) and cost efficacy (7). It has been applied for over a century (8) and aims at inducing tolerance to the sensitizer(s) via repeated injections of increasingly higher doses of allergen. Although not entirely understood, its efficacy is considered to involve induction of allergen-specific T regulatory cells, and reduced reactivity of several effector cell types, such as mast cells, basophils and eosinophils, as well as the induction of allergen-specific blocking IgG (9–12). The levels of allergen-specific IgE, however, remain relatively stable during the treatment (13, 14). Although levels of polyclonal allergen-specific antibodies of several isotypes during SIT have been thoroughly investigated using serological methods, very little is known about the sequences of the antibodies, the clonal dynamics of the B cell populations that produce them, and how the B cell populations are altered during the course of treatment.

Characterization of IgE repertoires in general has been addressed in several studies (reviewed by Gadermaier et al. (15)). These pioneering investigations have provided useful insights into the characteristics of IgE repertoires, such as their oligoclonal nature. They have begun to address the relationship between IgE repertoires in different tissues, and evolution of repertoires and single clones over time (16–19), albeit based on analysis of a limited number of transcripts from small numbers of samples and donors. More extensive data collection is important to advance our understanding of procedures like SIT that aim to modify disease by progressive modulation of the immune response. Furthermore, studies of the relationship between quality and complexity of IgE repertoires on biological outcomes have vastly enhanced our understanding of fundamental aspects of the allergic response (20, 21).

The development of highly efficient DNA sequencing technologies, and tools to analyze sequences encoding antibody variable domains now permits allergy researchers to address antibody responses in allergy in new ways. Wu et al. recently used this approach to identify seasonal changes in IgE-expressing total B cell repertoires in the blood and nasal biopsies of subjects with allergic rhinitis, and reported increased diversity and mutation of IgE transcripts during grass pollen season, as well as noting clonal lineages shared between blood and tissue (22). Another recent such study of total IgE repertoires of unknown specificity in allergic subjects characterized somatic mutation patterns and found decreased evidence for antigen selection in IgE compared to IgG (23). To extend the study of IgE-expressing B cell repertoires to include identification of allergen-specific cells, we have employed a combination of human recombinant antibody technology and high throughput sequencing (HTS) of antibody repertoires from primary B cell samples. Genes encoding immunoglobulin heavy chain variable (IGHV) domains were sequenced from peripheral blood and nasal tissue B cells of allergic donors that underwent SIT (n=8), and additional allergic donors who were not treated (n=8). Using phage display technology we identified a set (n=52) of distinct allergen-specific antibody IGHV domains from IgE-expressing B cells in the peripheral blood B cells of the SIT patients. HTS enabled detection of members of the allergen-specific B cell clones that expressed non-IgE isotypes, and tracking of the fate of allergen-specific clones during the course of SIT.

Material & Methods

Allergens

Recombinant Bet v 1.0101, Bet v 2.0101, Phl p 1.0102, Phl p 2.0202, Phl p 5.0101, Phl p 6.0101 and Art v 1.0101 were obtained from Biomay (Vienna, Austria) while recombinant Der p 1.0102, Der p 7.0101, purified natural Der p 1, Der p 2 and Fel d 1 were obtained from Indoor Biotechnologies (Cardiff, Wales).

Patient samples

Samples (blood and/or nasal biopsy) were obtained from 16 allergic individuals, 8 of which (donor number 1–8) underwent SIT. SIT donors were sampled during autumn immediately before the start of SIT (“0 month” time point), when an allergen dose of 10 000 SQ units had been reached, corresponding to approximately 7 weekly injections (still before the onset of natural exposure to common seasonal allergens), (“2 months” time point), and approximately one year after treatment start (“1 year” time point), while donors not undergoing SIT (donor number 9–16) were sampled at the initiation of the study and one year later (Figure 1). Details of the sample collection procedure are provided in the Supplementary Material and Methods section in the Online Repository. The study was approved by the regional ethical board at Lund University.

Figure 1.

Timing of SIT and sample collection. Blood samples at 2 months and 1 year timepoints were obtained approximately 1 week after the preceding SIT vaccine injection. Samples were not obtained from the non-vaccinated group at 2 months.

Determination of allergen-specific IgG4 and IgE levels

Allergen-specific IgG4 and IgE levels in serum samples from all 16 donors collected at the 0 month and 1 year time points were measured using the ISAC microarray system (Phadia Multiplexing (PMD), Vienna, Austria) containing 103 native or recombinant allergens. The analysis was performed as previously described (24).

Antibody heavy chain variable domain-encoding transcriptome analysis

Antibody heavy chain variable domain-encoding genes were sequenced using the 454 (Roche) platform using Titanium chemistry and subsequently analyzed as described in the Supplementary Material and Methods section in the Online Repository. Sequences analyzed in this study can be accessed via dbGAP record [in process].

Isolation and analysis of allergen-specific scFv generated from IgE-expressing B cells

Combinatorial single chain fragment variable (scFv) libraries based on IGHV-encoding transcript with origin in the IgE repertoire were constructed from cDNA samples derived from peripheral blood mononuclear cells collected at the 2 month time point from donors undergoing SIT and used for phage display-based selection of allergen-specific antibody fragments as outlined in the Supplementary Material and Methods section in the Online Repository.

Results

SIT induces increased production of allergen-specific IgG4 while allergen-specific IgE levels are unaltered

To assess whether or not SIT had any immunological impact in the subjects undergoing treatment, we determined the levels of allergen-specific IgE before initiation of SIT and after 1 year of SIT. As expected from an SIT procedure (13, 14, 25), the first year of vaccination induced increases in allergen-specific IgG4 against relevant allergens, while it had no or small effects on allergen-specific IgE (Supplementary Table EI and Supplementary Figure E1 in the Online Repository). No significant changes in allergen-specific IgG4 levels specific for a given allergen source were seen in allergic controls not undergoing SIT or in patients undergoing SIT with other extracts.

Isolation of allergen specific antibody fragments with origin in the human IgE repertoire

To study the fate of allergen-specific IgE clones during the course of SIT, we constructed combinatorial scFv libraries based on IGHV-genes expressed in the peripheral blood IgE+ B cell lineage repertoire at the 2 months sampling of the donors undergoing SIT. Binders specific for major allergens to which the patients were sensitized (Table I) were isolated using phage display selection. We successfully isolated a total of 52 allergen-specific binders. These binders, specific for 7 allergens, were retrieved from the 8 donor libraries, with each donor contributing a minimum of three binders (Table II). The isolated scFv were diverse in terms of IGHV, IGHD and IGHJ gene segment origin as well as CDRH3 lengths (15.4 ± 4.5 residues). All isolated binders were specific for the allergen they were selected against, with no or little reactivity to BSA (Supplementary Figure E2 in the Online Repository) or unrelated allergens (exemplified in Supplementary Figure E2 in the Online Repository), as determined by ELISA.

Table I.

Donor characteristics and allergen sources included in specific immunotherapy.

Donor ID	Age	Sex	Skin prick test1	Specific immunotherapy1	scFv-libraries selected on3
1	38	F	BP, GP2	BP	Bet v 1, Bet v 2
2	30	M	BP, GP, CD, DD	BP, GP, CD	Bet v 1, Bet v 2, Phl p 1, Phl p 2, Phl p 5, Phl p 6, Fel d 1
3	30	M	BP, GP	BP, GP	Bet v 1, Bet v 2, Phl p 1, Phl p 2, Phl p 5, Phl p 6
4	36	F	BP, GP, HDM, CD, DD	BP, GP, HDM	Bet v 1, Bet v 2, Phl p 1, Phl p 2, Phl p 5, Phl p 6, nDer p 14, rDer p 14, Der p 2, Der p 7
5	23	M	HDM, AF, CD, DD	HDM	nDer p 1, rDer p 1, Der p 2, Der p 7
6	41	F	BP, MP, HDM, MM	BP, MP, HDM	Bet v 1, Bet v 2, Art v 1, nDer p 1, rDer p 1, Der p 2, Der p 7
7	27	M	BP, GP, MP	BP, GP, MP	Bet v 1, Bet v 2, Phl p 1, Phl p 2, Phl p 5, Phl p 6, Art v 1
8	24	F	BP, GP, CD, DD, HD	BP, GP, CD	Bet v 1, Bet v 2, Phl p 1, Phl p 2, Phl p 5, Phl p 6, Fel d 1
9	25	M	BP, GP, MP, DD, CD	No	-
10	30	F	BP, GP, MP, DD,	No	-
11	23	M	BP, GP	No	-
12	24	F	HDM, BP	No	-
13	38	M	BP, GP, DD	No	-
14	22	M	BP, DD	No	-
15	24	F	BP, GP, HDM, CD, DD	No	-
16	47	M	BP, GP, CD, DD	No	-

¹AF, Aspergillus fumigatus; BP, birch pollen; CD, cat dander; DD, dog dander; GP, grass pollen; HD, horse dander; HDM, house dust mites; MM, mold mix; MP, mugwort pollen

²5 grass mixture containing Dactilis glomerata, Festuca pratensis, Lolium perenne, Phleum pratense, and Poa pratensis

³Allergen isoforms are clarified in the methods section.

⁴nDer p 1, natural Der p 1; rDer p 1, recombinant Der p 1

Table II.

Allergen-specific scFv isolated from combinatorial libraries based on IGHV with origin in the IgE repertoire.

Donor	Clone ID	Allergen	IGHV	IGHV mutations (%)	CDRH3	Identified by HTS
1	IT1-B215	Bet v 2	1–18	0	ARVRSSGYGNWFDP
1	IT1-B228	Bet v 2	1–69	0	ARDNGGEGY
1	IT1-B237	Bet v 2	3–21	4.2	ARGGTRYFAS
1	IT1-B248	Bet v 2	3–11	0	ARDRSPIAAARLAFDI
2	IT2-B12	Bet v 1	5–51	9.8	ARLGGGSRGWYYYYGMDV	Yes
2	IT2-B15	Bet v 1	3–48	5.3	ARGYYDGRKLRN
2	IT2-B116	Bet v 1	3–11	2.6	ARDAYFDWSLDY	Yes
2	IT2-B23	Bet v 2	3–66	5.0	ARGYYDGRKLRN	Yes
2	IT2-B227	Bet v 2	3–74	3.8	ARRDVAVVPGATGDNYYYGLDV	Yes
2	IT2-B239	Bet v 2	1–18	3.0	ARSLFLLSPKARGYYYYGMDV	Yes
2	IT2-P11	Phl p 1	3–23	4.5	ANLGFDY	Yes
2	IT2-P514	Phl p 5	1–69	8.7	ARALIPLESQAAD	Yes
2	IT2-P536	Phl p 5	1–69	5.7	ASMGRGYCSGDNCYNFDH	Yes
2	IT2-P64	Phl p 6	5–51	3.8	ARRFGGDWFGSLGWYYFDH	Yes
2	IT2-P614	Phl p 6	3–48	3.4	ARDRGTPWHYDGMDV	Yes
3	IT3-B11	Bet v 1	1–46	10.2	ASDIAEGLGQHLFDH
3	IT3-B116	Bet v 1	3–30	6.8	ASSLTSTGMGRY
3	IT3-B229	Bet v 2	1–69	12.1	ARDRKFYYDRSGVPYFDH
3	IT3-P11	Phl p 1	4-4	3.4	ARDGKNGSSDY	Yes
4	IT4-B119	Bet v 1	3–48	4.2	ARGPGYSSSWYGYYFDS	Yes
4	IT4-B148	Bet v 1	3–48	11.3	ASPPTSYDFWSDYSDYDYYYMDV	Yes
4	IT4-B225	Bet v 2	3–30	3.4	ARDLTGNFAN	Yes
4	IT4-B229	Bet v 2	3–20	9.8	ARDLIGNFAN
4	IT4-P517	Phl p 5	3–7	4.5	ARDSRQWWFHIEGDAFDI	Yes
4	IT4-P62	Phl p 6	3–33	7.6	ARDFFPLAVLSPPLGY	Yes
5	IT5-nD134	Der p 1	3–53	0.8	AREGGVAARPNPDAFDI	Yes
5	IT5-rD13	Der p 1	3–11	5.3	ARDGALVWFGNKSYGMDV	Yes
5	IT5-rD120	Der p 1	3–23	5.7	AKAHGFGSPGWGSGWHRKTPSRYPYYFDY	Yes
6	IT6-B15	Bet v 1	3–30	11.3	ARGRPRAYYYDDSGSKQYWDEHYFDY	Yes
6	IT6-B22	Bet v 2	3–30	10.6	ARDIFASQGAADY	Yes
6	IT6-nD12	Der p 1	3–53	8.4	ARLYYDFWSGHAYFFYYMDV	Yes
6	IT6-nD13	Der p 1	3–23	8.7	AKSERPIVATITGYYYYYMDV	Yes
6	IT6-nD16	Der p 1	3–30	9.1	AKAAYSYGMKNSLDF	Yes
6	IT6-nD17	Der p 1	3–30	17.0	TRDRSPVSGVLQH	Yes
6	IT6-rD111	Der p 1	3–64	6.0	VKNMVRGVITDAFEI	Yes
6	IT6-rD116	Der p 1	3–49	14.4	SRGLWFGKLWGPPREH	Yes
6	IT6-rD128	Der p 1	3–30	10.9	ARDEGGDSSGNH	Yes
7	IT7-B14	Bet v 1	1–2	13.6	ARGLRSQLWYLDV	Yes
7	IT7-B22	Bet v 2	3–23	0.4	APSRYCSGGSCYSGY
7	IT7-B210	Bet v 2	3–23	9.1	VKEKGWQQLPKGGHNWFDP	Yes
7	IT7-B214	Bet v 2	3–7	6.0	TRQSGWLPFKD	Yes
7	IT7-B215	Bet v 2	3–21	6.0	ARDLSYSGGD
7	IT7-B222	Bet v 2	3–30	0	AKAPRSSSFRYFQH
7	IT7-P51	Phl p 5	3–21	2.6	ARERSPWSEEAFDV	Yes
7	IT7-P517	Phl p 5	3–23	9.8	AKDGISEYCSGGSCHSRGWVYFDY	Yes
7	IT7-P61	Phl p 6	3–21	1.5	TRSPYYSGSGSYLDN	Yes
8	IT8-B13	Bet v 1	3–49	4.8	TRRATLYYDTSGYSYYFDY	Yes
8	IT8-B22	Bet v 2	1–18	0	ARGVGSRRQVYYYGMDV	Yes
8	IT8-B216	Bet v 2	3–11	0	ARVGRVRGVIRFDP
8	IT8-B243	Bet v 2	3–23	2.3	AKGGSGTISTP
8	IT8-P220	Phl p 2	4–31	7.1	ARGAGDFDS	Yes
8	IT8-P621	Phl p 6	3–30	9.4	ARGVGDYVWGPKGDY

High-throughput DNA sequencing of immunoglobulin repertoires during SIT

Transcripts encoding antibody heavy chain variable domains derived from the eight subjects undergoing SIT were sequenced using 454 sequencing technology. Blood samples obtained before initiation of SIT (0 months), during dose escalation (2 months) and after approximately 1 year of SIT were obtained for this purpose. Similarly, nasal biopsies obtained at the 0 month and 1 year time points were also analyzed. Productively-rearranged IGH repertoires amplified from genomic DNA template of all 16 donors showed overall similarity of IGHV and IGHJ gene segment usage, level of somatic hypermutation, and mean CDRH3 length and hydrophobicity both in relation to each other and to previously studied immunoglobulin repertoires from productive rearrangements in healthy subjects, indicating that SIT does not cause gross perturbations of B cell populations (Supplementary Figure E3–E5 in the Online Repository). Differences existed though in the level of mutations of several isotypes/subclasses, such as IgM, IgD and IgG3, between samples derived from blood and nasal biopsies, with the latter set of samples being more mutated (Supplementary Figure E5 in the Online Repository). These differences were, however, not related to whether the patient was engaged in SIT, or not. Although this might be a feature related to allergy we believe that it more likely is a general difference in the architecture of B cell populations between this type of tissue and blood. HTS libraries were additionally prepared from cDNA template so that the isotype encoded by each sequence could be identified. The IgE repertoires of these samples revealed relatively few (<150/donor) distinct IgE-encoding rearrangements, compared to rearrangements encoding other isotypes, before initiation of SIT despite the fact that the number of reads per unique sequence was substantially larger for the IgE population in comparison to other isotypes (Supplementary Tables EII–EIV in the Online Repository), in agreement with findings obtained by conventional sequencing approaches (15).

Detection and sequence characterization of clonal lineages containing allergen-specific antibodies

HTS approaches offer the possibility of a comprehensive detection of antibody sequences expressed by B cell clones in human clinical samples. By comparing the IGHV gene usage and sequences of the nucleotides encoding CDRH3 of antigen-specific scFv with the sequences obtained by HTS of total B cell repertoires, we were able to identify 2548 additional sequences that were members of the antigen-specific clones detected by phage display from the patient samples. These sequences represented additional members of 37 of the scFv IgE clones (1–230 sequences/scFv), encoding specificities for Bet v 1, Bet v 2, Phl p 1, Phl p 2, Phl p 5, Phl p 6, and Der p 1 (Table II). One of these sequences (IT2-P11) also appeared in sequences derived from subjects other than the individual from whom the phage displayed antibody was isolated; this likely represent a public rearrangement (26) as it had been established through a process involving addition of very few N nucleotides (Supplementary Figure E6 in the Online Repository). However, to avoid the possibility of PCR contamination affecting the results, this sequence was excluded from further analysis. HTS did not detect 15 of the phage-displayed IgE clones, suggesting that these clones may have been particularly rare in the IgE B cell repertoires, or that the experimental protocols for HTS compared to phage display library preparation and selection may have favored detection of some sequences compared to others (27).

Not surprisingly, sequences representing all of the scFv were identified by deep sequencing of IGH repertoires from the peripheral blood collected at 2 months, from the same RNA sample that was used to create the phage display libraries. To learn if variants of these clones had been present prior to the initiation of SIT and/or if they persisted in circulation over time, samples obtained at other points in time were also queried with respect to the presence of these CDRH3 (Figure 2). Sequences belonging to 4 and 13 of the scFv clones from 2 months were identified in the individuals from whom the phage displayed IgE was isolated, at 0 months and after 1 year of SIT, respectively (Figure 2). In particular, prior to SIT treatment of donor 6, two of the allergen-specific clones identified at 2 months were already present in the blood, and two allergen-specific clones were in the nasal biopsy. A large number of donors had allergen-specific clones also detected at one year of SIT (three in the blood of donor 2; three in the blood of donor 4; one in the blood of donor 5; and four in the biopsy and three in the blood of donor 6). These findings suggest that there is variability in the frequency of individual allergen-specific clones over time, with more clones persistently detectable after SIT has begun. The data from donor 6, however, clearly show that some allergen-specific B cell clones stimulated by SIT are part of the pre-treatment pool of IgE-expressing B cells in the allergic subject, and demonstrate that these clones can be found in circulation, as well as at the nasal tissue site affected by allergic reactions to aeroallergens. Our inability to detect other allergen-specific clones at other time points in the allergic subjects likely reflects the overall rarity of allergen-specific B cells in the blood, resulting in variability in their detection using standard volume clinical blood samples (28). To attempt to address this matter, blood samples were sequenced in biological duplicate. Among samples collected at 2 months, for which there also was positive identification of the IgE-producing clone set in blood samples at a different point, members of that clone set were identified in both duplicates in at least one of the time points in 9/11 cases, supporting the interpretation that more frequently-detected clones are present as larger clones of IgE-expressing B cells in the blood (Supplementary Table EV in the Online Repository). For clones that were detected by HTS in blood only at 2 months, it was possible to detect scFv-related heavy chain variable domain-encoding transcripts in both samples analyzed by HTS in only 3/20 instances (the 6 sequences derived from donor 7 could not be assessed in this manner as one of two duplicate samples failed to amplify in the sequencing assay). Despite the scarcity of the allergen-specific IgE clones in the blood, the persistent detection of clone members at different time points enables evaluation of clonal features such as isotype expression, mutation status, and tissue localization, to assess evolution of allergen-specific IgE during the SIT time course.

Figure 2.

Detection of sequences related to the isolated allergen-specific scFv over the time course of SIT, using HTS of antibody heavy chain rearrangements. Filled boxes indicate the time points at which such related sequences were also detected using HTS in the blood (red) or nasal biopsy (blue) samples. Sequences only identified after selection of scFv from antibody libraries but not by HTS are not shown.

Expression of other isotypes in allergen-specific clones containing IgE members

Although the vast majority of the allergen-specific transcripts identified by HTS in our experiments encoded IgE, both IgG- and IgA-expressing members of these clones were also identified in nasal biopsy tissues and blood. IgA or IgG-expressing members of four IgE scFv clones were identified in a nasal mucosa sample obtained one year after initiation of SIT. In addition IgG-encoding sequences related to four of the allergen-specific IgE clone sets were also detected in blood samples. These sequences give us an opportunity to evaluate the relationship between allergy-causing IgE and members of the same B cell clones expressing antibodies of other isotypes that may contribute to resolution of the allergic condition by SIT. It appears from our data that B cell clones containing IgE-expressing members as well as IgA members are enriched for allergen-specificity (Figure 3A). One interpretation of this finding is that SIT may stimulate clonal expansion, entry into peripheral circulation, and possibly isotype switching, of allergen-specific clones. Alternatively, clinically relevant, allergen-specific IgE B cell clones in each patient might be among the largest clones in their repertoires, and contain members expressing a variety of antibody isotypes, facilitating their detection by HTS.

Figure 3.

Analysis of isotype expression, tissue localization, and clonal persistence of B cells belonging to clonal lineages containing IgE-expressing members in SIT patients. (A) Isotype expression by allergen-specific B cell clones containing IgE members for which the allergen specificity was determined, compared to clones from SIT patients where the IgE specificity is not known. The clones of known allergen specificity are more likely to contain B cells expressing other IgA-switched antibodies. Significance was determined by Fisher’s exact test (two-tailed p-values * = 0.01 to 0.05, ** = 0.001 to 0.01, *** = 0.0001 to 0.001, **** = < 0.0001) Clones containing only IgE members are not shown on graph, but the difference was not significant with a two-tailed p-value of 0.077. Numbers above bars indicate the total number of clusters representing the given category. (B) Tissue and blood distribution of allergen-specific B cell clones containing IgE-expressing members. Clones of known allergen specificity were detected in both nasal biopsy and blood samples more frequently than clones whose allergen specificity was not identified. Data from all time points were included in this analysis. Clones in the ‘blood’ or ‘biopsy’ categories are also counted in the ‘blood and biopsy’ category. (C) Persistent detection of B cell from known allergen-specific clones versus clones of unknown specificity, during SIT. The clones whose allergen-specificity was identified were more likely to be identified in more than one time point during the SIT course. (D) Time points at which allergen-specific IgE clones were detected. Numbers on the x-axis indicate the time point in days.

Clones expressing allergen-specific IgE are present both in nasal biopsies and peripheral blood

To compare the repertoires of allergen-specific IgE in peripheral blood with those at the tissue sites of allergic disease, we searched for such sequences in nasal biopsies collected from subjects undergoing SIT. We were able to identify 5, 5 and 14 reads related to our selected clone sets encoding antibodies of IgE, IgG and IgA isotype, respectively, in one of the mucosal transcriptomes. These sequences represented substantially mutated heavy chain variable domains sequence variants of four mite allergen (Der p 1)-specific binders selected by phage display. Thus, despite the fact that each biopsy represents a very small, localized sample of nasal tissue, members of high-frequency IgE-expressing clones from the blood are present. Notably, allergen-specific IgE sequences are detected more frequently in both blood and tissue, compared to total IgE populations that are not of known allergen specificity, which tend to be seen in either blood or tissue alone (Figure 3B). This finding is consistent with clonal expansion of allergen-specific B cell clones induced by SIT, without a comparable effect on clones with no defined allergen specificity. In subjects not undergoing SIT, a higher proportion of their IgE-containing clones was found in biopsy tissue compared to the blood (Supplementary Figure E7 in the Online Repository), a consistent with the idea that SIT may increase the frequency of IgE-expressing B cells in the blood.

Allergen-specific IgE+ B cells show increased persistence in peripheral blood

We examined the time course of detection of the known allergen-specific B cell clones in the blood of each subject, compared to total IgE clones of unknown specificity, and observed that the allergen-specific clones were more frequently detected at two or more time points than the clones for which specificity was not determined. This phenomenon was most clear in the case of clones detected at 2 months and 1 year of SIT, but there were also clones detected prior to SIT as well as the 2 months and 1 year time points. In contrast, IgE sequences of unknown specificity were almost always seen at only a single time point (Figure 3C and 4D). These findings suggest that the B cells expressing IgE specific for allergens, to which the patients have been sensitized, and which would be expected to be stimulated by SIT treatment, are members of larger clones than those of the IgE-expressing B cells with unknown specificity.

Figure 4.

Antibody heavy chain somatic mutation trees describing putative clonal evolution relationships between members of allergen-specific B cell clones. Clone sets related to scFv IT5-rD13 (A), IT6-nD16 (B), IT6-rD128 (C), IT6-nD17 (D), IT6-rD111 (E) are shown. IgE sequences identified as allergen-specific in phage display experiments are represented by diamonds, while clone members identified by deep sequencing of nasal biopsies are shown as rectangles, and clone members identified by deep sequencing of peripheral blood mononuclear cell specimens are displayed as circles. The colors indicate the time point at which sequences were identified (red nodes = 0 months, blue = 2 months, green = 1 year of SIT). The symbols indicate the isotype expressed by the clone member. Nodes with no internal symbol represent IgE clone members, nodes with triangles represent IgG1 and nodes with a circled dot represent IgA members. Additional examples of trees are shown in Supplementary Figure E8 in the Online Repository.

Longitudinal tracking of allergen-specific antibody clone sets over time

To further assess if there were differences among the members of a given clone set related to allergen-specific IgE repertoires between different time points or sample types, we evaluated trees of somatic mutation relationships between IGHV domains of clone members (Figure 4). For instance, substantial, detectable mutational diversity existed at the 2 months time point within the clone set related to scFv IT5-rD13 (Figure 4A). After 1 year of SIT, clone members that were more mutated than those seen at the 2 month time point were present in circulation. Substantial mutational diversity was found within the clone set related to scFv IT6-nD16 (Figure 4B). In contrast to the clone IT5-rD13 findings, the clone members with the lowest mutational load were those seen at the 1 year SIT time point, and were expressed as IgG1, while clone members seen prior to SIT or at the 2 month time point were more mutated. These observations suggest that the observed clone members prior to SIT and at 2 months are only sparse samplings of the total allergen-specific clone at those time points, and that other unobserved clone members existed in the patient’s body that were less mutated, and had already switched their isotype to IgG1, or were in B cells that could later isotype switch to IgG1 (Figure 4C). Other clones (exemplified in Figure 4D–E) were also found in peripheral blood or in local tissue as differently mutated and/or class-switched variants. Clone members expressing IgA and IgE that were isolated from nasal biopsy tissue showed similar somatic mutation patterns, suggesting that isotype switching may have occurred locally in the nasal biopsy tissue. Overall, the data indicate that a number of mutational variants of IgE-expressing B cell clones persist in the patient’s body over time, and that some of these may undergo additional mutation during the first year of SIT. Of note, the clinical blood samples studied here represent only a sparse sampling of the patient’s entire peripheral blood B cell repertoire, and it is possible that variations in mutation levels of clones seen during SIT may in part reflect mobilization or preferential expansion of particular members of a pre-existing clonal lineage in response to the SIT antigen exposure.

Discussion

Attempts to study and understand the nature of human IgE repertoires have long been hampered by the limited numbers of IgE-expressing B cells present in clinical blood samples. Important features of allergen-specific IgE repertoires, such as their dynamics over time, and how repertoires in different tissue sites are related to each other in allergic disease or during SIT have therefore been difficult to characterize. Very recent reports indicate that IgE-expressing human B cells can be specifically isolated (29), but in contrast to studies of B cells expressing IgG, it has not yet been possible to isolate allergen-specific IgE-producing B cells using cell sorting or B cell immortalization technologies, as reviewed in Gadermaier et al. (15). Also very recently, an initial comparison of B cell repertoires in the blood and in nasal biopsies of pollen-allergic individuals has been reported, finding oligoclonality and clonal overlap between IgE repertoires in blood and tissue, and increased diversity and mutation of IgE lineages in pollen season, albeit without determining allergen specificity of the clonal lineages studied (22). By combining the strengths of HTS, proven in several studies of human antibody repertoires (30–35) with isolation of allergen-specific human IgE-derived antibody fragments from combinatorial antibody fragment libraries using phage display technology (36), and paired analysis of blood and nasal tissue samples from allergic subjects, we here provide an in-depth analysis of IgE repertoires during the first year of SIT. A caveat of our approach is that phage display libraries do not ensure native pairing of heavy and light immunoglobulin chains from individual B cells, and may favor selection of antibodies whose binding activity is primarily dependent on the heavy chain, and are of high affinity.

IgE responses in allergic individuals due to allergen exposure have in previous studies been shown to be rapid (37) and oligoclonal in nature (16, 19, 22, 38). Our results, derived from one of the most extensively sequenced and analyzed IgE transcriptomes to date, confirm the oligoclonal nature of IgE repertoires. We also observe substantial degree of somatic mutation in IgE transcripts sequenced at different time points (Supplementary Figure E5 in the Online Repository) and in selected antibody fragments (Table II). These findings do not definitively establish that IgE-producing B cells undergo somatic hypermutation, a controversial aspect of the biology of IgE-producing lymphocytes (17, 19, 39–42), although that would be one interpretation of the findings. Our data could also be consistent with a model in which different cells derived from a single B cell progenitor may undergo independent rounds of somatic hypermutation and subsequently and independently switch to IgE production. Other studies must be carried out to address at which stage mutations are introduced into the IgE-producing B cell population.

IgE-producing B cells are rare, in particular in peripheral blood, as indicated by our detection of only a single allergen-specific IgE-expressing B cell clone member in random sequencing of the 457,185 reads amplified from genomic DNA template. The low frequency of IgE+ B cells complicates any analysis of these repertoires, including those based on HTS. For instance, analysis of small blood samples will not always reproducibly detect rare clones. Indeed, the clones represented by the allergen-specific, selected scFv that were only detected at the 2 month time point (Figure 2) mostly contained very limited diversity, suggesting that the sequences originated from a single cell in the sample. HTS analysis of the 2 month time point samples did not always detect the allergen-specific IgE, expressing clones, underscoring their low frequency, although for some clones there could be false negative results in the HTS, as the primer sets used differ in the phage and HTS library protocols. Similarly, if the members of allergen-specific IgE-expressing clones remain at low frequencies at other time points, there will be a low probability of consistently detecting them in longitudinal sample sets. We did identify some allergen-specific clones that showed multiple mutational variants but were seen only at the 2 month time point, suggesting that the SIT allergen stimulus prior to this sample collection caused clonal expansion and possibly greater mobilization of such clone members into the blood. For example, sequences related to scFv IT2-B12, IT2-B23 and IT2-P614 (Supplementary Figure E8) were detected in both biological replicates obtained at the 2 months time point, and showed substantially different somatic mutation patterns. B cells from these clones were much rarer prior to SIT, and at 1 year of SIT, suggesting that some IgE-producing clone sets are recruited with different efficiencies into the circulating repertoire over time. Alternatively, there may be clone-specific differences in isotype-switching rates, long-term survival, or other factors affecting their frequency in the blood.

Interestingly, some allergen-specific B cell clones that were rare in blood could nevertheless be identified in nasal mucosa (Figure 4B–E), indicting that the IgE+ B cells in peripheral blood, although rare, show clonal overlap with the B cells implicated in a local immune response in tissue affected by allergic inflammation. In future studies, potentially using larger blood samples from allergic subjects, it will be of interest to further study the degree to which the circulating lymphocyte population accurately reflects the B cell clonotypes and phenotypes found in more difficult-to-access repertoires found at sites of local inflammation, like nasal mucosa.

Evolution of allergen-specific antibodies may occur at different sites. Several studies have demonstrated the presence of class switch recombination (CSR) and somatic hypermutation of Ig-encoding transcripts in local tissue, such as nasal or bronchial mucosa, contributing to the shaping of IgE repertoires found in such tissues (43–47). Whether the sequences we identified in nasal biopsies had been further evolved there or elsewhere is not known. The clone members identified in the tissue are clustered in the mutational tree diagrams of Figure 4, even when the tissue clone members express different isotypes, suggesting that they could have been developed by local cell division and isotype switching. These data would be strengthened if additional clones representing intermediate stages of an evolutionary path could be detected in tissue, to support a model of local hypermutation. However, all investigations of repertoires in biopsy samples of local tissue are affected by a range of procedural problems, including the difficulty of obtaining sufficient amounts of clinical material to study. Indeed, in prior reports, IgE-encoding transcripts found in different biopsies, or even in different parts of biopsies differ (39, 44). Nevertheless, the few allergen-specific clone members that we have identified in nasal biopsies firmly illustrates that members of the same allergen-specific antibody-producing clones can be found both in the circulation and in local tissue.

In the present investigation we have only assessed allergen-specific binders that are included in the IgE repertoire after initiation of SIT. We were able to identify examples of how clone members observed in blood and tissue change over time, and also detect members of these clones expressing IgG and IgA. It is of course also highly relevant to address allergen-specific repertoires of other isotypes that do not have an IgE component. Indeed, one of the proposed mechanisms of action of SIT is the induction of allergen specific antibodies, in particular those belonging to the IgG4 subclass. These may have entirely different clonal origins than those of the IgE repertoire, differences that may also translate into differences in allergen binding properties (summarized by Pomés (48)). Future studies should thus also focus on development of specific binders for instance from the IgG4 transcriptomes of samples like those collected in this study. As B cells producing antibodies of IgG class are more common than those producing IgE it may even be possible to use cell-sorting technologies (49–51) to isolate them. Such technology has proven effective in studies of B cells producing allergen-specific IgG, but not IgE, and holds promise to identify clones identical to those found in vivo. Studies of allergen-specific antibodies, in particular those of IgG4 subclass and their relation to IgE may substantially contribute to our understanding of the importance of their different properties (such as specificity and affinity) for protection against allergy. Eventually these approaches may even be implemented to clinically assess and benchmark vaccines used for SIT in efforts to better understand their mode of action.

Supplementary Material

1

Supplementary Figure E1

Changes in allergen-specific IgE (A–G) and IgG (H–N) between the 0 month and 1 year samples, as determined using the ISAC microarray system (Phadia). For donor 1–8 undergoing immunotherapy (filled lines) changes in allergen-specific antibody levels against allergens (Bet v 1 (A & H), Der p 1 (B & G), Der p 2 (C & J), Fel d 1 (D & K), Phl p 1 (E & L), Phl p 5 (F & M) and Phl p 6 (F & N)) included in the allergen extracts given to that donor are depicted by red lines. The control group (donor 9–16) is depicted by dashed lines.

Supplementary Figure E2

Specificity analysis of scFv IT6-nD17 (A), IT6-rD111 (B), IT2-B239 (C), and IT2-P536 (D) displayed on phage to diverse allergens and binding analysis of all isolated scFv (E, displayed on phage) with origin in the IgE repertoire of donors undergoing SIT (IT1-8), as determined by ELISA. All clones show minimal or no cross-reactivity to BSA. Mean values from duplicate runs are displayed.

Supplementary Figure E3

Allergic status and SIT are not associated with gross changes in total B cell antibody heavy chain repertoires. Analysis of immunoglobulin heavy chain gene rearrangements amplified from genomic DNA template of circulating B cells of the allergic subjects in this study, compared to healthy control subjects. The usage frequency of IGHV (a) and IGHJ (b) gene subgroups is shown. For each subgroup, the frequency is shown in the following order (from left to right): normal repertoires of other studies, samples obtained from non-vaccinated donors at time 0 (S1) and 1 year (S3), and samples obtained from vaccinated donors at time 0 (S1), 2 months (S2) and 1 year (S3). Frequency of mutation in the IGHV gene (c), the mean CDRH3 length (d) and the mean calculated hydrophobicity (e) is illustrated. The top and bottom panels in each figure section report unmutated sequences (top) and mutated sequences (bottom), respectively.

Supplementary Figure E4

Analysis of the CDR3 lengths of the immunoglobulin heavy chain gene rearrangements amplified from cDNA template. Each panel represents sequences derived from a different isotype, in the blood or nasal biopsy of allergic patients receiving or not receiving immunotherapy at different time points. Sequences are collapsed by unique sequences, defined as those sequences having the same V identity without allele, same J without allele and the same CDR3 sequence.

Supplementary Figure E5

Analysis of the V mutation levels of the immunoglobulin heavy chain gene rearrangements amplified from cDNA template. Each panel represents sequences derived from a different isotype in the blood or nasal biopsy of allergic patients receiving or not receiving immunotherapy at different time points. Sequences are collapsed by unique sequences, defined as those sequences having the same V identity without allele, same J without allele and the same CDR3 sequence. Significance of p<0.001 (*) was determined by pairwise T-test using Bonferroni correction

Supplementary Figure E6

Rearrangement of IT2-P11 as proposed by IMGT V-QUEST tool (10). The analysis suggests that the gene may represent a public rearrangement as it appears to have been established largely from individual IGHV and IGHJ genes without involvement of an IGHD gene and with the addition of only six N nucleotides (indicated by a horizontal line above the sequences). The part of the sequence encoding CDRH3 is indicated by a horizontal line below the gene sequences. Bases of the IGHV and IGHJ genes likely to have been trimmed of during the rearrangement process are not shown. Only those bases of IT2-P11 that were not encoded by PCR primers are shown. Bases showing identity between the IT2-P11 gene and its closest germline gene counterparts are highlighted with a grey background.

Supplementary Figure E7

Analysis of isotype expression, tissue localization, and clonal persistence of B cells belonging to clonal lineages containing IgE-expressing members in SIT vs Non-SIT patients. (A) Isotype expression by allergen-specific B cell clones containing IgE members. Clones containing only IgE members are not shown on graph, but the difference between groups was not significant. (B) Tissue and blood distribution of allergen-specific B cell clones containing IgE-expressing members. Non-SIT patients had a higher proportion of IgE clones detected in their biopsy samples. Significance was determined by Fisher’s exact test (two-tailed p-values * = 0.01 to 0.05, ** = 0.001 to 0.01, *** = 0.0001 to 0.001, **** = < 0.0001). (C) Shows samples in more than one time point during the SIT course. Numbers above bars indicate the total number of clusters representing the given category. Sequences from the 2-month time point were excluded since this sample was only collected for SIT patients.

Supplementary Figure E8

Trees describing putative evolutionary relationships between unique sequences belonging to different clone sets. Clone sets related to scFv IT2-B12 (A), IT2-B23 (B), IT2-P614 (C), IT4-B119 (D), IT4-B148 (E), IT2-B227 (F), IT4-P62 (G), IT6-B22 (H), IT2-P536 (I), IT2-B239 (J), IT6-nD12 (K) are shown. Phage sequences are represented by diamond nodes, while clones identified in the nasal biopsies are shown as rectangles, and clones derived from PBMC sequences are displayed as circles. The colors indicate the time points where clones were identified (red = time 0, pink = time 0 and 2 months, green = 1 year, purple = 2 months and 1 year, blue = 2 months). The panel on the right in each pair indicates what isotypes were represented in each lineage. Light blue and light purple nodes represent IgE and IgG2, respectively. The internal symbols indicate what isotypes were represented in each lineage. No internal symbol represents IgE clone members and while the IgG2 node seen in panel H is represented with an internal yellow x.

Key Findings.

IgE repertoire persistence and evolution hold promise as markers for monitoring specific immunotherapy of allergic disease

Abbreviations

IGHV

immunoglobulin heavy chain variable

HTS

high throughput sequencing

scFv

single chain fragment variable

SIT

specific immunotherapy