T Cell Epitope-Containing Domains of Ragweed Amb a 1 and Mugwort Art v 6 Modulate Immunologic Responses in Humans and Mice

Background Ragweed (Ambrosia artemisiifolia) and mugwort (Artemisia vulgaris) are the major cause of pollen allergy in late summer. Allergen-specific lymphocytes are crucial for immune modulation during immunotherapy. We sought to generate and pre-clinically characterise highly immunogenic domains of the homologous pectate lyases in ragweed (Amb a 1) and mugwort pollen (Art v 6) for immunotherapy. Methods Domains of Amb a 1 (Amb a 1α) and Art v 6 (Art v 6α) and a hybrid molecule, consisting of both domains, were designed, expressed in E. coli and purified. Human IgE reactivity and allergenicity were assessed by ELISA and mediator release experiments using ragweed and mugwort allergic patients. Moreover, T cell proliferation was determined. Blocking IgG antibodies and cytokine production in BALB/c mice were studied by ELISA and ELISPOT. Results The IgE binding capacity and in vitro allergenic activity of the Amb a 1 and Art v 6 domains and the hybrid were either greatly reduced or abolished. The recombinant proteins induced T cell proliferative responses comparable to those of the natural allergens, indicative of retained allergen-specific T cell response. Mice immunisation with the hypoallergens induced IL-4, IL-5, IL-13 and IFN-γ production after antigen-specific in vitro re-stimulation of splenocytes. Moreover, murine IgG antibodies that inhibited specific IgE binding of ragweed and mugwort pollen allergic patients were detected. Conclusion Accumulation of T cell epitopes and deletion of IgE reactive areas of Amb a 1 and Art v 6, modulated the immunologic properties of the allergen immuno-domains, leading to promising novel candidates for therapeutic approach.

Introduction mugwort pollen, or to both, and in vitro IgE to ragweed and/or mugwort pollen (ImmunoCAP; Phadia, Uppsala, Sweden) ( Table 1). Experiments using patients´blood samples were approved by the local ethics committee of the Medical University of Vienna, Austria (EK 712/2010) and all patients gave their written informed consent, including the guardians on behalf of the children enrolled in this study.

Purification of nAmb a 1 and nArt v 6
Natural Amb a 1 (nAmb a 1) and natural Art v 6 (nArt v 6) were purified from ragweed pollen and mugwort pollen extracts, respectively, as described elsewhere [11,14].
Cloning, expression, and purification of recombinant Amb a 1α, Art v 6α, as well as an Amb a 1α-Art v 6α hybrid molecule Recombinant Amb a 1.3 (GenBank Acession C53240) [11] clone R2 with 3 amino acid substitutions L23Y, F364L, H367R) and Art v 6.0101 (GenBank Acession AY904433) were obtained from Biomay (Vienna, Austria). These clones were used as template for the production of Amb a 1α and Art v 6α as C-terminal 6x His-tagged proteins by PCR. An E. coli codon-optimised synthetic gene coding for the hybrid molecule was synthesised by ATG Biosynthetics (Merzhausen, Germany). The recombinant hybrid protein consisted of head to tail fusion of rAmb a 1α and rArt v 6α, respectively, linked to a C-terminal 6-hexahistidine tag. Genes were cloned into the vector pET-28b (Novagen Inc, Madison, Wis) and constructs were transformed into E. coli BL21Star (DE3) cells (Invitrogen, Groningen, Netherlands). The correct sequence of the DNA inserts was confirmed by double-stranded DNA sequencing (ATG Biosynthetics GmbH, Merzhausen, Germany). Protein synthesis was induced with 0.5 mM IPTG (isopropyl b-D-thiogalactoside) at 16˚C overnight and cells were harvested by centrifugation (5000g, 20 min, 4˚C). The pellet was resuspended in lysis buffer (1:16) (20 mM Tris/HCl buffer, 0.5 M urea, 1 mM DTT, pH 9.5) containing protease inhibitor cocktail as per manufacturer's instructions (Sigma-Aldrich, St. Louis, MO) and subjected to three freeze-thaw cycles using liquid nitrogen. Cell debris was removed by means of centrifugation (16,000g for 30 minutes at 4˚C) and filtration.
Filtered lysate was loaded onto an anion-exchange column; DEAE (GE Healthcare, Munich Germany) for Amb a 1α purification and Q-Sepharose for Art v 6α purification, equilibrated with 20mM Tris pH 8.0, 1mM DTT, 0.5M urea; and Q-Sepharose FF for the hybrid molecule purification equilibrated with 20mM Tris/HCl pH 9.5, 1mM DTT, 6M urea. The proteins were eluted with a linear gradient of 0-1 M NaCl in equilibration buffer. The target proteins were subsequently purified by immobilized Ni-affinity chromatography (GE Healthcare) equilibrated in 20 mM Tris buffer pH 8.0, 0.5M NaCl, 20 mM imidazole, 0.5 M urea, 1 mM DTT for Amb a 1α and Art v 6α purification or 20 mM Tris buffer pH 9.5, 0.5M NaCl, 20 mM imidazole, 6 M urea, 1 mM DTT for the hybrid molecule purification, respectively. The proteins were eluted with decreasing imidazole concentrations (20 mM-0.5 M). Fractions containing the protein were pooled and dialysed against 20 mM sodium phosphate buffer (pH 8.0) containing 1 mM DTT. The hybrid molecule refolding was performed by stepwise decreasing the urea concentration during dialysis and the presence of 0.5M L-arginine and finally against 20 mM sodium phosphate buffer (pH 8.0) containing 1 mM DTT.
Endotoxin content of all recombinant proteins was < 1 ng/mg protein as determined by limulus amebocyte lysate (LAL) assay (Associates of Cape Cod. Inc, East Falmouth, Mass).

Circular dichroism (CD)
Far UV CD spectra were recorded at 0.1 mg/ml in 10 mM sodium phosphate pH 7.4 as described elsewhere [15].

Enzyme-linked immunosorbent assay (ELISA)
ELISA experiments were performed as previously described [16] with a few modifications. Maxisorp plates were coated with 50 μl of allergen at 2 μg/mL in PBS overnight at 4˚C. Bound IgE was detected with alkaline phosphatase-conjugated monoclonal antihuman IgE antibodies (BD Biosciences, Franklin Lakes, NJ) (1:3000) after incubation for 1.5 h at 37˚C and 1.5 h at RT. Determinations were performed in triplicates. For cross-inhibition assays 5 individual patients' sera were pre-incubated at 4˚C overnight with increasing concentrations of allergen.

Rat basophil leukaemia (RBL) cell mediator release assay
The allergenic potential of the allergens was measured by RBL assay as previously described [17]. Briefly, RBL-H2H3 cells were sensitised with 5 ragweed/mugwort pollen allergic sera (patients´number 24, 27, 31, 35, 38 Table 1) overnight. Degranulation was triggered by the addition of various concentrations of nAmb a 1, nArt v 6, Amb a 1α, Art v 6α and hybrid (10-0.0001 μg/ml). Release of β-hexosaminidase into the supernatant was measured by means of enzymatic cleavage of the fluorogenic substrate 4-methylumbelliferyl-N-acetyl-β-glucosaminide and expressed as the percentage of total enzyme content obtained by lysing the cells with 1% Triton X-100.

Amb a 1 and Art v 6-specific T cell reactivity
Peripheral blood mononuclear cells (PBMC) from ragweed/mugwort pollen allergic patients were isolated from heparinized blood by Ficoll-Hypaque (GE Healthcare Life Sciences) density gradient centrifugation. PBMCs were stimulated in triplicates with equimolar concentration of nArt v 6, nAmb a 1, Amb a 1α, Art v 6α or hybrid (62 pmol-7.5 pmol) for 6 days. T cell proliferation was measured by the incorporation of [ 3 H]thymidine (0.5 μ Ci/well) after 16 h by scintillation counting as described [11]. Stimulation indices (SI; ratio between cpm obtained in cultures stimulated with antigen and cpm obtained in control cultures without antigen) were determined as described [11].
Art v 6-and Amb a 1-specific T-cell lines (TCLs) were established from PBMCs in triplicates by using nArt v 6 or nAmb a 1 (2 μg/well), as the initial stimulus. The proliferative responses of TCL to nAmb a 1, nArt v 6, Amb a 1F, Art v 6α or hybrid (25 pmol) were assessed after 48 h by [ 3 H]thymidine uptake as described [11].

Animal experiments
Female BALB/c mice (Charles River Laboratories, Wilmington, Mass) 8-10 weeks old were immunised subcutaneously with 5 μg antigen adsorbed to Alugel-S (Serva, Heidelberg, Germany) or PBS given as two 50-μL subcutaneous injections administered bilaterally in the lumbar region and boosted on days 23, 38 and 52. Sera were collected on days 0, 31 and 64. Six animals per group were tested.
Serum IgG 1 and IgG 2a was analysed by ELISA as described [16]. In brief, plates were coated with 2 μg/ml antigen overnight. After washing and blocking, plates were incubated with individual mouse serum serial 5-fold dilutions, starting at 1:100, overnight at 4˚C. Rat anti-mouse IgG 1 and IgG 2a alkaline phosphatase (BD Biosciences, San Jose, CA) were added to detect bound IgG and the reaction developed with p-nitrophenyl phosphate (PNPP) (Sigma-Aldrich). Absorption was measured at 405 nm. For inhibition assay coated antigen was incubated with mouse serum (dilution 1:20) for 16 h at 4˚C. Individual human sera (n = 5) were added in a 1:10 dilution, and bound IgE was detected as described for direct ELISA.
RBL-H2H3 cells were incubated with a pool of 6 mice sera (1/60) for 2 h at 37˚C. After adding 0.3 μg/ml of each allergen, degranulation of RBL cells was induced and β-hexosaminidase activity in the supernatant was analysed as previously described [17].
Antigen-specific IL-4, IL-5, IL-10, IL-13 or IFN-γ-secreting splenic lymphocytes were detected by ELISPOT (eBioscience, San Diego, CA) according to the manufacturers' instructions. Animal experiments were conducted in accordance with EU guidelines 86/609/EWG and national legal regulations (TVG 2012). This study was approved by the Austrian Ministry of Science, Ref. II/3b (Gentechnik und Tierversuche), permission number BMWF-66-012/ 0010-II/3b/2013. Mice were kept in individually vented IVC modules in the SPF (special pathogen free) animal facilities at the University of Salzburg. Food and water was given ad libitum. Animals were kept in groups of five per IVC module at 20-24˚C and a relative humidity of 55%. Bedding and water was changed weekly. The light/dark cycles were set in 12 h intervals with a 30 min dim period from 0-200 lux and vice versa. Blood samples were taken by punctuation of the vena facialis. For analyses of splenocytes animals were sacrificed via cervical dislocation.

Statistical analysis
Data were analyzed by GraphPad Prism version 5 (Graph-Pad Software Inc., San Diego, CA, USA). Statistical analysis was performed using ANOVA and Bonferroni post test. P values < 0.05 were considered significant.
The purified proteins showed a mixture of α-helical and β-sheet secondary structure elements, and a folded state as determined by far-UV circular dichroism (Fig 2A).
Reduced human IgE reactivity and in vivo allergenic activity to Amb a 1α, Art v 6α and hybrid Compared to purified natural allergens, each of the Amb a 1α and Art v 6 α and hybrid molecules were virtually not recognized by specific IgE from 30 ragweed/mugwort-allergic patients (P< 0.001) (Fig 2B).
Furthermore, cross-inhibitory experiments with coated natural and recombinant proteins as inhibitors were performed. Self-inhibition of IgE binding by nAmb a 1 and nArt v 6 reached 90.5% and 82%, respectively (mean values) (Fig 2C). Inhibition of IgE binding to nAmb a 1 by preincubation with nArt v 6 ranged from 45 to 65%, whereas inhibition of IgE binding to Art v 6 by preincubation with Amb a 1 ranged from 60 to 80%. Compared to purified natural allergens, the domains' inhibitory capacity was significantly (P<0.001) reduced at 50 pmol (Fig  2C), 0.5 pmol and 0.005 pmol (results not shown).
The low IgE-binding capacity of the recombinant domains was confirmed by the lack of induction of mediator release even at the highest concentration (10 μg/ml). The natural proteins showed biological activity between 0.1 μg/ml and 10 μg/ml with a maximum mediator release of 60-70% (Fig 3).
Activation of human Amb a 1-or Art v 6-specific T lymphocytes by Amb a 1α, Art v 6α and hybrid Comparable proliferative responses in human PBMCs isolated from 10 ragweed/mugwort pollen-allergic donors were induced by equimolar amounts of the domains (62 pmol, Fig 4A,  Similarly, nAmb a 1-and nArt v 6-induced T cell lines (TCL) established from 14 ragweed/mugwort pollen-allergic patients and re-stimulated with natural or recombinant protein (25 pmol), displayed proliferation responses that were not significantly different between the natural and the constructs (Fig 4B and 4C). Amb a 1α, Art v 6α and the hybrid molecule induced responses in 7, 6 and 8 of 8 nAmb a 1-specific TCL, respectively (Fig 4B), and 5, 3, and 5 of 9 nArt v 6-specific TCL, respectively (Fig 4C), confirming that the relevant Amb a 1/Art v 6 T cell epitopes are located in the alpha chain domains.
Induction of murine Amb a 1-or Art v 6-specific blocking IgG antibodies by Amb a 1α, Art v 6α and hybrid To investigate the immunogenic properties of the recombinant proteins, mice were immunised with the natural or recombinant proteins, and the levels of serum IgG 1 and IgG 2a to Amb a 1, Art v 6 or Bet v 1 (as unrelated protein) were measured by ELISA. The domains induced IgG 1 and IgG 2a antibodies against nAmb a 1 and nArt v 6 (Fig 5A and 5B).
Cross-reactive IgG response was similar between the domains and nAmb a 1 but significantly stronger when compared to nArt v 6. Interestingly, the hybrid induced cross-reactive nArt v 6 IgG antibodies that were significantly higher than those induced by nAmb 1-immunised mice (Fig 5A). IgG 2a levels were comparable between proteins (Fig 5B). Lack of Bet v 1-specific IgG 1 and IgG 2a was observed (results no shown).
Degranulation by biologically active anti-Amb a 1 or Art v 6-specific IgE antibodies was confirmed by means of RBL-2H3 assay in sera of 6 mice immunized with nAmb a1, nArt v 6, Amb a 1α, Art v 6α or hybrid. The natural proteins showed biological activity with a maximum mediator release of 90% by nAmb a 1 and 82% by nArt v 6. Moreover, a significantly reduced capacity to induce mediator release by the variants compared to the natural proteins was also confirmed in our mouse model (P<0.001) (Fig 5C).
Mouse IgG antibodies induced by the recombinant domains partly inhibited the IgE binding of 5 ragweed/mugwort pollen-allergic patients' sera IgE to nAmb a 1 (30-35% mean inhibition) and nArt v 6 (17-21%), indicative of the presence of conserved antibody epitopes on the alpha domains (Fig 5D).

Retained T cell reactivity in mice
Splenocytes from immunised mice re-stimulated with each of the 5 proteins were evaluated for cytokine production by ELISPOT (Fig 6A and 6B). Recombinant proteins induced similar IL-5 production compared to the natural allergens (Fig 6A). Only Amb a 1α immunisation induced significantly higher IFN-γ levels than the natural allergens. Whereas the hybrid molecule induced similar IL-4 levels compared to the natural molecules, both Amb a 1α and Art v 6α immunisation induced significantly higher levels. Low and comparable levels of IL-13-or IL-10-secreting T cells were detected, whereas immunization with nAmb a 1 failed to induce IL-10-secreting cells (P<0.05) (Fig 6B). These results demonstrate that the alpha domains are able to evoke a TH1/TH2 type immune response.

Discussion
We hypothesised that the accumulation of T cell epitopes and the elimination of IgE domains can regulate the allergen specific T cell response to Amb a 1 and Art v 6 leading to a promising vaccine candidate. Our strategy focused on generating 3 molecules consisting of T cell epitopes containing immunodomains of the homologous PL Amb a 1, Art v 6 and an Amb a 1-Art v 6 hybrid. The hybrid strategy not only simplifies protein production, but has the advantage that a single molecule contains the immunologic information of relevant pollen allergens from different sources as demonstrate with hybrids from the Fagales family [18]. Since specific IgE levels are only partially predictive for clinical allergy [19], we determined the allergenic potency of the domains by mediator release assays using RBL cells passively sensitized with patients' serum IgE. We detected a reduction in the in vivo allergenicity for all domains, which correlated well with the low in vitro IgE-binding activity. This reduction in the allergenic potency and IgE reactivity of the constructs indicated that the deletion of the N-terminal part of either PL allergen could result in the disruption of IgE-binding epitopes [12].
The efficacy of AIT has been associated with induction of a regulatory T cell response, immune deviation to a balanced TH2/TH1 type, or both [20][21][22]. Hence, the preserved T cell immunogenicity is thought to be a requirement for successful AIT. We compared the T cell reactivity of equimolar concentrations of the natural and the recombinant proteins by stimulation of PBMC or TCLs from allergic individuals. From the experiments, it became evident that the recombinant molecules were able to induce an allergen-specific T lymphocyte activation. These results demonstrate a retained T cell reactivity, which led to the conclusion that Amb a 1α, Art v 6α, and the hybrid harbour the majority of the T cell epitope repertoire of the natural PL allergens.
The induction of allergen-specific IgG antibodies, able to block allergic patients' IgE binding and hence inhibit IgE-mediated effector cell degranulation, T-cell activation, and boosting of IgE production, is considered a hallmark for the clinical success of AIT [1][2][3]. Noteworthy, Amb a 1α, Art v 6α, and the hybrid contained antibody epitopes involved in the induction of IgG antibodies as demonstrated in our mouse model.
It is important to notice that in our animal model the strong reduction of allergenicity and the preservation of immunogenicity of the domains are in consonance with the reduced allergenic potency in humans that we observed. Further, the capacity of the murine Amb a 1α, Art v 6α, and hybrid antisera to significantly inhibit binding of human IgE to nAmb a 1 from 5 allergic patients has been demonstrated. However, whereas the blocking of allergic patients' IgE binding to nArt v 6 by antibodies induced with the domains was comparable to that of the natural allergens, antiserum from mice immunised with the domains had a significantly lower blocking capacity for Amb a 1-specific IgE. The fact that the domains represent only 2/3 of the sequence of Amb a 1 might be one explanation for this finding. Of note, a generally low inhibitory capacity of murine anti-nAmb a 1 or anti-nArt v 6 IgG (56% and 20%, respectively) was observed, which could be an indication that the murine immune response towards PL allergens is quite distinct from the human allergic response. These results could however be biased by the fact that an adjuvant was used in the murine immunization schedule resulting in differences in allergen presentation or processing by mice and humans, which are exposed in a completely different manner [23,24]. Consequently, future studies addressing allergen uptake and processing by dendritic cells in the presence of adjuvants would be beneficial for the design of effective allergy vaccines.
Besides epitope modifying effects the adjuvant ALUM may also bias the subsequent immune response on the T cell level. Here were measured a TH2-polarizing effect of ALUM, indicated by the production of IL-4, IL-5 and IL-13. Analyses of murine splenocytes revealed the induction of TH1 cells producing IFNγ by the recombinant proteins. In other words, a mixed TH1/TH2 immune response was elicited against Amb a 1α, Art v 6α, and hybrid compared to a clear TH2-driven response induced by the natural proteins.
Concomitant allergies to ragweed and mugwort pollen are a well-recognized problem in areas where both plants are endemic, where it is difficult to distinguish between co-sensitization and cross-reactivity [25]. Amb a 1 as well as the major mugwort pollen allergen Art v 1, a member of the defensin-like allergen family, have been suggested as marker allergens for the diagnostic workup [8,10]. Nevertheless, cross-reactive allergens such as members of the PL family represent a problem for patients and clinicians. It is still not clear whether a therapeutic intervention against Amb a 1 would lead to a simultaneous improvement of Art v 6 driven allergic reactions and vice versa. However, with our study results we suggest that a single allergen derivative might be sufficient for a concomitant treatment strategy of PL mediated allergies form the Asteraceae family.

Conclusion
In summary, the accumulation of T cell epitopes and the deletion of IgE reactive areas of the PL allergens Amb a 1 and Art v 6 resulted in the generation of highly immunogenic domains of both proteins. This led to the development of a hybrid molecule consisting of the combination of Amb a 1 and Art v 6 C-terminal regions. The domains were able to induce lymphoproliferative responses in ragweed/mugwort pollen-allergic patients´PBMCs and displayed low IgE reactivity and in vitro allergenicity. These results in humans correlated with results gained from a murine immunization model where Amb a 1α, Art v 6α, and hybrid were compared to wildtype allergens. Moreover, the single domains seemed sufficient to induce robust levels of IgG antibodies cross-reactive with the full-length WT allergens. Thus, based on our results we successfully generated candidates for AIT to redirect the allergic immune response supporting the idea of using only the alpha chain domain of Asteraceae PL allergens.