Suppression of IgE B Cells and IgE Binding to FcεRI by Gene Therapy with Single-Chain Anti-IgE

Mice

Eight- to 12-wk-old BALB/c mice were used in most experiments. C.129S2-Fcer1a^tm1Knt/J (FcεRI^−/−) mice were purchased from The Jackson Laboratory. Mice were bred and maintained in The Scripps Research Institute Animal Resources facility according to the Institutional Animal Care and Use guidelines.

Cells

293F cells were purchased from Invitrogen. Anti-IgE hybridoma cells R1E4 (9) and EM95 (10) were provided by Drs. D. Conrad (Virginia Commonwealth University, Richmond, VA) and F. Finkelman (University of Cincinnati, Cincinnati, OH). Anti-FcR hybridoma cells 2.4G2 (11) were purchased from American Type Culture Collection.

Development R1E4 single-chain variable fragment (scFv)

R1E4 variable region sequence was obtained using a 5′ RACE kit (Ambion) according to the manufacturer’s protocol. The following oligonucleotides were used for H chain and L chain 5′ RACE, respectively: 5′-AATAGCCCTTGACCAGGCATCC-3′ and 5′-CCAGTTGCTAAC TGTTCCGTGGA-3′.

H chain and L chain genes were PCR amplified, respectively (T34 and T31 for L chain and T35 and T33 for H chain), and R1E4 scFv (pUb-R1E4) was generated with nested PCR. To generate a membrane-tethered single chain Ab, we subcloned this fragment into SpeI- and XmaI-digested pUb-187.1 plasmid (12). This strategy fuses the single chain Fv with rat-IgG1 hinge, CH2 and CH3 domains, followed by a C-terminal MHC class I transmembrane/intracellular domain. The primers were as follows: T34, 5′-ctattaattattacaggtgcctgtgcaGACATTGTCTTGACCCAGTCT-3′; T31, 5′-ACCGCCAGAGCCACCTCCGCCTGAACCGCCTCCACCCCGTTTCAATTCCAGCTTGGTGC- 3′; T35, 5′-GCGGTTCAGGCGGAGGTGGCTCTGGCGGTGGCGGATCGCAGGTGACTCTGAAAGAGTCT-3′; and T33, 5′-cctcccgggtttct gggggctgttgtttcagcTGAGGAGACTGTGACCATGACTC-3′.

To decrease the antigenicity of R1E4, rat-IgG1 was replaced with mouse IgG1 and several amino acid reversion mutations were introduced to L chain framework regions (supplemental Fig. 1⁴). To generate a secreted form in the final construct (pR1E4), the MHC transmembrane and cytoplasmic domains were excised. The specificity of recombinant single-chain membrane R1E4 to IgE was confirmed by the transient transfection to 293F cells and flow cytometry analysis of IgE binding (Fig. 1⇓B). All of the plasmids used in in vivo injection were purified with an EndoFree Plasmid Maxi kit (Qiagen). For recombinant R1E4 (rR1E4), pR1E4 Ab-coding sequences were inserted into the pIRES-Zeocin hrGFP plasmid (pIRES-pR1E4). pIRES-pR1E4 (3 μg) was transfected into 293F cells with Lipofectamine2000 (Invitrogen) and stable cells were established. From this supernatant, rR1E4 was purified with a rProtein A column (GE Healthcare).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Generation and characterization of recombinant neutralizing anti-IgE single-chain Ab construct. A, The gene construct encoding rR1E4 (anti-IgE) showing intron/exon structure and selected features. Introns are depicted as thin lines. R1E4 scFv express under the control of human ubiquitin promoter. The R1E4 scFv was fused to hinge and Fc-encoding exons of the mouse IgG1 H chain. For membrane-bound anti-IgE expression, H-2K^b MHC class I transmembrane-coding sequences were included downstream. B, Control or recombinant plasmids expressing the membrane form of single-chain R1E4 were transiently cotransfected with enhanced GFP (EGFP) plasmid into 293 cells and the ability of the cells to bind to IgE was analyzed by flow cytometry.

Quantitative PCR (qPCR)

Total RNA was purified from 2 to 3 million spleen cells of control or pR1E4-treated mice using a RNEasy Plus kit (Qiagen). Reverse transcription was performed with a QuanteTect Reverse Transcription Kit (Qiagen) per the manufacturer’s protocol. IgHε mRNA was quantitated using SYBR GreenER qPCR Supermix (Invitrogen) with 7900HT (Applied Biosystems) and normalized with CD19 mRNA. Oligonucleotide primers used for IgE detection were 5′-acactcggagatgcccagatc-3′ and 5′-ggagcaccgttttgatacaggtc-3′ and for CD19 detection, 5′-aggtcattgcaaggtcagcagtgtg-3′ and 5′-ggcgtcactttgaagaatctcctg-3′.

Flow cytometry analysis

Erythrocyte-depleted cells were suspended in ice-cold staining buffer (HBSS buffer including 0.5 mM EDTA, 0.05 mM sodium azide, and 0.5% BSA) with appropriately titrated Abs. The following Abs were used: CD45R/B220 (RA3-6B2; BD Biosciences) (Pacific Blue), FcεRIα (MAR-1; eBioscience) (PE), IgE (23G3; eBioscience, or EM95) (FITC, PE, Alexa Fluor 647), CD49b (HMa2; BD Biosciences) (PE, allophycocyanin; biotin), c-kit (2B8; eBioscience) (allophycocyanin; biotin), CD4 (GK1.5; BD Biosciences) (PerCP-Cy5.5), CD8 (53-6.7; BD Biosciences) (PerCP-Cy5.5), and strptavidin-PE-Cy7 (eBioscience). For intracellular IgE staining, cells were incubated during surface staining with unlabeled anti-IgE (EM95); after fixation and permeabilization using a Cytofix/Cytoperm kit (BD Biosciences), cells were stained with labeled EM95 conjugate. These Abs were purchased from eBioscience or BD Biosciences as indicated. Propidium iodide (Invitrogen) was included in some experiments to exclude dead cells. To calculate the total FcεRI expression level on basophils based on IgE-binding capacity, FcγRs were preblocked with 2.4G2 for 10 min and the cells further incubated with purified IgE (IgELa) at 10 μg/ml for 30 min. Cells were then washed twice with FACS buffer and bound IgE was quantitated with anti-IgE conjugate. Data collection was done on an LSR II flow cytometer (BD Biosciences) and was analyzed using FlowJo software (Tree Star).

Hydrodynamic injection

Thirty micrograms of purified plasmid (pR1E4 or pUb control plasmid) was dissolved in 1.8 ml of TransIT-EE Delivery Solution (Mirus Bio) and injected via a tail vein. In the experiments depicted in Fig. 2⇓, 10 μg of a second plasmid driving human placental secreted alkaline phosphatase (pLIVE-SEAP; Mirus Bio) was coinjected, allowing one to monitor the efficiency of transfection by enzyme activity appearing in blood. All of the injected mice (five of five control and eight of eight pR1E4 treated) analyzed on day 13 after plasmid injection were alkaline phosphatase positive (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Effect of in vivo expression of secreted form of chimeric single-chain anti-IgE on markers of IgE expression. Two-month-old BALB/c mice were given pR1E4 or control plasmid i.v. and analyzed 13 days later. A, Free serum IgE level measured on day 13 after plasmid treatment. B, Total serum IgE concentration before treatment (□) and 13 days after treatment (▪). C, qPCR comparison of relative IgHε mRNA levels in the spleen of control or pR1E4-treated mice at day 13 of treatment. D and E, Flow cytometry analysis of IgE bound to peritoneal mast cells in control and pR1E4-treated mice. Peritoneal mast cells were defined as c-kit⁺FcεRI⁺CD4⁻CD8⁻B220⁻ cells. Mean fluorescence intensity (MFI) of binding by FcεRIα and anti-IgE Abs was monitored. F and G, Analysis of IgE bound to basophils in the spleen and bone marrow of control or treated individual mice. Basophils were identified as CD49b⁺FcεRI⁺CD4⁻CD8⁻B220⁻. H, Levels of surface IgE and FcεRI on basophils in the spleen (Sp) and bone marrow (BM) on day 13 after plasmid treatment. Results are means ± SD of eight mice receiving pR1E4 compared with five mice receiving empty vector. Statistical significance of differences between pR1E4-treated and control-treated mice was calculated using Student’s t test: *, p < 0.001 and #, p < 0.2.

IgE-eliciting immunizations

OVA (Sigma-Aldrich) was prepared with alum (Imject; Pierce) at a ratio of 10 μg of protein:100 μg of alum/mouse and was given i.p. Goat anti-mouse IgD (0.2 ml; eBioscience) was injected i.p.

IgE ELISA

An IgE ELISA quantitation kit was purchased from Bethyl and used according to the kit instructions. In experiments involving pR1E4, purified EM95 (anti-IgE) was coated on Nunc Maxisorp plates and bound serum IgE was detected with HRP- conjugated goat anti-mouse IgE (Bethyl). The presence of rR1E4 did not interfere with the total IgE concentration measurement because R1E4 and EM95 see nonoverlapping epitopes; R1E4 blocks IgE binding to FcεR1, whereas EM95 cannot. For quantitation of free IgE, rR1E4 was coated on the plate and bound IgE was detected with HRP-conjugated goat anti-mouse IgE. Color was developed with Ultra-TMB (Pierce).

Cell culture

Splenocytes (2 × 10⁵/ml) were cultured at 37°C/5% CO₂ in a final volume of 2.5 ml with 25 ng/ml IL-4, 1 μg/ml anti-CD40 (1C10; eBioscience) with or without 20 μg/ml rR1E4 in Advanced-RPMI 1640 medium supplemented with 5% FCS, 1× penicillin-streptomycin-glutamine (Life Technologies), 2 mM GlutaMAX-I (Life Technologies), and 55 μM 2-ME for 4 days. For Ca²⁺ mobilization analysis, splenocytes were cultured with 25 μg/ml IL-4 and 10 μg/ml anti-CD40. Culture supernatants were then tested for IgE concentration by ELISA; recovered cells were washed two times with FACS buffer before flow cytometry or Ca²⁺ mobilization analysis.

Calcium response

Cultured splenocytes were loaded with Fluo-4 (Invitrogen) per the manufacturer’s instructions. Calcium mobilization was induced by addition of 20 ng/ml mAbs at a cell concentration of 10⁶/ml in volume of 0.5 ml. In some experiments, cells were preincubated with 10 μg/ml anti-mouse IgG1 Ab for 30 min before challenge and fluorescence analysis. Analysis was conducted using the FL1 channel of an LSR-II flow cytometer and analyzed with FlowJo software.

Anaphylaxis

One hundred micrograms of EM95 was given per mouse retro-orbitally after anesthesia; rectal temperature was monitored with a RET-3 probe (Physitemp Instruments) for 60 min. In some experiments, 0.5 mg of 2.4G2 (anti-FcγRII/III) Ab was given to each mouse 24 h before challenge. In the case of rR1E4 injection, 10 μg of 2,4,6-trinitrophenyl (TNP)-BSA and 100 μg of IgE anti-TNP (IgELa) were given per mouse and 50 μg of rR1E4 was given retro-orbitally; rectal temperature was monitored for 60 min.

Development of recombinant single-chain anti-IgE

In contrast to passive anti-IgE administration, anti-IgE delivery via gene therapy has the potential to provide permanent therapy of IgE-mediated disease. To study the effects of anti-IgE gene therapy in the mouse (an intensively studied model for human allergy, asthma, and anaphylaxis (13, 14, 15)), we generated a chimeric single-chain anti-IgE gene based on neutralizing rat anti-mouse IgE mAb R1E4 (9). R1E4, like omalizumab (Genetech) in the human system, is unable to activate mast cells via IgE bound to FcεRI, but binds free IgE. R1E4 Ig H and L chain variable region codons were cloned, sequenced, and joined together with a short linker sequence, yielding a scFv gene (Fig. 1⇑A). The scFv gene was placed upstream of mouse IgG1 hinge and membrane proximal codons. Plasmids driving plasma membrane expression under the control of the human ubiquitin promoter (16) were prepared and validated (Fig. 1⇑B). A modified plasmid, called pR1E4, encoding a secreted version of the rR1E4 chimeric protein (rR1E4) was generated. The biological effects of in vivo expression of plasmids encoding membrane or secreted proteins was tested by the hydrodynamic (naked DNA) injection method, which leads to transient expression in the liver. Because tolerance of cognate B cells to protein Ags can be induced either by membrane expression on the liver (17) or by soluble protein (18, 19, 20), we also compared in vivo efficacy of plasmids encoding membrane and secreted forms of recombinant anti-IgE.

Neutralizing serum IgE in vivo

pR1E4 plasmid given to BALB/c mice led to single-chain anti-IgE secretion lasting at least 30 days that diminished at later times (data not shown). At day 13 after pR1E4 injection, levels of free IgE were reduced to ∼1% of those of control mice (Fig. 2⇑A). There was also a marked decline in the levels of total IgE, which includes both free IgE and IgE-rR1E4 complexes (Fig. 2⇑B). Plasmid encoding membrane-bound single-chain anti-IgE was also effective in reducing IgE levels, although the effect was less long-lasting (supplemental Fig. 2). For unknown reasons, plasmids encoding membrane-bound single-chain constructs were apparently somewhat toxic to expressing liver cells, even if they lacked specificity for IgE (data not shown). Therefore, additional experiments focused on the effects of the soluble form. Importantly, RNA analysis of the spleens of mice treated with pR1E4 revealed a >99% reduction in the levels of IgHε mRNA (Fig. 2⇑C). Consistent with the reduced IgE levels in pR1E4-treated mice, c-kit⁺ peritoneal mast cells and CD49b⁺ basophils in the bone marrow and spleens of anti-IgE-treated mice lacked detectable surface IgE, indicating that their FcεRs had lost bound IgE (Fig. 2⇑, D–H). Basophils were still present in these tissues as indicated by FcεRIα/CD49b double staining (Fig. 2⇑G). In treated mice, FcεRI expression levels in basophils appeared higher than in control plasmid-treated mice, however, total IgE-binding capacity of basophils was somewhat reduced (supplemental Fig. 3, A and B). This discrepancy was probably because anti-FcεRI mAb MAR-1 binds more tightly to FcεRI lacking bound IgE than to FcεRI carrying IgE (supplemental Fig. 3C). Reduced levels of bioactive IgE in pR1E4-treated mice did not lower the percentages or absolute numbers of mast cells and basophils (Table I⇓). Overall, these data support the conclusions that 1) gene therapy with recombinant single-chain anti-IgE is feasible and can effectively neutralize circulating IgE; 2) reduced levels of bioactive IgE have surprisingly minimal negative effects on mast cell and basophil survival, at least at the time point tested, and 3) recombinant single-chain anti-IgE treatment could actively suppress new IgE production by B cells.

Recombinant anti-IgE treatment of preimmunized mice

We next determined whether pR1E4 could be used to treat and suppress ongoing, established IgE responses. Basal serum IgE levels in mice were quite low (∼500 ng/ml in BALB/c mice and under 100 ng/ml in C57BL/6 mice). To evaluate the anti-IgE treatment, we used two different model responses: OVA immunization with alum adjuvant (OVA/alum), which is commonly used to elicit an Ag-specific IgE response, and anti-IgD treatment, a nonspecific B cell activator which elicits a stronger, but polyclonal IgE response (21, 22).

Ten days after receiving IgE-eliciting stimulus, mice were treated with pR1E4 or control plasmids and then analyzed 12 days later, i.e., 22 days after immunization (Fig. 3⇓). OVA/alum immunization increased total serum IgE levels manyfold above preimmune levels, a response that was essentially fully suppressed by treatment with pR1E4 (Fig. 3⇓A). IgHε mRNA qPCR analysis showed that pR1E4 treatment reduced IgHε mRNA levels in the spleen by 99% (Fig. 3⇓B). pR1E4 treatment also reduced the levels of the low-affinity IgE receptor CD23 (23) on follicular B cells by ∼50% (supplemental Fig. 4, A and B). We conclude that in mice with ongoing IgE responses, treatment with the soluble recombinant anti-IgE-producing plasmid pR1E4 could significantly (p = 0.0023) suppress levels of serum IgE, new IgE synthesis, and B cell-expressed CD23.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Effects of pR1E4 plasmid treatment on ongoing IgE responses induced by OVA/alum immunization or goat anti-mouse IgD treatment. Two-month-old BALB/c mice injected i.p. 10 days previously (day −10) with either 10 μg of OVA/alum or 200 μl of goat anti-mouse IgD serum were treated with control or pR1E4 plasmids i.v. on day 0 and analyzed on days 0 and 12 for suppression of the IgE response. A, Total serum IgE levels on day 0 (▪) and day 12 (□). B, Splenic IgHε mRNA levels assessed by qPCR. C–G, Flow cytometry analysis of IgE⁺ B cells in spleens of mice injected with anti-IgD and subsequently treated with control or pR1E4 plasmid. C, Analysis of the frequencies of B220^high and B220^lowIgE⁺ B cells among CD4⁻CD8⁻c-kit⁻CD49b⁻ gated viable spleen cells. D, Comparison of CD138 and IgD expression of B220^low and B220^high populations in control plasmid and pR1E4-treated mice, as gated in C. Left, B220^low cells from control plasmid-treated mouse; center, B220^high cells from control plasmid-treated mouse; and right, B220^high cells of pR1E4-treated mouse. E, Analysis of treated and control anti-IgD-injected mice for the presence of cells that expressed high levels of intracellular IgE. F, Summary of pR1E4-induced reductions in IgE⁺B220^high and IgE⁺B220^low subsets using the analysis in C. G, Quantitation of pR1E4-induced reduction in surface B220^highIgE⁺IgD⁻ B cells as in D. Significance was calculated using Student’s t test. *, p < 0.01 and #, p < 0.05. n = 4–6 mice/group.

Treatment of mice after anti-IgD injection served as a more stringent test of the ability to tolerize ongoing responses because IgE responses were >10-fold further enhanced over the OVA/alum-induced response. Anti-IgD injection induced ∼10 μg/ml IgE in sera in control mice, which could be markedly (99%) suppressed by subsequent in vivo expression of pR1E4 (Fig. 3⇑A). However, IgHε mRNA qPCR analysis revealed that mice given pR1E4 after anti-IgD treatment still had high levels 12 days later, suggesting that IgE mRNA-producing cells remained in the spleen (Fig. 3⇑B).

To identify the source of the IgE mRNA, we analyzed splenocytes from anti-IgD-treated mice for the presence of mIgE⁺ B cells (Fig. 3⇑, C–G). In control plasmid-treated mice, B220^high and B220^lowIgE⁺ B cells were detected (Fig. 3⇑C). B220^high B cells were mostly IgD⁺CD138⁻ cells and possibly represented background staining of naive B cells (Fig. 3⇑D, center). B220^low B cells were mostly IgD⁻CD138⁺ or CD138⁻ cells, indicating that they were IgE preplasma cells or B220^low memory B cells (Fig. 3⇑D, left). In pR1E4-treated mice, B220^highIgE⁺ B cells were detected, but most B220^lowIgE⁺ B cells were lost. The remaining B220^highIgE⁺ B cells in pR1E4-treated mice were IgD⁺CD138⁻ (Fig. 3⇑D), suggesting that pR1E4 blocked development of IgE memory B cells and preplasma cells. The frequency of the B220^lowIgE⁺ fraction declined in pR1E4-treated mice (p = 0.0024; Fig. 3⇑F). Intracellular staining for IgE (preblocked for surface staining) was conducted, revealing that plasma cells remained in the spleens of pR1E4-treated mice (Fig. 3⇑E). Total numbers of B220^low/intracellular IgE⁺ cells were on average lower in pR1E4-treated mice compared with control, but not to a statistically significant extent. We conclude that in anti-IgD- stimulated mice pR1E4 could neutralize circulating IgE and suppress the numbers of mIgE⁺ B cells, but could not suppress fully developed IgE plasma cells.

Anti-IgE can block IgE secretion in vitro

To determine mechanistically how soluble rR1E4 affects IgE B cells, we assessed the effects of R1E4 mAb on cultured naive B cells induced to switch to IgE by the addition of IL-4 and anti-CD40. Although the binding of R1E4 has been mapped to the ε CH3 domain (9), it was not clear whether this epitope would be accessible on mIgE⁺ B cells because on B cells mIg is oriented differently than in solution and is associated with Igα/β signal transducers. We found that mIgE⁺ B cells could be clearly identified with both R1E4 IgG and rR1E4 Ab (Fig. 4⇓A). Moreover, inclusion of rR1E4 both inhibited the secretion of IgE in the supernatant and also reduced the frequency of CD138⁺ and CD138⁻IgE⁺ B cells emerging in the cultures (Fig. 4⇓, B–D). Ca²⁺ signaling analysis indicated that rR1E4 Ab could induce Ca²⁺ mobilization in mIgE⁺ B cells upon supercross-linking by a second Ab (rat anti-mouse IgG1; Fig. 4⇓E). These data indicate that anti-IgE can block the development of IgE plasma cells in vitro and may in part explain the in vivo effects of this treatment.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

Ability of rR1E4 anti-IgE protein to affect the biological responses of IgE⁺ B cells generated in tissue culture. A, BALB/c spleen cells were induced to undergo IgE class switch by culture at 2 × 10⁵ cells/ml for 4 days with IL-4 (25 ng/ml) and anti-CD40 (10 μg/ml); binding by rR1E4 was examined. Right panel shows staining with rR1E4 (10 μg/ml) followed by FITC anti-mouse IgG1 and Alexa Fluor 647 anti-IgE (EM95). B–D, Spleen cells (2 × 10⁵ cells/ml) were stimulated for 4 days with IL-4 (25 ng/ml) and anti-CD40 (1 μg/ml) with or without 20 μg of rR1E4. B, IgE levels in culture supernatants on day 4 as measured by ELISA. C, Analysis of IgE⁺ B cell and IgE⁺ plasma cell frequencies at day 4 of culture, as determined by staining for intracellular IgE, B220, and CD138. D, Summary of rR1E4-induced changes in IgE subpopulations as shown in C. E, rR1E4-induced calcium flux in fluo-4-loaded B cells recovered at day 4 of culture. Top panel, B cells were treated with either anti-Igκ or rR1E4 alone. Lower panel, Cells were precultured with anti-IgG1 alone for 30 min (to obscure the response of IgG1 cells) and then stimulated with either anti-IgG1 or rR1E4, which is supercross-linked by free anti-IgG1. Results presented in B and D represent the means and SDs of four separate experiments. *, p < 10⁻⁴ and #, p < 0.05.

Anaphylaxis reaction

The ability of in vivo pR1E4 treatment to suppress allergic responses was tested by challenging mice on day 13 after treatment with an activating anti-IgE, EM95, which has the ability to induce mast cell degranulation and anaphylactic manifestations. Surprisingly, and notwithstanding their suppressed IgE levels, pR1E4-treated mice challenged in this way showed a systemic anaphylaxis reaction as indicated by rapidly lowered body temperature (Fig. 5⇓A). Recently, an alternative pathway for anaphylaxis has been elucidated in which IgG1-immune complexes activate FcγRIII on basophils, releasing platelet-activating factor (15, 24, 25, 26, 27). To test the possibility that this pathway was triggered by clustering of anti-IgE Abs upon challenge, pR1E4-treated or control mice were given FcγRII/III-blocking mAb 2.4G2 (28) 1 day before challenge. Indeed, anaphylaxis was blocked in pR1E4-treated but not in control mice receiving 2.4G2 (Fig. 5⇓B). To test whether mouse IgG1 R1E4 could induce anaphylaxis directly in a FcεRI-independent manner, we preinjected TNP-BSA and IgE anti-TNP into FcεRIα^−/− mice and challenged them 1 h later with rR1E4. As shown in Fig. 5⇓C, mice that received rR1E4 manifested an anaphylaxis reaction, indicating that rR1E4 itself triggered the reaction through FcγRs. Collectively, these data indicate that pR1E4 treatment suppresses FcεRI-mediated anaphylaxis, but permits and in fact promotes FcγRIII-mediated anaphylaxis.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Systemic anaphylaxis reactions in rR1E4-treated mice. A, Untreated BALB/c (wild-type (WT)) mice or pR1E4-treated (Sec) mice were challenged with 100 μg of activating anti-IgE mAb EM95 and rectal temperatures were measured over 60 min. B, Mice received 0.5 mg of Fc-blocking mAb 2.4G2 twenty-four hours before challenge with 100 μg of anti-IgE EM95. Untreated BALB/c (WT + Fc), pR1E4 treated (Sec + Fc). C, FcεRIα^−/− mice received 10 μg of TNP-BSA and 100 μg of anti-TNP IgE (IgELa) 1 h before challenge with 50 μg of rR1E4. Shown are means and SDs of the results obtained with the following numbers of mice per group. All groups included three mice, except pR1E4-treated mice in A which included four mice.