INTRODUCTION

Chimeric receptors that redirect effector cell function to tumor cells or HIV¹-infected cells have received much attention (1, 2, 3). In particular, chimeras of single chain Fv (scFv) monoclonal antibody (mAb) linked to FcRI or CD3 have been used to direct cytotoxic T lymphocytes (CTL) to tumor cells in vitro and in vivo (2, 4, 5, 6, 7). These receptors have proven utility; however, they suffer from two major drawbacks, the first being the need to generate and express individual chimeric antibody receptors for every tumor antigen targeted, and the second that a monospecific interaction between the receptor and tumor cells may be ineffective given that tumor cells have and continually develop heterogeneous antigen expression. By contrast, chimeras comprising Fc receptors (FcR) capable of binding antibodies of a variety of anti-tumor specificities may provide a more effective and universal means of redirecting cytotoxic lymphocytes to tumors. Given the high affinity of FcRI for IgE mAb and low serum levels of IgE, it is possible that redirection of effector cells using chimeric FcRI might provide a novel, effective anti-tumor strategy.

FcRI is the high affinity receptor for IgE and is found on mast cells, basophils, eosinophils, activated monocytes, and Langerhans cells (8, 9, 10, 11). It is a multimeric receptor complex consisting of a ligand binding  subunit, a  subunit, and a homodimer of two signal-transducing  subunits (8, 9, 12). The  and  subunits also play an important role in transport of the  subunit to the cell membrane (13, 14, 15). The  subunit of FcRI was found to be sufficient for IgE binding using a chimera of extracellular -chain linked to transmembrane and cytoplasmic tail of p55 interleukin-2 (IL-2) receptor (16). The IgE binding region was more closely mapped to the second extracellular domain of FcRI by using a series of chimeras that interchanged portions of FcRI and FcRIIa or FcRIII (17, 18).

The contribution of individual components of receptor complexes to functional receptor expression have also been determined for subunits of the T cell receptor (TCR). The TCR is a multimeric complex composed of an antigen-binding heterodimer (/- or /-chains) together with a set of invariant CD3 chains, , , , and  (19). The  chain is a 16-kDa protein with a 9-amino acid extracellular portion, a 21-amino acid transmembrane region, and a 113-amino acid cytoplasmic tail (20, 21).  usually exists as a 32-kDa homodimer but can also form heterodimers with CD3 (22, 23) or FcRI (24, 25) via a disulfide bond between cysteine residues present within the transmembrane region.  was initially shown to be play an important role in TCR-mediated signal transduction through the use of a -negative T cell hybridoma (26) and subsequently by -chain transfection studies (27, 28). The cytoplasmic portion of the -chain was demonstrated to be capable of signal transduction in the absence of other members of the TCR complex (29). Minimal sequence requirements for signal transduction via  were located to immunoreceptor tyrosine-based activation motifs using this approach (30, 31). CD3 contains three such immunoreceptor tyrosine-based activation motifs and FcRI one immunoreceptor tyrosine-based activation motif (31). Similarly, the important region for signaling via the  subunit of FcRI was located to the first 60 amino acids (32).

In this study, we have transiently expressed various FcRI chimeras in a simian kidney fibroblast cell line, COS-7, to investigate the structural determinants necessary to mediate functional IgE binding and subsequent signal transduction. The most effective chimeric FcRI receptors bound IgE and phagocytosed opsonized sheep red blood cells (SRBC) efficiently. The function of chimeric FcRI receptors was dependent upon an intact membrane proximal sequence and the type of transmembrane region employed. Expression of optimally functional chimeric FcRI receptor in a mouse CTL cell line conferred the ability to lyse target cells in an antibody-dependent manner.

MATERIALS AND METHODS

Relevant portions of molecules were amplified by polymerase chain reaction (PCR) from cDNA in the following plasmids: human  from pGEM3Z- (21), human FcRI from pKC3-E.1.1 (17), human FcRI (tryptophan 130 mutated to alanine 130, W130A) from pKC3-E.1.1, and human FcRIIa from M13mp18 (33). Following amplification, these portions were joined using PCR-splice overlap extension, cut with the appropriate restriction endonucleases, and cloned into the pCMV5 expression vector (34). A schematic representation and notation of each chimera is shown in Fig. 1. The sequences of the oligonucleotide primers used in the PCR reactions were as follows. In construct ^IIaIIa*,  was truncated at residue 59. 

Schematic representation of chimeric FcRI constructs.

: sense FcRI 5-GAGATCTAGACACAGTAACGACCAGGAG-3 (354); antisense FcRI 5-GGTAGCAGAGTACAGTAATGTTGAGGGGC-3; sense  5-CATTACTGTACTCTGCTACCTGCTGGAT-3; antisense  5-GAGACTCGAGGGCAGTTATAGGTCCCA-3 (357).

: sense FcRI (354); antisense FcRI 5-GTAGCAGAGTTGTAGCCAGTACTTCTCAC-3; sense  5-CTGGCTACAACTCTGCTACCTGCTGGAT-3; antisense  (357).

(C11S): sense FcRI (354); antisense FcRI 5-GTAGGAGAGTTGTAGCCAGTACTTCTCAC-3; sense  5-CTGGCTACAACTCTCCTACCTGCTGGAT-3; antisense  (357).

^IIaIIa: sense FcRI (354); antisense FcRI 5-GGCACTTGTACAGTAATGTTGAGGG-3; sense FcRIIa 5-CATTACTGTACAAGTGCCCAGCATG-3; antisense FcRIIa 5-CTTCACTCTGCAGTAGATCAAGGCCA-3 (487); sense  5-ATCTACTGCAGAGTGAAGTTCAGCAGG-3 (486); antisense  (357).

IIa: sense FcRI (354); antisense FcRI 5-CCTCATTGGTTGTAGCCAGTACTTCTC-3; sense FcRIIa 5-TGGCTACAACCAATGGGGATCATTGT-3; antisense FcRIIa (487); sense  (486); antisense  (357).

IIa: sense FcRI (354); antisense FcRI 5-CCACAATGATTACAGTAATGTTGAGGGGC-3; sense FcRIIa 5-ACATTACTGTAATCATTGTGGCTGTGCTC-3; antisense FcRIIa (487); sense  (486); antisense  (357).

^IIaIIa*: sense FcRI (354); antisense  5-CTCGAGAGATCTTTAGCTCCTGCTGAACTT-3.

^IIaIIa(W130A): sense FcRI (354); antisense FcRI 5-GGCACTTGTACAGTAATGTTGAGGG-3; sense FcRIIa 5-CATTACTGTACAAGTGCCCAGCATG-3; antisense FcRIIa (487); sense  (486); antisense  (357).

Each construct was used to transform Escherichia coli XL1-Blue (Stratagene) and fully sequenced using the T7 sequencing kit (Pharmacia, Uppsala, Sweden).

COS-7, a simian kidney fibroblast cell line, was maintained in Dulbecco's modified Eagle's medium (DMEM) (Life Technologies, Inc.) supplemented with 10% fetal calf serum, 5 × 10⁵ -mercaptoethanol, 2 m glutamine, 0.1 mg/ml streptomycin, and 100 units/ml penicillin. Cells were maintained at 37 °C with 10% CO₂. COS-7 cells were grown to 50% confluency (approximately 2 × 10⁵ cells in 2 ml per well), washed twice in DMEM without fetal calf serum, and transiently transfected by the DEAE-dextran method (35). Plasmid DNA (5 µg/ml) was incubated for 10 min at room temperature in DMEM containing 0.2 mg/ml DEAE-dextran (Pharmacia) and 1 m chloroquine (Sigma, Darmstadt, FRG). One ml of the mixture was added per well (6-well plates) or 10 ml per flask (175-cm² tissue culture flasks) (Falcon, Becton Dickinson, Plymouth, UK). Cells were incubated for 4 h at 37 °C, prior to media removal and 1-min 10% (v/v) dimethyl sulfoxide shock. Cells were then washed twice in DMEM and incubated for 48 h prior to harvest.

SRBC were washed three times in 0.15  NaCl, resuspended to 2 × 10⁸/ml, and labeled with trinitrophenyl (TNP) by incubating for 20 min in 0.05  2,4,6-trinitrobenzenesulfonic acid/phosphate-buffered saline (PBS, pH 7.2). After three washes, TNP-SRBC were resuspended to 1 × 10⁸/ml in a 1:5000 dilution of a mouse anti-TNP mAb (IgE ascites of ATCC cell line HB121) in PBS/0.5% bovine serum albumin (BSA) for 1 h at room temperature. Following 3 further washes, 2 × 10⁸ IgE-coated SRBC in 1 ml of PBS were added to transfected COS-7 cells in 6-well plates. The plates were incubated at 37 °C for 5 min, centrifuged (700 × g/5 min), and finally incubated (4 °C/30 min) before being microscopically examined for rosette formation. A rosette was defined as a COS-7 cell that bound more than 5 SRBC, and at least 400 COS-7 cells were evaluated in each well.

Transfected COS-7 cells in 6-well plates were stripped nonenzymatically, resuspended in 1 ml of a 1:1000 dilution of 3B4 anti-human FcRI (ascites mouse IgG₁ mAb) in PBS/0.5% BSA was added to transfected COS-7 cells in 6-well plates (above) and left at 4 °C for 1 h. Cells were then washed 3 times in PBS/0.5% BSA and incubated for 1 h at 4 °C in 1 ml of a 1:200 dilution of fluorescein isothiocyanate-labeled F(ab)₂ sheep anti-mouse Ig (Silenus, Hawthorn, Australia). After another 3 washes, fluorescence was assessed with a flow cytometer (FACScan, Becton Dickinson, San Jose, CA). The percentage of cells whose fluorescence intensity was clearly above that of cells treated with an isotype-matched control antibody (mouse IgG₁ mAb; anti-human granzyme B, 1D10) was recorded.

Twenty million COS-7 cell transfectants were washed 3 times in PBS and labeled in 200 µl of PBS with 5 µl of sodium [¹²⁵I]iodide (Amersham, UK) for 5 min in the presence of 1 mg/ml lactoperoxidase (Sigma) and hydrogen peroxide (BDH, Kilsyth, Australia) increasing in concentration every minute from 0.001% to 0.03%. After 3 additional washes, cells were lysed in 1 ml of lysis buffer (0.5% Nonidet P-40, 0.01  Tris, 0.15  NaCl, 1 m EDTA, 1 m phenylmethylsulfonyl fluoride, 0.2% aprotinin) for 30 min at 4 °C. Nuclei and cell debris were then pelleted by centrifugation (10,000 × g/15 min/4 °C), supernatant was decanted into a fresh tube and precleared with 50 µl of protein A-Sepharose (Pharmacia). The lysate was then divided into 2 aliquots and incubated overnight at 4 °C in the presence of 50 µl of Sepharose beads coupled to either 3B4 anti-FcRI mAb or to an isotype-matched control mAb (4H10, anti-human metase, mouse IgG₁). The beads were washed four times in lysis buffer, heated to 85 °C for 10 min, and the eluted proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (10% acrylamide gel). The gel was then dried and analyzed by autoradiography (Kodak).

Assays were performed using recombinant Fc portions of human IgE (obtained from Dr. Hannah Gould, King's College, London, UK), prepared by incubating 10 µg of IgE-Fc (in 50 µl of PBS, pH 7.4) with 500 µCi of Na¹²⁵I (Amersham) in the presence of 1 mg/ml chloramine T (Merck, Darmstadt, FRG) for 30 s on ice. Labeling reactions were stopped with 50 µl of 2.4 mg/ml sodium metabisulfite and 10 mg/ml tyrosine; free ¹²⁵I was removed from labeled protein by passage through a PD-10 Sephadex G-25 column (Pharmacia). COS-7 cells transiently transfected with chimeric cDNAs were harvested 60 h after transfection, washed twice in PBS/0.5% BSA, and resuspended at 2 × 10⁶/ml in PBS/0.5% BSA for use in binding assays. Fifty-µl aliquots of cells were incubated with 50 µl of serial dilutions of ¹²⁵I-Fc in PBS/0.5% BSA for 2 h at 4 °C. Cells were washed three times in PBS/0.5% BSA, resuspended in 100 µl, and counted in a -counter.

Phagocytosis of SRBC by transfected COS-7 cells was determined essentially as described in Hutchinson et al. (36). Briefly, COS-7 cells were prepared as for rosetting, then plates were incubated at 37 °C for 6 h and washed in PBS to remove nonadherent SRBC. COS-7 cells were then exposed to hypotonic shock for 20 s (PBS, 1 m NaCl, pH 2.5) resulting in lysis of exposed SRBC, but leaving COS-7 cells and any internalized SRBC intact. Cells were then fixed (0.5% glutaraldehyde in PBS, pH 7.4) and stained for myeloperoxidase activity (0.004% formaldehyde, 0.3 m o-dianisidine/PBS). The efficiency of phagocytosis was assessed microscopically and expressed as a phagocytic index calculated as the number of SRBC internalized per positive COS cell, where a cell containing 1 or more red cells was considered positive (and 100 were counted).

Generation of CTL Lines Stably Expressing Chimeric FcRI

CTL lines were generated by stable transfection of the mouse CTL cell line, CTLLR8 (37), with a construct containing FcRI driven by a 7.4-kilobase fragment of the murine perforin 5-flanking and promoter region including exon I, intron I, and part of exon II (38). The construct also contained the selectable marker ``neo'' driven by the thymidine kinase promoter. Transfection was performed by electroporation of 10⁷ cells in 0.25 ml of DMEM at 250 V and a capacitance of 960 microfarads (Gene Pulser, Bio-Rad). Stable transfected clones were derived by culture in complete DMEM containing 0.8 mg/ml Geneticin (Life Technologies, Inc.). FcRI-expressing clones were established by fluorescence-activated cell sorter analysis following treatment with 3B4 mAb and by their ability to form rosettes with IgE-coated SRBC. Parental CTLLR8 and the derived 3H2 clone were grown to midlog phase in complete DMEM supplemented with 10% culture supernatant from the phorbol 12-myristate 13-acetate-stimulated gibbon T cell leukemia MLA-144 as a source of IL-2.

The ability to redirect cytotoxicity using mouse CTLLR8 was assessed using a standard 4-h ⁵¹Cr release reverse ADCC assay. 3H2 cells were prepared by washing three times in DMEM and resuspension in 1 ml of complete media with a 1/500 dilution of 3B4 (anti-FcRI) or 4H10 (anti-human Met-1 isotype control) mAb. Following incubation at 4 °C for 30 min, cells were washed once in DMEM, counted, and resuspended to 10⁷ cells/ml. Mouse mastocytoma P815 (FcR⁺) target cells were washed in RPMI 1640 three times and then incubated for 1 h at 75 µCi of [⁵¹Cr]sodium chromate (Amersham, Bucks, UK) in 100 µl of RPMI 1640. After washing with complete medium, these cells were used as targets. The ⁵¹Cr-labeled target cells (2 × 10⁴ cells/ml) were mixed with parental CTLLR8 or 3H2 at various ratios in round-bottomed microtiter plates in a total volume of 200 µl for 4 h at 37 °C. The plates were then centifuged at 250 × g for 5 min, and 100-µl aliquots of the supernatants were assayed for radioactivity using a -counter. The spontaneous release of ⁵¹Cr was determined by incubating the target cells with medium alone, whereas the maximum release was determined by adding SDS to a final concentration of 5%. The percent specific lysis was calculated as follows: 100 × [(experimental ⁵¹Cr release  spontaneous ⁵¹Cr release)/(maximium ⁵¹Cr release  spontaneous ⁵¹Cr release)].

The ability to redirect cytotoxicity using mouse 3H2 cells was also assessed using a 4-h ⁵¹Cr release ADCC assay. An anti-Ly-2(CD8) IgE mAb was employed to direct 3H2 cytotoxicity. E3 (Ly-2⁺) target cells were washed in RPMI 1640 three times and then incubated for 1 h at 75 µCi of [⁵¹Cr]sodium chromate in 100 µl of RPMI 1640. After washing with complete medium, these E3 cells were labeled with chimeric (human/rat) YTS 169.4 anti-Ly-2 IgE mAb (39) for 30 min at 4 °C and washed twice in medium at 4 °C. Then, ⁵¹Cr-labeled target cells (2 × 10⁴ cells/ml) were mixed with 3H2 at various ratios in round-bottomed microtiter plates in a total volume of 200 µl for 4 h at 37 °C. The plates were then centifuged at 250 × g for 5 min, and 100-µl aliquots of the supernatants were assayed for radioactivity using a -counter. The spontaneous release and the percent specific lysis was calculated as above.

RESULTS

Construction and Expression of Chimeric FcRI

In order to construct a chimeric FcRI receptor capable of effector function in the absence of endogenous CD3 or FcRI signaling molecules, a number of chimeric gene constructs composed of various extracellular, transmembrane, and cytoplasmic domains were generated by PCR-splice overlap extension and transfected into COS-7 cells (Fig. 1). Expression was initially assessed by flow cytometry using the anti-human FcRI mAb, 3B4, that recognizes an epitope located in domain I of FcRI. The level of expression varied between chimeras, ranging from approximately 10 to 25% of cells, and no fluorescence was observed in COS-7 cells transfected with the CMV5 vector alone (Table I).

Table I. Flow cytometric analysis for chimeric FcRI expression in COS-7 cells COS-7 cells were transiently transfected with the constructs indicated. Expression of chimeric receptor was assessed as percentage of cells above a mean channel fluorescence of 280 ± 30 following treatment with 3B4 anti-FcRI mAb and fluorescein isothiocyanate-conjugated sheep anti-mouse Ig. Fluorescence intensity clearly above that (mean channel fluorescence > 280 ± 30) of cells treated with an isotype-matched control antibody (1D10; mouse IgG1; anti-human granzyme B) was calculated by subtracting the percentage of control antibody-treated cells. The percentage of positive immunofluorescent cells was calculated as the mean ± S.E. from three independent experiments. Chimera % positive cells  IIaIIaIIa 10.0  ± 1.2   9.5  ± 1.4   +  subunit 6.8  ± 1.0   11.2  ± 1.0   +  subunit 10.0  ± 1.1   (C11S) 13.9  ± 1.0   (C11S) +  subunit 10.9  ± 1.1  IIaIIa 20.7  ± 0.6  IIa 25.3  ± 1.8  IIa 14.4  ± 1.5  IIaIIa (W130A) 23.4  ± 1.2

The capacity of each construct to confer IgE binding ability on transfected COS-7 cells was assessed by a rosetting assay using IgE-coated SRBC (Fig. 2, A and B). Initially, a chimera was generated that included just the IgE binding domains of FcRI (lacking an 11-amino acid membrane-proximal region) fused to transmembrane and intracellular  (). Although this chimeric receptor was expressed on the cell surface (9.5%, Table I), it was unable to bind IgE as assessed by erythrocyte-antibody complex rosetting (Fig. 2, A and B). Domain II of FcRI is largely responsible for binding of IgE (17, 18, 40) and is normally separated from the cell membrane by the 11-amino acid membrane-proximal region. It was postulated that the lack of IgE binding by might be caused by steric hindrance between FcRI domain II and the cell membrane, and, therefore, a new construct was designed that included the membrane-proximal region.

The receptor was generated which includes the 11-amino acid membrane-proximal region of FcRI. COS-7 cells transfected with were able to form few rosettes (approximately 5%). Clearly, the presence of this region between the IgE binding domain II and the membrane permitted some IgE binding. Lysates of COS-7 cells transfected with were analyzed by SDS-PAGE following surface iodination and immunoprecipitation with 3B4 mAb. Chimeric molecules of the predicted 50 kDa were demonstrated under reducing conditions (Fig. 3); however, nonreducing conditions revealed only molecules of 120-200 kDa (Fig. 3). This suggested that the chimeric receptor existed as a multimeric complex on the cell surface.

Immunoprecipitation of chimeric FcRI from transfected COS-7 cells.

To determine whether IgE binding was affected by multimerization of chimeric receptor molecules, a construct was made in which Cys¹¹ of transmembrane  was mutated to Ser¹¹ ((C11S)). Cys¹¹ of  has previously been implicated in its dimerization (22, 23, 24). A minor increase in rosette formation was observed (up to 10%) (Fig. 2A). Immunoprecipitation and nonreducing SDS-PAGE demonstrated a ratio of ~60% 50-kDa monomer to ~40% 120-150-kDa multimeric complex (Fig. 3). Therefore, we hypothesize that other residues in the (C11S) chimera might also be responsible for multimer formation.

The transmembrane region of  was next substituted with that of human FcRIIa. Human FcRIIa normally exists as a monomer on the cell surface and can associate with the  subunit of FcRI although this interaction is not necessary for expression.² This chain has previously been used to express monomeric FcR chimeras (17, 36). The resulting construct was a tripartite chimera composed of extracellular FcRI, membrane-proximal and transmembrane FcRIIa, and intracellular  (^IIaIIa). This construct was expressed more effectively in COS-7 cells (Table I), and rosette formation by COS-7 cells transfected with ^IIaIIa was in excess of 50% (Fig. 2, A and B). Immunoprecipitation of ^IIaIIa demonstrated some residual multimeric receptor, but the majority of chimeric receptor existed as the 50-kDa monomer (Fig. 3). A construct simply made up of extracellular FcRI fused to transmembrane and cytoplasmic FcRIIa (^IIaIIaIIa) (Fig. 1) was expressed at lower levels (Table I), but also formed >50% rosettes (Fig. 2A) and existed predominantly as a 50-kDa monomer as determined by immunoprecipitation (data not shown).

Replacement of the membrane proximal (Ile¹⁷⁰  Gln¹⁸⁰) FcRIIa with the corresponding region (Gln¹⁷⁰  Ser¹⁷⁸) of FcRI (IIa) did not affect the ability of chimeric FcRI receptor to rosette IgE-coated SRBC (Fig. 2A). However, complete elimination of the membrane-proximal region, by linking Val¹⁶⁹ of FcRI directly to Pro¹⁷⁹ of FcRIIa (IIa), totally abolished rosette formation (Fig. 2A). Immunoprecipitation and nonreducing SDS-PAGE suggested that IIa does exist predominantly as a monomeric receptor (data not shown).

Affinity of Chimeric FcRI

The affinity of transfected FcRI chimeric receptors for IgE Fc was determined by Scatchard analysis. A biphasic binding curve was obtained for all the chimeric receptors (data not shown) with high affinity binding constants calculated and shown in Table II. The affinities of IIa (0.93 × 10^{9 1}) and ^IIaIIa (0.86 × 10^{9 1}) for IgE were less than that of wild type FcRI (2.2 × 10^{9 1}); however, that of (C11S) and were considerably less (<5 × 10^{7 1}). Mutation of FcRI Trp¹³⁰ to Ala¹³⁰ (^IIaIIa(W130A)) increased the affinity for IgE by 2-fold (Table II), consistent with a previous study by Hulett et al.³ that indicated that the affinity of FcRI could be enhanced by alterations in the second domain of the -chain. The greater affinity of ^IIaIIa(W130A) was evident despite similar expression (Table I) or rosette-forming ability (Fig. 2A) of this chimeric receptor and IIa or ^IIaIIa in COS-7 cells.

Table II. Scatchard analysis of affinity of chimeric FcRI expression in COS-7 cells Determination of affinity constants by Scatchard analysis. COS-7 cells transfected with the above chimeric receptors were incubated with various concentrations of 125I-labeled recombinant Fc portions of IgE. The specific ratio of bound/free IgE-Fc was plotted as a function of bound IgE-Fc. The slopes yielded the equilibrium constants listed. Chimera Affinity 1  IIaIIaIIa 2.20  ± 0.33 × 109   <5  × 107   <5  × 107   + subunit 1.29  ± 0.10 × 109   (C11S) <5  × 107  IIaIIa 0.86  ± 0.13 × 109  IIa 0.93  ± 0.25 × 109  IIa <5  × 107  IIaIIa (W130A) 1.64  ± 0.28 × 109

FcRI Chimeric Receptor-mediated Phagocytosis

In addition to mediate binding of IgE, it was important that chimeric receptors were able to trigger effector function. The ability to mediate phagocytosis of opsonized SRBC has been demonstrated for chimeric FcRI, FcRI-FcRIIa (36), FcRIII-, and FcRIII- (41). This process was shown to be dependent on phosphorylation of tyrosine residues in the cytoplasmic regions of  or . To determine whether chimeric FcRI was capable of mediating phagocytosis, transfected COS-7 cells were assessed for their ability to ingest IgE-coated SRBC (Table III and Fig. 4). The percentage of phagocytic positive COS-7 cells was found to correlate with rosette formation, and indices ranged from 0.0 to 9.7 (Table III). Not surprisingly, a truncated version of the IIa chimera (IIa*, see ``Materials and Methods'' and Fig. 1) lacking cytoplasmic , did not phagocytose IgE-coated SRBC (Table III), despite being expressed on COS cells (data not shown) and rosetting opsonized SRBC (Fig. 2A). Furthermore, COS-7 cells expressing ^IIaIIaIIa were able to phagocytose opsonized SRBC, consistent with previous reports that human FcRIIa can mediate phagocytosis in heterologous systems (42, 43). Overall, these data indicated that chimeric FcRI molecules were capable of effector function in the absence of endogenous CD3 or FcRI signaling molecules.

Table III. Phagocytic index of COS-7 cells expressing chimeric FcRI The efficiency of phagocytosis was expressed as a phagocytic index (PI) calculated as the number of SRBC internalized per positive COS cell, where a cell containing 1 or more red cells was considered positive (and 100 were counted). Chimera PI index Vector alone 0.0  ± 0.0   0.0  ± 0.0   1.5  ± 0.3   + -subunit 8.9  ± 1.1   (C11S) 1.6  ± 0.3 eIIaIIa 9.7  ± 2.1  IIa 9.1  ± 2.0  IIa 0.0  ± 0.0  IIaIIa (W130A) 8.6  ± 1.8  IIaIIa* 0.0  ± 0.0

Phagocytosis of opsonized SRBC by COS-7 cells transfected with chimeric FcRI.

Cotransfection with  Reconstitutes Chimeric FcRI Function

Chimeric FcRI gene constructs that did not confer COS-7 cells with optimal ability to rosette opsonized SRBC (i.e. and (C11S)) were cotransfected into COS-7 cells with a construct encoding the  subunit of the FcRI complex driven by the SV40 promoter. Importantly, cotransfection with  subunit had little effect upon the level of expression of any of these chimeras in COS-7 cells (Table I). Cotransfection greatly enhanced rosette formation (Fig. 2, A and B) and phagocytosis of opsonized SRBC (Table III). Interestingly, chimeric gene constructs that did not confer COS-7 cells with any ability to rosette opsonized SRBC (i.e. and IIa) were not affected by -chain cotransfection (Fig. 2A). Furthermore, steric hindrance caused by deletion of the extracellular membrane-proximal region, cannot be resolved by -chain cotransfection.

CTL Expressing Chimeric FcRI Lyse Targets in ADCC Assays

To characterize the ability of chimeric FcRI to function in cytotoxic lymphocytes, the mouse CTL cell line, CTLLR8, was stably transfected with ^IIaIIa and cloned by limiting dilution. Resultant clones were examined for cytotoxic potential in a reverse ADCC assay against FcR⁺ P815 targets using the 3B4 anti-FcRI mAb (mouse IgG₁). CTLLR8 clones 3H2 effectively lysed P815 in the presence of 3B4 mAb (>20%), but not in the presence of an isotype-matched negative control mAb (Fig. 5). By contrast, parental CTLLR8 displayed only a low level of direct cytotoxicity (<5%) toward P815 in the presence of either mAb (Fig. 5A).

R8-3H2 was also examined for cytotoxic potential in an IgE-dependent cellular cytotoxicity assay. The cytotoxicity of R8-3H2 was tested against mouse E3 (Ly-2⁺) thymoma targets in the absence or presence of a chimeric human/rat anti-Ly-2 IgE mAb (Fig. 5B). At three effector:target ratios, only E3 cells precoated with anti-Ly-2 IgE mAb were sensitive to lysis by R8-3H2, and up to 45% IgE-specific lysis was observed against the E3 target cells.

DISCUSSION

The data arising from the chimeric receptor studies presented herein suggest that the membrane-proximal (``stalk'') region of the FcRI chain plays a crucial role in the interaction with IgE, as its deletion was found to totally abrogate IgE binding. Previous studies have localized the IgE binding site of FcRI to multiple regions of the second extracellular domain (17, 18, 40) which based on molecular modelling map primarily to loop regions at the interface with domain 1 (44, 45). The stalk region of FcRI is situated distantly to the proposed binding site and as such does not appear to be directly involved in IgE binding. Indeed, the substitution of the FcRI stalk region with the corresponding region from FcRIIa had no apparent effect on IgE binding. Therefore, these data suggest that the FcRI stalk is acting as a spacer region to ensure correct topology of the receptor on the cell membrane such that the binding site is available to interact appropriately with IgE. A model of the interaction of FcRI with IgE has been proposed, suggesting that the extracellular region of FcRI lies across the cell membrane and interacts with the Fc portion of IgE principally through C3 (45, 46, 47, 48). IgE has been predicted to adopt a bent conformation that, upon interaction with FcRI, binds with its convex surface facing the cell membrane (48). The C3 and C4 domains of IgE are positioned closest to the membrane, which is consistent with the proposed model of FcRI-IgE interaction. It is possible that the removal of the FcRI stalk region may alter the proposed topology of the receptor such that it is no longer appropriately orientated on the cell surface to allow access of IgE to the binding region. Alternatively, the deletion of the FcRI stalk may simply be altering the conformation of the receptor resulting in disruption of the structural integrity of the IgE binding site. However, this scenario is unlikely, as replacement of the FcRI stalk region with that of FcRIIa, or a similar region of p55 IL-2 receptor (16), does not alter IgE binding. Other members of the immunoglobulin superfamily (49, 50) have sequences of amino acids that appear to function as hinge or spacer regions involved in improving ligand binding by providing flexibility or separation between domains.

Mutation of FcRI Trp¹³⁰  Ala¹³⁰ (^IIaIIa(W130A)) increased the affinity for IgE by 2-fold, consistent with the affinity of FcRI being modified by alterations in the second domain of the FcRI chain.³ The greater affinity of ^IIaIIa(W130A) was evident despite similar levels of cell surface expression or rosette formation of this chimeric receptor, IIa and ^IIaIIa in COS-7 cells. The ability of chimeric FcRI to be expressed and bind IgE was also critically dependent on the composition of the transmembrane region. Chimeras containing transmembrane FcRIIa were clearly more efficient than those with transmembrane . This increase in IgE binding affinity also correlated with the expression of monomeric receptor on the cell surface. The mechanism of formation of multimeric receptor is unknown, although it seems possible that close molecular association mediated by transmembrane  contributes to intermolecular disulfide bond formation. As there is only one cysteine residue present within the  portion of the receptor, any additional aggregation must occur through one or more of the four cysteines of FcRI. Homomultimers and/or heteromultimers have been observed previously for native and chimeric CD8 (29, 51). The inability of multimeric receptor to bind IgE may be due to steric hindrance of the binding domain of FcRI caused by the close association of two or more receptors. Alternatively it could be caused by the disruption of binding domain integrity due to the loss of one or more intramolecular disulfide bonds when cysteine residues paired intermolecularly.

Chimeras in which the extracellular domain of FcRI was attached to the transmembrane domain of FcRIIa and cytoplasmic domain of -chain conferred upon COS-7 cells the ability to efficiently phagocytose antibody-coated SRBC. These data paralleled similar experiments by Hutchinson et al. (36) that prepared chimeras of FcRI. Human FcRIIa has been reported to mediate phagocytosis in two different heterologous systems, 3T6 and COS-1 fibroblasts (42, 43). In the first of these studies, removal of the cytoplasmic tail of human FcRIIa abolished phagocytosis implying that this domain contains the phagocytic signal. In our study, the ^IIaIIaIIa chimeric receptor confirms that the cytoplasmic domain of FcRIIa contains the phagocytic signal. This chimera and those with an intact membrane-proximal region and transmembrane FcRIIa probably trigger efficient phagocytosis in COS-7 cells following IgE-mediated aggregation of tyrosine activation motifs. Clearly, truncation of the  cytoplasmic tail in an ^IIaIIa^* chimera abolished phagocytic function, thereby indicating that, in the absence of cytoplasmic IIa, the -chain was critical for effector function.

Chimeric receptors utilizing transmembrane and intracellular TCR- have previously been shown to successfully mediate CTL/NK cell-mediated lysis of surface Ig-bearing B-cell lines (30) and HIV gp120/41 expressing cells (3, 6, 52) in vitro and tumor growth in vivo (2). Despite this, expressing anti-tumor scFv in CTL suffers from two major disadvantages. Firstly, there is a need to generate and express individual scFv antibody receptors for every tumor antigen targeted, and, secondly, a monospecific interaction between the receptor and tumor may be ineffective given that tumor cells have and develop heterogeneous antigen expression. Herein, we have established the ability of ^IIaIIa chimeric receptor to mediate reverse- or IgE-ADCC of target cells when transfected into a mouse CTL cell line. Furthermore, the restoration of IgE binding by cotransfection with  and chimeric suggests that this receptor might also be functional in /-positive T-cells and NK cells. Given that specific anti-tumor IgE mAbs can simply be made by genetic engineering, these chimeric FcRI may provide a universal and alternative means of redirecting cytotoxic lymphocytes to tumors. Future efforts will be made to assess the versatility and efficacy of these IgE binding chimeric receptors to redirect killer cell function in mouse tumor models in vivo.